Update CHANGELOG for v1.1.0 (Eider) release

Change-Id: Ic429556e76bcc4f96a34e18835a153b07fe410a2

.mailmap

@@ -3,3 +3,6 @@ Johann Koenig <johannkoenig@google.com>
 Tero Rintaluoma <teror@google.com> <tero.rintaluoma@on2.com>
 Tom Finegan <tomfinegan@google.com>
 Ralph Giles <giles@xiph.org> <giles@entropywave.com>
+Ralph Giles <giles@xiph.org> <giles@mozilla.com>
+Alpha Lam <hclam@google.com> <hclam@chromium.org>
+Deb Mukherjee <debargha@google.com>

AUTHORS

@@ -6,10 +6,12 @@ Adrian Grange <agrange@google.com>
 Alex Converse <alex.converse@gmail.com>
 Alexis Ballier <aballier@gentoo.org>
 Alok Ahuja <waveletcoeff@gmail.com>
+Alpha Lam <hclam@google.com>
 Andoni Morales Alastruey <ylatuya@gmail.com>
 Andres Mejia <mcitadel@gmail.com>
 Aron Rosenberg <arosenberg@logitech.com>
 Attila Nagy <attilanagy@google.com>
+Deb Mukherjee <debargha@google.com>
 Fabio Pedretti <fabio.ped@libero.it>
 Frank Galligan <fgalligan@google.com>
 Fredrik Söderquist <fs@opera.com>
@@ -21,6 +23,7 @@ Henrik Lundin <hlundin@google.com>
 James Berry <jamesberry@google.com>
 James Zern <jzern@google.com>
 Jan Kratochvil <jan.kratochvil@redhat.com>
+Jeff Faust <jfaust@google.com>
 Jeff Muizelaar <jmuizelaar@mozilla.com>
 Jim Bankoski <jimbankoski@google.com>
 Johann Koenig <johannkoenig@google.com>
@@ -28,9 +31,11 @@ John Koleszar <jkoleszar@google.com>
 Joshua Bleecher Snyder <josh@treelinelabs.com>
 Justin Clift <justin@salasaga.org>
 Justin Lebar <justin.lebar@gmail.com>
+KO Myung-Hun <komh@chollian.net>
 Lou Quillio <louquillio@google.com>
 Luca Barbato <lu_zero@gentoo.org>
 Makoto Kato <makoto.kt@gmail.com>
+Marco Paniconi <marpan@google.com>
 Martin Ettl <ettl.martin78@googlemail.com>
 Michael Kohler <michaelkohler@live.com>
 Mike Hommey <mhommey@mozilla.com>
@@ -40,12 +45,15 @@ Patrik Westin <patrik.westin@gmail.com>
 Paul Wilkins <paulwilkins@google.com>
 Pavol Rusnak <stick@gk2.sk>
 Philip Jägenstedt <philipj@opera.com>
+Priit Laes <plaes@plaes.org>
 Rafael Ávila de Espíndola <rafael.espindola@gmail.com>
+Rafaël Carré <funman@videolan.org>
 Ralph Giles <giles@xiph.org>
 Ronald S. Bultje <rbultje@google.com>
 Scott LaVarnway <slavarnway@google.com>
 Stefan Holmer <holmer@google.com>
 Taekhyun Kim <takim@nvidia.com>
+Takanori MATSUURA <t.matsuu@gmail.com>
 Tero Rintaluoma <teror@google.com>
 Thijs Vermeir <thijsvermeir@gmail.com>
 Timothy B. Terriberry <tterribe@xiph.org>

CHANGELOG

@@ -1,3 +1,143 @@
+2012-05-09 v1.1.0 "Eider"
+  This introduces a number of enhancements, mostly focused on real-time
+  encoding. In addition, it fixes a decoder bug (first introduced in
+  Duclair) so all users of that release are encouraged to upgrade.
+
+  - Upgrading:
+    This release is ABI and API compatible with Duclair (v1.0.0). Users
+    of older releases should refer to the Upgrading notes in this
+    document for that release.
+
+    This release introduces a new temporal denoiser, controlled by the
+    VP8E_SET_NOISE_SENSITIVITY control. The temporal denoiser does not
+    currently take a strength parameter, so the control is effectively
+    a boolean - zero (off) or non-zero (on). For compatibility with
+    existing applications, the values accepted are the same as those
+    for the spatial denoiser (0-6). The temporal denoiser is enabled
+    by default, and the older spatial denoiser may be restored by
+    configuring with --disable-temporal-denoising. The temporal denoiser
+    is more computationally intensive than the spatial one.
+
+    This release removes support for a legacy, decode only API that was
+    supported, but deprecated, at the initial release of libvpx
+    (v0.9.0). This is not expected to have any impact. If you are
+    impacted, you can apply a reversion to commit 2bf8fb58 locally.
+    Please update to the latest libvpx API if you are affected.
+
+  - Enhancements:
+      Adds a motion compensated temporal denoiser to the encoder, which
+      gives higher quality than the older spatial denoiser. (See above
+      for notes on upgrading).
+
+      In addition, support for new compilers and platforms were added,
+      including:
+        improved support for XCode
+        Android x86 NDK build
+        OS/2 support
+        SunCC support
+
+      Changing resolution with vpx_codec_enc_config_set() is now
+      supported. Previously, reinitializing the codec was required to
+      change the input resolution.
+
+      The vpxenc application has initial support for producing multiple
+      encodes from the same input in one call. Resizing is not yet
+      supported, but varying other codec parameters is. Use -- to
+      delineate output streams. Options persist from one stream to the
+      next.
+
+      Also, the vpxenc application will now use a keyframe interval of
+      5 seconds by default. Use the --kf-max-dist option to override.
+
+  - Speed:
+      Decoder performance improved 2.5% versus Duclair. Encoder speed is
+      consistent with Duclair for most material. Two pass encoding of
+      slideshow-like material will see significant improvements.
+
+      Large realtime encoding speed gains at a small quality expense are
+      possible by configuring the on-the-fly bitpacking experiment with
+      --enable-onthefly-bitpacking. Realtime encoder can be up to 13%
+      faster (ARM) depending on the number of threads and bitrate
+      settings. This technique sees constant gain over the 5-16 speed
+      range. For VC style input the loss seen is up to 0.2dB. See commit
+      52cf4dca for further details.
+
+  - Quality:
+      On the whole, quality is consistent with the Duclair release. Some
+      tweaks:
+
+        Reduced blockiness in easy sections by applying a penalty to
+        intra modes.
+
+        Improved quality of static sections (like slideshows) with
+        two pass encoding.
+
+        Improved keyframe sizing with multiple temporal layers
+
+  - Bug Fixes:
+      Corrected alt-ref contribution to frame rate for visible updates
+      to the alt-ref buffer. This affected applications making manual
+      usage of the frame reference flags, or temporal layers.
+
+      Additional constraints were added to disable multi-frame quality
+      enhancement (MFQE) in sections of the frame where there is motion.
+      (#392)
+
+      Fixed corruption issues when vpx_codec_enc_config_set() was called
+      with spatial resampling enabled.
+
+      Fixed a decoder error introduced in Duclair where the segmentation
+      map was not being reinitialized on keyframes (#378)
+
+
+2012-01-27 v1.0.0 "Duclair"
+  Our fourth named release, focused on performance and features related to
+  real-time encoding. It also fixes a decoder crash bug introduced in
+  v0.9.7, so all users of that release are encouraged to upgrade.
+
+  - Upgrading:
+      This release is ABI incompatible with prior releases of libvpx, so the
+      "major" version number has been bumped to 1. You must recompile your
+      applications against the latest version of the libvpx headers. The
+      API remains compatible, and this should not require code changes in most
+      applications.
+
+  - Enhancements:
+      This release introduces several substantial new features to the encoder,
+      of particular interest to real time streaming applications.
+
+      Temporal scalability allows the encoder to produce a stream that can
+      be decimated to different frame rates, with independent rate targetting
+      for each substream.
+
+      Multiframe quality enhancement postprocessing can make visual quality
+      more consistent in the presence of frames that are substantially
+      different quality than the surrounding frames, as in the temporal
+      scalability case and in some forced keyframe scenarios.
+
+      Multiple-resolution encoding support allows the encoding of the
+      same content at different resolutions faster than encoding them
+      separately.
+
+  - Speed:
+      Optimization targets for this release included the decoder and the real-
+      time modes of the encoder. Decoder speed on x86 has improved 10.5% with
+      this release. Encoder improvements followed a curve where speeds 1-3
+      improved 4.0%-1.5%, speeds 4-8 improved <1%, and speeds 9-16 improved
+      1.5% to 10.5%, respectively. "Best" mode speed is consistent with the
+      Cayuga release.
+
+  - Quality:
+      Encoder quality in the single stream case is consistent with the Cayuga
+      release.
+
+  - Bug Fixes:
+      This release fixes an OOB read decoder crash bug present in v0.9.7
+      related to the clamping of motion vectors in SPLITMV blocks. This
+      behavior could be triggered by corrupt input or by starting
+      decoding from a P-frame.
+
+
 2011-08-15 v0.9.7-p1 "Cayuga" patch 1
   This is an incremental bugfix release against Cayuga. All users of that
   release are strongly encouraged to upgrade.

README

@@ -42,17 +42,13 @@ COMPILING THE APPLICATIONS/LIBRARIES:
   --help output of the configure script. As of this writing, the list of
   available targets is:
 
+    armv5te-android-gcc
     armv5te-linux-rvct
     armv5te-linux-gcc
-    armv5te-symbian-gcc
     armv6-darwin-gcc
     armv6-linux-rvct
     armv6-linux-gcc
-    armv6-symbian-gcc
-    iwmmxt-linux-rvct
-    iwmmxt-linux-gcc
-    iwmmxt2-linux-rvct
-    iwmmxt2-linux-gcc
+    armv7-android-gcc
     armv7-linux-rvct
     armv7-linux-gcc
     mips32-linux-gcc

args.c

@@ -57,7 +57,7 @@ int arg_match(struct arg *arg_, const struct arg_def *def, char **argv)
     }
     else if (def->long_name)
     {
-        int name_len = strlen(def->long_name);
+        const size_t name_len = strlen(def->long_name);
 
         if (strlen(arg.argv[0]) >= name_len + 2
             && arg.argv[0][1] == '-'

build/make/Android.mk

@@ -0,0 +1,199 @@
+##
+##  Copyright (c) 2012 The WebM project authors. All Rights Reserved.
+##
+##  Use of this source code is governed by a BSD-style license
+##  that can be found in the LICENSE file in the root of the source
+##  tree. An additional intellectual property rights grant can be found
+##  in the file PATENTS.  All contributing project authors may
+##  be found in the AUTHORS file in the root of the source tree.
+##
+
+#
+# This file is to be used for compiling libvpx for Android using the NDK.
+# In an Android project place a libvpx checkout in the jni directory.
+# Run the configure script from the jni directory.  Base libvpx
+# encoder/decoder configuration will look similar to:
+# ./libvpx/configure --target=armv7-android-gcc --disable-examples \
+#                    --sdk-path=/opt/android-ndk-r6b/
+#
+# When targeting Android, realtime-only is enabled by default.  This can
+# be overridden by adding the command line flag:
+#  --disable-realtime-only
+#
+# This will create .mk files that contain variables that contain the
+# source files to compile.
+#
+# Place an Android.mk file in the jni directory that references the
+# Android.mk file in the libvpx directory:
+# LOCAL_PATH := $(call my-dir)
+# include $(CLEAR_VARS)
+# include libvpx/build/make/Android.mk
+#
+# There are currently two TARGET_ARCH_ABI targets for ARM.
+# armeabi and armeabi-v7a.  armeabi-v7a is selected by creating an
+# Application.mk in the jni directory that contains:
+# APP_ABI := armeabi-v7a
+#
+# To change to building armeabi, run ./libvpx/configure again, but with
+# --target=arm5te-android-gcc and and modify the Application.mk file to
+# set APP_ABI := armeabi
+#
+# Running ndk-build will build libvpx and include it in your project.
+#
+
+CONFIG_DIR := $(LOCAL_PATH)
+LIBVPX_PATH := $(LOCAL_PATH)/libvpx
+ASM_CNV_PATH_LOCAL := $(TARGET_ARCH_ABI)/ads2gas
+ASM_CNV_PATH := $(LOCAL_PATH)/$(ASM_CNV_PATH_LOCAL)
+
+# Makefiles created by the libvpx configure process
+# This will need to be fixed to handle x86.
+ifeq ($(TARGET_ARCH_ABI),armeabi-v7a)
+  include $(CONFIG_DIR)/libs-armv7-android-gcc.mk
+else
+  include $(CONFIG_DIR)/libs-armv5te-android-gcc.mk
+endif
+
+# Rule that is normally in Makefile created by libvpx
+# configure.  Used to filter out source files based on configuration.
+enabled=$(filter-out $($(1)-no),$($(1)-yes))
+
+# Override the relative path that is defined by the libvpx
+# configure process
+SRC_PATH_BARE := $(LIBVPX_PATH)
+
+# Include the list of files to be built
+include $(LIBVPX_PATH)/libs.mk
+
+# Want arm, not thumb, optimized
+LOCAL_ARM_MODE := arm
+LOCAL_CFLAGS := -O3
+
+# -----------------------------------------------------------------------------
+# Template  : asm_offsets_template
+# Arguments : 1: assembly offsets file to be created
+#             2: c file to base assembly offsets on
+# Returns   : None
+# Usage     : $(eval $(call asm_offsets_template,<asmfile>, <srcfile>
+# Rationale : Create offsets at compile time using for structures that are
+#             defined in c, but used in assembly functions.
+# -----------------------------------------------------------------------------
+define asm_offsets_template
+
+_SRC:=$(2)
+_OBJ:=$(ASM_CNV_PATH)/$$(notdir $(2)).S
+
+_FLAGS = $$($$(my)CFLAGS) \
+          $$(call get-src-file-target-cflags,$(2)) \
+          $$(call host-c-includes,$$(LOCAL_C_INCLUDES) $$(CONFIG_DIR)) \
+          $$(LOCAL_CFLAGS) \
+          $$(NDK_APP_CFLAGS) \
+          $$(call host-c-includes,$$($(my)C_INCLUDES)) \
+          -DINLINE_ASM \
+          -S \
+
+_TEXT = "Compile $$(call get-src-file-text,$(2))"
+_CC   = $$(TARGET_CC)
+
+$$(eval $$(call ev-build-file))
+
+$(1) : $$(_OBJ) $(2)
+	@mkdir -p $$(dir $$@)
+	@grep $(OFFSET_PATTERN) $$< | tr -d '\#' | $(CONFIG_DIR)/$(ASM_CONVERSION) > $$@
+endef
+
+# Use ads2gas script to convert from RVCT format to GAS format.  This passes
+#  puts the processed file under $(ASM_CNV_PATH).  Local clean rule
+#  to handle removing these
+ASM_CNV_OFFSETS_DEPEND = $(ASM_CNV_PATH)/asm_com_offsets.asm
+ifeq ($(CONFIG_VP8_DECODER), yes)
+  ASM_CNV_OFFSETS_DEPEND += $(ASM_CNV_PATH)/asm_dec_offsets.asm
+endif
+ifeq ($(CONFIG_VP8_ENCODER), yes)
+  ASM_CNV_OFFSETS_DEPEND += $(ASM_CNV_PATH)/asm_enc_offsets.asm
+endif
+
+.PRECIOUS: %.asm.s
+$(ASM_CNV_PATH)/libvpx/%.asm.s: $(LIBVPX_PATH)/%.asm $(ASM_CNV_OFFSETS_DEPEND)
+	@mkdir -p $(dir $@)
+	@$(CONFIG_DIR)/$(ASM_CONVERSION) <$< > $@
+
+# For building vpx_rtcd.h, which has a rule in libs.mk
+TGT_ISA:=$(word 1, $(subst -, ,$(TOOLCHAIN)))
+target := libs
+
+LOCAL_SRC_FILES += vpx_config.c
+
+# Remove duplicate entries
+CODEC_SRCS_UNIQUE = $(sort $(CODEC_SRCS))
+
+# Pull out C files.  vpx_config.c is in the immediate directory and
+# so it does not need libvpx/ prefixed like the rest of the source files.
+CODEC_SRCS_C = $(filter %.c, $(CODEC_SRCS_UNIQUE))
+LOCAL_CODEC_SRCS_C = $(filter-out vpx_config.c, $(CODEC_SRCS_C))
+
+LOCAL_SRC_FILES += $(foreach file, $(LOCAL_CODEC_SRCS_C), libvpx/$(file))
+
+# Pull out assembly files, splitting NEON from the rest.  This is
+# done to specify that the NEON assembly files use NEON assembler flags.
+CODEC_SRCS_ASM_ALL = $(filter %.asm.s, $(CODEC_SRCS_UNIQUE))
+CODEC_SRCS_ASM = $(foreach v, \
+                 $(CODEC_SRCS_ASM_ALL), \
+                 $(if $(findstring neon,$(v)),,$(v)))
+CODEC_SRCS_ASM_ADS2GAS = $(patsubst %.s, \
+                         $(ASM_CNV_PATH_LOCAL)/libvpx/%.s, \
+                         $(CODEC_SRCS_ASM))
+LOCAL_SRC_FILES += $(CODEC_SRCS_ASM_ADS2GAS)
+
+ifeq ($(TARGET_ARCH_ABI),armeabi-v7a)
+  CODEC_SRCS_ASM_NEON = $(foreach v, \
+                        $(CODEC_SRCS_ASM_ALL),\
+                        $(if $(findstring neon,$(v)),$(v),))
+  CODEC_SRCS_ASM_NEON_ADS2GAS = $(patsubst %.s, \
+                                $(ASM_CNV_PATH_LOCAL)/libvpx/%.s, \
+                                $(CODEC_SRCS_ASM_NEON))
+  LOCAL_SRC_FILES += $(patsubst %.s, \
+                     %.s.neon, \
+                     $(CODEC_SRCS_ASM_NEON_ADS2GAS))
+endif
+
+LOCAL_CFLAGS += \
+    -DHAVE_CONFIG_H=vpx_config.h \
+    -I$(LIBVPX_PATH) \
+    -I$(ASM_CNV_PATH)
+
+LOCAL_MODULE := libvpx
+
+LOCAL_LDLIBS := -llog
+
+LOCAL_STATIC_LIBRARIES := cpufeatures
+
+$(foreach file, $(LOCAL_SRC_FILES), $(LOCAL_PATH)/$(file)): vpx_rtcd.h
+
+.PHONY: clean
+clean:
+	@echo "Clean: ads2gas files [$(TARGET_ARCH_ABI)]"
+	@$(RM) $(CODEC_SRCS_ASM_ADS2GAS) $(CODEC_SRCS_ASM_NEON_ADS2GAS)
+	@$(RM) $(patsubst %.asm, %.*, $(ASM_CNV_OFFSETS_DEPEND))
+	@$(RM) -r $(ASM_CNV_PATH)
+	@$(RM) $(CLEAN-OBJS)
+
+include $(BUILD_SHARED_LIBRARY)
+
+$(eval $(call asm_offsets_template,\
+    $(ASM_CNV_PATH)/asm_com_offsets.asm, \
+    $(LIBVPX_PATH)/vp8/common/asm_com_offsets.c))
+
+ifeq ($(CONFIG_VP8_DECODER), yes)
+  $(eval $(call asm_offsets_template,\
+    $(ASM_CNV_PATH)/asm_dec_offsets.asm, \
+    $(LIBVPX_PATH)/vp8/decoder/asm_dec_offsets.c))
+endif
+
+ifeq ($(CONFIG_VP8_ENCODER), yes)
+  $(eval $(call asm_offsets_template,\
+    $(ASM_CNV_PATH)/asm_enc_offsets.asm, \
+    $(LIBVPX_PATH)/vp8/encoder/asm_enc_offsets.c))
+endif
+
+$(call import-module,cpufeatures)

build/make/ads2gas.pl

@@ -129,11 +129,14 @@ while (<STDIN>)
     # ARM code
     s/\sARM/.arm/g;
 
+    # eabi_attributes numerical equivalents can be found in the
+    # "ARM IHI 0045C" document.
+
     # REQUIRE8 Stack is required to be 8-byte aligned
-    s/\sREQUIRE8/.eabi_attribute Tag_ABI_align_needed, 1/g;
+    s/\sREQUIRE8/.eabi_attribute 24, 1 \@Tag_ABI_align_needed/g;
 
     # PRESERVE8 Stack 8-byte align is preserved
-    s/\sPRESERVE8/.eabi_attribute Tag_ABI_align_preserved, 1/g;
+    s/\sPRESERVE8/.eabi_attribute 25, 1 \@Tag_ABI_align_preserved/g;
 
     # Use PROC and ENDP to give the symbols a .size directive.
     # This makes them show up properly in debugging tools like gdb and valgrind.

build/make/ads2gas_apple.pl

@@ -30,6 +30,8 @@ my @mapping_list = ("\$0", "\$1", "\$2", "\$3", "\$4", "\$5", "\$6", "\$7", "\$8
 
 my @incoming_array;
 
+my @imported_functions;
+
 # Perl trim function to remove whitespace from the start and end of the string
 sub trim($)
 {
@@ -132,7 +134,18 @@ while (<STDIN>)
     # Make function visible to linker, and make additional symbol with
     # prepended underscore
     s/EXPORT\s+\|([\$\w]*)\|/.globl _$1\n\t.globl $1/;
-    s/IMPORT\s+\|([\$\w]*)\|/.globl $1/;
+
+    # Prepend imported functions with _
+    if (s/IMPORT\s+\|([\$\w]*)\|/.globl $1/)
+    {
+        $function = trim($1);
+        push(@imported_functions, $function);
+    }
+
+    foreach $function (@imported_functions)
+    {
+        s/$function/_$function/;
+    }
 
     # No vertical bars required; make additional symbol with prepended
     # underscore

build/make/configure.sh

@@ -391,6 +391,8 @@ LDFLAGS = ${LDFLAGS}
 ASFLAGS = ${ASFLAGS}
 extralibs = ${extralibs}
 AS_SFX    = ${AS_SFX:-.asm}
+EXE_SFX   = ${EXE_SFX}
+RTCD_OPTIONS = ${RTCD_OPTIONS}
 EOF
 
     if enabled rvct; then cat >> $1 << EOF
@@ -454,9 +456,25 @@ process_common_cmdline() {
         ;;
         --enable-?*|--disable-?*)
         eval `echo "$opt" | sed 's/--/action=/;s/-/ option=/;s/-/_/g'`
-        echo "${CMDLINE_SELECT} ${ARCH_EXT_LIST}" | grep "^ *$option\$" >/dev/null || die_unknown $opt
+        if echo "${ARCH_EXT_LIST}" | grep "^ *$option\$" >/dev/null; then
+            [ $action = "disable" ] && RTCD_OPTIONS="${RTCD_OPTIONS}${opt} "
+        elif [ $action = "disable" ] && ! disabled $option ; then
+          echo "${CMDLINE_SELECT}" | grep "^ *$option\$" >/dev/null ||
+            die_unknown $opt
+        elif [ $action = "enable" ] && ! enabled $option ; then
+          echo "${CMDLINE_SELECT}" | grep "^ *$option\$" >/dev/null ||
+            die_unknown $opt
+        fi
         $action $option
         ;;
+        --require-?*)
+        eval `echo "$opt" | sed 's/--/action=/;s/-/ option=/;s/-/_/g'`
+        if echo "${ARCH_EXT_LIST}" none | grep "^ *$option\$" >/dev/null; then
+            RTCD_OPTIONS="${RTCD_OPTIONS}${opt} "
+        else
+            die_unknown $opt
+        fi
+        ;;
         --force-enable-?*|--force-disable-?*)
         eval `echo "$opt" | sed 's/--force-/action=/;s/-/ option=/;s/-/_/g'`
         $action $option
@@ -477,7 +495,11 @@ process_common_cmdline() {
         --libdir=*)
         libdir="${optval}"
         ;;
-        --libc|--as|--prefix|--libdir)
+        --sdk-path=*)
+        [ -d "${optval}" ] || die "Not a directory: ${optval}"
+        sdk_path="${optval}"
+        ;;
+        --libc|--as|--prefix|--libdir|--sdk-path)
         die "Option ${opt} requires argument"
         ;;
         --help|-h) show_help
@@ -522,6 +544,7 @@ setup_gnu_toolchain() {
     STRIP=${STRIP:-${CROSS}strip}
     NM=${NM:-${CROSS}nm}
         AS_SFX=.s
+        EXE_SFX=
 }
 
 process_common_toolchain() {
@@ -561,6 +584,14 @@ process_common_toolchain() {
                 tgt_isa=x86_64
                 tgt_os=darwin10
                 ;;
+            *darwin11*)
+                tgt_isa=x86_64
+                tgt_os=darwin11
+                ;;
+            *darwin12*)
+                tgt_isa=x86_64
+                tgt_os=darwin12
+                ;;
             *mingw32*|*cygwin*)
                 [ -z "$tgt_isa" ] && tgt_isa=x86
                 tgt_os=win32
@@ -571,6 +602,9 @@ process_common_toolchain() {
             *solaris2.10)
                 tgt_os=solaris
                 ;;
+            *os2*)
+                tgt_os=os2
+                ;;
         esac
 
         if [ -n "$tgt_isa" ] && [ -n "$tgt_os" ]; then
@@ -599,8 +633,8 @@ process_common_toolchain() {
 
     # Enable the architecture family
     case ${tgt_isa} in
-        arm*|iwmmxt*) enable arm;;
-    mips*)        enable mips;;
+        arm*) enable arm;;
+        mips*) enable mips;;
     esac
 
     # PIC is probably what we want when building shared libs
@@ -608,35 +642,51 @@ process_common_toolchain() {
 
     # Handle darwin variants. Newer SDKs allow targeting older
     # platforms, so find the newest SDK available.
-    if [ -d "/Developer/SDKs/MacOSX10.4u.sdk" ]; then
-        osx_sdk_dir="/Developer/SDKs/MacOSX10.4u.sdk"
-    fi
-    if [ -d "/Developer/SDKs/MacOSX10.5.sdk" ]; then
-        osx_sdk_dir="/Developer/SDKs/MacOSX10.5.sdk"
-    fi
-    if [ -d "/Developer/SDKs/MacOSX10.6.sdk" ]; then
-        osx_sdk_dir="/Developer/SDKs/MacOSX10.6.sdk"
+    case ${toolchain} in
+        *-darwin*)
+            if [ -z "${DEVELOPER_DIR}" ]; then
+                DEVELOPER_DIR=`xcode-select -print-path 2> /dev/null`
+                [ $? -ne 0 ] && OSX_SKIP_DIR_CHECK=1
+            fi
+            if [ -z "${OSX_SKIP_DIR_CHECK}" ]; then
+                OSX_SDK_ROOTS="${DEVELOPER_DIR}/SDKs"
+                OSX_SDK_VERSIONS="MacOSX10.4u.sdk MacOSX10.5.sdk MacOSX10.6.sdk"
+                OSX_SDK_VERSIONS="${OSX_SDK_VERSIONS} MacOSX10.7.sdk"
+                for v in ${OSX_SDK_VERSIONS}; do
+                    if [ -d "${OSX_SDK_ROOTS}/${v}" ]; then
+                        osx_sdk_dir="${OSX_SDK_ROOTS}/${v}"
+                    fi
+                done
+            fi
+            ;;
+    esac
+
+    if [ -d "${osx_sdk_dir}" ]; then
+        add_cflags  "-isysroot ${osx_sdk_dir}"
+        add_ldflags "-isysroot ${osx_sdk_dir}"
     fi
 
     case ${toolchain} in
         *-darwin8-*)
-            add_cflags  "-isysroot ${osx_sdk_dir}"
             add_cflags  "-mmacosx-version-min=10.4"
-            add_ldflags "-isysroot ${osx_sdk_dir}"
             add_ldflags "-mmacosx-version-min=10.4"
             ;;
         *-darwin9-*)
-            add_cflags  "-isysroot ${osx_sdk_dir}"
             add_cflags  "-mmacosx-version-min=10.5"
-            add_ldflags "-isysroot ${osx_sdk_dir}"
             add_ldflags "-mmacosx-version-min=10.5"
             ;;
         *-darwin10-*)
-            add_cflags  "-isysroot ${osx_sdk_dir}"
             add_cflags  "-mmacosx-version-min=10.6"
-            add_ldflags "-isysroot ${osx_sdk_dir}"
             add_ldflags "-mmacosx-version-min=10.6"
             ;;
+        *-darwin11-*)
+            add_cflags  "-mmacosx-version-min=10.7"
+            add_ldflags "-mmacosx-version-min=10.7"
+            ;;
+        *-darwin12-*)
+            add_cflags  "-mmacosx-version-min=10.8"
+            add_ldflags "-mmacosx-version-min=10.8"
+            ;;
     esac
 
     # Handle Solaris variants. Solaris 10 needs -lposix4
@@ -652,40 +702,44 @@ process_common_toolchain() {
 
     # Process ARM architecture variants
     case ${toolchain} in
-    arm*|iwmmxt*)
-    # on arm, isa versions are supersets
-    enabled armv7a && soft_enable armv7 ### DEBUG
-    enabled armv7 && soft_enable armv6
-    enabled armv7 || enabled armv6 && soft_enable armv5te
-    enabled armv7 || enabled armv6 && soft_enable fast_unaligned
-    enabled iwmmxt2 && soft_enable iwmmxt
-    enabled iwmmxt && soft_enable armv5te
+    arm*)
+        # on arm, isa versions are supersets
+        case ${tgt_isa} in
+        armv7)
+            soft_enable neon
+            soft_enable media
+            soft_enable edsp
+            soft_enable fast_unaligned
+            ;;
+        armv6)
+            soft_enable media
+            soft_enable edsp
+            soft_enable fast_unaligned
+            ;;
+        armv5te)
+            soft_enable edsp
+            ;;
+        esac
 
-    asm_conversion_cmd="cat"
+        asm_conversion_cmd="cat"
 
         case ${tgt_cc} in
         gcc)
-        if enabled iwmmxt || enabled iwmmxt2
-            then
-                CROSS=${CROSS:-arm-iwmmxt-linux-gnueabi-}
-            elif enabled symbian; then
-                CROSS=${CROSS:-arm-none-symbianelf-}
-            else
-                CROSS=${CROSS:-arm-none-linux-gnueabi-}
-            fi
+            CROSS=${CROSS:-arm-none-linux-gnueabi-}
             link_with_cc=gcc
             setup_gnu_toolchain
             arch_int=${tgt_isa##armv}
             arch_int=${arch_int%%te}
             check_add_asflags --defsym ARCHITECTURE=${arch_int}
             tune_cflags="-mtune="
-        if enabled iwmmxt || enabled iwmmxt2
-            then
-                check_add_asflags -mcpu=${tgt_isa}
-            elif enabled armv7
-            then
-                check_add_cflags -march=armv7-a -mcpu=cortex-a8 -mfpu=neon -mfloat-abi=softfp  #-ftree-vectorize
-                check_add_asflags -mcpu=cortex-a8 -mfpu=neon -mfloat-abi=softfp  #-march=armv7-a
+            if [ ${tgt_isa} == "armv7" ]; then
+                if enabled neon
+                then
+                    check_add_cflags -mfpu=neon #-ftree-vectorize
+                    check_add_asflags -mfpu=neon
+                fi
+                check_add_cflags -march=armv7-a -mcpu=cortex-a8 -mfloat-abi=softfp
+                check_add_asflags -mcpu=cortex-a8 -mfloat-abi=softfp  #-march=armv7-a
             else
                 check_add_cflags -march=${tgt_isa}
                 check_add_asflags -march=${tgt_isa}
@@ -703,10 +757,14 @@ process_common_toolchain() {
             tune_cflags="--cpu="
             tune_asflags="--cpu="
             if [ -z "${tune_cpu}" ]; then
-            if enabled armv7
-                then
-                    check_add_cflags --cpu=Cortex-A8 --fpu=softvfp+vfpv3
-                    check_add_asflags --cpu=Cortex-A8 --fpu=softvfp+vfpv3
+                if [ ${tgt_isa} == "armv7" ]; then
+                    if enabled neon
+                    then
+                        check_add_cflags --fpu=softvfp+vfpv3
+                        check_add_asflags --fpu=softvfp+vfpv3
+                    fi
+                    check_add_cflags --cpu=Cortex-A8
+                    check_add_asflags --cpu=Cortex-A8
                 else
                     check_add_cflags --cpu=${tgt_isa##armv}
                     check_add_asflags --cpu=${tgt_isa##armv}
@@ -727,12 +785,51 @@ process_common_toolchain() {
             disable multithread
             disable os_support
             ;;
+
+        android*)
+            SDK_PATH=${sdk_path}
+            COMPILER_LOCATION=`find "${SDK_PATH}" \
+                               -name "arm-linux-androideabi-gcc*" -print -quit`
+            TOOLCHAIN_PATH=${COMPILER_LOCATION%/*}/arm-linux-androideabi-
+            CC=${TOOLCHAIN_PATH}gcc
+            AR=${TOOLCHAIN_PATH}ar
+            LD=${TOOLCHAIN_PATH}gcc
+            AS=${TOOLCHAIN_PATH}as
+            STRIP=${TOOLCHAIN_PATH}strip
+            NM=${TOOLCHAIN_PATH}nm
+
+            if [ -z "${alt_libc}" ]; then
+                alt_libc=`find "${SDK_PATH}" -name arch-arm -print | \
+                          awk '{n = split($0,a,"/"); \
+                                split(a[n-1],b,"-"); \
+                                print $0 " " b[2]}' | \
+                          sort -g -k 2 | \
+                          awk '{ print $1 }' | tail -1`
+            fi
+
+            add_cflags "--sysroot=${alt_libc}"
+            add_ldflags "--sysroot=${alt_libc}"
+
+            add_cflags "-I${SDK_PATH}/sources/android/cpufeatures/"
+
+            enable pic
+            soft_enable realtime_only
+            if [ ${tgt_isa} == "armv7" ]; then
+                enable runtime_cpu_detect
+            fi
+          ;;
+
         darwin*)
-            SDK_PATH=/Developer/Platforms/iPhoneOS.platform/Developer
+            if [ -z "${sdk_path}" ]; then
+                SDK_PATH=`xcode-select -print-path 2> /dev/null`
+                SDK_PATH=${SDK_PATH}/Platforms/iPhoneOS.platform/Developer
+            else
+                SDK_PATH=${sdk_path}
+            fi
             TOOLCHAIN_PATH=${SDK_PATH}/usr/bin
             CC=${TOOLCHAIN_PATH}/gcc
             AR=${TOOLCHAIN_PATH}/ar
-            LD=${TOOLCHAIN_PATH}/arm-apple-darwin10-gcc-4.2.1
+            LD=${TOOLCHAIN_PATH}/arm-apple-darwin10-llvm-gcc-4.2
             AS=${TOOLCHAIN_PATH}/as
             STRIP=${TOOLCHAIN_PATH}/strip
             NM=${TOOLCHAIN_PATH}/nm
@@ -746,13 +843,14 @@ process_common_toolchain() {
             add_cflags -arch ${tgt_isa}
             add_ldflags -arch_only ${tgt_isa}
 
-            add_cflags  "-isysroot ${SDK_PATH}/SDKs/iPhoneOS4.3.sdk"
+            if [ -z "${alt_libc}" ]; then
+                alt_libc=${SDK_PATH}/SDKs/iPhoneOS5.1.sdk
+            fi
 
-            # This should be overridable
-            alt_libc=${SDK_PATH}/SDKs/iPhoneOS4.3.sdk
+            add_cflags  "-isysroot ${alt_libc}"
 
             # Add the paths for the alternate libc
-            for d in usr/include usr/include/gcc/darwin/4.2/ usr/lib/gcc/arm-apple-darwin10/4.2.1/include/; do
+            for d in usr/include; do
                 try_dir="${alt_libc}/${d}"
                 [ -d "${try_dir}" ] && add_cflags -I"${try_dir}"
             done
@@ -789,19 +887,6 @@ process_common_toolchain() {
             fi
         ;;
 
-        symbian*)
-            enable symbian
-            # Add the paths for the alternate libc
-            for d in include/libc; do
-                try_dir="${alt_libc}/${d}"
-                [ -d "${try_dir}" ] && add_cflags -I"${try_dir}"
-            done
-            for d in release/armv5/urel; do
-                try_dir="${alt_libc}/${d}"
-                [ -d "${try_dir}" ] && add_ldflags -L"${try_dir}"
-            done
-            add_cflags -DIMPORT_C=
-
         esac
     ;;
     mips*)
@@ -856,6 +941,9 @@ process_common_toolchain() {
                 LD=${LD:-${CROSS}gcc}
                 CROSS=${CROSS:-g}
                 ;;
+            os2)
+                AS=${AS:-nasm}
+                ;;
         esac
 
         AS="${alt_as:-${AS:-auto}}"
@@ -926,6 +1014,11 @@ process_common_toolchain() {
                 # enabled icc && ! enabled pic && add_cflags -fno-pic -mdynamic-no-pic
                 enabled icc && ! enabled pic && add_cflags -fno-pic
             ;;
+            os2)
+                add_asflags -f aout
+                enabled debug && add_asflags -g
+                EXE_SFX=.exe
+            ;;
             *) log "Warning: Unknown os $tgt_os while setting up $AS flags"
             ;;
         esac
@@ -988,6 +1081,7 @@ EOF
     if enabled multithread; then
         case ${toolchain} in
             *-win*);;
+            *-android-gcc);;
             *) check_header pthread.h && add_extralibs -lpthread
         esac
     fi

build/make/gen_asm_deps.sh

@@ -42,7 +42,7 @@ done
 
 [ -n "$srcfile" ] || show_help
 sfx=${sfx:-asm}
-includes=$(egrep -i "include +\"?+[a-z0-9_/]+\.${sfx}" $srcfile |
+includes=$(LC_ALL=C egrep -i "include +\"?+[a-z0-9_/]+\.${sfx}" $srcfile |
            perl -p -e "s;.*?([a-z0-9_/]+.${sfx}).*;\1;")
 #" restore editor state
 for inc in ${includes}; do

build/make/rtcd.sh

@@ -0,0 +1,330 @@
+#!/bin/sh
+self=$0
+
+usage() {
+  cat <<EOF >&2
+Usage: $self [options] FILE
+
+Reads the Run Time CPU Detections definitions from FILE and generates a
+C header file on stdout.
+
+Options:
+  --arch=ARCH   Architecture to generate defs for (required)
+  --disable-EXT Disable support for EXT extensions
+  --require-EXT Require support for EXT extensions
+  --sym=SYMBOL  Unique symbol to use for RTCD initialization function
+  --config=FILE File with CONFIG_FOO=yes lines to parse
+EOF
+  exit 1
+}
+
+die() {
+  echo "$@" >&2
+  exit 1
+}
+
+die_argument_required() {
+  die "Option $opt requires argument"
+}
+
+for opt; do
+  optval="${opt#*=}"
+  case "$opt" in
+    --arch) die_argument_required;;
+    --arch=*) arch=${optval};;
+    --disable-*) eval "disable_${opt#--disable-}=true";;
+    --require-*) REQUIRES="${REQUIRES}${opt#--require-} ";;
+    --sym) die_argument_required;;
+    --sym=*) symbol=${optval};;
+    --config=*) config_file=${optval};;
+    -h|--help)
+      usage
+      ;;
+    -*)
+      die "Unrecognized option: ${opt%%=*}"
+      ;;
+    *)
+      defs_file="$defs_file $opt"
+      ;;
+  esac
+  shift
+done
+for f in $defs_file; do [ -f "$f" ] || usage; done
+[ -n "$arch" ] || usage
+
+# Import the configuration
+[ -f "$config_file" ] && eval $(grep CONFIG_ "$config_file")
+
+#
+# Routines for the RTCD DSL to call
+#
+prototype() {
+  local rtyp
+  case "$1" in
+    unsigned) rtyp="$1 "; shift;;
+  esac
+  rtyp="${rtyp}$1"
+  local fn="$2"
+  local args="$3"
+
+  eval "${2}_rtyp='$rtyp'"
+  eval "${2}_args='$3'"
+  ALL_FUNCS="$ALL_FUNCS $fn"
+  specialize $fn c
+}
+
+specialize() {
+  local fn="$1"
+  shift
+  for opt in "$@"; do
+    eval "${fn}_${opt}=${fn}_${opt}"
+  done
+}
+
+require() {
+  for fn in $ALL_FUNCS; do
+    for opt in "$@"; do
+      local ofn=$(eval "echo \$${fn}_${opt}")
+      [ -z "$ofn" ] && continue
+
+      # if we already have a default, then we can disable it, as we know
+      # we can do better.
+      local best=$(eval "echo \$${fn}_default")
+      local best_ofn=$(eval "echo \$${best}")
+      [ -n "$best" ] && [ "$best_ofn" != "$ofn" ] && eval "${best}_link=false"
+      eval "${fn}_default=${fn}_${opt}"
+      eval "${fn}_${opt}_link=true"
+    done
+  done
+}
+
+forward_decls() {
+  ALL_FORWARD_DECLS="$ALL_FORWARD_DECLS $1"
+}
+
+#
+# Include the user's directives
+#
+for f in $defs_file; do
+  . $f
+done
+
+#
+# Process the directives according to the command line
+#
+process_forward_decls() {
+  for fn in $ALL_FORWARD_DECLS; do
+    eval $fn
+  done
+}
+
+determine_indirection() {
+  [ "$CONFIG_RUNTIME_CPU_DETECT" = "yes" ] || require $ALL_ARCHS
+  for fn in $ALL_FUNCS; do
+    local n=""
+    local rtyp="$(eval "echo \$${fn}_rtyp")"
+    local args="$(eval "echo \"\$${fn}_args\"")"
+    local dfn="$(eval "echo \$${fn}_default")"
+    dfn=$(eval "echo \$${dfn}")
+    for opt in "$@"; do
+      local ofn=$(eval "echo \$${fn}_${opt}")
+      [ -z "$ofn" ] && continue
+      local link=$(eval "echo \$${fn}_${opt}_link")
+      [ "$link" = "false" ] && continue
+      n="${n}x"
+    done
+    if [ "$n" = "x" ]; then
+      eval "${fn}_indirect=false"
+    else
+      eval "${fn}_indirect=true"
+    fi
+  done
+}
+
+declare_function_pointers() {
+  for fn in $ALL_FUNCS; do
+    local rtyp="$(eval "echo \$${fn}_rtyp")"
+    local args="$(eval "echo \"\$${fn}_args\"")"
+    local dfn="$(eval "echo \$${fn}_default")"
+    dfn=$(eval "echo \$${dfn}")
+    for opt in "$@"; do
+      local ofn=$(eval "echo \$${fn}_${opt}")
+      [ -z "$ofn" ] && continue
+      echo "$rtyp ${ofn}($args);"
+    done
+    if [ "$(eval "echo \$${fn}_indirect")" = "false" ]; then
+      echo "#define ${fn} ${dfn}"
+    else
+      echo "RTCD_EXTERN $rtyp (*${fn})($args);"
+    fi
+    echo
+  done
+}
+
+set_function_pointers() {
+  for fn in $ALL_FUNCS; do
+    local n=""
+    local rtyp="$(eval "echo \$${fn}_rtyp")"
+    local args="$(eval "echo \"\$${fn}_args\"")"
+    local dfn="$(eval "echo \$${fn}_default")"
+    dfn=$(eval "echo \$${dfn}")
+    if $(eval "echo \$${fn}_indirect"); then
+      echo "    $fn = $dfn;"
+      for opt in "$@"; do
+        local ofn=$(eval "echo \$${fn}_${opt}")
+        [ -z "$ofn" ] && continue
+        [ "$ofn" = "$dfn" ] && continue;
+        local link=$(eval "echo \$${fn}_${opt}_link")
+        [ "$link" = "false" ] && continue
+        local cond="$(eval "echo \$have_${opt}")"
+        echo "    if (${cond}) $fn = $ofn;"
+      done
+    fi
+    echo
+  done
+}
+
+filter() {
+  local filtered
+  for opt in "$@"; do
+    [ -z $(eval "echo \$disable_${opt}") ] && filtered="$filtered $opt"
+  done
+  echo $filtered
+}
+
+#
+# Helper functions for generating the arch specific RTCD files
+#
+common_top() {
+  local outfile_basename=$(basename ${symbol:-rtcd.h})
+  local include_guard=$(echo $outfile_basename | tr '[a-z]' '[A-Z]' | tr -c '[A-Z]' _)
+  cat <<EOF
+#ifndef ${include_guard}
+#define ${include_guard}
+
+#ifdef RTCD_C
+#define RTCD_EXTERN
+#else
+#define RTCD_EXTERN extern
+#endif
+
+$(process_forward_decls)
+
+$(declare_function_pointers c $ALL_ARCHS)
+EOF
+}
+
+common_bottom() {
+  cat <<EOF
+#endif
+EOF
+}
+
+x86() {
+  determine_indirection c $ALL_ARCHS
+
+  # Assign the helper variable for each enabled extension
+  for opt in $ALL_ARCHS; do
+    local uc=$(echo $opt | tr '[a-z]' '[A-Z]')
+    eval "have_${opt}=\"flags & HAS_${uc}\""
+  done
+
+  cat <<EOF
+$(common_top)
+void ${symbol:-rtcd}(void);
+
+#ifdef RTCD_C
+#include "vpx_ports/x86.h"
+void ${symbol:-rtcd}(void)
+{
+    int flags = x86_simd_caps();
+
+    (void)flags;
+
+$(set_function_pointers c $ALL_ARCHS)
+}
+#endif
+$(common_bottom)
+EOF
+}
+
+arm() {
+  determine_indirection c $ALL_ARCHS
+
+  # Assign the helper variable for each enabled extension
+  for opt in $ALL_ARCHS; do
+    local uc=$(echo $opt | tr '[a-z]' '[A-Z]')
+    eval "have_${opt}=\"flags & HAS_${uc}\""
+  done
+
+  cat <<EOF
+$(common_top)
+#include "vpx_config.h"
+
+void ${symbol:-rtcd}(void);
+
+#ifdef RTCD_C
+#include "vpx_ports/arm.h"
+void ${symbol:-rtcd}(void)
+{
+    int flags = arm_cpu_caps();
+
+    (void)flags;
+
+$(set_function_pointers c $ALL_ARCHS)
+}
+#endif
+$(common_bottom)
+EOF
+}
+
+
+unoptimized() {
+  determine_indirection c
+  cat <<EOF
+$(common_top)
+#include "vpx_config.h"
+
+void ${symbol:-rtcd}(void);
+
+#ifdef RTCD_C
+void ${symbol:-rtcd}(void)
+{
+$(set_function_pointers c)
+}
+#endif
+$(common_bottom)
+EOF
+
+}
+#
+# Main Driver
+#
+require c
+case $arch in
+  x86)
+    ALL_ARCHS=$(filter mmx sse sse2 sse3 ssse3 sse4_1)
+    x86
+    ;;
+  x86_64)
+    ALL_ARCHS=$(filter mmx sse sse2 sse3 ssse3 sse4_1)
+    REQUIRES=${REQUIRES:-mmx sse sse2}
+    require $(filter $REQUIRES)
+    x86
+    ;;
+  armv5te)
+    ALL_ARCHS=$(filter edsp)
+    arm
+    ;;
+  armv6)
+    ALL_ARCHS=$(filter edsp media)
+    arm
+    ;;
+  armv7)
+    ALL_ARCHS=$(filter edsp media neon)
+    arm
+    ;;
+  *)
+    unoptimized
+    ;;
+esac

configure

@@ -25,6 +25,7 @@ Advanced options:
   ${toggle_unit_tests}            build unit tests
   --libc=PATH                     path to alternate libc
   --as={yasm|nasm|auto}           use specified assembler [auto, yasm preferred]
+  --sdk-path=PATH                 path to root of sdk (iOS, android builds only)
   ${toggle_fast_unaligned}        don't use unaligned accesses, even when
                                   supported by hardware [auto]
   ${toggle_codec_srcs}            in/exclude codec library source code
@@ -38,6 +39,7 @@ Advanced options:
   ${toggle_multithread}           multithreaded encoding and decoding
   ${toggle_spatial_resampling}    spatial sampling (scaling) support
   ${toggle_realtime_only}         enable this option while building for real-time encoding
+  ${toggle_onthefly_bitpacking}   enable on-the-fly bitpacking in real-time encoding
   ${toggle_error_concealment}     enable this option to get a decoder which is able to conceal losses
   ${toggle_runtime_cpu_detect}    runtime cpu detection
   ${toggle_shared}                shared library support
@@ -45,6 +47,7 @@ Advanced options:
   ${toggle_small}                 favor smaller size over speed
   ${toggle_postproc_visualizer}   macro block / block level visualizers
   ${toggle_multi_res_encoding}    enable multiple-resolution encoding
+  ${toggle_temporal_denoising}    enable temporal denoising and disable the spatial denoiser
 
 Codecs:
   Codecs can be selectively enabled or disabled individually, or by family:
@@ -80,19 +83,15 @@ EOF
 
 # all_platforms is a list of all supported target platforms. Maintain
 # alphabetically by architecture, generic-gnu last.
+all_platforms="${all_platforms} armv5te-android-gcc"
 all_platforms="${all_platforms} armv5te-linux-rvct"
 all_platforms="${all_platforms} armv5te-linux-gcc"
 all_platforms="${all_platforms} armv5te-none-rvct"
-all_platforms="${all_platforms} armv5te-symbian-gcc"
 all_platforms="${all_platforms} armv6-darwin-gcc"
 all_platforms="${all_platforms} armv6-linux-rvct"
 all_platforms="${all_platforms} armv6-linux-gcc"
 all_platforms="${all_platforms} armv6-none-rvct"
-all_platforms="${all_platforms} armv6-symbian-gcc"
-all_platforms="${all_platforms} iwmmxt-linux-rvct"
-all_platforms="${all_platforms} iwmmxt-linux-gcc"
-all_platforms="${all_platforms} iwmmxt2-linux-rvct"
-all_platforms="${all_platforms} iwmmxt2-linux-gcc"
+all_platforms="${all_platforms} armv7-android-gcc"   #neon Cortex-A8
 all_platforms="${all_platforms} armv7-darwin-gcc"    #neon Cortex-A8
 all_platforms="${all_platforms} armv7-linux-rvct"    #neon Cortex-A8
 all_platforms="${all_platforms} armv7-linux-gcc"     #neon Cortex-A8
@@ -110,8 +109,11 @@ all_platforms="${all_platforms} x86-darwin8-icc"
 all_platforms="${all_platforms} x86-darwin9-gcc"
 all_platforms="${all_platforms} x86-darwin9-icc"
 all_platforms="${all_platforms} x86-darwin10-gcc"
+all_platforms="${all_platforms} x86-darwin11-gcc"
+all_platforms="${all_platforms} x86-darwin12-gcc"
 all_platforms="${all_platforms} x86-linux-gcc"
 all_platforms="${all_platforms} x86-linux-icc"
+all_platforms="${all_platforms} x86-os2-gcc"
 all_platforms="${all_platforms} x86-solaris-gcc"
 all_platforms="${all_platforms} x86-win32-gcc"
 all_platforms="${all_platforms} x86-win32-vs7"
@@ -119,6 +121,8 @@ all_platforms="${all_platforms} x86-win32-vs8"
 all_platforms="${all_platforms} x86-win32-vs9"
 all_platforms="${all_platforms} x86_64-darwin9-gcc"
 all_platforms="${all_platforms} x86_64-darwin10-gcc"
+all_platforms="${all_platforms} x86_64-darwin11-gcc"
+all_platforms="${all_platforms} x86_64-darwin12-gcc"
 all_platforms="${all_platforms} x86_64-linux-gcc"
 all_platforms="${all_platforms} x86_64-linux-icc"
 all_platforms="${all_platforms} x86_64-solaris-gcc"
@@ -127,6 +131,9 @@ all_platforms="${all_platforms} x86_64-win64-vs8"
 all_platforms="${all_platforms} x86_64-win64-vs9"
 all_platforms="${all_platforms} universal-darwin8-gcc"
 all_platforms="${all_platforms} universal-darwin9-gcc"
+all_platforms="${all_platforms} universal-darwin10-gcc"
+all_platforms="${all_platforms} universal-darwin11-gcc"
+all_platforms="${all_platforms} universal-darwin12-gcc"
 all_platforms="${all_platforms} generic-gnu"
 
 # all_targets is a list of all targets that can be configured
@@ -165,6 +172,7 @@ enable md5
 enable spatial_resampling
 enable multithread
 enable os_support
+enable temporal_denoising
 
 [ -d ${source_path}/../include ] && enable alt_tree_layout
 for d in vp8; do
@@ -178,6 +186,8 @@ else
 # customer environment
 [ -f ${source_path}/../include/vpx/vp8cx.h ] && CODECS="${CODECS} vp8_encoder"
 [ -f ${source_path}/../include/vpx/vp8dx.h ] && CODECS="${CODECS} vp8_decoder"
+[ -f ${source_path}/../include/vpx/vp8cx.h ] || disable vp8_encoder
+[ -f ${source_path}/../include/vpx/vp8dx.h ] || disable vp8_decoder
 
 [ -f ${source_path}/../lib/*/*mt.lib ] && soft_enable static_msvcrt
 fi
@@ -194,11 +204,9 @@ ARCH_LIST="
     ppc64
 "
 ARCH_EXT_LIST="
-    armv5te
-    armv6
-    armv7
-    iwmmxt
-    iwmmxt2
+    edsp
+    media
+    neon
 
     mips32
 
@@ -256,6 +264,7 @@ CONFIG_LIST="
     static_msvcrt
     spatial_resampling
     realtime_only
+    onthefly_bitpacking
     error_concealment
     shared
     static
@@ -264,6 +273,7 @@ CONFIG_LIST="
     os_support
     unit_tests
     multi_res_encoding
+    temporal_denoising
 "
 CMDLINE_SELECT="
     extra_warnings
@@ -300,6 +310,7 @@ CMDLINE_SELECT="
     mem_tracker
     spatial_resampling
     realtime_only
+    onthefly_bitpacking
     error_concealment
     shared
     static
@@ -307,6 +318,7 @@ CMDLINE_SELECT="
     postproc_visualizer
     unit_tests
     multi_res_encoding
+    temporal_denoising
 "
 
 process_cmdline() {
@@ -487,11 +499,20 @@ process_toolchain() {
     case $toolchain in
         universal-darwin*)
             local darwin_ver=${tgt_os##darwin}
-            fat_bin_archs="$fat_bin_archs ppc32-${tgt_os}-gcc"
 
-            # Intel
-            fat_bin_archs="$fat_bin_archs x86-${tgt_os}-${tgt_cc}"
-            if [ $darwin_ver -gt 8 ]; then
+            # Snow Leopard (10.6/darwin10) dropped support for PPC
+            # Include PPC support for all prior versions
+            if [ $darwin_ver -lt 10 ]; then
+                fat_bin_archs="$fat_bin_archs ppc32-${tgt_os}-gcc"
+            fi
+
+            # Tiger (10.4/darwin8) brought support for x86
+            if [ $darwin_ver -ge 8 ]; then
+                fat_bin_archs="$fat_bin_archs x86-${tgt_os}-${tgt_cc}"
+            fi
+
+            # Leopard (10.5/darwin9) brought 64 bit support
+            if [ $darwin_ver -ge 9 ]; then
                 fat_bin_archs="$fat_bin_archs x86_64-${tgt_os}-${tgt_cc}"
             fi
             ;;
@@ -507,6 +528,10 @@ process_toolchain() {
         check_add_cflags -Wpointer-arith
         check_add_cflags -Wtype-limits
         check_add_cflags -Wcast-qual
+        check_add_cflags -Wimplicit-function-declaration
+        check_add_cflags -Wuninitialized
+        check_add_cflags -Wunused-variable
+        check_add_cflags -Wunused-but-set-variable        
         enabled extra_warnings || check_add_cflags -Wno-unused-function
     fi
 

docs.mk

@@ -21,9 +21,6 @@ CODEC_DOX :=    mainpage.dox \
 		usage_dx.dox \
 
 # Other doxy files sourced in Markdown
-TXT_DOX-$(CONFIG_VP8)          += vp8_api1_migration.dox
-vp8_api1_migration.dox.DESC     = VP8 API 1.x Migration
-
 TXT_DOX = $(call enabled,TXT_DOX)
 
 %.dox: %.txt

examples.mk

@@ -32,6 +32,7 @@ vpxenc.SRCS                 += args.c args.h y4minput.c y4minput.h
 vpxenc.SRCS                 += tools_common.c tools_common.h
 vpxenc.SRCS                 += vpx_ports/mem_ops.h
 vpxenc.SRCS                 += vpx_ports/mem_ops_aligned.h
+vpxenc.SRCS                 += vpx_ports/vpx_timer.h
 vpxenc.SRCS                 += libmkv/EbmlIDs.h
 vpxenc.SRCS                 += libmkv/EbmlWriter.c
 vpxenc.SRCS                 += libmkv/EbmlWriter.h
@@ -102,6 +103,7 @@ vp8_multi_resolution_encoder.SRCS  \
                          += third_party/libyuv/include/libyuv/basic_types.h  \
                             third_party/libyuv/include/libyuv/cpu_id.h  \
                             third_party/libyuv/include/libyuv/scale.h  \
+                            third_party/libyuv/source/row.h \
                             third_party/libyuv/source/scale.c  \
                             third_party/libyuv/source/cpu_id.c
 vp8_multi_resolution_encoder.GUID         = 04f8738e-63c8-423b-90fa-7c2703a374de
@@ -167,12 +169,12 @@ $(eval $(if $(filter universal%,$(TOOLCHAIN)),LIPO_OBJS,BUILD_OBJS):=yes)
 # Create build/install dependencies for all examples. The common case
 # is handled here. The MSVS case is handled below.
 NOT_MSVS = $(if $(CONFIG_MSVS),,yes)
-DIST-BINS-$(NOT_MSVS)      += $(addprefix bin/,$(ALL_EXAMPLES:.c=))
-INSTALL-BINS-$(NOT_MSVS)   += $(addprefix bin/,$(UTILS:.c=))
+DIST-BINS-$(NOT_MSVS)      += $(addprefix bin/,$(ALL_EXAMPLES:.c=$(EXE_SFX)))
+INSTALL-BINS-$(NOT_MSVS)   += $(addprefix bin/,$(UTILS:.c=$(EXE_SFX)))
 DIST-SRCS-yes              += $(ALL_SRCS)
 INSTALL-SRCS-yes           += $(UTIL_SRCS)
 OBJS-$(NOT_MSVS)           += $(if $(BUILD_OBJS),$(call objs,$(ALL_SRCS)))
-BINS-$(NOT_MSVS)           += $(addprefix $(BUILD_PFX),$(ALL_EXAMPLES:.c=))
+BINS-$(NOT_MSVS)           += $(addprefix $(BUILD_PFX),$(ALL_EXAMPLES:.c=$(EXE_SFX)))
 
 
 # Instantiate linker template for all examples.
@@ -182,7 +184,7 @@ $(foreach bin,$(BINS-yes),\
     $(if $(BUILD_OBJS),$(eval $(bin):\
         $(LIB_PATH)/lib$(CODEC_LIB)$(CODEC_LIB_SUF)))\
     $(if $(BUILD_OBJS),$(eval $(call linker_template,$(bin),\
-        $(call objs,$($(notdir $(bin)).SRCS)) \
+        $(call objs,$($(notdir $(bin:$(EXE_SFX)=)).SRCS)) \
         -l$(CODEC_LIB) $(addprefix -l,$(CODEC_EXTRA_LIBS))\
         )))\
     $(if $(LIPO_OBJS),$(eval $(call lipo_bin_template,$(bin))))\

examples/postproc.txt

@@ -58,7 +58,7 @@ if(frame_cnt%30 == 1) {
     if(vpx_codec_control(&codec, VP8_SET_POSTPROC, &pp))
         die_codec(&codec, "Failed to turn off postproc");
 } else if(frame_cnt%30 == 16) {
-    vp8_postproc_cfg_t  pp = {VP8_DEBLOCK | VP8_DEMACROBLOCK, 4, 0};
+    vp8_postproc_cfg_t  pp = {VP8_DEBLOCK | VP8_DEMACROBLOCK | VP8_MFQE, 4, 0};
 
     if(vpx_codec_control(&codec, VP8_SET_POSTPROC, &pp))
         die_codec(&codec, "Failed to turn on postproc");

libs.mk

@@ -17,6 +17,7 @@ else
   ASM:=.asm
 endif
 
+CODEC_SRCS-yes += CHANGELOG
 CODEC_SRCS-yes += libs.mk
 
 include $(SRC_PATH_BARE)/vpx/vpx_codec.mk
@@ -34,9 +35,9 @@ ifeq ($(CONFIG_VP8_ENCODER),yes)
   include $(SRC_PATH_BARE)/$(VP8_PREFIX)vp8cx.mk
   CODEC_SRCS-yes += $(addprefix $(VP8_PREFIX),$(call enabled,VP8_CX_SRCS))
   CODEC_EXPORTS-yes += $(addprefix $(VP8_PREFIX),$(VP8_CX_EXPORTS))
-  CODEC_SRCS-yes += $(VP8_PREFIX)vp8cx.mk vpx/vp8.h vpx/vp8cx.h vpx/vp8e.h
+  CODEC_SRCS-yes += $(VP8_PREFIX)vp8cx.mk vpx/vp8.h vpx/vp8cx.h
   CODEC_SRCS-$(ARCH_ARM) += $(VP8_PREFIX)vp8cx_arm.mk
-  INSTALL-LIBS-yes += include/vpx/vp8.h include/vpx/vp8e.h include/vpx/vp8cx.h
+  INSTALL-LIBS-yes += include/vpx/vp8.h include/vpx/vp8cx.h
   INSTALL_MAPS += include/vpx/% $(SRC_PATH_BARE)/$(VP8_PREFIX)/%
   CODEC_DOC_SRCS += vpx/vp8.h vpx/vp8cx.h
   CODEC_DOC_SECTIONS += vp8 vp8_encoder
@@ -48,7 +49,6 @@ ifeq ($(CONFIG_VP8_DECODER),yes)
   CODEC_SRCS-yes += $(addprefix $(VP8_PREFIX),$(call enabled,VP8_DX_SRCS))
   CODEC_EXPORTS-yes += $(addprefix $(VP8_PREFIX),$(VP8_DX_EXPORTS))
   CODEC_SRCS-yes += $(VP8_PREFIX)vp8dx.mk vpx/vp8.h vpx/vp8dx.h
-  CODEC_SRCS-$(ARCH_ARM) += $(VP8_PREFIX)vp8dx_arm.mk
   INSTALL-LIBS-yes += include/vpx/vp8.h include/vpx/vp8dx.h
   INSTALL_MAPS += include/vpx/% $(SRC_PATH_BARE)/$(VP8_PREFIX)/%
   CODEC_DOC_SRCS += vpx/vp8.h vpx/vp8dx.h
@@ -90,6 +90,7 @@ endif
 $(eval $(if $(filter universal%,$(TOOLCHAIN)),LIPO_LIBVPX,BUILD_LIBVPX):=yes)
 
 CODEC_SRCS-$(BUILD_LIBVPX) += build/make/version.sh
+CODEC_SRCS-$(BUILD_LIBVPX) += build/make/rtcd.sh
 CODEC_SRCS-$(BUILD_LIBVPX) += vpx/vpx_integer.h
 CODEC_SRCS-$(BUILD_LIBVPX) += vpx_ports/asm_offsets.h
 CODEC_SRCS-$(BUILD_LIBVPX) += vpx_ports/vpx_timer.h
@@ -114,7 +115,6 @@ INSTALL-LIBS-yes += include/vpx/vpx_integer.h
 INSTALL-LIBS-yes += include/vpx/vpx_codec_impl_top.h
 INSTALL-LIBS-yes += include/vpx/vpx_codec_impl_bottom.h
 INSTALL-LIBS-$(CONFIG_DECODERS) += include/vpx/vpx_decoder.h
-INSTALL-LIBS-$(CONFIG_DECODERS) += include/vpx/vpx_decoder_compat.h
 INSTALL-LIBS-$(CONFIG_ENCODERS) += include/vpx/vpx_encoder.h
 ifeq ($(CONFIG_EXTERNAL_BUILD),yes)
 ifeq ($(CONFIG_MSVS),yes)
@@ -183,6 +183,7 @@ vpx.vcproj: $(CODEC_SRCS) vpx.def
 PROJECTS-$(BUILD_LIBVPX) += vpx.vcproj
 
 vpx.vcproj: vpx_config.asm
+vpx.vcproj: vpx_rtcd.h
 
 endif
 else
@@ -232,7 +233,7 @@ vpx.pc: config.mk libs.mk
 	$(qexec)echo '# pkg-config file from libvpx $(VERSION_STRING)' > $@
 	$(qexec)echo 'prefix=$(PREFIX)' >> $@
 	$(qexec)echo 'exec_prefix=$${prefix}' >> $@
-	$(qexec)echo 'libdir=$${prefix}/lib' >> $@
+	$(qexec)echo 'libdir=$${prefix}/$(LIBSUBDIR)' >> $@
 	$(qexec)echo 'includedir=$${prefix}/include' >> $@
 	$(qexec)echo '' >> $@
 	$(qexec)echo 'Name: vpx' >> $@
@@ -279,19 +280,21 @@ $(filter %$(ASM).o,$(OBJS-yes)): $(BUILD_PFX)vpx_config.asm
 # Calculate platform- and compiler-specific offsets for hand coded assembly
 #
 
+OFFSET_PATTERN:='^[a-zA-Z0-9_]* EQU'
+
 ifeq ($(filter icc gcc,$(TGT_CC)), $(TGT_CC))
     $(BUILD_PFX)asm_com_offsets.asm: $(BUILD_PFX)$(VP8_PREFIX)common/asm_com_offsets.c.S
-	grep -w EQU $< | tr -d '$$\#' $(ADS2GAS) > $@
+	LC_ALL=C grep $(OFFSET_PATTERN) $< | tr -d '$$\#' $(ADS2GAS) > $@
     $(BUILD_PFX)$(VP8_PREFIX)common/asm_com_offsets.c.S: $(VP8_PREFIX)common/asm_com_offsets.c
     CLEAN-OBJS += $(BUILD_PFX)asm_com_offsets.asm $(BUILD_PFX)$(VP8_PREFIX)common/asm_com_offsets.c.S
 
     $(BUILD_PFX)asm_enc_offsets.asm: $(BUILD_PFX)$(VP8_PREFIX)encoder/asm_enc_offsets.c.S
-	grep -w EQU $< | tr -d '$$\#' $(ADS2GAS) > $@
+	LC_ALL=C grep $(OFFSET_PATTERN) $< | tr -d '$$\#' $(ADS2GAS) > $@
     $(BUILD_PFX)$(VP8_PREFIX)encoder/asm_enc_offsets.c.S: $(VP8_PREFIX)encoder/asm_enc_offsets.c
     CLEAN-OBJS += $(BUILD_PFX)asm_enc_offsets.asm $(BUILD_PFX)$(VP8_PREFIX)encoder/asm_enc_offsets.c.S
 
     $(BUILD_PFX)asm_dec_offsets.asm: $(BUILD_PFX)$(VP8_PREFIX)decoder/asm_dec_offsets.c.S
-	grep -w EQU $< | tr -d '$$\#' $(ADS2GAS) > $@
+	LC_ALL=C grep $(OFFSET_PATTERN) $< | tr -d '$$\#' $(ADS2GAS) > $@
     $(BUILD_PFX)$(VP8_PREFIX)decoder/asm_dec_offsets.c.S: $(VP8_PREFIX)decoder/asm_dec_offsets.c
     CLEAN-OBJS += $(BUILD_PFX)asm_dec_offsets.asm $(BUILD_PFX)$(VP8_PREFIX)decoder/asm_dec_offsets.c.S
 else
@@ -322,6 +325,18 @@ endif
 $(shell $(SRC_PATH_BARE)/build/make/version.sh "$(SRC_PATH_BARE)" $(BUILD_PFX)vpx_version.h)
 CLEAN-OBJS += $(BUILD_PFX)vpx_version.h
 
+#
+# Rule to generate runtime cpu detection files
+#
+$(OBJS-yes:.o=.d): $(BUILD_PFX)vpx_rtcd.h
+$(BUILD_PFX)vpx_rtcd.h: $(SRC_PATH_BARE)/$(sort $(filter %rtcd_defs.sh,$(CODEC_SRCS)))
+	@echo "    [CREATE] $@"
+	$(qexec)$(SRC_PATH_BARE)/build/make/rtcd.sh --arch=$(TGT_ISA) \
+          --sym=vpx_rtcd \
+          --config=$(target)$(if $(FAT_ARCHS),,-$(TOOLCHAIN)).mk \
+          $(RTCD_OPTIONS) $^ > $@
+CLEAN-OBJS += $(BUILD_PFX)vpx_rtcd.h
+
 CODEC_DOC_SRCS += vpx/vpx_codec.h \
                   vpx/vpx_decoder.h \
                   vpx/vpx_encoder.h \

mainpage.dox

@@ -12,8 +12,12 @@
 
   This distribution of the WebM VP8 Codec SDK includes the following support:
 
-  \if vp8_encoder    - \ref vp8_encoder   \endif
-  \if vp8_decoder    - \ref vp8_decoder   \endif
+  \if vp8_encoder
+  - \ref vp8_encoder
+  \endif
+  \if vp8_decoder
+  - \ref vp8_decoder
+  \endif
 
 
   \section main_startpoints Starting Points
@@ -24,8 +28,12 @@
   - Read the \ref samples "sample code" for examples of how to interact with the
     codec.
   - \ref codec reference
-    \if encoder - \ref encoder reference \endif
-    \if decoder - \ref decoder reference \endif
+    \if encoder
+    - \ref encoder reference
+    \endif
+    \if decoder
+    - \ref decoder reference
+    \endif
 
   \section main_support Support Options & FAQ
   The WebM project is an open source project supported by its community. For

tools/ftfy.sh

@@ -0,0 +1,160 @@
+#!/bin/sh
+self="$0"
+dirname_self=$(dirname "$self")
+
+usage() {
+  cat <<EOF >&2
+Usage: $self [option]
+
+This script applies a whitespace transformation to the commit at HEAD. If no
+options are given, then the modified files are left in the working tree.
+
+Options:
+  -h, --help     Shows this message
+  -n, --dry-run  Shows a diff of the changes to be made.
+  --amend        Squashes the changes into the commit at HEAD
+                     This option will also reformat the commit message.
+  --commit       Creates a new commit containing only the whitespace changes
+  --msg-only     Reformat the commit message only, ignore the patch itself.
+
+EOF
+  rm -f ${CLEAN_FILES}
+  exit 1
+}
+
+
+log() {
+  echo "${self##*/}: $@" >&2
+}
+
+
+vpx_style() {
+  astyle --style=bsd --min-conditional-indent=0 --break-blocks \
+         --pad-oper --pad-header --unpad-paren \
+         --align-pointer=name \
+         --indent-preprocessor --convert-tabs --indent-labels \
+         --suffix=none --quiet "$@"
+  sed -i 's/[[:space:]]\{1,\},/,/g' "$@"
+}
+
+
+apply() {
+  [ $INTERSECT_RESULT -ne 0 ] && patch -p1 < "$1"
+}
+
+
+commit() {
+  LAST_CHANGEID=$(git show | awk '/Change-Id:/{print $2}')
+  if [ -z "$LAST_CHANGEID" ]; then
+    log "HEAD doesn't have a Change-Id, unable to generate a new commit"
+    exit 1
+  fi
+
+  # Build a deterministic Change-Id from the parent's
+  NEW_CHANGEID=${LAST_CHANGEID}-styled
+  NEW_CHANGEID=I$(echo $NEW_CHANGEID | git hash-object --stdin)
+
+  # Commit, preserving authorship from the parent commit.
+  git commit -a -C HEAD > /dev/null
+  git commit --amend -F- << EOF
+Cosmetic: Fix whitespace in change ${LAST_CHANGEID:0:9}
+
+Change-Id: ${NEW_CHANGEID}
+EOF
+}
+
+
+show_commit_msg_diff() {
+  if [ $DIFF_MSG_RESULT -ne 0 ]; then
+    log "Modified commit message:"
+    diff -u "$ORIG_COMMIT_MSG" "$NEW_COMMIT_MSG" | tail -n +3
+  fi
+}
+
+
+amend() {
+  show_commit_msg_diff
+  if [ $DIFF_MSG_RESULT -ne 0 ] || [ $INTERSECT_RESULT -ne 0 ]; then
+    git commit -a --amend -F "$NEW_COMMIT_MSG"
+  fi
+}
+
+
+diff_msg() {
+  git log -1 --format=%B > "$ORIG_COMMIT_MSG"
+  "${dirname_self}"/wrap-commit-msg.py \
+      < "$ORIG_COMMIT_MSG" > "$NEW_COMMIT_MSG"
+  cmp -s "$ORIG_COMMIT_MSG" "$NEW_COMMIT_MSG"
+  DIFF_MSG_RESULT=$?
+}
+
+
+# Temporary files
+ORIG_DIFF=orig.diff.$$
+MODIFIED_DIFF=modified.diff.$$
+FINAL_DIFF=final.diff.$$
+ORIG_COMMIT_MSG=orig.commit-msg.$$
+NEW_COMMIT_MSG=new.commit-msg.$$
+CLEAN_FILES="${ORIG_DIFF} ${MODIFIED_DIFF} ${FINAL_DIFF}"
+CLEAN_FILES="${CLEAN_FILES} ${ORIG_COMMIT_MSG} ${NEW_COMMIT_MSG}"
+
+# Preconditions
+[ $# -lt 2 ] || usage
+
+# Check that astyle supports pad-header and align-pointer=name
+if ! astyle --pad-header --align-pointer=name < /dev/null; then
+  log "Install astyle v1.24 or newer"
+  exit 1
+fi
+
+if ! git diff --quiet HEAD; then
+  log "Working tree is dirty, commit your changes first"
+  exit 1
+fi
+
+# Need to be in the root
+cd "$(git rev-parse --show-toplevel)"
+
+# Collect the original diff
+git show > "${ORIG_DIFF}"
+
+# Apply the style guide on new and modified files and collect its diff
+for f in $(git diff HEAD^ --name-only -M90 --diff-filter=AM \
+           | grep '\.[ch]$'); do
+  case "$f" in
+    third_party/*) continue;;
+    nestegg/*) continue;;
+  esac
+  vpx_style "$f"
+done
+git diff --no-color --no-ext-diff > "${MODIFIED_DIFF}"
+
+# Intersect the two diffs
+"${dirname_self}"/intersect-diffs.py \
+    "${ORIG_DIFF}" "${MODIFIED_DIFF}" > "${FINAL_DIFF}"
+INTERSECT_RESULT=$?
+git reset --hard >/dev/null
+
+# Fixup the commit message
+diff_msg
+
+# Handle options
+if [ -n "$1" ]; then
+  case "$1" in
+    -h|--help) usage;;
+    -n|--dry-run) cat "${FINAL_DIFF}"; show_commit_msg_diff;;
+    --commit) apply "${FINAL_DIFF}"; commit;;
+    --amend) apply "${FINAL_DIFF}"; amend;;
+    --msg-only) amend;;
+    *) usage;;
+  esac
+else
+  apply "${FINAL_DIFF}"
+  if ! git diff --quiet; then
+    log "Formatting changes applied, verify and commit."
+    log "See also: http://www.webmproject.org/code/contribute/conventions/"
+    git diff --stat
+  fi
+fi
+
+rm -f ${CLEAN_FILES}

tools/intersect-diffs.py

@@ -0,0 +1,188 @@
+#!/usr/bin/env python
+##  Copyright (c) 2012 The WebM project authors. All Rights Reserved.
+##
+##  Use of this source code is governed by a BSD-style license
+##  that can be found in the LICENSE file in the root of the source
+##  tree. An additional intellectual property rights grant can be found
+##  in the file PATENTS.  All contributing project authors may
+##  be found in the AUTHORS file in the root of the source tree.
+##
+"""Calculates the "intersection" of two unified diffs.
+
+Given two diffs, A and B, it finds all hunks in B that had non-context lines
+in A and prints them to stdout. This is useful to determine the hunks in B that
+are relevant to A. The resulting file can be applied with patch(1) on top of A.
+"""
+
+__author__ = "jkoleszar@google.com"
+
+import re
+import sys
+
+
+class DiffLines(object):
+    """A container for one half of a diff."""
+
+    def __init__(self, filename, offset, length):
+        self.filename = filename
+        self.offset = offset
+        self.length = length
+        self.lines = []
+        self.delta_line_nums = []
+
+    def Append(self, line):
+        l = len(self.lines)
+        if line[0] != " ":
+            self.delta_line_nums.append(self.offset + l)
+        self.lines.append(line[1:])
+        assert l+1 <= self.length
+
+    def Complete(self):
+        return len(self.lines) == self.length
+
+    def __contains__(self, item):
+        return item >= self.offset and item <= self.offset + self.length - 1
+
+
+class DiffHunk(object):
+    """A container for one diff hunk, consisting of two DiffLines."""
+
+    def __init__(self, header, file_a, file_b, start_a, len_a, start_b, len_b):
+        self.header = header
+        self.left = DiffLines(file_a, start_a, len_a)
+        self.right = DiffLines(file_b, start_b, len_b)
+        self.lines = []
+
+    def Append(self, line):
+        """Adds a line to the DiffHunk and its DiffLines children."""
+        if line[0] == "-":
+            self.left.Append(line)
+        elif line[0] == "+":
+            self.right.Append(line)
+        elif line[0] == " ":
+            self.left.Append(line)
+            self.right.Append(line)
+        else:
+            assert False, ("Unrecognized character at start of diff line "
+                           "%r" % line[0])
+        self.lines.append(line)
+
+    def Complete(self):
+        return self.left.Complete() and self.right.Complete()
+
+    def __repr__(self):
+        return "DiffHunk(%s, %s, len %d)" % (
+            self.left.filename, self.right.filename,
+            max(self.left.length, self.right.length))
+
+
+def ParseDiffHunks(stream):
+    """Walk a file-like object, yielding DiffHunks as they're parsed."""
+
+    file_regex = re.compile(r"(\+\+\+|---) (\S+)")
+    range_regex = re.compile(r"@@ -(\d+)(,(\d+))? \+(\d+)(,(\d+))?")
+    hunk = None
+    while True:
+        line = stream.readline()
+        if not line:
+            break
+
+        if hunk is None:
+            # Parse file names
+            diff_file = file_regex.match(line)
+            if diff_file:
+              if line.startswith("---"):
+                  a_line = line
+                  a = diff_file.group(2)
+                  continue
+              if line.startswith("+++"):
+                  b_line = line
+                  b = diff_file.group(2)
+                  continue
+
+            # Parse offset/lengths
+            diffrange = range_regex.match(line)
+            if diffrange:
+                if diffrange.group(2):
+                    start_a = int(diffrange.group(1))
+                    len_a = int(diffrange.group(3))
+                else:
+                    start_a = 1
+                    len_a = int(diffrange.group(1))
+
+                if diffrange.group(5):
+                    start_b = int(diffrange.group(4))
+                    len_b = int(diffrange.group(6))
+                else:
+                    start_b = 1
+                    len_b = int(diffrange.group(4))
+
+                header = [a_line, b_line, line]
+                hunk = DiffHunk(header, a, b, start_a, len_a, start_b, len_b)
+        else:
+            # Add the current line to the hunk
+            hunk.Append(line)
+
+            # See if the whole hunk has been parsed. If so, yield it and prepare
+            # for the next hunk.
+            if hunk.Complete():
+                yield hunk
+                hunk = None
+
+    # Partial hunks are a parse error
+    assert hunk is None
+
+
+def FormatDiffHunks(hunks):
+    """Re-serialize a list of DiffHunks."""
+    r = []
+    last_header = None
+    for hunk in hunks:
+        this_header = hunk.header[0:2]
+        if last_header != this_header:
+            r.extend(hunk.header)
+            last_header = this_header
+        else:
+            r.extend(hunk.header[2])
+        r.extend(hunk.lines)
+        r.append("\n")
+    return "".join(r)
+
+
+def ZipHunks(rhs_hunks, lhs_hunks):
+    """Join two hunk lists on filename."""
+    for rhs_hunk in rhs_hunks:
+        rhs_file = rhs_hunk.right.filename.split("/")[1:]
+
+        for lhs_hunk in lhs_hunks:
+            lhs_file = lhs_hunk.left.filename.split("/")[1:]
+            if lhs_file != rhs_file:
+                continue
+            yield (rhs_hunk, lhs_hunk)
+
+
+def main():
+    old_hunks = [x for x in ParseDiffHunks(open(sys.argv[1], "r"))]
+    new_hunks = [x for x in ParseDiffHunks(open(sys.argv[2], "r"))]
+    out_hunks = []
+
+    # Join the right hand side of the older diff with the left hand side of the
+    # newer diff.
+    for old_hunk, new_hunk in ZipHunks(old_hunks, new_hunks):
+        if new_hunk in out_hunks:
+            continue
+        old_lines = old_hunk.right
+        new_lines = new_hunk.left
+
+        # Determine if this hunk overlaps any non-context line from the other
+        for i in old_lines.delta_line_nums:
+            if i in new_lines:
+                out_hunks.append(new_hunk)
+                break
+
+    if out_hunks:
+        print FormatDiffHunks(out_hunks)
+        sys.exit(1)
+
+if __name__ == "__main__":
+    main()

tools/wrap-commit-msg.py

@@ -0,0 +1,70 @@
+#!/usr/bin/env python
+##  Copyright (c) 2012 The WebM project authors. All Rights Reserved.
+##
+##  Use of this source code is governed by a BSD-style license
+##  that can be found in the LICENSE file in the root of the source
+##  tree. An additional intellectual property rights grant can be found
+##  in the file PATENTS.  All contributing project authors may
+##  be found in the AUTHORS file in the root of the source tree.
+##
+"""Wraps paragraphs of text, preserving manual formatting
+
+This is like fold(1), but has the special convention of not modifying lines
+that start with whitespace. This allows you to intersperse blocks with
+special formatting, like code blocks, with written prose. The prose will
+be wordwrapped, and the manual formatting will be preserved.
+
+ * This won't handle the case of a bulleted (or ordered) list specially, so
+   manual wrapping must be done.
+
+Occasionally it's useful to put something with explicit formatting that
+doesn't look at all like a block of text inline.
+
+  indicator = has_leading_whitespace(line);
+  if (indicator)
+    preserve_formatting(line);
+
+The intent is that this docstring would make it through the transform
+and still be legible and presented as it is in the source. If additional
+cases are handled, update this doc to describe the effect.
+"""
+
+__author__ = "jkoleszar@google.com"
+import textwrap
+import sys
+
+def wrap(text):
+    if text:
+        return textwrap.fill(text, break_long_words=False) + '\n'
+    return ""
+
+
+def main(fileobj):
+    text = ""
+    output = ""
+    while True:
+        line = fileobj.readline()
+        if not line:
+            break
+
+        if line.lstrip() == line:
+            text += line
+        else:
+            output += wrap(text)
+            text=""
+            output += line
+    output += wrap(text)
+
+    # Replace the file or write to stdout.
+    if fileobj == sys.stdin:
+        fileobj = sys.stdout
+    else:
+        fileobj.seek(0)
+        fileobj.truncate(0)
+    fileobj.write(output)
+
+if __name__ == "__main__":
+    if len(sys.argv) > 1:
+        main(open(sys.argv[1], "r+"))
+    else:
+        main(sys.stdin)

tools_common.c

@@ -9,15 +9,21 @@
  */
 #include <stdio.h>
 #include "tools_common.h"
-#ifdef _WIN32
+#if defined(_WIN32) || defined(__OS2__)
 #include <io.h>
 #include <fcntl.h>
+
+#ifdef __OS2__
+#define _setmode    setmode
+#define _fileno     fileno
+#define _O_BINARY   O_BINARY
+#endif
 #endif
 
 FILE* set_binary_mode(FILE *stream)
 {
     (void)stream;
-#ifdef _WIN32
+#if defined(_WIN32) || defined(__OS2__)
     _setmode(_fileno(stream), _O_BINARY);
 #endif
     return stream;

usage.dox

@@ -1,6 +1,6 @@
 /*!\page usage Usage
 
-    The vpx Multi-Format codec SDK provides a unified interface amongst its
+    The vpx multi-format codec SDK provides a unified interface amongst its
     supported codecs. This abstraction allows applications using this SDK to
     easily support multiple video formats with minimal code duplication or
     "special casing." This section describes the interface common to all codecs.
@@ -14,8 +14,12 @@
 
     Fore more information on decoder and encoder specific usage, see the
     following pages:
-    \if decoder - \subpage usage_decode \endif
-    \if decoder - \subpage usage_encode \endif
+    \if decoder
+    - \subpage usage_decode
+    \endif
+    \if decoder
+    - \subpage usage_encode
+    \endif
 
     \section usage_types Important Data Types
     There are two important data structures to consider in this interface.

vp8/common/alloccommon.c

@@ -37,12 +37,15 @@ static void update_mode_info_border(MODE_INFO *mi, int rows, int cols)
 void vp8_de_alloc_frame_buffers(VP8_COMMON *oci)
 {
     int i;
-
     for (i = 0; i < NUM_YV12_BUFFERS; i++)
         vp8_yv12_de_alloc_frame_buffer(&oci->yv12_fb[i]);
 
     vp8_yv12_de_alloc_frame_buffer(&oci->temp_scale_frame);
+#if CONFIG_POSTPROC
     vp8_yv12_de_alloc_frame_buffer(&oci->post_proc_buffer);
+    if (oci->post_proc_buffer_int_used)
+        vp8_yv12_de_alloc_frame_buffer(&oci->post_proc_buffer_int);
+#endif
 
     vpx_free(oci->above_context);
     vpx_free(oci->mip);
@@ -95,12 +98,18 @@ int vp8_alloc_frame_buffers(VP8_COMMON *oci, int width, int height)
         return 1;
     }
 
+#if CONFIG_POSTPROC
     if (vp8_yv12_alloc_frame_buffer(&oci->post_proc_buffer, width, height, VP8BORDERINPIXELS) < 0)
     {
         vp8_de_alloc_frame_buffers(oci);
         return 1;
     }
 
+    oci->post_proc_buffer_int_used = 0;
+    vpx_memset(&oci->postproc_state, 0, sizeof(oci->postproc_state));
+    vpx_memset((&oci->post_proc_buffer)->buffer_alloc,128,(&oci->post_proc_buffer)->frame_size);
+#endif
+
     oci->mb_rows = height >> 4;
     oci->mb_cols = width >> 4;
     oci->MBs = oci->mb_rows * oci->mb_cols;
@@ -199,7 +208,7 @@ void vp8_create_common(VP8_COMMON *oci)
     oci->clr_type = REG_YUV;
     oci->clamp_type = RECON_CLAMP_REQUIRED;
 
-    /* Initialise reference frame sign bias structure to defaults */
+    /* Initialize reference frame sign bias structure to defaults */
     vpx_memset(oci->ref_frame_sign_bias, 0, sizeof(oci->ref_frame_sign_bias));
 
     /* Default disable buffer to buffer copying */
@@ -211,13 +220,3 @@ void vp8_remove_common(VP8_COMMON *oci)
 {
     vp8_de_alloc_frame_buffers(oci);
 }
-
-void vp8_initialize_common()
-{
-    vp8_coef_tree_initialize();
-
-    vp8_entropy_mode_init();
-
-    vp8_init_scan_order_mask();
-
-}

vp8/common/arm/arm_systemdependent.c

@@ -1,104 +0,0 @@
-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-
-#include "vpx_config.h"
-#include "vpx_ports/arm.h"
-#include "vp8/common/g_common.h"
-#include "vp8/common/pragmas.h"
-#include "vp8/common/subpixel.h"
-#include "vp8/common/loopfilter.h"
-#include "vp8/common/recon.h"
-#include "vp8/common/idct.h"
-#include "vp8/common/onyxc_int.h"
-
-void vp8_arch_arm_common_init(VP8_COMMON *ctx)
-{
-#if CONFIG_RUNTIME_CPU_DETECT
-    VP8_COMMON_RTCD *rtcd = &ctx->rtcd;
-    int flags = arm_cpu_caps();
-    rtcd->flags = flags;
-
-    /* Override default functions with fastest ones for this CPU. */
-#if HAVE_ARMV5TE
-    if (flags & HAS_EDSP)
-    {
-    }
-#endif
-
-#if HAVE_ARMV6
-    if (flags & HAS_MEDIA)
-    {
-        rtcd->subpix.sixtap16x16   = vp8_sixtap_predict16x16_armv6;
-        rtcd->subpix.sixtap8x8     = vp8_sixtap_predict8x8_armv6;
-        rtcd->subpix.sixtap8x4     = vp8_sixtap_predict8x4_armv6;
-        rtcd->subpix.sixtap4x4     = vp8_sixtap_predict_armv6;
-        rtcd->subpix.bilinear16x16 = vp8_bilinear_predict16x16_armv6;
-        rtcd->subpix.bilinear8x8   = vp8_bilinear_predict8x8_armv6;
-        rtcd->subpix.bilinear8x4   = vp8_bilinear_predict8x4_armv6;
-        rtcd->subpix.bilinear4x4   = vp8_bilinear_predict4x4_armv6;
-
-        rtcd->idct.idct16       = vp8_short_idct4x4llm_v6_dual;
-        rtcd->idct.iwalsh16     = vp8_short_inv_walsh4x4_v6;
-
-        rtcd->loopfilter.normal_mb_v = vp8_loop_filter_mbv_armv6;
-        rtcd->loopfilter.normal_b_v  = vp8_loop_filter_bv_armv6;
-        rtcd->loopfilter.normal_mb_h = vp8_loop_filter_mbh_armv6;
-        rtcd->loopfilter.normal_b_h  = vp8_loop_filter_bh_armv6;
-        rtcd->loopfilter.simple_mb_v =
-                vp8_loop_filter_simple_vertical_edge_armv6;
-        rtcd->loopfilter.simple_b_v  = vp8_loop_filter_bvs_armv6;
-        rtcd->loopfilter.simple_mb_h =
-                vp8_loop_filter_simple_horizontal_edge_armv6;
-        rtcd->loopfilter.simple_b_h  = vp8_loop_filter_bhs_armv6;
-
-        rtcd->recon.copy16x16   = vp8_copy_mem16x16_v6;
-        rtcd->recon.copy8x8     = vp8_copy_mem8x8_v6;
-        rtcd->recon.copy8x4     = vp8_copy_mem8x4_v6;
-        rtcd->recon.intra4x4_predict = vp8_intra4x4_predict_armv6;
-    }
-#endif
-
-#if HAVE_ARMV7
-    if (flags & HAS_NEON)
-    {
-        rtcd->subpix.sixtap16x16   = vp8_sixtap_predict16x16_neon;
-        rtcd->subpix.sixtap8x8     = vp8_sixtap_predict8x8_neon;
-        rtcd->subpix.sixtap8x4     = vp8_sixtap_predict8x4_neon;
-        rtcd->subpix.sixtap4x4     = vp8_sixtap_predict_neon;
-        rtcd->subpix.bilinear16x16 = vp8_bilinear_predict16x16_neon;
-        rtcd->subpix.bilinear8x8   = vp8_bilinear_predict8x8_neon;
-        rtcd->subpix.bilinear8x4   = vp8_bilinear_predict8x4_neon;
-        rtcd->subpix.bilinear4x4   = vp8_bilinear_predict4x4_neon;
-
-        rtcd->idct.idct16       = vp8_short_idct4x4llm_neon;
-        rtcd->idct.iwalsh16     = vp8_short_inv_walsh4x4_neon;
-
-        rtcd->loopfilter.normal_mb_v = vp8_loop_filter_mbv_neon;
-        rtcd->loopfilter.normal_b_v  = vp8_loop_filter_bv_neon;
-        rtcd->loopfilter.normal_mb_h = vp8_loop_filter_mbh_neon;
-        rtcd->loopfilter.normal_b_h  = vp8_loop_filter_bh_neon;
-        rtcd->loopfilter.simple_mb_v = vp8_loop_filter_mbvs_neon;
-        rtcd->loopfilter.simple_b_v  = vp8_loop_filter_bvs_neon;
-        rtcd->loopfilter.simple_mb_h = vp8_loop_filter_mbhs_neon;
-        rtcd->loopfilter.simple_b_h  = vp8_loop_filter_bhs_neon;
-
-        rtcd->recon.copy16x16   = vp8_copy_mem16x16_neon;
-        rtcd->recon.copy8x8     = vp8_copy_mem8x8_neon;
-        rtcd->recon.copy8x4     = vp8_copy_mem8x4_neon;
-        rtcd->recon.build_intra_predictors_mby =
-            vp8_build_intra_predictors_mby_neon;
-        rtcd->recon.build_intra_predictors_mby_s =
-            vp8_build_intra_predictors_mby_s_neon;
-    }
-#endif
-
-#endif
-}

vp8/common/arm/armv6/idct_blk_v6.c

@@ -9,8 +9,7 @@
  */
 
 #include "vpx_config.h"
-#include "vp8/common/idct.h"
-#include "vp8/decoder/dequantize.h"
+#include "vpx_rtcd.h"
 
 
 void vp8_dequant_idct_add_y_block_v6(short *q, short *dq,

vp8/common/arm/armv6/vp8_variance16x16_armv6.asm

@@ -144,7 +144,7 @@ loop
     ldr     r6, [sp, #40]       ; get address of sse
     mul     r0, r8, r8          ; sum * sum
     str     r11, [r6]           ; store sse
-    sub     r0, r11, r0, asr #8 ; return (sse - ((sum * sum) >> 8))
+    sub     r0, r11, r0, lsr #8 ; return (sse - ((sum * sum) >> 8))
 
     ldmfd   sp!, {r4-r12, pc}
 

vp8/common/arm/armv6/vp8_variance_halfpixvar16x16_h_armv6.asm

@@ -169,7 +169,7 @@ loop
     ldr     r6, [sp, #40]       ; get address of sse
     mul     r0, r8, r8          ; sum * sum
     str     r11, [r6]           ; store sse
-    sub     r0, r11, r0, asr #8 ; return (sse - ((sum * sum) >> 8))
+    sub     r0, r11, r0, lsr #8 ; return (sse - ((sum * sum) >> 8))
 
     ldmfd   sp!, {r4-r12, pc}
 

vp8/common/arm/armv6/vp8_variance_halfpixvar16x16_hv_armv6.asm

@@ -210,7 +210,7 @@ loop
     ldr     r6, [sp, #40]       ; get address of sse
     mul     r0, r8, r8          ; sum * sum
     str     r11, [r6]           ; store sse
-    sub     r0, r11, r0, asr #8 ; return (sse - ((sum * sum) >> 8))
+    sub     r0, r11, r0, lsr #8 ; return (sse - ((sum * sum) >> 8))
 
     ldmfd   sp!, {r4-r12, pc}
 

vp8/common/arm/armv6/vp8_variance_halfpixvar16x16_v_armv6.asm

@@ -171,7 +171,7 @@ loop
     ldr     r6, [sp, #40]       ; get address of sse
     mul     r0, r8, r8          ; sum * sum
     str     r11, [r6]           ; store sse
-    sub     r0, r11, r0, asr #8 ; return (sse - ((sum * sum) >> 8))
+    sub     r0, r11, r0, lsr #8 ; return (sse - ((sum * sum) >> 8))
 
     ldmfd   sp!, {r4-r12, pc}
 

vp8/common/arm/bilinearfilter_arm.c

@@ -8,10 +8,10 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
+#include "vpx_config.h"
+#include "vpx_rtcd.h"
 #include <math.h>
 #include "vp8/common/filter.h"
-#include "vp8/common/subpixel.h"
 #include "bilinearfilter_arm.h"
 
 void vp8_filter_block2d_bil_armv6

vp8/common/arm/dequantize_arm.c

@@ -10,36 +10,32 @@
 
 
 #include "vpx_config.h"
-#include "vp8/decoder/dequantize.h"
-#include "vp8/common/idct.h"
-#include "vpx_mem/vpx_mem.h"
+#include "vp8/common/blockd.h"
 
-#if HAVE_ARMV7
+#if HAVE_NEON
 extern void vp8_dequantize_b_loop_neon(short *Q, short *DQC, short *DQ);
 #endif
 
-#if HAVE_ARMV6
+#if HAVE_MEDIA
 extern void vp8_dequantize_b_loop_v6(short *Q, short *DQC, short *DQ);
 #endif
 
-#if HAVE_ARMV7
+#if HAVE_NEON
 
-void vp8_dequantize_b_neon(BLOCKD *d)
+void vp8_dequantize_b_neon(BLOCKD *d, short *DQC)
 {
     short *DQ  = d->dqcoeff;
     short *Q   = d->qcoeff;
-    short *DQC = d->dequant;
 
     vp8_dequantize_b_loop_neon(Q, DQC, DQ);
 }
 #endif
 
-#if HAVE_ARMV6
-void vp8_dequantize_b_v6(BLOCKD *d)
+#if HAVE_MEDIA
+void vp8_dequantize_b_v6(BLOCKD *d, short *DQC)
 {
     short *DQ  = d->dqcoeff;
     short *Q   = d->qcoeff;
-    short *DQC = d->dequant;
 
     vp8_dequantize_b_loop_v6(Q, DQC, DQ);
 }

vp8/common/arm/filter_arm.c

@@ -10,9 +10,9 @@
 
 
 #include "vpx_config.h"
+#include "vpx_rtcd.h"
 #include <math.h>
 #include "vp8/common/filter.h"
-#include "vp8/common/subpixel.h"
 #include "vpx_ports/mem.h"
 
 extern void vp8_filter_block2d_first_pass_armv6
@@ -86,8 +86,8 @@ extern void vp8_filter_block2d_second_pass_only_armv6
     const short *vp8_filter
 );
 
-#if HAVE_ARMV6
-void vp8_sixtap_predict_armv6
+#if HAVE_MEDIA
+void vp8_sixtap_predict4x4_armv6
 (
     unsigned char  *src_ptr,
     int  src_pixels_per_line,

vp8/common/arm/idct_arm.h

@@ -1,51 +0,0 @@
-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-
-#ifndef IDCT_ARM_H
-#define IDCT_ARM_H
-
-#if HAVE_ARMV6
-extern prototype_idct(vp8_short_idct4x4llm_v6_dual);
-extern prototype_idct_scalar_add(vp8_dc_only_idct_add_v6);
-extern prototype_second_order(vp8_short_inv_walsh4x4_1_v6);
-extern prototype_second_order(vp8_short_inv_walsh4x4_v6);
-
-#if !CONFIG_RUNTIME_CPU_DETECT
-#undef  vp8_idct_idct16
-#define vp8_idct_idct16 vp8_short_idct4x4llm_v6_dual
-
-#undef  vp8_idct_idct1_scalar_add
-#define vp8_idct_idct1_scalar_add vp8_dc_only_idct_add_v6
-
-#undef  vp8_idct_iwalsh16
-#define vp8_idct_iwalsh16 vp8_short_inv_walsh4x4_v6
-#endif
-#endif
-
-#if HAVE_ARMV7
-extern prototype_idct(vp8_short_idct4x4llm_neon);
-extern prototype_idct_scalar_add(vp8_dc_only_idct_add_neon);
-extern prototype_second_order(vp8_short_inv_walsh4x4_1_neon);
-extern prototype_second_order(vp8_short_inv_walsh4x4_neon);
-
-#if !CONFIG_RUNTIME_CPU_DETECT
-#undef  vp8_idct_idct16
-#define vp8_idct_idct16 vp8_short_idct4x4llm_neon
-
-#undef  vp8_idct_idct1_scalar_add
-#define vp8_idct_idct1_scalar_add vp8_dc_only_idct_add_neon
-
-#undef  vp8_idct_iwalsh16
-#define vp8_idct_iwalsh16 vp8_short_inv_walsh4x4_neon
-#endif
-#endif
-
-#endif

vp8/common/arm/loopfilter_arm.c

@@ -10,17 +10,22 @@
 
 
 #include "vpx_config.h"
+#include "vpx_rtcd.h"
 #include "vp8/common/loopfilter.h"
 #include "vp8/common/onyxc_int.h"
 
-#if HAVE_ARMV6
+#define prototype_loopfilter(sym) \
+    void sym(unsigned char *src, int pitch, const unsigned char *blimit,\
+             const unsigned char *limit, const unsigned char *thresh, int count)
+
+#if HAVE_MEDIA
 extern prototype_loopfilter(vp8_loop_filter_horizontal_edge_armv6);
 extern prototype_loopfilter(vp8_loop_filter_vertical_edge_armv6);
 extern prototype_loopfilter(vp8_mbloop_filter_horizontal_edge_armv6);
 extern prototype_loopfilter(vp8_mbloop_filter_vertical_edge_armv6);
 #endif
 
-#if HAVE_ARMV7
+#if HAVE_NEON
 typedef void loopfilter_y_neon(unsigned char *src, int pitch,
         unsigned char blimit, unsigned char limit, unsigned char thresh);
 typedef void loopfilter_uv_neon(unsigned char *u, int pitch,
@@ -38,8 +43,8 @@ extern loopfilter_uv_neon vp8_mbloop_filter_horizontal_edge_uv_neon;
 extern loopfilter_uv_neon vp8_mbloop_filter_vertical_edge_uv_neon;
 #endif
 
-#if HAVE_ARMV6
-/*ARMV6 loopfilter functions*/
+#if HAVE_MEDIA
+/* ARMV6/MEDIA loopfilter functions*/
 /* Horizontal MB filtering */
 void vp8_loop_filter_mbh_armv6(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
                                int y_stride, int uv_stride, loop_filter_info *lfi)
@@ -113,7 +118,7 @@ void vp8_loop_filter_bvs_armv6(unsigned char *y_ptr, int y_stride,
 }
 #endif
 
-#if HAVE_ARMV7
+#if HAVE_NEON
 /* NEON loopfilter functions */
 /* Horizontal MB filtering */
 void vp8_loop_filter_mbh_neon(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,

vp8/common/arm/loopfilter_arm.h

@@ -1,93 +0,0 @@
-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-
-#ifndef LOOPFILTER_ARM_H
-#define LOOPFILTER_ARM_H
-
-#include "vpx_config.h"
-
-#if HAVE_ARMV6
-extern prototype_loopfilter_block(vp8_loop_filter_mbv_armv6);
-extern prototype_loopfilter_block(vp8_loop_filter_bv_armv6);
-extern prototype_loopfilter_block(vp8_loop_filter_mbh_armv6);
-extern prototype_loopfilter_block(vp8_loop_filter_bh_armv6);
-extern prototype_simple_loopfilter(vp8_loop_filter_bvs_armv6);
-extern prototype_simple_loopfilter(vp8_loop_filter_bhs_armv6);
-extern prototype_simple_loopfilter(vp8_loop_filter_simple_horizontal_edge_armv6);
-extern prototype_simple_loopfilter(vp8_loop_filter_simple_vertical_edge_armv6);
-
-#if !CONFIG_RUNTIME_CPU_DETECT
-#undef  vp8_lf_normal_mb_v
-#define vp8_lf_normal_mb_v vp8_loop_filter_mbv_armv6
-
-#undef  vp8_lf_normal_b_v
-#define vp8_lf_normal_b_v vp8_loop_filter_bv_armv6
-
-#undef  vp8_lf_normal_mb_h
-#define vp8_lf_normal_mb_h vp8_loop_filter_mbh_armv6
-
-#undef  vp8_lf_normal_b_h
-#define vp8_lf_normal_b_h vp8_loop_filter_bh_armv6
-
-#undef  vp8_lf_simple_mb_v
-#define vp8_lf_simple_mb_v vp8_loop_filter_simple_vertical_edge_armv6
-
-#undef  vp8_lf_simple_b_v
-#define vp8_lf_simple_b_v vp8_loop_filter_bvs_armv6
-
-#undef  vp8_lf_simple_mb_h
-#define vp8_lf_simple_mb_h vp8_loop_filter_simple_horizontal_edge_armv6
-
-#undef  vp8_lf_simple_b_h
-#define vp8_lf_simple_b_h vp8_loop_filter_bhs_armv6
-#endif /* !CONFIG_RUNTIME_CPU_DETECT */
-
-#endif /* HAVE_ARMV6 */
-
-#if HAVE_ARMV7
-extern prototype_loopfilter_block(vp8_loop_filter_mbv_neon);
-extern prototype_loopfilter_block(vp8_loop_filter_bv_neon);
-extern prototype_loopfilter_block(vp8_loop_filter_mbh_neon);
-extern prototype_loopfilter_block(vp8_loop_filter_bh_neon);
-extern prototype_simple_loopfilter(vp8_loop_filter_mbvs_neon);
-extern prototype_simple_loopfilter(vp8_loop_filter_bvs_neon);
-extern prototype_simple_loopfilter(vp8_loop_filter_mbhs_neon);
-extern prototype_simple_loopfilter(vp8_loop_filter_bhs_neon);
-
-#if !CONFIG_RUNTIME_CPU_DETECT
-#undef  vp8_lf_normal_mb_v
-#define vp8_lf_normal_mb_v vp8_loop_filter_mbv_neon
-
-#undef  vp8_lf_normal_b_v
-#define vp8_lf_normal_b_v vp8_loop_filter_bv_neon
-
-#undef  vp8_lf_normal_mb_h
-#define vp8_lf_normal_mb_h vp8_loop_filter_mbh_neon
-
-#undef  vp8_lf_normal_b_h
-#define vp8_lf_normal_b_h vp8_loop_filter_bh_neon
-
-#undef  vp8_lf_simple_mb_v
-#define vp8_lf_simple_mb_v vp8_loop_filter_mbvs_neon
-
-#undef  vp8_lf_simple_b_v
-#define vp8_lf_simple_b_v vp8_loop_filter_bvs_neon
-
-#undef  vp8_lf_simple_mb_h
-#define vp8_lf_simple_mb_h vp8_loop_filter_mbhs_neon
-
-#undef  vp8_lf_simple_b_h
-#define vp8_lf_simple_b_h vp8_loop_filter_bhs_neon
-#endif /* !CONFIG_RUNTIME_CPU_DETECT */
-
-#endif /* HAVE_ARMV7 */
-
-#endif /* LOOPFILTER_ARM_H */

vp8/common/arm/neon/idct_blk_neon.c

@@ -9,8 +9,7 @@
  */
 
 #include "vpx_config.h"
-#include "vp8/common/idct.h"
-#include "vp8/decoder/dequantize.h"
+#include "vpx_rtcd.h"
 
 /* place these declarations here because we don't want to maintain them
  * outside of this scope

vp8/common/arm/neon/sixtappredict4x4_neon.asm

@@ -9,7 +9,7 @@
 ;
 
 
-    EXPORT  |vp8_sixtap_predict_neon|
+    EXPORT  |vp8_sixtap_predict4x4_neon|
     ARM
     REQUIRE8
     PRESERVE8
@@ -33,7 +33,7 @@ filter4_coeff
 ; stack(r4) unsigned char *dst_ptr,
 ; stack(lr) int  dst_pitch
 
-|vp8_sixtap_predict_neon| PROC
+|vp8_sixtap_predict4x4_neon| PROC
     push            {r4, lr}
 
     adr             r12, filter4_coeff

vp8/common/arm/neon/variance_neon.asm

@@ -77,14 +77,14 @@ variance16x16_neon_loop
     ;vmov.32        r1, d1[0]
     ;mul            r0, r0, r0
     ;str            r1, [r12]
-    ;sub            r0, r1, r0, asr #8
+    ;sub            r0, r1, r0, lsr #8
 
-    ;sum is in [-255x256, 255x256]. sumxsum is 32-bit. Shift to right should
-    ;have sign-bit exension, which is vshr.s. Have to use s32 to make it right.
+    ; while sum is signed, sum * sum is always positive and must be treated as
+    ; unsigned to avoid propagating the sign bit.
     vmull.s32       q5, d0, d0
     vst1.32         {d1[0]}, [r12]              ;store sse
-    vshr.s32        d10, d10, #8
-    vsub.s32        d0, d1, d10
+    vshr.u32        d10, d10, #8
+    vsub.u32        d0, d1, d10
 
     vmov.32         r0, d0[0]                   ;return
     bx              lr
@@ -145,8 +145,8 @@ variance16x8_neon_loop
 
     vmull.s32       q5, d0, d0
     vst1.32         {d1[0]}, [r12]              ;store sse
-    vshr.s32        d10, d10, #7
-    vsub.s32        d0, d1, d10
+    vshr.u32        d10, d10, #7
+    vsub.u32        d0, d1, d10
 
     vmov.32         r0, d0[0]                   ;return
     bx              lr
@@ -200,8 +200,8 @@ variance8x16_neon_loop
 
     vmull.s32       q5, d0, d0
     vst1.32         {d1[0]}, [r12]              ;store sse
-    vshr.s32        d10, d10, #7
-    vsub.s32        d0, d1, d10
+    vshr.u32        d10, d10, #7
+    vsub.u32        d0, d1, d10
 
     vmov.32         r0, d0[0]                   ;return
     bx              lr
@@ -265,8 +265,8 @@ variance8x8_neon_loop
 
     vmull.s32       q5, d0, d0
     vst1.32         {d1[0]}, [r12]              ;store sse
-    vshr.s32        d10, d10, #6
-    vsub.s32        d0, d1, d10
+    vshr.u32        d10, d10, #6
+    vsub.u32        d0, d1, d10
 
     vmov.32         r0, d0[0]                   ;return
     bx              lr

vp8/common/arm/neon/vp8_subpixelvariance16x16_neon.asm

@@ -9,6 +9,11 @@
 ;
 
 
+bilinear_taps_coeff
+    DCD     128, 0, 112, 16, 96, 32, 80, 48, 64, 64, 48, 80, 32, 96, 16, 112
+
+;-----------------
+
     EXPORT  |vp8_sub_pixel_variance16x16_neon_func|
     ARM
     REQUIRE8
@@ -27,7 +32,7 @@
 |vp8_sub_pixel_variance16x16_neon_func| PROC
     push            {r4-r6, lr}
 
-    ldr             r12, _BilinearTaps_coeff_
+    adr             r12, bilinear_taps_coeff
     ldr             r4, [sp, #16]           ;load *dst_ptr from stack
     ldr             r5, [sp, #20]           ;load dst_pixels_per_line from stack
     ldr             r6, [sp, #24]           ;load *sse from stack
@@ -405,8 +410,8 @@ sub_pixel_variance16x16_neon_loop
 
     vmull.s32       q5, d0, d0
     vst1.32         {d1[0]}, [r6]               ;store sse
-    vshr.s32        d10, d10, #8
-    vsub.s32        d0, d1, d10
+    vshr.u32        d10, d10, #8
+    vsub.u32        d0, d1, d10
 
     add             sp, sp, #528
     vmov.32         r0, d0[0]                   ;return
@@ -415,11 +420,4 @@ sub_pixel_variance16x16_neon_loop
 
     ENDP
 
-;-----------------
-
-_BilinearTaps_coeff_
-    DCD     bilinear_taps_coeff
-bilinear_taps_coeff
-    DCD     128, 0, 112, 16, 96, 32, 80, 48, 64, 64, 48, 80, 32, 96, 16, 112
-
     END

vp8/common/arm/neon/vp8_subpixelvariance16x16s_neon.asm

@@ -112,8 +112,8 @@ vp8_filt_fpo16x16s_4_0_loop_neon
 
     vmull.s32       q5, d0, d0
     vst1.32         {d1[0]}, [lr]               ;store sse
-    vshr.s32        d10, d10, #8
-    vsub.s32        d0, d1, d10
+    vshr.u32        d10, d10, #8
+    vsub.u32        d0, d1, d10
 
     vmov.32         r0, d0[0]                   ;return
     pop             {pc}
@@ -208,8 +208,8 @@ vp8_filt_spo16x16s_0_4_loop_neon
 
     vmull.s32       q5, d0, d0
     vst1.32         {d1[0]}, [lr]               ;store sse
-    vshr.s32        d10, d10, #8
-    vsub.s32        d0, d1, d10
+    vshr.u32        d10, d10, #8
+    vsub.u32        d0, d1, d10
 
     vmov.32         r0, d0[0]                   ;return
     pop             {pc}
@@ -327,8 +327,8 @@ vp8_filt16x16s_4_4_loop_neon
 
     vmull.s32       q5, d0, d0
     vst1.32         {d1[0]}, [lr]               ;store sse
-    vshr.s32        d10, d10, #8
-    vsub.s32        d0, d1, d10
+    vshr.u32        d10, d10, #8
+    vsub.u32        d0, d1, d10
 
     vmov.32         r0, d0[0]                   ;return
     pop             {pc}
@@ -560,8 +560,8 @@ sub_pixel_variance16x16s_neon_loop
 
     vmull.s32       q5, d0, d0
     vst1.32         {d1[0]}, [lr]               ;store sse
-    vshr.s32        d10, d10, #8
-    vsub.s32        d0, d1, d10
+    vshr.u32        d10, d10, #8
+    vsub.u32        d0, d1, d10
 
     add             sp, sp, #256
     vmov.32         r0, d0[0]                   ;return

vp8/common/arm/neon/vp8_subpixelvariance8x8_neon.asm

@@ -27,7 +27,7 @@
 |vp8_sub_pixel_variance8x8_neon| PROC
     push            {r4-r5, lr}
 
-    ldr             r12, _BilinearTaps_coeff_
+    adr             r12, bilinear_taps_coeff
     ldr             r4, [sp, #12]           ;load *dst_ptr from stack
     ldr             r5, [sp, #16]           ;load dst_pixels_per_line from stack
     ldr             lr, [sp, #20]           ;load *sse from stack
@@ -206,8 +206,8 @@ sub_pixel_variance8x8_neon_loop
 
     vmull.s32       q5, d0, d0
     vst1.32         {d1[0]}, [lr]               ;store sse
-    vshr.s32        d10, d10, #6
-    vsub.s32        d0, d1, d10
+    vshr.u32        d10, d10, #6
+    vsub.u32        d0, d1, d10
 
     vmov.32         r0, d0[0]                   ;return
     pop             {r4-r5, pc}
@@ -216,8 +216,6 @@ sub_pixel_variance8x8_neon_loop
 
 ;-----------------
 
-_BilinearTaps_coeff_
-    DCD     bilinear_taps_coeff
 bilinear_taps_coeff
     DCD     128, 0, 112, 16, 96, 32, 80, 48, 64, 64, 48, 80, 32, 96, 16, 112
 

vp8/common/arm/recon_arm.h

@@ -1,65 +0,0 @@
-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-
-#ifndef RECON_ARM_H
-#define RECON_ARM_H
-
-#if HAVE_ARMV6
-
-extern prototype_copy_block(vp8_copy_mem8x8_v6);
-extern prototype_copy_block(vp8_copy_mem8x4_v6);
-extern prototype_copy_block(vp8_copy_mem16x16_v6);
-extern prototype_intra4x4_predict(vp8_intra4x4_predict_armv6);
-
-#if !CONFIG_RUNTIME_CPU_DETECT
-#undef  vp8_recon_copy8x8
-#define vp8_recon_copy8x8 vp8_copy_mem8x8_v6
-
-#undef  vp8_recon_copy8x4
-#define vp8_recon_copy8x4 vp8_copy_mem8x4_v6
-
-#undef  vp8_recon_copy16x16
-#define vp8_recon_copy16x16 vp8_copy_mem16x16_v6
-
-#undef  vp8_recon_intra4x4_predict
-#define vp8_recon_intra4x4_predict vp8_intra4x4_predict_armv6
-#endif
-#endif
-
-#if HAVE_ARMV7
-
-extern prototype_copy_block(vp8_copy_mem8x8_neon);
-extern prototype_copy_block(vp8_copy_mem8x4_neon);
-extern prototype_copy_block(vp8_copy_mem16x16_neon);
-
-extern prototype_build_intra_predictors(vp8_build_intra_predictors_mby_neon);
-extern prototype_build_intra_predictors(vp8_build_intra_predictors_mby_s_neon);
-
-#if !CONFIG_RUNTIME_CPU_DETECT
-#undef  vp8_recon_copy8x8
-#define vp8_recon_copy8x8 vp8_copy_mem8x8_neon
-
-#undef  vp8_recon_copy8x4
-#define vp8_recon_copy8x4 vp8_copy_mem8x4_neon
-
-#undef  vp8_recon_copy16x16
-#define vp8_recon_copy16x16 vp8_copy_mem16x16_neon
-
-#undef  vp8_recon_build_intra_predictors_mby
-#define vp8_recon_build_intra_predictors_mby vp8_build_intra_predictors_mby_neon
-
-#undef  vp8_recon_build_intra_predictors_mby_s
-#define vp8_recon_build_intra_predictors_mby_s vp8_build_intra_predictors_mby_s_neon
-
-#endif
-#endif
-
-#endif

vp8/common/arm/reconintra_arm.c

@@ -10,12 +10,11 @@
 
 
 #include "vpx_config.h"
+#include "vpx_rtcd.h"
 #include "vp8/common/blockd.h"
-#include "vp8/common/reconintra.h"
 #include "vpx_mem/vpx_mem.h"
-#include "vp8/common/recon.h"
 
-#if HAVE_ARMV7
+#if HAVE_NEON
 extern void vp8_build_intra_predictors_mby_neon_func(
     unsigned char *y_buffer,
     unsigned char *ypred_ptr,
@@ -35,10 +34,7 @@ void vp8_build_intra_predictors_mby_neon(MACROBLOCKD *x)
 
     vp8_build_intra_predictors_mby_neon_func(y_buffer, ypred_ptr, y_stride, mode, Up, Left);
 }
-#endif
 
-
-#if HAVE_ARMV7
 extern void vp8_build_intra_predictors_mby_s_neon_func(
     unsigned char *y_buffer,
     unsigned char *ypred_ptr,

vp8/common/arm/subpixel_arm.h

@@ -1,89 +0,0 @@
-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-
-#ifndef SUBPIXEL_ARM_H
-#define SUBPIXEL_ARM_H
-
-#if HAVE_ARMV6
-extern prototype_subpixel_predict(vp8_sixtap_predict16x16_armv6);
-extern prototype_subpixel_predict(vp8_sixtap_predict8x8_armv6);
-extern prototype_subpixel_predict(vp8_sixtap_predict8x4_armv6);
-extern prototype_subpixel_predict(vp8_sixtap_predict_armv6);
-extern prototype_subpixel_predict(vp8_bilinear_predict16x16_armv6);
-extern prototype_subpixel_predict(vp8_bilinear_predict8x8_armv6);
-extern prototype_subpixel_predict(vp8_bilinear_predict8x4_armv6);
-extern prototype_subpixel_predict(vp8_bilinear_predict4x4_armv6);
-
-#if !CONFIG_RUNTIME_CPU_DETECT
-#undef  vp8_subpix_sixtap16x16
-#define vp8_subpix_sixtap16x16 vp8_sixtap_predict16x16_armv6
-
-#undef  vp8_subpix_sixtap8x8
-#define vp8_subpix_sixtap8x8 vp8_sixtap_predict8x8_armv6
-
-#undef  vp8_subpix_sixtap8x4
-#define vp8_subpix_sixtap8x4 vp8_sixtap_predict8x4_armv6
-
-#undef  vp8_subpix_sixtap4x4
-#define vp8_subpix_sixtap4x4 vp8_sixtap_predict_armv6
-
-#undef  vp8_subpix_bilinear16x16
-#define vp8_subpix_bilinear16x16 vp8_bilinear_predict16x16_armv6
-
-#undef  vp8_subpix_bilinear8x8
-#define vp8_subpix_bilinear8x8 vp8_bilinear_predict8x8_armv6
-
-#undef  vp8_subpix_bilinear8x4
-#define vp8_subpix_bilinear8x4 vp8_bilinear_predict8x4_armv6
-
-#undef  vp8_subpix_bilinear4x4
-#define vp8_subpix_bilinear4x4 vp8_bilinear_predict4x4_armv6
-#endif
-#endif
-
-#if HAVE_ARMV7
-extern prototype_subpixel_predict(vp8_sixtap_predict16x16_neon);
-extern prototype_subpixel_predict(vp8_sixtap_predict8x8_neon);
-extern prototype_subpixel_predict(vp8_sixtap_predict8x4_neon);
-extern prototype_subpixel_predict(vp8_sixtap_predict_neon);
-extern prototype_subpixel_predict(vp8_bilinear_predict16x16_neon);
-extern prototype_subpixel_predict(vp8_bilinear_predict8x8_neon);
-extern prototype_subpixel_predict(vp8_bilinear_predict8x4_neon);
-extern prototype_subpixel_predict(vp8_bilinear_predict4x4_neon);
-
-#if !CONFIG_RUNTIME_CPU_DETECT
-#undef  vp8_subpix_sixtap16x16
-#define vp8_subpix_sixtap16x16 vp8_sixtap_predict16x16_neon
-
-#undef  vp8_subpix_sixtap8x8
-#define vp8_subpix_sixtap8x8 vp8_sixtap_predict8x8_neon
-
-#undef  vp8_subpix_sixtap8x4
-#define vp8_subpix_sixtap8x4 vp8_sixtap_predict8x4_neon
-
-#undef  vp8_subpix_sixtap4x4
-#define vp8_subpix_sixtap4x4 vp8_sixtap_predict_neon
-
-#undef  vp8_subpix_bilinear16x16
-#define vp8_subpix_bilinear16x16 vp8_bilinear_predict16x16_neon
-
-#undef  vp8_subpix_bilinear8x8
-#define vp8_subpix_bilinear8x8 vp8_bilinear_predict8x8_neon
-
-#undef  vp8_subpix_bilinear8x4
-#define vp8_subpix_bilinear8x4 vp8_bilinear_predict8x4_neon
-
-#undef  vp8_subpix_bilinear4x4
-#define vp8_subpix_bilinear4x4 vp8_bilinear_predict4x4_neon
-#endif
-#endif
-
-#endif

vp8/common/arm/variance_arm.c

@@ -9,11 +9,12 @@
  */
 
 #include "vpx_config.h"
-#include "vp8/encoder/variance.h"
+#include "vpx_rtcd.h"
+#include "vp8/common/variance.h"
 #include "vp8/common/filter.h"
-#include "vp8/common/arm/bilinearfilter_arm.h"
 
-#if HAVE_ARMV6
+#if HAVE_MEDIA
+#include "vp8/common/arm/bilinearfilter_arm.h"
 
 unsigned int vp8_sub_pixel_variance8x8_armv6
 (
@@ -91,10 +92,21 @@ unsigned int vp8_sub_pixel_variance16x16_armv6
     return var;
 }
 
-#endif /* HAVE_ARMV6 */
+#endif /* HAVE_MEDIA */
 
 
-#if HAVE_ARMV7
+#if HAVE_NEON
+
+extern unsigned int vp8_sub_pixel_variance16x16_neon_func
+(
+    const unsigned char  *src_ptr,
+    int  src_pixels_per_line,
+    int  xoffset,
+    int  yoffset,
+    const unsigned char *dst_ptr,
+    int dst_pixels_per_line,
+    unsigned int *sse
+);
 
 unsigned int vp8_sub_pixel_variance16x16_neon
 (

vp8/common/asm_com_offsets.c

@@ -15,6 +15,10 @@
 #include "vpx_scale/yv12config.h"
 #include "vp8/common/blockd.h"
 
+#if CONFIG_POSTPROC
+#include "postproc.h"
+#endif /* CONFIG_POSTPROC */
+
 BEGIN
 
 /* vpx_scale */
@@ -30,12 +34,17 @@ DEFINE(yv12_buffer_config_v_buffer,             offsetof(YV12_BUFFER_CONFIG, v_b
 DEFINE(yv12_buffer_config_border,               offsetof(YV12_BUFFER_CONFIG, border));
 DEFINE(VP8BORDERINPIXELS_VAL,                   VP8BORDERINPIXELS);
 
+#if CONFIG_POSTPROC
+/* mfqe.c / filter_by_weight */
+DEFINE(MFQE_PRECISION_VAL,                      MFQE_PRECISION);
+#endif /* CONFIG_POSTPROC */
+
 END
 
 /* add asserts for any offset that is not supported by assembly code */
 /* add asserts for any size that is not supported by assembly code */
 
-#if HAVE_ARMV6
+#if HAVE_MEDIA
 /* switch case in vp8_intra4x4_predict_armv6 is based on these enumerated values */
 ct_assert(B_DC_PRED, B_DC_PRED == 0);
 ct_assert(B_TM_PRED, B_TM_PRED == 1);
@@ -49,7 +58,14 @@ ct_assert(B_HD_PRED, B_HD_PRED == 8);
 ct_assert(B_HU_PRED, B_HU_PRED == 9);
 #endif
 
-#if HAVE_ARMV7
+#if HAVE_NEON
 /* vp8_yv12_extend_frame_borders_neon makes several assumptions based on this */
 ct_assert(VP8BORDERINPIXELS_VAL, VP8BORDERINPIXELS == 32)
 #endif
+
+#if HAVE_SSE2
+#if CONFIG_POSTPROC
+/* vp8_filter_by_weight16x16 and 8x8 */
+ct_assert(MFQE_PRECISION_VAL, MFQE_PRECISION == 4)
+#endif /* CONFIG_POSTPROC */
+#endif /* HAVE_SSE2 */

vp8/common/bigend.h

@@ -1,32 +0,0 @@
-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-
-#ifndef _bigend_h
-#define _bigend_h
-
-#if defined(__cplusplus)
-extern "C" {
-#endif
-
-#define invert2(x) ( (((x)>>8)&0x00ff) | (((x)<<8)&0xff00) )
-#define invert4(x) ( ((invert2(x)&0x0000ffff)<<16) | (invert2((x>>16))&0x0000ffff) )
-
-#define high_byte(x) (unsigned char)x
-#define mid2Byte(x) (unsigned char)(x >> 8)
-#define mid1Byte(x) (unsigned char)(x >> 16)
-#define low_byte(x) (unsigned char)(x >> 24)
-
-#define SWAPENDS 1
-
-#if defined(__cplusplus)
-}
-#endif
-#endif

vp8/common/blockd.h

@@ -18,12 +18,8 @@ void vpx_log(const char *format, ...);
 #include "vpx_scale/yv12config.h"
 #include "mv.h"
 #include "treecoder.h"
-#include "subpixel.h"
 #include "vpx_ports/mem.h"
 
-#define TRUE    1
-#define FALSE   0
-
 /*#define DCPRED 1*/
 #define DCPREDSIMTHRESH 0
 #define DCPREDCNTTHRESH 3
@@ -154,14 +150,15 @@ typedef enum
 
 typedef struct
 {
-    MB_PREDICTION_MODE mode, uv_mode;
-    MV_REFERENCE_FRAME ref_frame;
+    uint8_t mode, uv_mode;
+    uint8_t ref_frame;
+    uint8_t is_4x4;
     int_mv mv;
 
-    unsigned char partitioning;
-    unsigned char mb_skip_coeff;                                /* does this mb has coefficients at all, 1=no coefficients, 0=need decode tokens */
-    unsigned char need_to_clamp_mvs;
-    unsigned char segment_id;                  /* Which set of segmentation parameters should be used for this MB */
+    uint8_t partitioning;
+    uint8_t mb_skip_coeff;                                /* does this mb has coefficients at all, 1=no coefficients, 0=need decode tokens */
+    uint8_t need_to_clamp_mvs;
+    uint8_t segment_id;                  /* Which set of segmentation parameters should be used for this MB */
 } MB_MODE_INFO;
 
 typedef struct
@@ -182,36 +179,33 @@ typedef struct
 } LOWER_RES_INFO;
 #endif
 
-typedef struct
+typedef struct blockd
 {
     short *qcoeff;
     short *dqcoeff;
     unsigned char  *predictor;
-    short *diff;
     short *dequant;
 
-    /* 16 Y blocks, 4 U blocks, 4 V blocks each with 16 entries */
-    unsigned char **base_pre;
-    int pre;
-    int pre_stride;
-
-    unsigned char **base_dst;
-    int dst;
-    int dst_stride;
-
+    int offset;
     char *eob;
 
     union b_mode_info bmi;
 } BLOCKD;
 
-typedef struct MacroBlockD
+typedef void (*vp8_subpix_fn_t)(unsigned char *src, int src_pitch, int xofst, int yofst, unsigned char *dst, int dst_pitch);
+
+typedef struct macroblockd
 {
-    DECLARE_ALIGNED(16, short, diff[400]);      /* from idct diff */
     DECLARE_ALIGNED(16, unsigned char,  predictor[384]);
     DECLARE_ALIGNED(16, short, qcoeff[400]);
     DECLARE_ALIGNED(16, short, dqcoeff[400]);
     DECLARE_ALIGNED(16, char,  eobs[25]);
 
+    DECLARE_ALIGNED(16, short,  dequant_y1[16]);
+    DECLARE_ALIGNED(16, short,  dequant_y1_dc[16]);
+    DECLARE_ALIGNED(16, short,  dequant_y2[16]);
+    DECLARE_ALIGNED(16, short,  dequant_uv[16]);
+
     /* 16 Y blocks, 4 U, 4 V, 1 DC 2nd order block, each with 16 entries. */
     BLOCKD block[25];
     int fullpixel_mask;
@@ -222,11 +216,21 @@ typedef struct MacroBlockD
     MODE_INFO *mode_info_context;
     int mode_info_stride;
 
+#if CONFIG_TEMPORAL_DENOISING
+    MB_PREDICTION_MODE best_sse_inter_mode;
+    int_mv best_sse_mv;
+    unsigned char need_to_clamp_best_mvs;
+#endif
+
     FRAME_TYPE frame_type;
 
     int up_available;
     int left_available;
 
+    unsigned char *recon_above[3];
+    unsigned char *recon_left[3];
+    int recon_left_stride[2];
+
     /* Y,U,V,Y2 */
     ENTROPY_CONTEXT_PLANES *above_context;
     ENTROPY_CONTEXT_PLANES *left_context;
@@ -265,11 +269,8 @@ typedef struct MacroBlockD
     int mb_to_top_edge;
     int mb_to_bottom_edge;
 
-    int ref_frame_cost[MAX_REF_FRAMES];
 
 
-    unsigned int frames_since_golden;
-    unsigned int frames_till_alt_ref_frame;
     vp8_subpix_fn_t  subpixel_predict;
     vp8_subpix_fn_t  subpixel_predict8x4;
     vp8_subpix_fn_t  subpixel_predict8x8;
@@ -286,10 +287,6 @@ typedef struct MacroBlockD
      */
     DECLARE_ALIGNED(32, unsigned char, y_buf[22*32]);
 #endif
-
-#if CONFIG_RUNTIME_CPU_DETECT
-    struct VP8_COMMON_RTCD  *rtcd;
-#endif
 } MACROBLOCKD;
 
 

vp8/common/common.h

@@ -18,8 +18,6 @@
 
 #include "vpx_mem/vpx_mem.h"
 
-#include "common_types.h"
-
 /* Only need this for fixed-size arrays, for structs just assign. */
 
 #define vp8_copy( Dest, Src) { \

vp8/common/dequantize.c

@@ -10,16 +10,15 @@
 
 
 #include "vpx_config.h"
-#include "dequantize.h"
-#include "vp8/common/idct.h"
+#include "vpx_rtcd.h"
+#include "vp8/common/blockd.h"
 #include "vpx_mem/vpx_mem.h"
 
-void vp8_dequantize_b_c(BLOCKD *d)
+void vp8_dequantize_b_c(BLOCKD *d, short *DQC)
 {
     int i;
     short *DQ  = d->dqcoeff;
     short *Q   = d->qcoeff;
-    short *DQC = d->dequant;
 
     for (i = 0; i < 16; i++)
     {

vp8/common/dma_desc.h

@@ -1,125 +0,0 @@
-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-
-#ifndef _dma_desc_h
-#define _dma_desc_h
-
-#if defined(__cplusplus)
-extern "C" {
-#endif
-
-
-#define NDSIZE_LG   0x00000900  // Next Descriptor Size
-#define NDSIZE_SM   0x00000800  // Next Descriptor Size
-#define NDSIZE_7    0x00000700  // Next Descriptor Size
-#define NDSIZE_6    0x00000600  // Next Descriptor Size
-#define NDSIZE_5    0x00000500  // Next Descriptor Size
-#define NDSIZE_4    0x00000400  // Next Descriptor Size
-#define NDSIZE_3    0x00000300  // Next Descriptor Size
-#define NDSIZE_2    0x00000200  // Next Descriptor Size
-#define NDSIZE_1    0x00000100  // Next Descriptor Size
-
-#define FLOW_STOP       0x0000
-#define FLOW_AUTO       0x1000
-#define FLOW_DESC_AR    0x4000
-#define FLOW_DESC_SM    0x6000
-#define FLOW_DESC_LG    0x7000
-
-    typedef struct
-    {
-        unsigned int ndp;
-        //unsigned short ndpl;
-        //unsigned short ndph;
-        unsigned int sa;
-        //unsigned short sal;
-        //unsigned short sah;
-
-        unsigned short dmacfg;
-        unsigned short xcnt;
-        unsigned short xmod;
-        unsigned short ycnt;
-        unsigned short ymod;
-
-    } LARGE_DESC;
-
-    typedef struct
-    {
-        unsigned short ndpl;
-        unsigned short sal;
-        unsigned short sah;
-        unsigned short dmacfg;
-        unsigned short xcnt;
-        unsigned short xmod;
-        unsigned short ycnt;
-        unsigned short ymod;
-    } SMALL_DESC;
-
-    typedef struct
-    {
-        unsigned short sal;
-        unsigned short sah;
-        unsigned short dmacfg;
-        unsigned short xcnt;
-        unsigned short xmod;
-        unsigned short ycnt;
-        unsigned short ymod;
-    } ARRAY_DESC_7;
-
-    typedef struct
-    {
-        unsigned short sal;
-        unsigned short sah;
-        unsigned short dmacfg;
-        unsigned short xcnt;
-        unsigned short xmod;
-        unsigned short ycnt;
-    } ARRAY_DESC_6;
-
-    typedef struct
-    {
-        unsigned short sal;
-        unsigned short sah;
-        unsigned short dmacfg;
-        unsigned short xcnt;
-        unsigned short xmod;
-    } ARRAY_DESC_5;
-
-    typedef struct
-    {
-        unsigned short sal;
-        unsigned short sah;
-        unsigned short dmacfg;
-        unsigned short xcnt;
-    } ARRAY_DESC_4;
-
-    typedef struct
-    {
-        unsigned short sal;
-        unsigned short sah;
-        unsigned short dmacfg;
-    } ARRAY_DESC_3;
-
-    typedef struct
-    {
-        unsigned short sal;
-        unsigned short sah;
-    } ARRAY_DESC_2;
-
-    typedef struct
-    {
-        unsigned short sal;
-    } ARRAY_DESC_1;
-
-#if defined(__cplusplus)
-}
-#endif
-
-#endif //_dma_desc_h

vp8/common/duck_io.h

@@ -1,116 +0,0 @@
-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-
-#ifndef _duck_io_h
-#define _duck_io_h
-
-#if defined(__cplusplus)
-extern "C" {
-#endif
-
-#if defined (_WIN32)
-    typedef __int64 int64_t;
-#elif defined(__MWERKS__)
-    typedef long long int64_t;
-#elif defined(__APPLE__) || defined(__POWERPC)
-#include <ppc/types.h>
-#else
-    typedef long long int64_t;
-#endif
-
-    typedef struct
-    {
-        int64_t  offset;     // offset to start from
-        int    blocking;    // non-zero for blocking
-    } re_open_t;
-
-
-    typedef enum
-    {
-        SAL_ERR_MAX                 = -10,
-        SAL_ERROR                   = -11, // Default error
-        SAL_ERR_WSASTARTUP          = -12,
-        SAL_ERR_SOCKET_CREATE       = -13,
-        SAL_ERR_RESOLVING_HOSTNAME  = -14,
-        SAL_ERR_SERVER_CONNECTION   = -15,
-        SAL_ERR_SENDING_DATA        = -16,
-        SAL_ERR_RECEIVING_DATA      = -17,
-        SAL_ERR_404_FILE_NOT_FOUND  = -18,
-        SAL_ERR_PARSING_HTTP_HEADER = -19,
-        SAL_ERR_PARSING_CONTENT_LEN = -20,
-        SAL_ERR_CONNECTION_TIMEOUT  = -21,
-        SAL_ERR_FILE_OPEN_FAILED    = -22,
-        SAL_ERR_MIN                 = -23
-    } SAL_ERR; /* EMH 1-15-03 */
-
-
-    typedef struct sal_err_map_temp
-    {
-        SAL_ERR code;
-        const char *decode;
-
-    } sal_err_map_t;
-
-
-    static char *sal_err_text(SAL_ERR e)
-    {
-        int t;
-        const sal_err_map_t g_sal_err_map[] =
-        {
-            {   SAL_ERR_WSASTARTUP,             "Error with WSAStartup"         },
-            {   SAL_ERR_SOCKET_CREATE,          "Error creating socket"         },
-            {   SAL_ERR_RESOLVING_HOSTNAME,     "Error resolving hostname"      },
-            {   SAL_ERR_SERVER_CONNECTION,      "Error connecting to server"    },
-            {   SAL_ERR_SENDING_DATA,           "Error sending data"            },
-            {   SAL_ERR_RECEIVING_DATA,         "Error receiving data"          },
-            {   SAL_ERR_404_FILE_NOT_FOUND,     "Error file not found "         },
-            {   SAL_ERR_PARSING_HTTP_HEADER,    "Error parsing http header"     },
-            {   SAL_ERR_PARSING_CONTENT_LEN,    "Error parsing content length"  },
-            {   SAL_ERR_CONNECTION_TIMEOUT,     "Error Connection timed out"    },
-            {   SAL_ERR_FILE_OPEN_FAILED,       "Error opening file"            }
-        };
-
-        for (t = 0; t < sizeof(g_sal_err_map) / sizeof(sal_err_map_t); t++)
-        {
-            if (e == g_sal_err_map[t].code)
-                return (char *) g_sal_err_map[t].decode;
-        }
-
-        return 0;
-    }
-
-
-
-
-
-
-
-    int duck_open(const char *fname, unsigned long user_data);
-
-    void duck_close(int ghndl);
-
-    int duck_read(int ghndl, unsigned char *buf, int nbytes);
-
-    int64_t duck_seek(int g_hndl, int64_t offs, int origin);
-
-    int duck_read_finished(int han, int flag); /* FWG 7-9-99 */
-
-    int duck_name(int handle, char name[], size_t max_len); /* EMH 9-23-03 */
-
-    int duck_read_blocking(int handle, unsigned char *buffer, int bytes); /* EMH 9-23-03 */
-
-    int64_t duck_available_data(int handle); /* EMH 10-23-03 */
-
-#if defined(__cplusplus)
-}
-#endif
-
-#endif

vp8/common/entropy.c

@@ -8,23 +8,11 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
-
-#include <stdio.h>
-
 #include "entropy.h"
-#include "string.h"
 #include "blockd.h"
 #include "onyxc_int.h"
 #include "vpx_mem/vpx_mem.h"
 
-#define uchar unsigned char     /* typedefs can clash */
-#define uint  unsigned int
-
-typedef const uchar cuchar;
-typedef const uint cuint;
-
-typedef vp8_prob Prob;
-
 #include "coefupdateprobs.h"
 
 DECLARE_ALIGNED(16, const unsigned char, vp8_norm[256]) =
@@ -47,10 +35,11 @@ DECLARE_ALIGNED(16, const unsigned char, vp8_norm[256]) =
     0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
 };
 
-DECLARE_ALIGNED(16, cuchar, vp8_coef_bands[16]) =
+DECLARE_ALIGNED(16, const unsigned char, vp8_coef_bands[16]) =
 { 0, 1, 2, 3, 6, 4, 5, 6, 6, 6, 6, 6, 6, 6, 6, 7};
 
-DECLARE_ALIGNED(16, cuchar, vp8_prev_token_class[MAX_ENTROPY_TOKENS]) =
+DECLARE_ALIGNED(16, const unsigned char,
+                vp8_prev_token_class[MAX_ENTROPY_TOKENS]) =
 { 0, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0};
 
 DECLARE_ALIGNED(16, const int, vp8_default_zig_zag1d[16]) =
@@ -69,7 +58,26 @@ DECLARE_ALIGNED(16, const short, vp8_default_inv_zig_zag[16]) =
    10, 11, 15, 16
 };
 
-DECLARE_ALIGNED(16, short, vp8_default_zig_zag_mask[16]);
+/* vp8_default_zig_zag_mask generated with:
+
+    void vp8_init_scan_order_mask()
+    {
+        int i;
+
+        for (i = 0; i < 16; i++)
+        {
+            vp8_default_zig_zag_mask[vp8_default_zig_zag1d[i]] = 1 << i;
+        }
+
+    }
+*/
+DECLARE_ALIGNED(16, const short, vp8_default_zig_zag_mask[16]) =
+{
+     1,    2,    32,     64,
+     4,   16,   128,   4096,
+     8,  256,  2048,   8192,
+   512, 1024, 16384, -32768
+};
 
 const int vp8_mb_feature_data_bits[MB_LVL_MAX] = {7, 6};
 
@@ -90,56 +98,72 @@ const vp8_tree_index vp8_coef_tree[ 22] =     /* corresponding _CONTEXT_NODEs */
     -DCT_VAL_CATEGORY5, -DCT_VAL_CATEGORY6   /* 10 = CAT_FIVE */
 };
 
-struct vp8_token_struct vp8_coef_encodings[MAX_ENTROPY_TOKENS];
+/* vp8_coef_encodings generated with:
+    vp8_tokens_from_tree(vp8_coef_encodings, vp8_coef_tree);
+*/
+const vp8_token vp8_coef_encodings[MAX_ENTROPY_TOKENS] =
+{
+    {2, 2},
+    {6, 3},
+    {28, 5},
+    {58, 6},
+    {59, 6},
+    {60, 6},
+    {61, 6},
+    {124, 7},
+    {125, 7},
+    {126, 7},
+    {127, 7},
+    {0, 1}
+};
 
 /* Trees for extra bits.  Probabilities are constant and
    do not depend on previously encoded bits */
 
-static const Prob Pcat1[] = { 159};
-static const Prob Pcat2[] = { 165, 145};
-static const Prob Pcat3[] = { 173, 148, 140};
-static const Prob Pcat4[] = { 176, 155, 140, 135};
-static const Prob Pcat5[] = { 180, 157, 141, 134, 130};
-static const Prob Pcat6[] =
+static const vp8_prob Pcat1[] = { 159};
+static const vp8_prob Pcat2[] = { 165, 145};
+static const vp8_prob Pcat3[] = { 173, 148, 140};
+static const vp8_prob Pcat4[] = { 176, 155, 140, 135};
+static const vp8_prob Pcat5[] = { 180, 157, 141, 134, 130};
+static const vp8_prob Pcat6[] =
 { 254, 254, 243, 230, 196, 177, 153, 140, 133, 130, 129};
 
-static vp8_tree_index cat1[2], cat2[4], cat3[6], cat4[8], cat5[10], cat6[22];
 
-void vp8_init_scan_order_mask()
-{
-    int i;
+/* tree index tables generated with:
 
-    for (i = 0; i < 16; i++)
+    void init_bit_tree(vp8_tree_index *p, int n)
     {
-        vp8_default_zig_zag_mask[vp8_default_zig_zag1d[i]] = 1 << i;
+        int i = 0;
+
+        while (++i < n)
+        {
+            p[0] = p[1] = i << 1;
+            p += 2;
+        }
+
+        p[0] = p[1] = 0;
     }
 
-}
-
-static void init_bit_tree(vp8_tree_index *p, int n)
-{
-    int i = 0;
-
-    while (++i < n)
+    void init_bit_trees()
     {
-        p[0] = p[1] = i << 1;
-        p += 2;
+        init_bit_tree(cat1, 1);
+        init_bit_tree(cat2, 2);
+        init_bit_tree(cat3, 3);
+        init_bit_tree(cat4, 4);
+        init_bit_tree(cat5, 5);
+        init_bit_tree(cat6, 11);
     }
+*/
 
-    p[0] = p[1] = 0;
-}
+static const vp8_tree_index cat1[2] = { 0, 0 };
+static const vp8_tree_index cat2[4] = { 2, 2, 0, 0 };
+static const vp8_tree_index cat3[6] = { 2, 2, 4, 4, 0, 0 };
+static const vp8_tree_index cat4[8] = { 2, 2, 4, 4, 6, 6, 0, 0 };
+static const vp8_tree_index cat5[10] = { 2, 2, 4, 4, 6, 6, 8, 8, 0, 0 };
+static const vp8_tree_index cat6[22] = { 2, 2, 4, 4, 6, 6, 8, 8, 10, 10, 12, 12,
+                                        14, 14, 16, 16, 18, 18, 20, 20, 0, 0 };
 
-static void init_bit_trees()
-{
-    init_bit_tree(cat1, 1);
-    init_bit_tree(cat2, 2);
-    init_bit_tree(cat3, 3);
-    init_bit_tree(cat4, 4);
-    init_bit_tree(cat5, 5);
-    init_bit_tree(cat6, 11);
-}
-
-vp8_extra_bit_struct vp8_extra_bits[12] =
+const vp8_extra_bit_struct vp8_extra_bits[12] =
 {
     { 0, 0, 0, 0},
     { 0, 0, 0, 1},
@@ -163,8 +187,3 @@ void vp8_default_coef_probs(VP8_COMMON *pc)
                    sizeof(default_coef_probs));
 }
 
-void vp8_coef_tree_initialize()
-{
-    init_bit_trees();
-    vp8_tokens_from_tree(vp8_coef_encodings, vp8_coef_tree);
-}

vp8/common/entropy.h

@@ -35,7 +35,7 @@
 
 extern const vp8_tree_index vp8_coef_tree[];
 
-extern struct vp8_token_struct vp8_coef_encodings[MAX_ENTROPY_TOKENS];
+extern const struct vp8_token_struct vp8_coef_encodings[MAX_ENTROPY_TOKENS];
 
 typedef struct
 {
@@ -45,7 +45,7 @@ typedef struct
     int base_val;
 } vp8_extra_bit_struct;
 
-extern vp8_extra_bit_struct vp8_extra_bits[12];    /* indexed by token value */
+extern const vp8_extra_bit_struct vp8_extra_bits[12];    /* indexed by token value */
 
 #define PROB_UPDATE_BASELINE_COST   7
 
@@ -94,7 +94,7 @@ void vp8_default_coef_probs(struct VP8Common *);
 
 extern DECLARE_ALIGNED(16, const int, vp8_default_zig_zag1d[16]);
 extern DECLARE_ALIGNED(16, const short, vp8_default_inv_zig_zag[16]);
-extern short vp8_default_zig_zag_mask[16];
+extern DECLARE_ALIGNED(16, const short, vp8_default_zig_zag_mask[16]);
 extern const int vp8_mb_feature_data_bits[MB_LVL_MAX];
 
 void vp8_coef_tree_initialize(void);

vp8/common/entropymode.c

@@ -8,22 +8,13 @@
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
+#define USE_PREBUILT_TABLES
 
 #include "entropymode.h"
 #include "entropy.h"
 #include "vpx_mem/vpx_mem.h"
 
-static const unsigned int kf_y_mode_cts[VP8_YMODES] = { 1607, 915, 812, 811, 5455};
-static const unsigned int y_mode_cts  [VP8_YMODES] = { 8080, 1908, 1582, 1007, 5874};
-
-static const unsigned int uv_mode_cts  [VP8_UV_MODES] = { 59483, 13605, 16492, 4230};
-static const unsigned int kf_uv_mode_cts[VP8_UV_MODES] = { 5319, 1904, 1703, 674};
-
-static const unsigned int bmode_cts[VP8_BINTRAMODES] =
-{
-    43891, 17694, 10036, 3920, 3363, 2546, 5119, 3221, 2471, 1723
-};
-
+#include "vp8_entropymodedata.h"
 
 int vp8_mv_cont(const int_mv *l, const int_mv *a)
 {
@@ -59,7 +50,7 @@ const vp8_prob vp8_sub_mv_ref_prob2 [SUBMVREF_COUNT][VP8_SUBMVREFS-1] =
 
 
 
-vp8_mbsplit vp8_mbsplits [VP8_NUMMBSPLITS] =
+const vp8_mbsplit vp8_mbsplits [VP8_NUMMBSPLITS] =
 {
     {
         0,  0,  0,  0,
@@ -84,7 +75,7 @@ vp8_mbsplit vp8_mbsplits [VP8_NUMMBSPLITS] =
         4,  5,  6,  7,
         8,  9,  10, 11,
         12, 13, 14, 15,
-    },
+    }
 };
 
 const int vp8_mbsplit_count [VP8_NUMMBSPLITS] = { 2, 2, 4, 16};
@@ -155,17 +146,6 @@ const vp8_tree_index vp8_sub_mv_ref_tree[6] =
     -ZERO4X4, -NEW4X4
 };
 
-
-struct vp8_token_struct vp8_bmode_encodings   [VP8_BINTRAMODES];
-struct vp8_token_struct vp8_ymode_encodings   [VP8_YMODES];
-struct vp8_token_struct vp8_kf_ymode_encodings [VP8_YMODES];
-struct vp8_token_struct vp8_uv_mode_encodings  [VP8_UV_MODES];
-struct vp8_token_struct vp8_mbsplit_encodings [VP8_NUMMBSPLITS];
-
-struct vp8_token_struct vp8_mv_ref_encoding_array    [VP8_MVREFS];
-struct vp8_token_struct vp8_sub_mv_ref_encoding_array [VP8_SUBMVREFS];
-
-
 const vp8_tree_index vp8_small_mvtree [14] =
 {
     2, 8,
@@ -177,89 +157,21 @@ const vp8_tree_index vp8_small_mvtree [14] =
     -6, -7
 };
 
-struct vp8_token_struct vp8_small_mvencodings [8];
-
 void vp8_init_mbmode_probs(VP8_COMMON *x)
 {
-    unsigned int bct [VP8_YMODES] [2];      /* num Ymodes > num UV modes */
-
-    vp8_tree_probs_from_distribution(
-        VP8_YMODES, vp8_ymode_encodings, vp8_ymode_tree,
-        x->fc.ymode_prob, bct, y_mode_cts,
-        256, 1
-    );
-    vp8_tree_probs_from_distribution(
-        VP8_YMODES, vp8_kf_ymode_encodings, vp8_kf_ymode_tree,
-        x->kf_ymode_prob, bct, kf_y_mode_cts,
-        256, 1
-    );
-    vp8_tree_probs_from_distribution(
-        VP8_UV_MODES, vp8_uv_mode_encodings, vp8_uv_mode_tree,
-        x->fc.uv_mode_prob, bct, uv_mode_cts,
-        256, 1
-    );
-    vp8_tree_probs_from_distribution(
-        VP8_UV_MODES, vp8_uv_mode_encodings, vp8_uv_mode_tree,
-        x->kf_uv_mode_prob, bct, kf_uv_mode_cts,
-        256, 1
-    );
+    vpx_memcpy(x->fc.ymode_prob, vp8_ymode_prob, sizeof(vp8_ymode_prob));
+    vpx_memcpy(x->kf_ymode_prob, vp8_kf_ymode_prob, sizeof(vp8_kf_ymode_prob));
+    vpx_memcpy(x->fc.uv_mode_prob, vp8_uv_mode_prob, sizeof(vp8_uv_mode_prob));
+    vpx_memcpy(x->kf_uv_mode_prob, vp8_kf_uv_mode_prob, sizeof(vp8_kf_uv_mode_prob));
     vpx_memcpy(x->fc.sub_mv_ref_prob, sub_mv_ref_prob, sizeof(sub_mv_ref_prob));
 }
 
-
-static void intra_bmode_probs_from_distribution(
-    vp8_prob p [VP8_BINTRAMODES-1],
-    unsigned int branch_ct [VP8_BINTRAMODES-1] [2],
-    const unsigned int events [VP8_BINTRAMODES]
-)
-{
-    vp8_tree_probs_from_distribution(
-        VP8_BINTRAMODES, vp8_bmode_encodings, vp8_bmode_tree,
-        p, branch_ct, events,
-        256, 1
-    );
-}
-
 void vp8_default_bmode_probs(vp8_prob p [VP8_BINTRAMODES-1])
 {
-    unsigned int branch_ct [VP8_BINTRAMODES-1] [2];
-    intra_bmode_probs_from_distribution(p, branch_ct, bmode_cts);
+    vpx_memcpy(p, vp8_bmode_prob, sizeof(vp8_bmode_prob));
 }
 
 void vp8_kf_default_bmode_probs(vp8_prob p [VP8_BINTRAMODES] [VP8_BINTRAMODES] [VP8_BINTRAMODES-1])
 {
-    unsigned int branch_ct [VP8_BINTRAMODES-1] [2];
-
-    int i = 0;
-
-    do
-    {
-        int j = 0;
-
-        do
-        {
-            intra_bmode_probs_from_distribution(
-                p[i][j], branch_ct, vp8_kf_default_bmode_counts[i][j]);
-
-        }
-        while (++j < VP8_BINTRAMODES);
-    }
-    while (++i < VP8_BINTRAMODES);
-}
-
-
-void vp8_entropy_mode_init()
-{
-    vp8_tokens_from_tree(vp8_bmode_encodings,   vp8_bmode_tree);
-    vp8_tokens_from_tree(vp8_ymode_encodings,   vp8_ymode_tree);
-    vp8_tokens_from_tree(vp8_kf_ymode_encodings, vp8_kf_ymode_tree);
-    vp8_tokens_from_tree(vp8_uv_mode_encodings,  vp8_uv_mode_tree);
-    vp8_tokens_from_tree(vp8_mbsplit_encodings, vp8_mbsplit_tree);
-
-    vp8_tokens_from_tree_offset(vp8_mv_ref_encoding_array,
-                                vp8_mv_ref_tree, NEARESTMV);
-    vp8_tokens_from_tree_offset(vp8_sub_mv_ref_encoding_array,
-                                vp8_sub_mv_ref_tree, LEFT4X4);
-
-    vp8_tokens_from_tree(vp8_small_mvencodings, vp8_small_mvtree);
+    vpx_memcpy(p, vp8_kf_bmode_prob, sizeof(vp8_kf_bmode_prob));
 }

vp8/common/entropymode.h

@@ -52,22 +52,20 @@ extern const vp8_tree_index  vp8_mbsplit_tree[];
 extern const vp8_tree_index  vp8_mv_ref_tree[];
 extern const vp8_tree_index  vp8_sub_mv_ref_tree[];
 
-extern struct vp8_token_struct vp8_bmode_encodings   [VP8_BINTRAMODES];
-extern struct vp8_token_struct vp8_ymode_encodings   [VP8_YMODES];
-extern struct vp8_token_struct vp8_kf_ymode_encodings [VP8_YMODES];
-extern struct vp8_token_struct vp8_uv_mode_encodings  [VP8_UV_MODES];
-extern struct vp8_token_struct vp8_mbsplit_encodings  [VP8_NUMMBSPLITS];
+extern const struct vp8_token_struct vp8_bmode_encodings[VP8_BINTRAMODES];
+extern const struct vp8_token_struct vp8_ymode_encodings[VP8_YMODES];
+extern const struct vp8_token_struct vp8_kf_ymode_encodings[VP8_YMODES];
+extern const struct vp8_token_struct vp8_uv_mode_encodings[VP8_UV_MODES];
+extern const struct vp8_token_struct vp8_mbsplit_encodings[VP8_NUMMBSPLITS];
 
 /* Inter mode values do not start at zero */
 
-extern struct vp8_token_struct vp8_mv_ref_encoding_array    [VP8_MVREFS];
-extern struct vp8_token_struct vp8_sub_mv_ref_encoding_array [VP8_SUBMVREFS];
+extern const struct vp8_token_struct vp8_mv_ref_encoding_array[VP8_MVREFS];
+extern const struct vp8_token_struct vp8_sub_mv_ref_encoding_array[VP8_SUBMVREFS];
 
 extern const vp8_tree_index vp8_small_mvtree[];
 
-extern struct vp8_token_struct vp8_small_mvencodings [8];
-
-void vp8_entropy_mode_init(void);
+extern const struct vp8_token_struct vp8_small_mvencodings[8];
 
 void vp8_init_mbmode_probs(VP8_COMMON *x);
 

vp8/common/filter.c

@@ -149,7 +149,7 @@ static void filter_block2d
 }
 
 
-void vp8_sixtap_predict_c
+void vp8_sixtap_predict4x4_c
 (
     unsigned char  *src_ptr,
     int   src_pixels_per_line,

vp8/common/findnearmv.c

@@ -134,13 +134,51 @@ void vp8_find_near_mvs
     best_mv->as_int = near_mvs[0].as_int;
     nearest->as_int = near_mvs[CNT_NEAREST].as_int;
     nearby->as_int = near_mvs[CNT_NEAR].as_int;
-
-    //TODO: move clamp outside findnearmv
-    vp8_clamp_mv2(nearest, xd);
-    vp8_clamp_mv2(nearby, xd);
-    vp8_clamp_mv2(best_mv, xd);
 }
 
+
+static void invert_and_clamp_mvs(int_mv *inv, int_mv *src, MACROBLOCKD *xd)
+{
+    inv->as_mv.row = src->as_mv.row * -1;
+    inv->as_mv.col = src->as_mv.col * -1;
+    vp8_clamp_mv2(inv, xd);
+    vp8_clamp_mv2(src, xd);
+}
+
+
+int vp8_find_near_mvs_bias
+(
+    MACROBLOCKD *xd,
+    const MODE_INFO *here,
+    int_mv mode_mv_sb[2][MB_MODE_COUNT],
+    int_mv best_mv_sb[2],
+    int cnt[4],
+    int refframe,
+    int *ref_frame_sign_bias
+)
+{
+    int sign_bias = ref_frame_sign_bias[refframe];
+
+    vp8_find_near_mvs(xd,
+                      here,
+                      &mode_mv_sb[sign_bias][NEARESTMV],
+                      &mode_mv_sb[sign_bias][NEARMV],
+                      &best_mv_sb[sign_bias],
+                      cnt,
+                      refframe,
+                      ref_frame_sign_bias);
+
+    invert_and_clamp_mvs(&mode_mv_sb[!sign_bias][NEARESTMV],
+                         &mode_mv_sb[sign_bias][NEARESTMV], xd);
+    invert_and_clamp_mvs(&mode_mv_sb[!sign_bias][NEARMV],
+                         &mode_mv_sb[sign_bias][NEARMV], xd);
+    invert_and_clamp_mvs(&best_mv_sb[!sign_bias],
+                         &best_mv_sb[sign_bias], xd);
+
+    return sign_bias;
+}
+
+
 vp8_prob *vp8_mv_ref_probs(
     vp8_prob p[VP8_MVREFS-1], const int near_mv_ref_ct[4]
 )

vp8/common/findnearmv.h

@@ -60,10 +60,10 @@ static unsigned int vp8_check_mv_bounds(int_mv *mv, int mb_to_left_edge,
                                 int mb_to_bottom_edge)
 {
     unsigned int need_to_clamp;
-    need_to_clamp = (mv->as_mv.col < mb_to_left_edge) ? 1 : 0;
-    need_to_clamp |= (mv->as_mv.col > mb_to_right_edge) ? 1 : 0;
-    need_to_clamp |= (mv->as_mv.row < mb_to_top_edge) ? 1 : 0;
-    need_to_clamp |= (mv->as_mv.row > mb_to_bottom_edge) ? 1 : 0;
+    need_to_clamp = (mv->as_mv.col < mb_to_left_edge);
+    need_to_clamp |= (mv->as_mv.col > mb_to_right_edge);
+    need_to_clamp |= (mv->as_mv.row < mb_to_top_edge);
+    need_to_clamp |= (mv->as_mv.row > mb_to_bottom_edge);
     return need_to_clamp;
 }
 
@@ -77,6 +77,19 @@ void vp8_find_near_mvs
     int *ref_frame_sign_bias
 );
 
+
+int vp8_find_near_mvs_bias
+(
+    MACROBLOCKD *xd,
+    const MODE_INFO *here,
+    int_mv mode_mv_sb[2][MB_MODE_COUNT],
+    int_mv best_mv_sb[2],
+    int cnt[4],
+    int refframe,
+    int *ref_frame_sign_bias
+);
+
+
 vp8_prob *vp8_mv_ref_probs(
     vp8_prob p[VP8_MVREFS-1], const int near_mv_ref_ct[4]
 );

vp8/common/g_common.h

@@ -1,21 +0,0 @@
-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-
-extern void (*vp8_clear_system_state)(void);
-extern void (*vp8_plane_add_noise)(unsigned char *Start, unsigned int Width, unsigned int Height, int Pitch, int DPitch, int q);
-extern void (*de_interlace)
-(
-    unsigned char *src_ptr,
-    unsigned char *dst_ptr,
-    int Width,
-    int Height,
-    int Stride
-);

vp8/common/generic/systemdependent.c

@@ -10,31 +10,33 @@
 
 
 #include "vpx_config.h"
-#include "vp8/common/g_common.h"
-#include "vp8/common/subpixel.h"
-#include "vp8/common/loopfilter.h"
-#include "vp8/common/recon.h"
-#include "vp8/common/idct.h"
+#include "vpx_rtcd.h"
+#if ARCH_ARM
+#include "vpx_ports/arm.h"
+#elif ARCH_X86 || ARCH_X86_64
+#include "vpx_ports/x86.h"
+#endif
 #include "vp8/common/onyxc_int.h"
 
 #if CONFIG_MULTITHREAD
-#if HAVE_UNISTD_H
+#if HAVE_UNISTD_H && !defined(__OS2__)
 #include <unistd.h>
 #elif defined(_WIN32)
 #include <windows.h>
 typedef void (WINAPI *PGNSI)(LPSYSTEM_INFO);
+#elif defined(__OS2__)
+#define INCL_DOS
+#define INCL_DOSSPINLOCK
+#include <os2.h>
 #endif
 #endif
 
-extern void vp8_arch_x86_common_init(VP8_COMMON *ctx);
-extern void vp8_arch_arm_common_init(VP8_COMMON *ctx);
-
 #if CONFIG_MULTITHREAD
 static int get_cpu_count()
 {
     int core_count = 16;
 
-#if HAVE_UNISTD_H
+#if HAVE_UNISTD_H && !defined(__OS2__)
 #if defined(_SC_NPROCESSORS_ONLN)
     core_count = sysconf(_SC_NPROCESSORS_ONLN);
 #elif defined(_SC_NPROC_ONLN)
@@ -57,6 +59,21 @@ static int get_cpu_count()
 
         core_count = sysinfo.dwNumberOfProcessors;
     }
+#elif defined(__OS2__)
+    {
+        ULONG proc_id;
+        ULONG status;
+
+        core_count = 0;
+        for (proc_id = 1; ; proc_id++)
+        {
+            if (DosGetProcessorStatus(proc_id, &status))
+                break;
+
+            if (status == PROC_ONLINE)
+                core_count++;
+        }
+    }
 #else
     /* other platforms */
 #endif
@@ -65,70 +82,69 @@ static int get_cpu_count()
 }
 #endif
 
+
+#if HAVE_PTHREAD_H
+#include <pthread.h>
+static void once(void (*func)(void))
+{
+    static pthread_once_t lock = PTHREAD_ONCE_INIT;
+    pthread_once(&lock, func);
+}
+
+
+#elif defined(_WIN32)
+static void once(void (*func)(void))
+{
+    /* Using a static initializer here rather than InitializeCriticalSection()
+     * since there's no race-free context in which to execute it. Protecting
+     * it with an atomic op like InterlockedCompareExchangePointer introduces
+     * an x86 dependency, and InitOnceExecuteOnce requires Vista.
+     */
+    static CRITICAL_SECTION lock = {(void *)-1, -1, 0, 0, 0, 0};
+    static int done;
+
+    EnterCriticalSection(&lock);
+
+    if (!done)
+    {
+        func();
+        done = 1;
+    }
+
+    LeaveCriticalSection(&lock);
+}
+
+
+#else
+/* No-op version that performs no synchronization. vpx_rtcd() is idempotent,
+ * so as long as your platform provides atomic loads/stores of pointers
+ * no synchronization is strictly necessary.
+ */
+
+static void once(void (*func)(void))
+{
+    static int done;
+
+    if(!done)
+    {
+        func();
+        done = 1;
+    }
+}
+#endif
+
+
 void vp8_machine_specific_config(VP8_COMMON *ctx)
 {
-#if CONFIG_RUNTIME_CPU_DETECT
-    VP8_COMMON_RTCD *rtcd = &ctx->rtcd;
-
-    rtcd->idct.idct16       = vp8_short_idct4x4llm_c;
-    rtcd->idct.idct1_scalar_add = vp8_dc_only_idct_add_c;
-    rtcd->idct.iwalsh1      = vp8_short_inv_walsh4x4_1_c;
-    rtcd->idct.iwalsh16     = vp8_short_inv_walsh4x4_c;
-
-    rtcd->recon.copy16x16   = vp8_copy_mem16x16_c;
-    rtcd->recon.copy8x8     = vp8_copy_mem8x8_c;
-    rtcd->recon.copy8x4     = vp8_copy_mem8x4_c;
-
-    rtcd->recon.build_intra_predictors_mby =
-        vp8_build_intra_predictors_mby;
-    rtcd->recon.build_intra_predictors_mby_s =
-        vp8_build_intra_predictors_mby_s;
-    rtcd->recon.build_intra_predictors_mbuv =
-        vp8_build_intra_predictors_mbuv;
-    rtcd->recon.build_intra_predictors_mbuv_s =
-        vp8_build_intra_predictors_mbuv_s;
-    rtcd->recon.intra4x4_predict =
-        vp8_intra4x4_predict_c;
-
-    rtcd->subpix.sixtap16x16   = vp8_sixtap_predict16x16_c;
-    rtcd->subpix.sixtap8x8     = vp8_sixtap_predict8x8_c;
-    rtcd->subpix.sixtap8x4     = vp8_sixtap_predict8x4_c;
-    rtcd->subpix.sixtap4x4     = vp8_sixtap_predict_c;
-    rtcd->subpix.bilinear16x16 = vp8_bilinear_predict16x16_c;
-    rtcd->subpix.bilinear8x8   = vp8_bilinear_predict8x8_c;
-    rtcd->subpix.bilinear8x4   = vp8_bilinear_predict8x4_c;
-    rtcd->subpix.bilinear4x4   = vp8_bilinear_predict4x4_c;
-
-    rtcd->loopfilter.normal_mb_v = vp8_loop_filter_mbv_c;
-    rtcd->loopfilter.normal_b_v  = vp8_loop_filter_bv_c;
-    rtcd->loopfilter.normal_mb_h = vp8_loop_filter_mbh_c;
-    rtcd->loopfilter.normal_b_h  = vp8_loop_filter_bh_c;
-    rtcd->loopfilter.simple_mb_v = vp8_loop_filter_simple_vertical_edge_c;
-    rtcd->loopfilter.simple_b_v  = vp8_loop_filter_bvs_c;
-    rtcd->loopfilter.simple_mb_h = vp8_loop_filter_simple_horizontal_edge_c;
-    rtcd->loopfilter.simple_b_h  = vp8_loop_filter_bhs_c;
-
-#if CONFIG_POSTPROC || (CONFIG_VP8_ENCODER && CONFIG_INTERNAL_STATS)
-    rtcd->postproc.down             = vp8_mbpost_proc_down_c;
-    rtcd->postproc.across           = vp8_mbpost_proc_across_ip_c;
-    rtcd->postproc.downacross       = vp8_post_proc_down_and_across_c;
-    rtcd->postproc.addnoise         = vp8_plane_add_noise_c;
-    rtcd->postproc.blend_mb_inner   = vp8_blend_mb_inner_c;
-    rtcd->postproc.blend_mb_outer   = vp8_blend_mb_outer_c;
-    rtcd->postproc.blend_b          = vp8_blend_b_c;
-#endif
-
-#endif
-
-#if ARCH_X86 || ARCH_X86_64
-    vp8_arch_x86_common_init(ctx);
-#endif
-
-#if ARCH_ARM
-    vp8_arch_arm_common_init(ctx);
-#endif
-
 #if CONFIG_MULTITHREAD
     ctx->processor_core_count = get_cpu_count();
 #endif /* CONFIG_MULTITHREAD */
+
+#if ARCH_ARM
+    ctx->cpu_caps = arm_cpu_caps();
+#elif ARCH_X86 || ARCH_X86_64
+    ctx->cpu_caps = x86_simd_caps();
+#endif
+
+    once(vpx_rtcd);
 }

vp8/common/idct.h

@@ -1,80 +0,0 @@
-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-
-#ifndef __INC_IDCT_H
-#define __INC_IDCT_H
-
-#define prototype_second_order(sym) \
-    void sym(short *input, short *output)
-
-#define prototype_idct(sym) \
-    void sym(short *input, unsigned char *pred, int pitch, unsigned char *dst, \
-             int dst_stride)
-
-#define prototype_idct_scalar_add(sym) \
-    void sym(short input, \
-            unsigned char *pred, int pred_stride, \
-            unsigned char *dst, \
-            int dst_stride)
-
-#if ARCH_X86 || ARCH_X86_64
-#include "x86/idct_x86.h"
-#endif
-
-#if ARCH_ARM
-#include "arm/idct_arm.h"
-#endif
-
-#ifndef vp8_idct_idct16
-#define vp8_idct_idct16 vp8_short_idct4x4llm_c
-#endif
-extern prototype_idct(vp8_idct_idct16);
-/* add this prototype to prevent compiler warning about implicit
- * declaration of vp8_short_idct4x4llm_c function in dequantize.c
- * when building, for example, neon optimized version */
-extern prototype_idct(vp8_short_idct4x4llm_c);
-
-#ifndef vp8_idct_idct1_scalar_add
-#define vp8_idct_idct1_scalar_add vp8_dc_only_idct_add_c
-#endif
-extern prototype_idct_scalar_add(vp8_idct_idct1_scalar_add);
-
-
-#ifndef vp8_idct_iwalsh1
-#define vp8_idct_iwalsh1 vp8_short_inv_walsh4x4_1_c
-#endif
-extern prototype_second_order(vp8_idct_iwalsh1);
-
-#ifndef vp8_idct_iwalsh16
-#define vp8_idct_iwalsh16 vp8_short_inv_walsh4x4_c
-#endif
-extern prototype_second_order(vp8_idct_iwalsh16);
-
-typedef prototype_idct((*vp8_idct_fn_t));
-typedef prototype_idct_scalar_add((*vp8_idct_scalar_add_fn_t));
-typedef prototype_second_order((*vp8_second_order_fn_t));
-
-typedef struct
-{
-    vp8_idct_fn_t            idct16;
-    vp8_idct_scalar_add_fn_t idct1_scalar_add;
-
-    vp8_second_order_fn_t iwalsh1;
-    vp8_second_order_fn_t iwalsh16;
-} vp8_idct_rtcd_vtable_t;
-
-#if CONFIG_RUNTIME_CPU_DETECT
-#define IDCT_INVOKE(ctx,fn) (ctx)->fn
-#else
-#define IDCT_INVOKE(ctx,fn) vp8_idct_##fn
-#endif
-
-#endif

vp8/common/idct_blk.c

@@ -9,8 +9,7 @@
  */
 
 #include "vpx_config.h"
-#include "vp8/common/idct.h"
-#include "dequantize.h"
+#include "vpx_rtcd.h"
 
 void vp8_dequant_idct_add_c(short *input, short *dq,
                             unsigned char *dest, int stride);

vp8/common/idctllm_test.cc

@@ -0,0 +1,31 @@
+/*
+ *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
+ *
+ *  Use of this source code is governed by a BSD-style license
+ *  that can be found in the LICENSE file in the root of the source
+ *  tree. An additional intellectual property rights grant can be found
+ *  in the file PATENTS.  All contributing project authors may
+ *  be found in the AUTHORS file in the root of the source tree.
+ */
+
+
+ extern "C" {
+    void vp8_short_idct4x4llm_c(short *input, unsigned char *pred_ptr,
+                            int pred_stride, unsigned char *dst_ptr,
+                            int dst_stride);
+}
+
+#include "vpx_config.h"
+#include "idctllm_test.h"
+namespace
+{
+
+INSTANTIATE_TEST_CASE_P(C, IDCTTest,
+                        ::testing::Values(vp8_short_idct4x4llm_c));
+
+} // namespace
+
+int main(int argc, char **argv) {
+  ::testing::InitGoogleTest(&argc, argv);
+  return RUN_ALL_TESTS();
+}

vp8/common/idctllm_test.h

@@ -0,0 +1,113 @@
+/*
+ *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
+ *
+ *  Use of this source code is governed by a BSD-style license
+ *  that can be found in the LICENSE file in the root of the source
+ *  tree. An additional intellectual property rights grant can be found
+ *  in the file PATENTS.  All contributing project authors may
+ *  be found in the AUTHORS file in the root of the source tree.
+ */
+
+
+ #include "third_party/googletest/src/include/gtest/gtest.h"
+typedef void (*idct_fn_t)(short *input, unsigned char *pred_ptr,
+                            int pred_stride, unsigned char *dst_ptr,
+                            int dst_stride);
+namespace {
+class IDCTTest : public ::testing::TestWithParam<idct_fn_t>
+{
+  protected:
+    virtual void SetUp()
+    {
+        int i;
+
+        UUT = GetParam();
+        memset(input, 0, sizeof(input));
+        /* Set up guard blocks */
+        for(i=0; i<256; i++)
+            output[i] = ((i&0xF)<4&&(i<64))?0:-1;
+    }
+
+    idct_fn_t UUT;
+    short input[16];
+    unsigned char output[256];
+    unsigned char predict[256];
+};
+
+TEST_P(IDCTTest, TestGuardBlocks)
+{
+    int i;
+
+    for(i=0; i<256; i++)
+        if((i&0xF) < 4 && i<64)
+            EXPECT_EQ(0, output[i]) << i;
+        else
+            EXPECT_EQ(255, output[i]);
+}
+
+TEST_P(IDCTTest, TestAllZeros)
+{
+    int i;
+
+    UUT(input, output, 16, output, 16);
+
+    for(i=0; i<256; i++)
+        if((i&0xF) < 4 && i<64)
+            EXPECT_EQ(0, output[i]) << "i==" << i;
+        else
+            EXPECT_EQ(255, output[i]) << "i==" << i;
+}
+
+TEST_P(IDCTTest, TestAllOnes)
+{
+    int i;
+
+    input[0] = 4;
+    UUT(input, output, 16, output, 16);
+
+    for(i=0; i<256; i++)
+        if((i&0xF) < 4 && i<64)
+            EXPECT_EQ(1, output[i]) << "i==" << i;
+        else
+            EXPECT_EQ(255, output[i]) << "i==" << i;
+}
+
+TEST_P(IDCTTest, TestAddOne)
+{
+    int i;
+
+    for(i=0; i<256; i++)
+        predict[i] = i;
+
+    input[0] = 4;
+    UUT(input, predict, 16, output, 16);
+
+    for(i=0; i<256; i++)
+        if((i&0xF) < 4 && i<64)
+            EXPECT_EQ(i+1, output[i]) << "i==" << i;
+        else
+            EXPECT_EQ(255, output[i]) << "i==" << i;
+}
+
+TEST_P(IDCTTest, TestWithData)
+{
+    int i;
+
+    for(i=0; i<16; i++)
+        input[i] = i;
+
+    UUT(input, output, 16, output, 16);
+
+    for(i=0; i<256; i++)
+        if((i&0xF) > 3 || i>63)
+            EXPECT_EQ(255, output[i]) << "i==" << i;
+        else if(i == 0)
+            EXPECT_EQ(11, output[i]) << "i==" << i;
+        else if(i == 34)
+            EXPECT_EQ(1, output[i]) << "i==" << i;
+        else if(i == 2 || i == 17 || i == 32)
+            EXPECT_EQ(3, output[i]) << "i==" << i;
+        else
+            EXPECT_EQ(0, output[i]) << "i==" << i;
+}
+}

vp8/common/invtrans.c

@@ -1,56 +0,0 @@
-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-
-#include "invtrans.h"
-
-
-void vp8_inverse_transform_b(const vp8_idct_rtcd_vtable_t *rtcd, BLOCKD *b,
-                             int pitch)
-{
-    if (*b->eob > 1)
-    {
-        IDCT_INVOKE(rtcd, idct16)(b->dqcoeff, b->predictor, pitch,
-              *(b->base_dst) + b->dst, b->dst_stride);
-    }
-    else
-    {
-        IDCT_INVOKE(rtcd, idct1_scalar_add)(b->dqcoeff[0], b->predictor, pitch,
-                         *(b->base_dst) + b->dst, b->dst_stride);
-    }
-
-}
-
-void vp8_inverse_transform_mby(const vp8_idct_rtcd_vtable_t *rtcd, MACROBLOCKD *x)
-{
-    int i;
-
-    if(x->mode_info_context->mbmi.mode != SPLITMV)
-    {
-        /* do 2nd order transform on the dc block */
-        IDCT_INVOKE(rtcd, iwalsh16)(x->block[24].dqcoeff, x->dqcoeff);
-    }
-
-    for (i = 0; i < 16; i++)
-    {
-        vp8_inverse_transform_b(rtcd, &x->block[i], 16);
-    }
-
-}
-void vp8_inverse_transform_mbuv(const vp8_idct_rtcd_vtable_t *rtcd, MACROBLOCKD *x)
-{
-    int i;
-
-    for (i = 16; i < 24; i++)
-    {
-        vp8_inverse_transform_b(rtcd, &x->block[i], 8);
-    }
-
-}

vp8/common/invtrans.h

@@ -13,11 +13,50 @@
 #define __INC_INVTRANS_H
 
 #include "vpx_config.h"
-#include "idct.h"
+#include "vpx_rtcd.h"
 #include "blockd.h"
-extern void vp8_inverse_transform_b(const vp8_idct_rtcd_vtable_t *rtcd, BLOCKD *b, int pitch);
-extern void vp8_inverse_transform_mb(const vp8_idct_rtcd_vtable_t *rtcd, MACROBLOCKD *x);
-extern void vp8_inverse_transform_mby(const vp8_idct_rtcd_vtable_t *rtcd, MACROBLOCKD *x);
-extern void vp8_inverse_transform_mbuv(const vp8_idct_rtcd_vtable_t *rtcd, MACROBLOCKD *x);
+#include "onyxc_int.h"
 
+#if CONFIG_MULTITHREAD
+#include "vpx_mem/vpx_mem.h"
+#endif
+
+static void eob_adjust(char *eobs, short *diff)
+{
+    /* eob adjust.... the idct can only skip if both the dc and eob are zero */
+    int js;
+    for(js = 0; js < 16; js++)
+    {
+        if((eobs[js] == 0) && (diff[0] != 0))
+            eobs[js]++;
+        diff+=16;
+    }
+}
+
+static void vp8_inverse_transform_mby(MACROBLOCKD *xd)
+{
+    short *DQC = xd->dequant_y1;
+
+    if (xd->mode_info_context->mbmi.mode != SPLITMV)
+    {
+        /* do 2nd order transform on the dc block */
+        if (xd->eobs[24] > 1)
+        {
+            vp8_short_inv_walsh4x4
+                (&xd->block[24].dqcoeff[0], xd->qcoeff);
+        }
+        else
+        {
+            vp8_short_inv_walsh4x4_1
+                (&xd->block[24].dqcoeff[0], xd->qcoeff);
+        }
+        eob_adjust(xd->eobs, xd->qcoeff);
+
+        DQC = xd->dequant_y1_dc;
+    }
+    vp8_dequant_idct_add_y_block
+                    (xd->qcoeff, DQC,
+                     xd->dst.y_buffer,
+                     xd->dst.y_stride, xd->eobs);
+}
 #endif

vp8/common/littlend.h

@@ -1,33 +0,0 @@
-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-
-#ifndef _littlend_h
-#define _littlend_h
-
-#if defined(__cplusplus)
-extern "C" {
-#endif
-
-#define invert2(x) (x)
-#define invert4(x) (x)
-
-#define low_byte(x) (unsigned char)x
-#define mid1Byte(x) (unsigned char)(x >> 8)
-#define mid2Byte(x) (unsigned char)(x >> 16)
-#define high_byte(x) (unsigned char)(x >> 24)
-
-#define SWAPENDS 0
-
-#if defined(__cplusplus)
-}
-#endif
-
-#endif

vp8/common/loopfilter.c

@@ -10,96 +10,13 @@
 
 
 #include "vpx_config.h"
+#include "vpx_rtcd.h"
 #include "loopfilter.h"
 #include "onyxc_int.h"
 #include "vpx_mem/vpx_mem.h"
 
 typedef unsigned char uc;
 
-prototype_loopfilter(vp8_loop_filter_horizontal_edge_c);
-prototype_loopfilter(vp8_loop_filter_vertical_edge_c);
-prototype_loopfilter(vp8_mbloop_filter_horizontal_edge_c);
-prototype_loopfilter(vp8_mbloop_filter_vertical_edge_c);
-
-prototype_simple_loopfilter(vp8_loop_filter_simple_horizontal_edge_c);
-prototype_simple_loopfilter(vp8_loop_filter_simple_vertical_edge_c);
-
-/* Horizontal MB filtering */
-void vp8_loop_filter_mbh_c(unsigned char *y_ptr, unsigned char *u_ptr,
-                           unsigned char *v_ptr, int y_stride, int uv_stride,
-                           loop_filter_info *lfi)
-{
-    vp8_mbloop_filter_horizontal_edge_c(y_ptr, y_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 2);
-
-    if (u_ptr)
-        vp8_mbloop_filter_horizontal_edge_c(u_ptr, uv_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 1);
-
-    if (v_ptr)
-        vp8_mbloop_filter_horizontal_edge_c(v_ptr, uv_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 1);
-}
-
-/* Vertical MB Filtering */
-void vp8_loop_filter_mbv_c(unsigned char *y_ptr, unsigned char *u_ptr,
-                           unsigned char *v_ptr, int y_stride, int uv_stride,
-                           loop_filter_info *lfi)
-{
-    vp8_mbloop_filter_vertical_edge_c(y_ptr, y_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 2);
-
-    if (u_ptr)
-        vp8_mbloop_filter_vertical_edge_c(u_ptr, uv_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 1);
-
-    if (v_ptr)
-        vp8_mbloop_filter_vertical_edge_c(v_ptr, uv_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 1);
-}
-
-/* Horizontal B Filtering */
-void vp8_loop_filter_bh_c(unsigned char *y_ptr, unsigned char *u_ptr,
-                          unsigned char *v_ptr, int y_stride, int uv_stride,
-                          loop_filter_info *lfi)
-{
-    vp8_loop_filter_horizontal_edge_c(y_ptr + 4 * y_stride, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
-    vp8_loop_filter_horizontal_edge_c(y_ptr + 8 * y_stride, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
-    vp8_loop_filter_horizontal_edge_c(y_ptr + 12 * y_stride, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
-
-    if (u_ptr)
-        vp8_loop_filter_horizontal_edge_c(u_ptr + 4 * uv_stride, uv_stride, lfi->blim, lfi->lim, lfi->hev_thr, 1);
-
-    if (v_ptr)
-        vp8_loop_filter_horizontal_edge_c(v_ptr + 4 * uv_stride, uv_stride, lfi->blim, lfi->lim, lfi->hev_thr, 1);
-}
-
-void vp8_loop_filter_bhs_c(unsigned char *y_ptr, int y_stride,
-                           const unsigned char *blimit)
-{
-    vp8_loop_filter_simple_horizontal_edge_c(y_ptr + 4 * y_stride, y_stride, blimit);
-    vp8_loop_filter_simple_horizontal_edge_c(y_ptr + 8 * y_stride, y_stride, blimit);
-    vp8_loop_filter_simple_horizontal_edge_c(y_ptr + 12 * y_stride, y_stride, blimit);
-}
-
-/* Vertical B Filtering */
-void vp8_loop_filter_bv_c(unsigned char *y_ptr, unsigned char *u_ptr,
-                          unsigned char *v_ptr, int y_stride, int uv_stride,
-                          loop_filter_info *lfi)
-{
-    vp8_loop_filter_vertical_edge_c(y_ptr + 4, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
-    vp8_loop_filter_vertical_edge_c(y_ptr + 8, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
-    vp8_loop_filter_vertical_edge_c(y_ptr + 12, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
-
-    if (u_ptr)
-        vp8_loop_filter_vertical_edge_c(u_ptr + 4, uv_stride, lfi->blim, lfi->lim, lfi->hev_thr, 1);
-
-    if (v_ptr)
-        vp8_loop_filter_vertical_edge_c(v_ptr + 4, uv_stride, lfi->blim, lfi->lim, lfi->hev_thr, 1);
-}
-
-void vp8_loop_filter_bvs_c(unsigned char *y_ptr, int y_stride,
-                           const unsigned char *blimit)
-{
-    vp8_loop_filter_simple_vertical_edge_c(y_ptr + 4, y_stride, blimit);
-    vp8_loop_filter_simple_vertical_edge_c(y_ptr + 8, y_stride, blimit);
-    vp8_loop_filter_simple_vertical_edge_c(y_ptr + 12, y_stride, blimit);
-}
-
 static void lf_init_lut(loop_filter_info_n *lfi)
 {
     int filt_lvl;
@@ -293,6 +210,8 @@ void vp8_loop_filter_frame
 
     int mb_row;
     int mb_col;
+    int mb_rows = cm->mb_rows;
+    int mb_cols = cm->mb_cols;
 
     int filter_level;
 
@@ -300,6 +219,8 @@ void vp8_loop_filter_frame
 
     /* Point at base of Mb MODE_INFO list */
     const MODE_INFO *mode_info_context = cm->mi;
+    int post_y_stride = post->y_stride;
+    int post_uv_stride = post->uv_stride;
 
     /* Initialize the loop filter for this frame. */
     vp8_loop_filter_frame_init(cm, mbd, cm->filter_level);
@@ -310,23 +231,23 @@ void vp8_loop_filter_frame
     v_ptr = post->v_buffer;
 
     /* vp8_filter each macro block */
-    for (mb_row = 0; mb_row < cm->mb_rows; mb_row++)
+    if (cm->filter_type == NORMAL_LOOPFILTER)
     {
-        for (mb_col = 0; mb_col < cm->mb_cols; mb_col++)
+        for (mb_row = 0; mb_row < mb_rows; mb_row++)
         {
-            int skip_lf = (mode_info_context->mbmi.mode != B_PRED &&
-                            mode_info_context->mbmi.mode != SPLITMV &&
-                            mode_info_context->mbmi.mb_skip_coeff);
-
-            const int mode_index = lfi_n->mode_lf_lut[mode_info_context->mbmi.mode];
-            const int seg = mode_info_context->mbmi.segment_id;
-            const int ref_frame = mode_info_context->mbmi.ref_frame;
-
-            filter_level = lfi_n->lvl[seg][ref_frame][mode_index];
-
-            if (filter_level)
+            for (mb_col = 0; mb_col < mb_cols; mb_col++)
             {
-                if (cm->filter_type == NORMAL_LOOPFILTER)
+                int skip_lf = (mode_info_context->mbmi.mode != B_PRED &&
+                                mode_info_context->mbmi.mode != SPLITMV &&
+                                mode_info_context->mbmi.mb_skip_coeff);
+
+                const int mode_index = lfi_n->mode_lf_lut[mode_info_context->mbmi.mode];
+                const int seg = mode_info_context->mbmi.segment_id;
+                const int ref_frame = mode_info_context->mbmi.ref_frame;
+
+                filter_level = lfi_n->lvl[seg][ref_frame][mode_index];
+
+                if (filter_level)
                 {
                     const int hev_index = lfi_n->hev_thr_lut[frame_type][filter_level];
                     lfi.mblim = lfi_n->mblim[filter_level];
@@ -335,55 +256,88 @@ void vp8_loop_filter_frame
                     lfi.hev_thr = lfi_n->hev_thr[hev_index];
 
                     if (mb_col > 0)
-                        LF_INVOKE(&cm->rtcd.loopfilter, normal_mb_v)
-                        (y_ptr, u_ptr, v_ptr, post->y_stride, post->uv_stride, &lfi);
+                        vp8_loop_filter_mbv
+                        (y_ptr, u_ptr, v_ptr, post_y_stride, post_uv_stride, &lfi);
 
                     if (!skip_lf)
-                        LF_INVOKE(&cm->rtcd.loopfilter, normal_b_v)
-                        (y_ptr, u_ptr, v_ptr, post->y_stride, post->uv_stride, &lfi);
+                        vp8_loop_filter_bv
+                        (y_ptr, u_ptr, v_ptr, post_y_stride, post_uv_stride, &lfi);
 
                     /* don't apply across umv border */
                     if (mb_row > 0)
-                        LF_INVOKE(&cm->rtcd.loopfilter, normal_mb_h)
-                        (y_ptr, u_ptr, v_ptr, post->y_stride, post->uv_stride, &lfi);
+                        vp8_loop_filter_mbh
+                        (y_ptr, u_ptr, v_ptr, post_y_stride, post_uv_stride, &lfi);
 
                     if (!skip_lf)
-                        LF_INVOKE(&cm->rtcd.loopfilter, normal_b_h)
-                        (y_ptr, u_ptr, v_ptr, post->y_stride, post->uv_stride, &lfi);
+                        vp8_loop_filter_bh
+                        (y_ptr, u_ptr, v_ptr, post_y_stride, post_uv_stride, &lfi);
                 }
-                else
-                {
-                    if (mb_col > 0)
-                        LF_INVOKE(&cm->rtcd.loopfilter, simple_mb_v)
-                        (y_ptr, post->y_stride, lfi_n->mblim[filter_level]);
 
-                    if (!skip_lf)
-                        LF_INVOKE(&cm->rtcd.loopfilter, simple_b_v)
-                        (y_ptr, post->y_stride, lfi_n->blim[filter_level]);
+                y_ptr += 16;
+                u_ptr += 8;
+                v_ptr += 8;
 
-                    /* don't apply across umv border */
-                    if (mb_row > 0)
-                        LF_INVOKE(&cm->rtcd.loopfilter, simple_mb_h)
-                        (y_ptr, post->y_stride, lfi_n->mblim[filter_level]);
-
-                    if (!skip_lf)
-                        LF_INVOKE(&cm->rtcd.loopfilter, simple_b_h)
-                        (y_ptr, post->y_stride, lfi_n->blim[filter_level]);
-                }
+                mode_info_context++;     /* step to next MB */
             }
+            y_ptr += post_y_stride  * 16 - post->y_width;
+            u_ptr += post_uv_stride *  8 - post->uv_width;
+            v_ptr += post_uv_stride *  8 - post->uv_width;
 
-            y_ptr += 16;
-            u_ptr += 8;
-            v_ptr += 8;
+            mode_info_context++;         /* Skip border mb */
 
-            mode_info_context++;     /* step to next MB */
         }
+    }
+    else /* SIMPLE_LOOPFILTER */
+    {
+        for (mb_row = 0; mb_row < mb_rows; mb_row++)
+        {
+            for (mb_col = 0; mb_col < mb_cols; mb_col++)
+            {
+                int skip_lf = (mode_info_context->mbmi.mode != B_PRED &&
+                                mode_info_context->mbmi.mode != SPLITMV &&
+                                mode_info_context->mbmi.mb_skip_coeff);
 
-        y_ptr += post->y_stride  * 16 - post->y_width;
-        u_ptr += post->uv_stride *  8 - post->uv_width;
-        v_ptr += post->uv_stride *  8 - post->uv_width;
+                const int mode_index = lfi_n->mode_lf_lut[mode_info_context->mbmi.mode];
+                const int seg = mode_info_context->mbmi.segment_id;
+                const int ref_frame = mode_info_context->mbmi.ref_frame;
 
-        mode_info_context++;         /* Skip border mb */
+                filter_level = lfi_n->lvl[seg][ref_frame][mode_index];
+                if (filter_level)
+                {
+                    const unsigned char * mblim = lfi_n->mblim[filter_level];
+                    const unsigned char * blim = lfi_n->blim[filter_level];
+
+                    if (mb_col > 0)
+                        vp8_loop_filter_simple_mbv
+                        (y_ptr, post_y_stride, mblim);
+
+                    if (!skip_lf)
+                        vp8_loop_filter_simple_bv
+                        (y_ptr, post_y_stride, blim);
+
+                    /* don't apply across umv border */
+                    if (mb_row > 0)
+                        vp8_loop_filter_simple_mbh
+                        (y_ptr, post_y_stride, mblim);
+
+                    if (!skip_lf)
+                        vp8_loop_filter_simple_bh
+                        (y_ptr, post_y_stride, blim);
+                }
+
+                y_ptr += 16;
+                u_ptr += 8;
+                v_ptr += 8;
+
+                mode_info_context++;     /* step to next MB */
+            }
+            y_ptr += post_y_stride  * 16 - post->y_width;
+            u_ptr += post_uv_stride *  8 - post->uv_width;
+            v_ptr += post_uv_stride *  8 - post->uv_width;
+
+            mode_info_context++;         /* Skip border mb */
+
+        }
     }
 }
 
@@ -446,39 +400,39 @@ void vp8_loop_filter_frame_yonly
                     lfi.hev_thr = lfi_n->hev_thr[hev_index];
 
                     if (mb_col > 0)
-                        LF_INVOKE(&cm->rtcd.loopfilter, normal_mb_v)
+                        vp8_loop_filter_mbv
                         (y_ptr, 0, 0, post->y_stride, 0, &lfi);
 
                     if (!skip_lf)
-                        LF_INVOKE(&cm->rtcd.loopfilter, normal_b_v)
+                        vp8_loop_filter_bv
                         (y_ptr, 0, 0, post->y_stride, 0, &lfi);
 
                     /* don't apply across umv border */
                     if (mb_row > 0)
-                        LF_INVOKE(&cm->rtcd.loopfilter, normal_mb_h)
+                        vp8_loop_filter_mbh
                         (y_ptr, 0, 0, post->y_stride, 0, &lfi);
 
                     if (!skip_lf)
-                        LF_INVOKE(&cm->rtcd.loopfilter, normal_b_h)
+                        vp8_loop_filter_bh
                         (y_ptr, 0, 0, post->y_stride, 0, &lfi);
                 }
                 else
                 {
                     if (mb_col > 0)
-                        LF_INVOKE(&cm->rtcd.loopfilter, simple_mb_v)
+                        vp8_loop_filter_simple_mbv
                         (y_ptr, post->y_stride, lfi_n->mblim[filter_level]);
 
                     if (!skip_lf)
-                        LF_INVOKE(&cm->rtcd.loopfilter, simple_b_v)
+                        vp8_loop_filter_simple_bv
                         (y_ptr, post->y_stride, lfi_n->blim[filter_level]);
 
                     /* don't apply across umv border */
                     if (mb_row > 0)
-                        LF_INVOKE(&cm->rtcd.loopfilter, simple_mb_h)
+                        vp8_loop_filter_simple_mbh
                         (y_ptr, post->y_stride, lfi_n->mblim[filter_level]);
 
                     if (!skip_lf)
-                        LF_INVOKE(&cm->rtcd.loopfilter, simple_b_h)
+                        vp8_loop_filter_simple_bh
                         (y_ptr, post->y_stride, lfi_n->blim[filter_level]);
                 }
             }
@@ -578,35 +532,35 @@ void vp8_loop_filter_partial_frame
                     lfi.hev_thr = lfi_n->hev_thr[hev_index];
 
                     if (mb_col > 0)
-                        LF_INVOKE(&cm->rtcd.loopfilter, normal_mb_v)
+                        vp8_loop_filter_mbv
                         (y_ptr, 0, 0, post->y_stride, 0, &lfi);
 
                     if (!skip_lf)
-                        LF_INVOKE(&cm->rtcd.loopfilter, normal_b_v)
+                        vp8_loop_filter_bv
                         (y_ptr, 0, 0, post->y_stride, 0, &lfi);
 
-                    LF_INVOKE(&cm->rtcd.loopfilter, normal_mb_h)
+                    vp8_loop_filter_mbh
                         (y_ptr, 0, 0, post->y_stride, 0, &lfi);
 
                     if (!skip_lf)
-                        LF_INVOKE(&cm->rtcd.loopfilter, normal_b_h)
+                        vp8_loop_filter_bh
                         (y_ptr, 0, 0, post->y_stride, 0, &lfi);
                 }
                 else
                 {
                     if (mb_col > 0)
-                        LF_INVOKE(&cm->rtcd.loopfilter, simple_mb_v)
+                        vp8_loop_filter_simple_mbv
                         (y_ptr, post->y_stride, lfi_n->mblim[filter_level]);
 
                     if (!skip_lf)
-                        LF_INVOKE(&cm->rtcd.loopfilter, simple_b_v)
+                        vp8_loop_filter_simple_bv
                         (y_ptr, post->y_stride, lfi_n->blim[filter_level]);
 
-                    LF_INVOKE(&cm->rtcd.loopfilter, simple_mb_h)
+                    vp8_loop_filter_simple_mbh
                         (y_ptr, post->y_stride, lfi_n->mblim[filter_level]);
 
                     if (!skip_lf)
-                        LF_INVOKE(&cm->rtcd.loopfilter, simple_b_h)
+                        vp8_loop_filter_simple_bh
                         (y_ptr, post->y_stride, lfi_n->blim[filter_level]);
                 }
             }

vp8/common/loopfilter.h

@@ -14,6 +14,7 @@
 
 #include "vpx_ports/mem.h"
 #include "vpx_config.h"
+#include "vpx_rtcd.h"
 
 #define MAX_LOOP_FILTER             63
 /* fraction of total macroblock rows to be used in fast filter level picking */
@@ -46,7 +47,7 @@ typedef struct
     unsigned char mode_lf_lut[10];
 } loop_filter_info_n;
 
-typedef struct
+typedef struct loop_filter_info
 {
     const unsigned char * mblim;
     const unsigned char * blim;
@@ -55,86 +56,6 @@ typedef struct
 } loop_filter_info;
 
 
-#define prototype_loopfilter(sym) \
-    void sym(unsigned char *src, int pitch, const unsigned char *blimit,\
-             const unsigned char *limit, const unsigned char *thresh, int count)
-
-#define prototype_loopfilter_block(sym) \
-    void sym(unsigned char *y, unsigned char *u, unsigned char *v, \
-             int ystride, int uv_stride, loop_filter_info *lfi)
-
-#define prototype_simple_loopfilter(sym) \
-    void sym(unsigned char *y, int ystride, const unsigned char *blimit)
-
-#if ARCH_X86 || ARCH_X86_64
-#include "x86/loopfilter_x86.h"
-#endif
-
-#if ARCH_ARM
-#include "arm/loopfilter_arm.h"
-#endif
-
-#ifndef vp8_lf_normal_mb_v
-#define vp8_lf_normal_mb_v vp8_loop_filter_mbv_c
-#endif
-extern prototype_loopfilter_block(vp8_lf_normal_mb_v);
-
-#ifndef vp8_lf_normal_b_v
-#define vp8_lf_normal_b_v vp8_loop_filter_bv_c
-#endif
-extern prototype_loopfilter_block(vp8_lf_normal_b_v);
-
-#ifndef vp8_lf_normal_mb_h
-#define vp8_lf_normal_mb_h vp8_loop_filter_mbh_c
-#endif
-extern prototype_loopfilter_block(vp8_lf_normal_mb_h);
-
-#ifndef vp8_lf_normal_b_h
-#define vp8_lf_normal_b_h vp8_loop_filter_bh_c
-#endif
-extern prototype_loopfilter_block(vp8_lf_normal_b_h);
-
-#ifndef vp8_lf_simple_mb_v
-#define vp8_lf_simple_mb_v vp8_loop_filter_simple_vertical_edge_c
-#endif
-extern prototype_simple_loopfilter(vp8_lf_simple_mb_v);
-
-#ifndef vp8_lf_simple_b_v
-#define vp8_lf_simple_b_v vp8_loop_filter_bvs_c
-#endif
-extern prototype_simple_loopfilter(vp8_lf_simple_b_v);
-
-#ifndef vp8_lf_simple_mb_h
-#define vp8_lf_simple_mb_h vp8_loop_filter_simple_horizontal_edge_c
-#endif
-extern prototype_simple_loopfilter(vp8_lf_simple_mb_h);
-
-#ifndef vp8_lf_simple_b_h
-#define vp8_lf_simple_b_h vp8_loop_filter_bhs_c
-#endif
-extern prototype_simple_loopfilter(vp8_lf_simple_b_h);
-
-typedef prototype_loopfilter_block((*vp8_lf_block_fn_t));
-typedef prototype_simple_loopfilter((*vp8_slf_block_fn_t));
-
-typedef struct
-{
-    vp8_lf_block_fn_t  normal_mb_v;
-    vp8_lf_block_fn_t  normal_b_v;
-    vp8_lf_block_fn_t  normal_mb_h;
-    vp8_lf_block_fn_t  normal_b_h;
-    vp8_slf_block_fn_t  simple_mb_v;
-    vp8_slf_block_fn_t  simple_b_v;
-    vp8_slf_block_fn_t  simple_mb_h;
-    vp8_slf_block_fn_t  simple_b_h;
-} vp8_loopfilter_rtcd_vtable_t;
-
-#if CONFIG_RUNTIME_CPU_DETECT
-#define LF_INVOKE(ctx,fn) (ctx)->fn
-#else
-#define LF_INVOKE(ctx,fn) vp8_lf_##fn
-#endif
-
 typedef void loop_filter_uvfunction
 (
     unsigned char *u,   /* source pointer */
@@ -147,22 +68,22 @@ typedef void loop_filter_uvfunction
 
 /* assorted loopfilter functions which get used elsewhere */
 struct VP8Common;
-struct MacroBlockD;
+struct macroblockd;
 
 void vp8_loop_filter_init(struct VP8Common *cm);
 
 void vp8_loop_filter_frame_init(struct VP8Common *cm,
-                                struct MacroBlockD *mbd,
+                                struct macroblockd *mbd,
                                 int default_filt_lvl);
 
-void vp8_loop_filter_frame(struct VP8Common *cm, struct MacroBlockD *mbd);
+void vp8_loop_filter_frame(struct VP8Common *cm, struct macroblockd *mbd);
 
 void vp8_loop_filter_partial_frame(struct VP8Common *cm,
-                                   struct MacroBlockD *mbd,
+                                   struct macroblockd *mbd,
                                    int default_filt_lvl);
 
 void vp8_loop_filter_frame_yonly(struct VP8Common *cm,
-                                 struct MacroBlockD *mbd,
+                                 struct macroblockd *mbd,
                                  int default_filt_lvl);
 
 void vp8_loop_filter_update_sharpness(loop_filter_info_n *lfi,

vp8/common/loopfilter_filters.c

@@ -15,7 +15,7 @@
 
 typedef unsigned char uc;
 
-static __inline signed char vp8_signed_char_clamp(int t)
+static signed char vp8_signed_char_clamp(int t)
 {
     t = (t < -128 ? -128 : t);
     t = (t > 127 ? 127 : t);
@@ -24,9 +24,9 @@ static __inline signed char vp8_signed_char_clamp(int t)
 
 
 /* should we apply any filter at all ( 11111111 yes, 00000000 no) */
-static __inline signed char vp8_filter_mask(uc limit, uc blimit,
-                                     uc p3, uc p2, uc p1, uc p0,
-                                     uc q0, uc q1, uc q2, uc q3)
+static signed char vp8_filter_mask(uc limit, uc blimit,
+                            uc p3, uc p2, uc p1, uc p0,
+                            uc q0, uc q1, uc q2, uc q3)
 {
     signed char mask = 0;
     mask |= (abs(p3 - p2) > limit);
@@ -40,7 +40,7 @@ static __inline signed char vp8_filter_mask(uc limit, uc blimit,
 }
 
 /* is there high variance internal edge ( 11111111 yes, 00000000 no) */
-static __inline signed char vp8_hevmask(uc thresh, uc p1, uc p0, uc q0, uc q1)
+static signed char vp8_hevmask(uc thresh, uc p1, uc p0, uc q0, uc q1)
 {
     signed char hev = 0;
     hev  |= (abs(p1 - p0) > thresh) * -1;
@@ -48,7 +48,7 @@ static __inline signed char vp8_hevmask(uc thresh, uc p1, uc p0, uc q0, uc q1)
     return hev;
 }
 
-static __inline void vp8_filter(signed char mask, uc hev, uc *op1,
+static void vp8_filter(signed char mask, uc hev, uc *op1,
         uc *op0, uc *oq0, uc *oq1)
 
 {
@@ -158,7 +158,7 @@ void vp8_loop_filter_vertical_edge_c
     while (++i < count * 8);
 }
 
-static __inline void vp8_mbfilter(signed char mask, uc hev,
+static void vp8_mbfilter(signed char mask, uc hev,
                            uc *op2, uc *op1, uc *op0, uc *oq0, uc *oq1, uc *oq2)
 {
     signed char s, u;
@@ -279,7 +279,7 @@ void vp8_mbloop_filter_vertical_edge_c
 }
 
 /* should we apply any filter at all ( 11111111 yes, 00000000 no) */
-static __inline signed char vp8_simple_filter_mask(uc blimit, uc p1, uc p0, uc q0, uc q1)
+static signed char vp8_simple_filter_mask(uc blimit, uc p1, uc p0, uc q0, uc q1)
 {
 /* Why does this cause problems for win32?
  * error C2143: syntax error : missing ';' before 'type'
@@ -289,7 +289,7 @@ static __inline signed char vp8_simple_filter_mask(uc blimit, uc p1, uc p0, uc q
     return mask;
 }
 
-static __inline void vp8_simple_filter(signed char mask, uc *op1, uc *op0, uc *oq0, uc *oq1)
+static void vp8_simple_filter(signed char mask, uc *op1, uc *op0, uc *oq0, uc *oq1)
 {
     signed char vp8_filter, Filter1, Filter2;
     signed char p1 = (signed char) * op1 ^ 0x80;
@@ -352,3 +352,79 @@ void vp8_loop_filter_simple_vertical_edge_c
     while (++i < 16);
 
 }
+
+/* Horizontal MB filtering */
+void vp8_loop_filter_mbh_c(unsigned char *y_ptr, unsigned char *u_ptr,
+                           unsigned char *v_ptr, int y_stride, int uv_stride,
+                           loop_filter_info *lfi)
+{
+    vp8_mbloop_filter_horizontal_edge_c(y_ptr, y_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 2);
+
+    if (u_ptr)
+        vp8_mbloop_filter_horizontal_edge_c(u_ptr, uv_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 1);
+
+    if (v_ptr)
+        vp8_mbloop_filter_horizontal_edge_c(v_ptr, uv_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 1);
+}
+
+/* Vertical MB Filtering */
+void vp8_loop_filter_mbv_c(unsigned char *y_ptr, unsigned char *u_ptr,
+                           unsigned char *v_ptr, int y_stride, int uv_stride,
+                           loop_filter_info *lfi)
+{
+    vp8_mbloop_filter_vertical_edge_c(y_ptr, y_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 2);
+
+    if (u_ptr)
+        vp8_mbloop_filter_vertical_edge_c(u_ptr, uv_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 1);
+
+    if (v_ptr)
+        vp8_mbloop_filter_vertical_edge_c(v_ptr, uv_stride, lfi->mblim, lfi->lim, lfi->hev_thr, 1);
+}
+
+/* Horizontal B Filtering */
+void vp8_loop_filter_bh_c(unsigned char *y_ptr, unsigned char *u_ptr,
+                          unsigned char *v_ptr, int y_stride, int uv_stride,
+                          loop_filter_info *lfi)
+{
+    vp8_loop_filter_horizontal_edge_c(y_ptr + 4 * y_stride, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
+    vp8_loop_filter_horizontal_edge_c(y_ptr + 8 * y_stride, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
+    vp8_loop_filter_horizontal_edge_c(y_ptr + 12 * y_stride, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
+
+    if (u_ptr)
+        vp8_loop_filter_horizontal_edge_c(u_ptr + 4 * uv_stride, uv_stride, lfi->blim, lfi->lim, lfi->hev_thr, 1);
+
+    if (v_ptr)
+        vp8_loop_filter_horizontal_edge_c(v_ptr + 4 * uv_stride, uv_stride, lfi->blim, lfi->lim, lfi->hev_thr, 1);
+}
+
+void vp8_loop_filter_bhs_c(unsigned char *y_ptr, int y_stride,
+                           const unsigned char *blimit)
+{
+    vp8_loop_filter_simple_horizontal_edge_c(y_ptr + 4 * y_stride, y_stride, blimit);
+    vp8_loop_filter_simple_horizontal_edge_c(y_ptr + 8 * y_stride, y_stride, blimit);
+    vp8_loop_filter_simple_horizontal_edge_c(y_ptr + 12 * y_stride, y_stride, blimit);
+}
+
+/* Vertical B Filtering */
+void vp8_loop_filter_bv_c(unsigned char *y_ptr, unsigned char *u_ptr,
+                          unsigned char *v_ptr, int y_stride, int uv_stride,
+                          loop_filter_info *lfi)
+{
+    vp8_loop_filter_vertical_edge_c(y_ptr + 4, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
+    vp8_loop_filter_vertical_edge_c(y_ptr + 8, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
+    vp8_loop_filter_vertical_edge_c(y_ptr + 12, y_stride, lfi->blim, lfi->lim, lfi->hev_thr, 2);
+
+    if (u_ptr)
+        vp8_loop_filter_vertical_edge_c(u_ptr + 4, uv_stride, lfi->blim, lfi->lim, lfi->hev_thr, 1);
+
+    if (v_ptr)
+        vp8_loop_filter_vertical_edge_c(v_ptr + 4, uv_stride, lfi->blim, lfi->lim, lfi->hev_thr, 1);
+}
+
+void vp8_loop_filter_bvs_c(unsigned char *y_ptr, int y_stride,
+                           const unsigned char *blimit)
+{
+    vp8_loop_filter_simple_vertical_edge_c(y_ptr + 4, y_stride, blimit);
+    vp8_loop_filter_simple_vertical_edge_c(y_ptr + 8, y_stride, blimit);
+    vp8_loop_filter_simple_vertical_edge_c(y_ptr + 12, y_stride, blimit);
+}

vp8/common/mbpitch.c

@@ -11,74 +11,6 @@
 
 #include "blockd.h"
 
-typedef enum
-{
-    PRED = 0,
-    DEST = 1
-} BLOCKSET;
-
-static void setup_block
-(
-    BLOCKD *b,
-    int mv_stride,
-    unsigned char **base,
-    int Stride,
-    int offset,
-    BLOCKSET bs
-)
-{
-
-    if (bs == DEST)
-    {
-        b->dst_stride = Stride;
-        b->dst = offset;
-        b->base_dst = base;
-    }
-    else
-    {
-        b->pre_stride = Stride;
-        b->pre = offset;
-        b->base_pre = base;
-    }
-
-}
-
-
-static void setup_macroblock(MACROBLOCKD *x, BLOCKSET bs)
-{
-    int block;
-
-    unsigned char **y, **u, **v;
-
-    if (bs == DEST)
-    {
-        y = &x->dst.y_buffer;
-        u = &x->dst.u_buffer;
-        v = &x->dst.v_buffer;
-    }
-    else
-    {
-        y = &x->pre.y_buffer;
-        u = &x->pre.u_buffer;
-        v = &x->pre.v_buffer;
-    }
-
-    for (block = 0; block < 16; block++) /* y blocks */
-    {
-        setup_block(&x->block[block], x->dst.y_stride, y, x->dst.y_stride,
-                        (block >> 2) * 4 * x->dst.y_stride + (block & 3) * 4, bs);
-    }
-
-    for (block = 16; block < 20; block++) /* U and V blocks */
-    {
-        setup_block(&x->block[block], x->dst.uv_stride, u, x->dst.uv_stride,
-                        ((block - 16) >> 1) * 4 * x->dst.uv_stride + (block & 1) * 4, bs);
-
-        setup_block(&x->block[block+4], x->dst.uv_stride, v, x->dst.uv_stride,
-                        ((block - 16) >> 1) * 4 * x->dst.uv_stride + (block & 1) * 4, bs);
-    }
-}
-
 void vp8_setup_block_dptrs(MACROBLOCKD *x)
 {
     int r, c;
@@ -87,7 +19,6 @@ void vp8_setup_block_dptrs(MACROBLOCKD *x)
     {
         for (c = 0; c < 4; c++)
         {
-            x->block[r*4+c].diff      = &x->diff[r * 4 * 16 + c * 4];
             x->block[r*4+c].predictor = x->predictor + r * 4 * 16 + c * 4;
         }
     }
@@ -96,7 +27,6 @@ void vp8_setup_block_dptrs(MACROBLOCKD *x)
     {
         for (c = 0; c < 2; c++)
         {
-            x->block[16+r*2+c].diff      = &x->diff[256 + r * 4 * 8 + c * 4];
             x->block[16+r*2+c].predictor = x->predictor + 256 + r * 4 * 8 + c * 4;
 
         }
@@ -106,14 +36,11 @@ void vp8_setup_block_dptrs(MACROBLOCKD *x)
     {
         for (c = 0; c < 2; c++)
         {
-            x->block[20+r*2+c].diff      = &x->diff[320+ r * 4 * 8 + c * 4];
             x->block[20+r*2+c].predictor = x->predictor + 320 + r * 4 * 8 + c * 4;
 
         }
     }
 
-    x->block[24].diff = &x->diff[384];
-
     for (r = 0; r < 25; r++)
     {
         x->block[r].qcoeff  = x->qcoeff  + r * 16;
@@ -124,8 +51,18 @@ void vp8_setup_block_dptrs(MACROBLOCKD *x)
 
 void vp8_build_block_doffsets(MACROBLOCKD *x)
 {
+    int block;
 
-    /* handle the destination pitch features */
-    setup_macroblock(x, DEST);
-    setup_macroblock(x, PRED);
+    for (block = 0; block < 16; block++) /* y blocks */
+    {
+        x->block[block].offset =
+            (block >> 2) * 4 * x->dst.y_stride + (block & 3) * 4;
+    }
+
+    for (block = 16; block < 20; block++) /* U and V blocks */
+    {
+        x->block[block+4].offset =
+        x->block[block].offset =
+            ((block - 16) >> 1) * 4 * x->dst.uv_stride + (block & 1) * 4;
+    }
 }

vp8/common/mfqe.c

@@ -0,0 +1,385 @@
+/*
+ *  Copyright (c) 2012 The WebM project authors. All Rights Reserved.
+ *
+ *  Use of this source code is governed by a BSD-style license
+ *  that can be found in the LICENSE file in the root of the source
+ *  tree. An additional intellectual property rights grant can be found
+ *  in the file PATENTS.  All contributing project authors may
+ *  be found in the AUTHORS file in the root of the source tree.
+ */
+
+
+/* MFQE: Multiframe Quality Enhancement
+ * In rate limited situations keyframes may cause significant visual artifacts
+ * commonly referred to as "popping." This file implements a postproccesing
+ * algorithm which blends data from the preceeding frame when there is no
+ * motion and the q from the previous frame is lower which indicates that it is
+ * higher quality.
+ */
+
+#include "postproc.h"
+#include "variance.h"
+#include "vpx_mem/vpx_mem.h"
+#include "vpx_rtcd.h"
+#include "vpx_scale/yv12config.h"
+
+#include <limits.h>
+#include <stdlib.h>
+
+static void filter_by_weight(unsigned char *src, int src_stride,
+                             unsigned char *dst, int dst_stride,
+                             int block_size, int src_weight)
+{
+    int dst_weight = (1 << MFQE_PRECISION) - src_weight;
+    int rounding_bit = 1 << (MFQE_PRECISION - 1);
+    int r, c;
+
+    for (r = 0; r < block_size; r++)
+    {
+        for (c = 0; c < block_size; c++)
+        {
+            dst[c] = (src[c] * src_weight +
+                      dst[c] * dst_weight +
+                      rounding_bit) >> MFQE_PRECISION;
+        }
+        src += src_stride;
+        dst += dst_stride;
+    }
+}
+
+void vp8_filter_by_weight16x16_c(unsigned char *src, int src_stride,
+                                 unsigned char *dst, int dst_stride,
+                                 int src_weight)
+{
+    filter_by_weight(src, src_stride, dst, dst_stride, 16, src_weight);
+}
+
+void vp8_filter_by_weight8x8_c(unsigned char *src, int src_stride,
+                               unsigned char *dst, int dst_stride,
+                               int src_weight)
+{
+    filter_by_weight(src, src_stride, dst, dst_stride, 8, src_weight);
+}
+
+void vp8_filter_by_weight4x4_c(unsigned char *src, int src_stride,
+                               unsigned char *dst, int dst_stride,
+                               int src_weight)
+{
+    filter_by_weight(src, src_stride, dst, dst_stride, 4, src_weight);
+}
+
+static void apply_ifactor(unsigned char *y_src,
+                          int y_src_stride,
+                          unsigned char *y_dst,
+                          int y_dst_stride,
+                          unsigned char *u_src,
+                          unsigned char *v_src,
+                          int uv_src_stride,
+                          unsigned char *u_dst,
+                          unsigned char *v_dst,
+                          int uv_dst_stride,
+                          int block_size,
+                          int src_weight)
+{
+    if (block_size == 16)
+    {
+        vp8_filter_by_weight16x16(y_src, y_src_stride, y_dst, y_dst_stride, src_weight);
+        vp8_filter_by_weight8x8(u_src, uv_src_stride, u_dst, uv_dst_stride, src_weight);
+        vp8_filter_by_weight8x8(v_src, uv_src_stride, v_dst, uv_dst_stride, src_weight);
+    }
+    else /* if (block_size == 8) */
+    {
+        vp8_filter_by_weight8x8(y_src, y_src_stride, y_dst, y_dst_stride, src_weight);
+        vp8_filter_by_weight4x4(u_src, uv_src_stride, u_dst, uv_dst_stride, src_weight);
+        vp8_filter_by_weight4x4(v_src, uv_src_stride, v_dst, uv_dst_stride, src_weight);
+    }
+}
+
+static unsigned int int_sqrt(unsigned int x)
+{
+    unsigned int y = x;
+    unsigned int guess;
+    int p = 1;
+    while (y>>=1) p++;
+    p>>=1;
+
+    guess=0;
+    while (p>=0)
+    {
+        guess |= (1<<p);
+        if (x<guess*guess)
+            guess -= (1<<p);
+        p--;
+    }
+    /* choose between guess or guess+1 */
+    return guess+(guess*guess+guess+1<=x);
+}
+
+#define USE_SSD
+static void multiframe_quality_enhance_block
+(
+    int blksize, /* Currently only values supported are 16, 8 */
+    int qcurr,
+    int qprev,
+    unsigned char *y,
+    unsigned char *u,
+    unsigned char *v,
+    int y_stride,
+    int uv_stride,
+    unsigned char *yd,
+    unsigned char *ud,
+    unsigned char *vd,
+    int yd_stride,
+    int uvd_stride
+)
+{
+    static const unsigned char VP8_ZEROS[16]=
+    {
+         0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
+    };
+    int uvblksize = blksize >> 1;
+    int qdiff = qcurr - qprev;
+
+    int i;
+    unsigned char *up;
+    unsigned char *udp;
+    unsigned char *vp;
+    unsigned char *vdp;
+
+    unsigned int act, actd, sad, usad, vsad, sse, thr, thrsq, actrisk;
+
+    if (blksize == 16)
+    {
+        actd = (vp8_variance16x16(yd, yd_stride, VP8_ZEROS, 0, &sse)+128)>>8;
+        act = (vp8_variance16x16(y, y_stride, VP8_ZEROS, 0, &sse)+128)>>8;
+#ifdef USE_SSD
+        sad = (vp8_variance16x16(y, y_stride, yd, yd_stride, &sse));
+        sad = (sse + 128)>>8;
+        usad = (vp8_variance8x8(u, uv_stride, ud, uvd_stride, &sse));
+        usad = (sse + 32)>>6;
+        vsad = (vp8_variance8x8(v, uv_stride, vd, uvd_stride, &sse));
+        vsad = (sse + 32)>>6;
+#else
+        sad = (vp8_sad16x16(y, y_stride, yd, yd_stride, INT_MAX)+128)>>8;
+        usad = (vp8_sad8x8(u, uv_stride, ud, uvd_stride, INT_MAX)+32)>>6;
+        vsad = (vp8_sad8x8(v, uv_stride, vd, uvd_stride, INT_MAX)+32)>>6;
+#endif
+    }
+    else /* if (blksize == 8) */
+    {
+        actd = (vp8_variance8x8(yd, yd_stride, VP8_ZEROS, 0, &sse)+32)>>6;
+        act = (vp8_variance8x8(y, y_stride, VP8_ZEROS, 0, &sse)+32)>>6;
+#ifdef USE_SSD
+        sad = (vp8_variance8x8(y, y_stride, yd, yd_stride, &sse));
+        sad = (sse + 32)>>6;
+        usad = (vp8_variance4x4(u, uv_stride, ud, uvd_stride, &sse));
+        usad = (sse + 8)>>4;
+        vsad = (vp8_variance4x4(v, uv_stride, vd, uvd_stride, &sse));
+        vsad = (sse + 8)>>4;
+#else
+        sad = (vp8_sad8x8(y, y_stride, yd, yd_stride, INT_MAX)+32)>>6;
+        usad = (vp8_sad4x4(u, uv_stride, ud, uvd_stride, INT_MAX)+8)>>4;
+        vsad = (vp8_sad4x4(v, uv_stride, vd, uvd_stride, INT_MAX)+8)>>4;
+#endif
+    }
+
+    actrisk = (actd > act * 5);
+
+    /* thr = qdiff/8 + log2(act) + log4(qprev) */
+    thr = (qdiff >> 3);
+    while (actd >>= 1) thr++;
+    while (qprev >>= 2) thr++;
+
+#ifdef USE_SSD
+    thrsq = thr * thr;
+    if (sad < thrsq &&
+        /* additional checks for color mismatch and excessive addition of
+         * high-frequencies */
+        4 * usad < thrsq && 4 * vsad < thrsq && !actrisk)
+#else
+    if (sad < thr &&
+        /* additional checks for color mismatch and excessive addition of
+         * high-frequencies */
+        2 * usad < thr && 2 * vsad < thr && !actrisk)
+#endif
+    {
+        int ifactor;
+#ifdef USE_SSD
+        /* TODO: optimize this later to not need sqr root */
+        sad = int_sqrt(sad);
+#endif
+        ifactor = (sad << MFQE_PRECISION) / thr;
+        ifactor >>= (qdiff >> 5);
+
+        if (ifactor)
+        {
+            apply_ifactor(y, y_stride, yd, yd_stride,
+                          u, v, uv_stride,
+                          ud, vd, uvd_stride,
+                          blksize, ifactor);
+        }
+    }
+    else  /* else implicitly copy from previous frame */
+    {
+        if (blksize == 16)
+        {
+            vp8_copy_mem16x16(y, y_stride, yd, yd_stride);
+            vp8_copy_mem8x8(u, uv_stride, ud, uvd_stride);
+            vp8_copy_mem8x8(v, uv_stride, vd, uvd_stride);
+        }
+        else  /* if (blksize == 8) */
+        {
+            vp8_copy_mem8x8(y, y_stride, yd, yd_stride);
+            for (up = u, udp = ud, i = 0; i < uvblksize; ++i, up += uv_stride, udp += uvd_stride)
+                vpx_memcpy(udp, up, uvblksize);
+            for (vp = v, vdp = vd, i = 0; i < uvblksize; ++i, vp += uv_stride, vdp += uvd_stride)
+                vpx_memcpy(vdp, vp, uvblksize);
+        }
+    }
+}
+
+static int qualify_inter_mb(const MODE_INFO *mode_info_context, int *map)
+{
+    if (mode_info_context->mbmi.mb_skip_coeff)
+        map[0] = map[1] = map[2] = map[3] = 1;
+    else if (mode_info_context->mbmi.mode==SPLITMV)
+    {
+        static int ndx[4][4] =
+        {
+            {0, 1, 4, 5},
+            {2, 3, 6, 7},
+            {8, 9, 12, 13},
+            {10, 11, 14, 15}
+        };
+        int i, j;
+        for (i=0; i<4; ++i)
+        {
+            map[i] = 1;
+            for (j=0; j<4 && map[j]; ++j)
+                map[i] &= (mode_info_context->bmi[ndx[i][j]].mv.as_mv.row <= 2 &&
+                           mode_info_context->bmi[ndx[i][j]].mv.as_mv.col <= 2);
+        }
+    }
+    else
+    {
+        map[0] = map[1] = map[2] = map[3] =
+            (mode_info_context->mbmi.mode > B_PRED &&
+             abs(mode_info_context->mbmi.mv.as_mv.row) <= 2 &&
+             abs(mode_info_context->mbmi.mv.as_mv.col) <= 2);
+    }
+    return (map[0]+map[1]+map[2]+map[3]);
+}
+
+void vp8_multiframe_quality_enhance
+(
+    VP8_COMMON *cm
+)
+{
+    YV12_BUFFER_CONFIG *show = cm->frame_to_show;
+    YV12_BUFFER_CONFIG *dest = &cm->post_proc_buffer;
+
+    FRAME_TYPE frame_type = cm->frame_type;
+    /* Point at base of Mb MODE_INFO list has motion vectors etc */
+    const MODE_INFO *mode_info_context = cm->mi;
+    int mb_row;
+    int mb_col;
+    int totmap, map[4];
+    int qcurr = cm->base_qindex;
+    int qprev = cm->postproc_state.last_base_qindex;
+
+    unsigned char *y_ptr, *u_ptr, *v_ptr;
+    unsigned char *yd_ptr, *ud_ptr, *vd_ptr;
+
+    /* Set up the buffer pointers */
+    y_ptr = show->y_buffer;
+    u_ptr = show->u_buffer;
+    v_ptr = show->v_buffer;
+    yd_ptr = dest->y_buffer;
+    ud_ptr = dest->u_buffer;
+    vd_ptr = dest->v_buffer;
+
+    /* postprocess each macro block */
+    for (mb_row = 0; mb_row < cm->mb_rows; mb_row++)
+    {
+        for (mb_col = 0; mb_col < cm->mb_cols; mb_col++)
+        {
+            /* if motion is high there will likely be no benefit */
+            if (frame_type == INTER_FRAME) totmap = qualify_inter_mb(mode_info_context, map);
+            else totmap = (frame_type == KEY_FRAME ? 4 : 0);
+            if (totmap)
+            {
+                if (totmap < 4)
+                {
+                    int i, j;
+                    for (i=0; i<2; ++i)
+                        for (j=0; j<2; ++j)
+                        {
+                            if (map[i*2+j])
+                            {
+                                multiframe_quality_enhance_block(8, qcurr, qprev,
+                                                                 y_ptr + 8*(i*show->y_stride+j),
+                                                                 u_ptr + 4*(i*show->uv_stride+j),
+                                                                 v_ptr + 4*(i*show->uv_stride+j),
+                                                                 show->y_stride,
+                                                                 show->uv_stride,
+                                                                 yd_ptr + 8*(i*dest->y_stride+j),
+                                                                 ud_ptr + 4*(i*dest->uv_stride+j),
+                                                                 vd_ptr + 4*(i*dest->uv_stride+j),
+                                                                 dest->y_stride,
+                                                                 dest->uv_stride);
+                            }
+                            else
+                            {
+                                /* copy a 8x8 block */
+                                int k;
+                                unsigned char *up = u_ptr + 4*(i*show->uv_stride+j);
+                                unsigned char *udp = ud_ptr + 4*(i*dest->uv_stride+j);
+                                unsigned char *vp = v_ptr + 4*(i*show->uv_stride+j);
+                                unsigned char *vdp = vd_ptr + 4*(i*dest->uv_stride+j);
+                                vp8_copy_mem8x8(y_ptr + 8*(i*show->y_stride+j), show->y_stride,
+                                                yd_ptr + 8*(i*dest->y_stride+j), dest->y_stride);
+                                for (k = 0; k < 4; ++k, up += show->uv_stride, udp += dest->uv_stride,
+                                                        vp += show->uv_stride, vdp += dest->uv_stride)
+                                {
+                                    vpx_memcpy(udp, up, 4);
+                                    vpx_memcpy(vdp, vp, 4);
+                                }
+                            }
+                        }
+                }
+                else /* totmap = 4 */
+                {
+                    multiframe_quality_enhance_block(16, qcurr, qprev, y_ptr,
+                                                     u_ptr, v_ptr,
+                                                     show->y_stride,
+                                                     show->uv_stride,
+                                                     yd_ptr, ud_ptr, vd_ptr,
+                                                     dest->y_stride,
+                                                     dest->uv_stride);
+                }
+            }
+            else
+            {
+                vp8_copy_mem16x16(y_ptr, show->y_stride, yd_ptr, dest->y_stride);
+                vp8_copy_mem8x8(u_ptr, show->uv_stride, ud_ptr, dest->uv_stride);
+                vp8_copy_mem8x8(v_ptr, show->uv_stride, vd_ptr, dest->uv_stride);
+            }
+            y_ptr += 16;
+            u_ptr += 8;
+            v_ptr += 8;
+            yd_ptr += 16;
+            ud_ptr += 8;
+            vd_ptr += 8;
+            mode_info_context++;     /* step to next MB */
+        }
+
+        y_ptr += show->y_stride  * 16 - 16 * cm->mb_cols;
+        u_ptr += show->uv_stride *  8 - 8 * cm->mb_cols;
+        v_ptr += show->uv_stride *  8 - 8 * cm->mb_cols;
+        yd_ptr += dest->y_stride  * 16 - 16 * cm->mb_cols;
+        ud_ptr += dest->uv_stride *  8 - 8 * cm->mb_cols;
+        vd_ptr += dest->uv_stride *  8 - 8 * cm->mb_cols;
+
+        mode_info_context++;         /* Skip border mb */
+    }
+}

vp8/common/modecontext.c

@@ -1,146 +0,0 @@
-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-
-#include "entropymode.h"
-
-const unsigned int vp8_kf_default_bmode_counts [VP8_BINTRAMODES] [VP8_BINTRAMODES] [VP8_BINTRAMODES] =
-{
-    {
-        /*Above Mode :  0*/
-        { 43438,   2195,    470,    316,    615,    171,    217,    412,    124,    160, }, /* left_mode 0 */
-        {  5722,   2751,    296,    291,     81,     68,     80,    101,    100,    170, }, /* left_mode 1 */
-        {  1629,    201,    307,     25,     47,     16,     34,     72,     19,     28, }, /* left_mode 2 */
-        {   332,    266,     36,    500,     20,     65,     23,     14,    154,    106, }, /* left_mode 3 */
-        {   450,     97,     10,     24,    117,     10,      2,     12,      8,     71, }, /* left_mode 4 */
-        {   384,     49,     29,     44,     12,    162,     51,      5,     87,     42, }, /* left_mode 5 */
-        {   495,     53,    157,     27,     14,     57,    180,     17,     17,     34, }, /* left_mode 6 */
-        {   695,     64,     62,      9,     27,      5,      3,    147,     10,     26, }, /* left_mode 7 */
-        {   230,     54,     20,    124,     16,    125,     29,     12,    283,     37, }, /* left_mode 8 */
-        {   260,     87,     21,    120,     32,     16,     33,     16,     33,    203, }, /* left_mode 9 */
-    },
-    {
-        /*Above Mode :  1*/
-        {  3934,   2573,    355,    137,    128,     87,    133,    117,     37,     27, }, /* left_mode 0 */
-        {  1036,   1929,    278,    135,     27,     37,     48,     55,     41,     91, }, /* left_mode 1 */
-        {   223,    256,    253,     15,     13,      9,     28,     64,      3,      3, }, /* left_mode 2 */
-        {   120,    129,     17,    316,     15,     11,      9,      4,     53,     74, }, /* left_mode 3 */
-        {   129,     58,      6,     11,     38,      2,      0,      5,      2,     67, }, /* left_mode 4 */
-        {    53,     22,     11,     16,      8,     26,     14,      3,     19,     12, }, /* left_mode 5 */
-        {    59,     26,     61,     11,      4,      9,     35,     13,      8,      8, }, /* left_mode 6 */
-        {   101,     52,     40,      8,      5,      2,      8,     59,      2,     20, }, /* left_mode 7 */
-        {    48,     34,     10,     52,      8,     15,      6,      6,     63,     20, }, /* left_mode 8 */
-        {    96,     48,     22,     63,     11,     14,      5,      8,      9,     96, }, /* left_mode 9 */
-    },
-    {
-        /*Above Mode :  2*/
-        {   709,    461,    506,     36,     27,     33,    151,     98,     24,      6, }, /* left_mode 0 */
-        {   201,    375,    442,     27,     13,      8,     46,     58,      6,     19, }, /* left_mode 1 */
-        {   122,    140,    417,      4,     13,      3,     33,     59,      4,      2, }, /* left_mode 2 */
-        {    36,     17,     22,     16,      6,      8,     12,     17,      9,     21, }, /* left_mode 3 */
-        {    51,     15,      7,      1,     14,      0,      4,      5,      3,     22, }, /* left_mode 4 */
-        {    18,     11,     30,      9,      7,     20,     11,      5,      2,      6, }, /* left_mode 5 */
-        {    38,     21,    103,      9,      4,     12,     79,     13,      2,      5, }, /* left_mode 6 */
-        {    64,     17,     66,      2,     12,      4,      2,     65,      4,      5, }, /* left_mode 7 */
-        {    14,      7,      7,     16,      3,     11,      4,     13,     15,     16, }, /* left_mode 8 */
-        {    36,      8,     32,      9,      9,      4,     14,      7,      6,     24, }, /* left_mode 9 */
-    },
-    {
-        /*Above Mode :  3*/
-        {  1340,    173,     36,    119,     30,     10,     13,     10,     20,     26, }, /* left_mode 0 */
-        {   156,    293,     26,    108,      5,     16,      2,      4,     23,     30, }, /* left_mode 1 */
-        {    60,     34,     13,      7,      3,      3,      0,      8,      4,      5, }, /* left_mode 2 */
-        {    72,     64,      1,    235,      3,      9,      2,      7,     28,     38, }, /* left_mode 3 */
-        {    29,     14,      1,      3,      5,      0,      2,      2,      5,     13, }, /* left_mode 4 */
-        {    22,      7,      4,     11,      2,      5,      1,      2,      6,      4, }, /* left_mode 5 */
-        {    18,     14,      5,      6,      4,      3,     14,      0,      9,      2, }, /* left_mode 6 */
-        {    41,     10,      7,      1,      2,      0,      0,     10,      2,      1, }, /* left_mode 7 */
-        {    23,     19,      2,     33,      1,      5,      2,      0,     51,      8, }, /* left_mode 8 */
-        {    33,     26,      7,     53,      3,      9,      3,      3,      9,     19, }, /* left_mode 9 */
-    },
-    {
-        /*Above Mode :  4*/
-        {   410,    165,     43,     31,     66,     15,     30,     54,      8,     17, }, /* left_mode 0 */
-        {   115,     64,     27,     18,     30,      7,     11,     15,      4,     19, }, /* left_mode 1 */
-        {    31,     23,     25,      1,      7,      2,      2,     10,      0,      5, }, /* left_mode 2 */
-        {    17,      4,      1,      6,      8,      2,      7,      5,      5,     21, }, /* left_mode 3 */
-        {   120,     12,      1,      2,     83,      3,      0,      4,      1,     40, }, /* left_mode 4 */
-        {     4,      3,      1,      2,      1,      2,      5,      0,      3,      6, }, /* left_mode 5 */
-        {    10,      2,     13,      6,      6,      6,      8,      2,      4,      5, }, /* left_mode 6 */
-        {    58,     10,      5,      1,     28,      1,      1,     33,      1,      9, }, /* left_mode 7 */
-        {     8,      2,      1,      4,      2,      5,      1,      1,      2,     10, }, /* left_mode 8 */
-        {    76,      7,      5,      7,     18,      2,      2,      0,      5,     45, }, /* left_mode 9 */
-    },
-    {
-        /*Above Mode :  5*/
-        {   444,     46,     47,     20,     14,    110,     60,     14,     60,      7, }, /* left_mode 0 */
-        {    59,     57,     25,     18,      3,     17,     21,      6,     14,      6, }, /* left_mode 1 */
-        {    24,     17,     20,      6,      4,     13,      7,      2,      3,      2, }, /* left_mode 2 */
-        {    13,     11,      5,     14,      4,      9,      2,      4,     15,      7, }, /* left_mode 3 */
-        {     8,      5,      2,      1,      4,      0,      1,      1,      2,     12, }, /* left_mode 4 */
-        {    19,      5,      5,      7,      4,     40,      6,      3,     10,      4, }, /* left_mode 5 */
-        {    16,      5,      9,      1,      1,     16,     26,      2,     10,      4, }, /* left_mode 6 */
-        {    11,      4,      8,      1,      1,      4,      4,      5,      4,      1, }, /* left_mode 7 */
-        {    15,      1,      3,      7,      3,     21,      7,      1,     34,      5, }, /* left_mode 8 */
-        {    18,      5,      1,      3,      4,      3,      7,      1,      2,      9, }, /* left_mode 9 */
-    },
-    {
-        /*Above Mode :  6*/
-        {   476,    149,     94,     13,     14,     77,    291,     27,     23,      3, }, /* left_mode 0 */
-        {    79,     83,     42,     14,      2,     12,     63,      2,      4,     14, }, /* left_mode 1 */
-        {    43,     36,     55,      1,      3,      8,     42,     11,      5,      1, }, /* left_mode 2 */
-        {     9,      9,      6,     16,      1,      5,      6,      3,     11,     10, }, /* left_mode 3 */
-        {    10,      3,      1,      3,     10,      1,      0,      1,      1,      4, }, /* left_mode 4 */
-        {    14,      6,     15,      5,      1,     20,     25,      2,      5,      0, }, /* left_mode 5 */
-        {    28,      7,     51,      1,      0,      8,    127,      6,      2,      5, }, /* left_mode 6 */
-        {    13,      3,      3,      2,      3,      1,      2,      8,      1,      2, }, /* left_mode 7 */
-        {    10,      3,      3,      3,      3,      8,      2,      2,      9,      3, }, /* left_mode 8 */
-        {    13,      7,     11,      4,      0,      4,      6,      2,      5,      8, }, /* left_mode 9 */
-    },
-    {
-        /*Above Mode :  7*/
-        {   376,    135,    119,      6,     32,      8,     31,    224,      9,      3, }, /* left_mode 0 */
-        {    93,     60,     54,      6,     13,      7,      8,     92,      2,     12, }, /* left_mode 1 */
-        {    74,     36,     84,      0,      3,      2,      9,     67,      2,      1, }, /* left_mode 2 */
-        {    19,      4,      4,      8,      8,      2,      4,      7,      6,     16, }, /* left_mode 3 */
-        {    51,      7,      4,      1,     77,      3,      0,     14,      1,     15, }, /* left_mode 4 */
-        {     7,      7,      5,      7,      4,      7,      4,      5,      0,      3, }, /* left_mode 5 */
-        {    18,      2,     19,      2,      2,      4,     12,     11,      1,      2, }, /* left_mode 6 */
-        {   129,      6,     27,      1,     21,      3,      0,    189,      0,      6, }, /* left_mode 7 */
-        {     9,      1,      2,      8,      3,      7,      0,      5,      3,      3, }, /* left_mode 8 */
-        {    20,      4,      5,     10,      4,      2,      7,     17,      3,     16, }, /* left_mode 9 */
-    },
-    {
-        /*Above Mode :  8*/
-        {   617,     68,     34,     79,     11,     27,     25,     14,     75,     13, }, /* left_mode 0 */
-        {    51,     82,     21,     26,      6,     12,     13,      1,     26,     16, }, /* left_mode 1 */
-        {    29,      9,     12,     11,      3,      7,      1,     10,      2,      2, }, /* left_mode 2 */
-        {    17,     19,     11,     74,      4,      3,      2,      0,     58,     13, }, /* left_mode 3 */
-        {    10,      1,      1,      3,      4,      1,      0,      2,      1,      8, }, /* left_mode 4 */
-        {    14,      4,      5,      5,      1,     13,      2,      0,     27,      8, }, /* left_mode 5 */
-        {    10,      3,      5,      4,      1,      7,      6,      4,      5,      1, }, /* left_mode 6 */
-        {    10,      2,      6,      2,      1,      1,      1,      4,      2,      1, }, /* left_mode 7 */
-        {    14,      8,      5,     23,      2,     12,      6,      2,    117,      5, }, /* left_mode 8 */
-        {     9,      6,      2,     19,      1,      6,      3,      2,      9,      9, }, /* left_mode 9 */
-    },
-    {
-        /*Above Mode :  9*/
-        {   680,     73,     22,     38,     42,      5,     11,      9,      6,     28, }, /* left_mode 0 */
-        {   113,    112,     21,     22,     10,      2,      8,      4,      6,     42, }, /* left_mode 1 */
-        {    44,     20,     24,      6,      5,      4,      3,      3,      1,      2, }, /* left_mode 2 */
-        {    40,     23,      7,     71,      5,      2,      4,      1,      7,     22, }, /* left_mode 3 */
-        {    85,      9,      4,      4,     17,      2,      0,      3,      2,     23, }, /* left_mode 4 */
-        {    13,      4,      2,      6,      1,      7,      0,      1,      7,      6, }, /* left_mode 5 */
-        {    26,      6,      8,      3,      2,      3,      8,      1,      5,      4, }, /* left_mode 6 */
-        {    54,      8,      9,      6,      7,      0,      1,     11,      1,      3, }, /* left_mode 7 */
-        {     9,     10,      4,     13,      2,      5,      4,      2,     14,      8, }, /* left_mode 8 */
-        {    92,      9,      5,     19,     15,      3,      3,      1,      6,     58, }, /* left_mode 9 */
-    },
-};

vp8/common/mv.h

@@ -19,7 +19,7 @@ typedef struct
     short col;
 } MV;
 
-typedef union
+typedef union int_mv
 {
     uint32_t  as_int;
     MV        as_mv;

vp8/common/onyx.h

@@ -22,9 +22,9 @@ extern "C"
 #include "vpx/vp8cx.h"
 #include "vpx/vpx_encoder.h"
 #include "vpx_scale/yv12config.h"
-#include "type_aliases.h"
 #include "ppflags.h"
-    typedef int *VP8_PTR;
+
+    struct VP8_COMP;
 
     /* Create/destroy static data structures. */
 
@@ -60,19 +60,19 @@ extern "C"
         MODE_BESTQUALITY    = 0x2,
         MODE_FIRSTPASS      = 0x3,
         MODE_SECONDPASS     = 0x4,
-        MODE_SECONDPASS_BEST = 0x5,
+        MODE_SECONDPASS_BEST = 0x5
     } MODE;
 
     typedef enum
     {
         FRAMEFLAGS_KEY    = 1,
         FRAMEFLAGS_GOLDEN = 2,
-        FRAMEFLAGS_ALTREF = 4,
+        FRAMEFLAGS_ALTREF = 4
     } FRAMETYPE_FLAGS;
 
 
 #include <assert.h>
-    static __inline void Scale2Ratio(int mode, int *hr, int *hs)
+    static void Scale2Ratio(int mode, int *hr, int *hs)
     {
         switch (mode)
         {
@@ -147,10 +147,14 @@ extern "C"
         int over_shoot_pct;
 
         // buffering parameters
-        int64_t starting_buffer_level;  // in seconds
+        int64_t starting_buffer_level;  // in bytes
         int64_t optimal_buffer_level;
         int64_t maximum_buffer_size;
 
+        int64_t starting_buffer_level_in_ms;  // in milli-seconds
+        int64_t optimal_buffer_level_in_ms;
+        int64_t maximum_buffer_size_in_ms;
+
         // controlling quality
         int fixed_q;
         int worst_allowed_q;
@@ -203,10 +207,10 @@ extern "C"
 
         // Temporal scaling parameters
         unsigned int number_of_layers;
-        unsigned int target_bitrate[MAX_PERIODICITY];
-        unsigned int rate_decimator[MAX_PERIODICITY];
+        unsigned int target_bitrate[VPX_TS_MAX_PERIODICITY];
+        unsigned int rate_decimator[VPX_TS_MAX_PERIODICITY];
         unsigned int periodicity;
-        unsigned int layer_id[MAX_PERIODICITY];
+        unsigned int layer_id[VPX_TS_MAX_PERIODICITY];
 
 #if CONFIG_MULTI_RES_ENCODING
         /* Number of total resolutions encoded */
@@ -226,27 +230,27 @@ extern "C"
 
     void vp8_initialize();
 
-    VP8_PTR vp8_create_compressor(VP8_CONFIG *oxcf);
-    void vp8_remove_compressor(VP8_PTR *comp);
+    struct VP8_COMP* vp8_create_compressor(VP8_CONFIG *oxcf);
+    void vp8_remove_compressor(struct VP8_COMP* *comp);
 
-    void vp8_init_config(VP8_PTR onyx, VP8_CONFIG *oxcf);
-    void vp8_change_config(VP8_PTR onyx, VP8_CONFIG *oxcf);
+    void vp8_init_config(struct VP8_COMP* onyx, VP8_CONFIG *oxcf);
+    void vp8_change_config(struct VP8_COMP* onyx, VP8_CONFIG *oxcf);
 
 // receive a frames worth of data caller can assume that a copy of this frame is made
 // and not just a copy of the pointer..
-    int vp8_receive_raw_frame(VP8_PTR comp, unsigned int frame_flags, YV12_BUFFER_CONFIG *sd, int64_t time_stamp, int64_t end_time_stamp);
-    int vp8_get_compressed_data(VP8_PTR comp, unsigned int *frame_flags, unsigned long *size, unsigned char *dest, unsigned char *dest_end, int64_t *time_stamp, int64_t *time_end, int flush);
-    int vp8_get_preview_raw_frame(VP8_PTR comp, YV12_BUFFER_CONFIG *dest, vp8_ppflags_t *flags);
+    int vp8_receive_raw_frame(struct VP8_COMP* comp, unsigned int frame_flags, YV12_BUFFER_CONFIG *sd, int64_t time_stamp, int64_t end_time_stamp);
+    int vp8_get_compressed_data(struct VP8_COMP* comp, unsigned int *frame_flags, unsigned long *size, unsigned char *dest, unsigned char *dest_end, int64_t *time_stamp, int64_t *time_end, int flush);
+    int vp8_get_preview_raw_frame(struct VP8_COMP* comp, YV12_BUFFER_CONFIG *dest, vp8_ppflags_t *flags);
 
-    int vp8_use_as_reference(VP8_PTR comp, int ref_frame_flags);
-    int vp8_update_reference(VP8_PTR comp, int ref_frame_flags);
-    int vp8_get_reference(VP8_PTR comp, VP8_REFFRAME ref_frame_flag, YV12_BUFFER_CONFIG *sd);
-    int vp8_set_reference(VP8_PTR comp, VP8_REFFRAME ref_frame_flag, YV12_BUFFER_CONFIG *sd);
-    int vp8_update_entropy(VP8_PTR comp, int update);
-    int vp8_set_roimap(VP8_PTR comp, unsigned char *map, unsigned int rows, unsigned int cols, int delta_q[4], int delta_lf[4], unsigned int threshold[4]);
-    int vp8_set_active_map(VP8_PTR comp, unsigned char *map, unsigned int rows, unsigned int cols);
-    int vp8_set_internal_size(VP8_PTR comp, VPX_SCALING horiz_mode, VPX_SCALING vert_mode);
-    int vp8_get_quantizer(VP8_PTR c);
+    int vp8_use_as_reference(struct VP8_COMP* comp, int ref_frame_flags);
+    int vp8_update_reference(struct VP8_COMP* comp, int ref_frame_flags);
+    int vp8_get_reference(struct VP8_COMP* comp, VP8_REFFRAME ref_frame_flag, YV12_BUFFER_CONFIG *sd);
+    int vp8_set_reference(struct VP8_COMP* comp, VP8_REFFRAME ref_frame_flag, YV12_BUFFER_CONFIG *sd);
+    int vp8_update_entropy(struct VP8_COMP* comp, int update);
+    int vp8_set_roimap(struct VP8_COMP* comp, unsigned char *map, unsigned int rows, unsigned int cols, int delta_q[4], int delta_lf[4], unsigned int threshold[4]);
+    int vp8_set_active_map(struct VP8_COMP* comp, unsigned char *map, unsigned int rows, unsigned int cols);
+    int vp8_set_internal_size(struct VP8_COMP* comp, VPX_SCALING horiz_mode, VPX_SCALING vert_mode);
+    int vp8_get_quantizer(struct VP8_COMP* c);
 
 #ifdef __cplusplus
 }

vp8/common/onyxc_int.h

@@ -13,12 +13,11 @@
 #define __INC_VP8C_INT_H
 
 #include "vpx_config.h"
+#include "vpx_rtcd.h"
 #include "vpx/internal/vpx_codec_internal.h"
 #include "loopfilter.h"
 #include "entropymv.h"
 #include "entropy.h"
-#include "idct.h"
-#include "recon.h"
 #if CONFIG_POSTPROC
 #include "postproc.h"
 #endif
@@ -27,10 +26,6 @@
 #include "header.h"
 /*#endif*/
 
-/* Create/destroy static data structures. */
-
-void vp8_initialize_common(void);
-
 #define MINQ 0
 #define MAXQ 127
 #define QINDEX_RANGE (MAXQ + 1)
@@ -70,30 +65,14 @@ typedef enum
     BILINEAR = 1
 } INTERPOLATIONFILTERTYPE;
 
-typedef struct VP8_COMMON_RTCD
-{
-#if CONFIG_RUNTIME_CPU_DETECT
-    vp8_idct_rtcd_vtable_t        idct;
-    vp8_recon_rtcd_vtable_t       recon;
-    vp8_subpix_rtcd_vtable_t      subpix;
-    vp8_loopfilter_rtcd_vtable_t  loopfilter;
-#if CONFIG_POSTPROC
-    vp8_postproc_rtcd_vtable_t    postproc;
-#endif
-    int                           flags;
-#else
-    int unused;
-#endif
-} VP8_COMMON_RTCD;
-
 typedef struct VP8Common
 
 {
     struct vpx_internal_error_info  error;
 
-    DECLARE_ALIGNED(16, short, Y1dequant[QINDEX_RANGE][16]);
-    DECLARE_ALIGNED(16, short, Y2dequant[QINDEX_RANGE][16]);
-    DECLARE_ALIGNED(16, short, UVdequant[QINDEX_RANGE][16]);
+    DECLARE_ALIGNED(16, short, Y1dequant[QINDEX_RANGE][2]);
+    DECLARE_ALIGNED(16, short, Y2dequant[QINDEX_RANGE][2]);
+    DECLARE_ALIGNED(16, short, UVdequant[QINDEX_RANGE][2]);
 
     int Width;
     int Height;
@@ -109,9 +88,13 @@ typedef struct VP8Common
     int fb_idx_ref_cnt[NUM_YV12_BUFFERS];
     int new_fb_idx, lst_fb_idx, gld_fb_idx, alt_fb_idx;
 
-    YV12_BUFFER_CONFIG post_proc_buffer;
     YV12_BUFFER_CONFIG temp_scale_frame;
 
+#if CONFIG_POSTPROC
+    YV12_BUFFER_CONFIG post_proc_buffer;
+    YV12_BUFFER_CONFIG post_proc_buffer_int;
+    int post_proc_buffer_int_used;
+#endif
 
     FRAME_TYPE last_frame_type;  /* Save last frame's frame type for motion search. */
     FRAME_TYPE frame_type;
@@ -199,15 +182,13 @@ typedef struct VP8Common
     double bitrate;
     double framerate;
 
-#if CONFIG_RUNTIME_CPU_DETECT
-    VP8_COMMON_RTCD rtcd;
-#endif
 #if CONFIG_MULTITHREAD
     int processor_core_count;
 #endif
 #if CONFIG_POSTPROC
     struct postproc_state  postproc_state;
 #endif
+    int cpu_caps;
 } VP8_COMMON;
 
 #endif

vp8/common/onyxd.h

@@ -18,13 +18,13 @@
 extern "C"
 {
 #endif
-#include "type_aliases.h"
 #include "vpx_scale/yv12config.h"
 #include "ppflags.h"
 #include "vpx_ports/mem.h"
 #include "vpx/vpx_codec.h"
 
-    typedef void   *VP8D_PTR;
+    struct VP8D_COMP;
+
     typedef struct
     {
         int     Width;
@@ -49,19 +49,19 @@ extern "C"
 
     void vp8dx_initialize(void);
 
-    void vp8dx_set_setting(VP8D_PTR comp, VP8D_SETTING oxst, int x);
+    void vp8dx_set_setting(struct VP8D_COMP* comp, VP8D_SETTING oxst, int x);
 
-    int vp8dx_get_setting(VP8D_PTR comp, VP8D_SETTING oxst);
+    int vp8dx_get_setting(struct VP8D_COMP* comp, VP8D_SETTING oxst);
 
-    int vp8dx_receive_compressed_data(VP8D_PTR comp, unsigned long size, const unsigned char *dest, int64_t time_stamp);
-    int vp8dx_get_raw_frame(VP8D_PTR comp, YV12_BUFFER_CONFIG *sd, int64_t *time_stamp, int64_t *time_end_stamp, vp8_ppflags_t *flags);
+    int vp8dx_receive_compressed_data(struct VP8D_COMP* comp, unsigned long size, const unsigned char *dest, int64_t time_stamp);
+    int vp8dx_get_raw_frame(struct VP8D_COMP* comp, YV12_BUFFER_CONFIG *sd, int64_t *time_stamp, int64_t *time_end_stamp, vp8_ppflags_t *flags);
 
-    vpx_codec_err_t vp8dx_get_reference(VP8D_PTR comp, VP8_REFFRAME ref_frame_flag, YV12_BUFFER_CONFIG *sd);
-    vpx_codec_err_t vp8dx_set_reference(VP8D_PTR comp, VP8_REFFRAME ref_frame_flag, YV12_BUFFER_CONFIG *sd);
+    vpx_codec_err_t vp8dx_get_reference(struct VP8D_COMP* comp, VP8_REFFRAME ref_frame_flag, YV12_BUFFER_CONFIG *sd);
+    vpx_codec_err_t vp8dx_set_reference(struct VP8D_COMP* comp, VP8_REFFRAME ref_frame_flag, YV12_BUFFER_CONFIG *sd);
 
-    VP8D_PTR vp8dx_create_decompressor(VP8D_CONFIG *oxcf);
+    struct VP8D_COMP* vp8dx_create_decompressor(VP8D_CONFIG *oxcf);
 
-    void vp8dx_remove_decompressor(VP8D_PTR comp);
+    void vp8dx_remove_decompressor(struct VP8D_COMP* comp);
 
 #ifdef __cplusplus
 }

vp8/common/postproc.c

@@ -10,12 +10,14 @@
 
 
 #include "vpx_config.h"
+#include "vpx_rtcd.h"
 #include "vpx_scale/yv12config.h"
 #include "postproc.h"
-#include "vpx_scale/yv12extend.h"
+#include "common.h"
 #include "vpx_scale/vpxscale.h"
 #include "systemdependent.h"
 
+#include <limits.h>
 #include <math.h>
 #include <stdlib.h>
 #include <stdio.h>
@@ -121,7 +123,6 @@ const short vp8_rv[] =
     0, 9, 5, 5, 11, 10, 13, 9, 10, 13,
 };
 
-
 extern void vp8_blit_text(const char *msg, unsigned char *address, const int pitch);
 extern void vp8_blit_line(int x0, int x1, int y0, int y1, unsigned char *image, const int pitch);
 /***********************************************************************************************************
@@ -175,6 +176,12 @@ void vp8_post_proc_down_and_across_c
         p_src = dst_ptr;
         p_dst = dst_ptr;
 
+        for (i = -8; i<0; i++)
+          p_src[i]=p_src[0];
+
+        for (i = cols; i<cols+8; i++)
+          p_src[i]=p_src[cols-1];
+
         for (i = 0; i < 8; i++)
             d[i] = p_src[i];
 
@@ -225,12 +232,19 @@ void vp8_mbpost_proc_across_ip_c(unsigned char *src, int pitch, int rows, int co
     unsigned char *s = src;
     unsigned char d[16];
 
-
     for (r = 0; r < rows; r++)
     {
         int sumsq = 0;
         int sum   = 0;
 
+        for (i = -8; i<0; i++)
+          s[i]=s[0];
+
+        // 17 avoids valgrind warning - we buffer values in c in d
+        // and only write them when we've read 8 ahead...
+        for (i = cols; i<cols+17; i++)
+          s[i]=s[cols-1];
+
         for (i = -8; i <= 6; i++)
         {
             sumsq += s[i] * s[i];
@@ -269,7 +283,7 @@ void vp8_mbpost_proc_down_c(unsigned char *dst, int pitch, int rows, int cols, i
     int r, c, i;
     const short *rv3 = &vp8_rv[63&rand()];
 
-    for (c = 0; c < cols; c++)
+    for (c = 0; c < cols; c++ )
     {
         unsigned char *s = &dst[c];
         int sumsq = 0;
@@ -277,6 +291,14 @@ void vp8_mbpost_proc_down_c(unsigned char *dst, int pitch, int rows, int cols, i
         unsigned char d[16];
         const short *rv2 = rv3 + ((c * 17) & 127);
 
+        for (i = -8; i < 0; i++)
+          s[i*pitch]=s[0];
+
+        // 17 avoids valgrind warning - we buffer values in c in d
+        // and only write them when we've read 8 ahead...
+        for (i = rows; i < rows+17; i++)
+          s[i*pitch]=s[(rows-1)*pitch];
+
         for (i = -8; i <= 6; i++)
         {
             sumsq += s[i*pitch] * s[i*pitch];
@@ -305,46 +327,44 @@ static void vp8_deblock_and_de_macro_block(YV12_BUFFER_CONFIG         *source,
         YV12_BUFFER_CONFIG         *post,
         int                         q,
         int                         low_var_thresh,
-        int                         flag,
-        vp8_postproc_rtcd_vtable_t *rtcd)
+        int                         flag)
 {
     double level = 6.0e-05 * q * q * q - .0067 * q * q + .306 * q + .0065;
     int ppl = (int)(level + .5);
     (void) low_var_thresh;
     (void) flag;
 
-    POSTPROC_INVOKE(rtcd, downacross)(source->y_buffer, post->y_buffer, source->y_stride,  post->y_stride, source->y_height, source->y_width,  ppl);
-    POSTPROC_INVOKE(rtcd, across)(post->y_buffer, post->y_stride, post->y_height, post->y_width, q2mbl(q));
-    POSTPROC_INVOKE(rtcd, down)(post->y_buffer, post->y_stride, post->y_height, post->y_width, q2mbl(q));
+    vp8_post_proc_down_and_across(source->y_buffer, post->y_buffer, source->y_stride,  post->y_stride, source->y_height, source->y_width,  ppl);
+    vp8_mbpost_proc_across_ip(post->y_buffer, post->y_stride, post->y_height, post->y_width, q2mbl(q));
+    vp8_mbpost_proc_down(post->y_buffer, post->y_stride, post->y_height, post->y_width, q2mbl(q));
 
-    POSTPROC_INVOKE(rtcd, downacross)(source->u_buffer, post->u_buffer, source->uv_stride, post->uv_stride, source->uv_height, source->uv_width, ppl);
-    POSTPROC_INVOKE(rtcd, downacross)(source->v_buffer, post->v_buffer, source->uv_stride, post->uv_stride, source->uv_height, source->uv_width, ppl);
+    vp8_post_proc_down_and_across(source->u_buffer, post->u_buffer, source->uv_stride, post->uv_stride, source->uv_height, source->uv_width, ppl);
+    vp8_post_proc_down_and_across(source->v_buffer, post->v_buffer, source->uv_stride, post->uv_stride, source->uv_height, source->uv_width, ppl);
 
 }
 
 void vp8_deblock(YV12_BUFFER_CONFIG         *source,
-                        YV12_BUFFER_CONFIG         *post,
-                        int                         q,
-                        int                         low_var_thresh,
-                        int                         flag,
-                        vp8_postproc_rtcd_vtable_t *rtcd)
+                 YV12_BUFFER_CONFIG         *post,
+                 int                         q,
+                 int                         low_var_thresh,
+                 int                         flag)
 {
     double level = 6.0e-05 * q * q * q - .0067 * q * q + .306 * q + .0065;
     int ppl = (int)(level + .5);
     (void) low_var_thresh;
     (void) flag;
 
-    POSTPROC_INVOKE(rtcd, downacross)(source->y_buffer, post->y_buffer, source->y_stride,  post->y_stride, source->y_height, source->y_width,   ppl);
-    POSTPROC_INVOKE(rtcd, downacross)(source->u_buffer, post->u_buffer, source->uv_stride, post->uv_stride,  source->uv_height, source->uv_width, ppl);
-    POSTPROC_INVOKE(rtcd, downacross)(source->v_buffer, post->v_buffer, source->uv_stride, post->uv_stride, source->uv_height, source->uv_width, ppl);
+    vp8_post_proc_down_and_across(source->y_buffer, post->y_buffer, source->y_stride,  post->y_stride, source->y_height, source->y_width,   ppl);
+    vp8_post_proc_down_and_across(source->u_buffer, post->u_buffer, source->uv_stride, post->uv_stride,  source->uv_height, source->uv_width, ppl);
+    vp8_post_proc_down_and_across(source->v_buffer, post->v_buffer, source->uv_stride, post->uv_stride, source->uv_height, source->uv_width, ppl);
 }
 
+#if !(CONFIG_TEMPORAL_DENOISING)
 void vp8_de_noise(YV12_BUFFER_CONFIG         *source,
                   YV12_BUFFER_CONFIG         *post,
                   int                         q,
                   int                         low_var_thresh,
-                  int                         flag,
-                  vp8_postproc_rtcd_vtable_t *rtcd)
+                  int                         flag)
 {
     double level = 6.0e-05 * q * q * q - .0067 * q * q + .306 * q + .0065;
     int ppl = (int)(level + .5);
@@ -352,7 +372,7 @@ void vp8_de_noise(YV12_BUFFER_CONFIG         *source,
     (void) low_var_thresh;
     (void) flag;
 
-    POSTPROC_INVOKE(rtcd, downacross)(
+    vp8_post_proc_down_and_across(
         source->y_buffer + 2 * source->y_stride + 2,
         source->y_buffer + 2 * source->y_stride + 2,
         source->y_stride,
@@ -360,14 +380,14 @@ void vp8_de_noise(YV12_BUFFER_CONFIG         *source,
         source->y_height - 4,
         source->y_width - 4,
         ppl);
-    POSTPROC_INVOKE(rtcd, downacross)(
+    vp8_post_proc_down_and_across(
         source->u_buffer + 2 * source->uv_stride + 2,
         source->u_buffer + 2 * source->uv_stride + 2,
         source->uv_stride,
         source->uv_stride,
         source->uv_height - 4,
         source->uv_width - 4, ppl);
-    POSTPROC_INVOKE(rtcd, downacross)(
+    vp8_post_proc_down_and_across(
         source->v_buffer + 2 * source->uv_stride + 2,
         source->v_buffer + 2 * source->uv_stride + 2,
         source->uv_stride,
@@ -376,6 +396,7 @@ void vp8_de_noise(YV12_BUFFER_CONFIG         *source,
         source->uv_width - 4, ppl);
 
 }
+#endif
 
 double vp8_gaussian(double sigma, double mu, double x)
 {
@@ -383,9 +404,6 @@ double vp8_gaussian(double sigma, double mu, double x)
            (exp(-(x - mu) * (x - mu) / (2 * sigma * sigma)));
 }
 
-extern void (*vp8_clear_system_state)(void);
-
-
 static void fillrd(struct postproc_state *state, int q, int a)
 {
     char char_dist[300];
@@ -671,13 +689,7 @@ static void constrain_line (int x0, int *x1, int y0, int *y1, int width, int hei
     }
 }
 
-
-#if CONFIG_RUNTIME_CPU_DETECT
-#define RTCD_VTABLE(oci) (&(oci)->rtcd.postproc)
-#else
-#define RTCD_VTABLE(oci) NULL
-#endif
-
+#if CONFIG_POSTPROC
 int vp8_post_proc_frame(VP8_COMMON *oci, YV12_BUFFER_CONFIG *dest, vp8_ppflags_t *ppflags)
 {
     int q = oci->filter_level * 10 / 6;
@@ -699,28 +711,79 @@ int vp8_post_proc_frame(VP8_COMMON *oci, YV12_BUFFER_CONFIG *dest, vp8_ppflags_t
         dest->y_width = oci->Width;
         dest->y_height = oci->Height;
         dest->uv_height = dest->y_height / 2;
+        oci->postproc_state.last_base_qindex = oci->base_qindex;
+        oci->postproc_state.last_frame_valid = 1;
         return 0;
+    }
 
+    /* Allocate post_proc_buffer_int if needed */
+    if ((flags & VP8D_MFQE) && !oci->post_proc_buffer_int_used)
+    {
+        if ((flags & VP8D_DEBLOCK) || (flags & VP8D_DEMACROBLOCK))
+        {
+            int width = (oci->Width + 15) & ~15;
+            int height = (oci->Height + 15) & ~15;
+
+            if (vp8_yv12_alloc_frame_buffer(&oci->post_proc_buffer_int,
+                                            width, height, VP8BORDERINPIXELS))
+                vpx_internal_error(&oci->error, VPX_CODEC_MEM_ERROR,
+                                   "Failed to allocate MFQE framebuffer");
+
+            oci->post_proc_buffer_int_used = 1;
+
+            // insure that postproc is set to all 0's so that post proc
+            // doesn't pull random data in from edge
+            vpx_memset((&oci->post_proc_buffer_int)->buffer_alloc,128,(&oci->post_proc_buffer)->frame_size);
+
+        }
     }
 
 #if ARCH_X86||ARCH_X86_64
     vpx_reset_mmx_state();
 #endif
 
-    if (flags & VP8D_DEMACROBLOCK)
+    if ((flags & VP8D_MFQE) &&
+         oci->postproc_state.last_frame_valid &&
+         oci->current_video_frame >= 2 &&
+         oci->base_qindex - oci->postproc_state.last_base_qindex >= 10)
+    {
+        vp8_multiframe_quality_enhance(oci);
+        if (((flags & VP8D_DEBLOCK) || (flags & VP8D_DEMACROBLOCK)) &&
+            oci->post_proc_buffer_int_used)
+        {
+            vp8_yv12_copy_frame(&oci->post_proc_buffer, &oci->post_proc_buffer_int);
+            if (flags & VP8D_DEMACROBLOCK)
+            {
+                vp8_deblock_and_de_macro_block(&oci->post_proc_buffer_int, &oci->post_proc_buffer,
+                                               q + (deblock_level - 5) * 10, 1, 0);
+            }
+            else if (flags & VP8D_DEBLOCK)
+            {
+                vp8_deblock(&oci->post_proc_buffer_int, &oci->post_proc_buffer,
+                            q, 1, 0);
+            }
+        }
+        /* Move partially towards the base q of the previous frame */
+        oci->postproc_state.last_base_qindex = (3*oci->postproc_state.last_base_qindex + oci->base_qindex)>>2;
+    }
+    else if (flags & VP8D_DEMACROBLOCK)
     {
         vp8_deblock_and_de_macro_block(oci->frame_to_show, &oci->post_proc_buffer,
-                                       q + (deblock_level - 5) * 10, 1, 0, RTCD_VTABLE(oci));
+                                       q + (deblock_level - 5) * 10, 1, 0);
+        oci->postproc_state.last_base_qindex = oci->base_qindex;
     }
     else if (flags & VP8D_DEBLOCK)
     {
         vp8_deblock(oci->frame_to_show, &oci->post_proc_buffer,
-                    q, 1, 0, RTCD_VTABLE(oci));
+                    q, 1, 0);
+        oci->postproc_state.last_base_qindex = oci->base_qindex;
     }
     else
     {
-        vp8_yv12_copy_frame_ptr(oci->frame_to_show, &oci->post_proc_buffer);
+        vp8_yv12_copy_frame(oci->frame_to_show, &oci->post_proc_buffer);
+        oci->postproc_state.last_base_qindex = oci->base_qindex;
     }
+    oci->postproc_state.last_frame_valid = 1;
 
     if (flags & VP8D_ADDNOISE)
     {
@@ -730,7 +793,7 @@ int vp8_post_proc_frame(VP8_COMMON *oci, YV12_BUFFER_CONFIG *dest, vp8_ppflags_t
             fillrd(&oci->postproc_state, 63 - q, noise_level);
         }
 
-        POSTPROC_INVOKE(RTCD_VTABLE(oci), addnoise)
+        vp8_plane_add_noise
         (oci->post_proc_buffer.y_buffer,
          oci->postproc_state.noise,
          oci->postproc_state.blackclamp,
@@ -1028,7 +1091,7 @@ int vp8_post_proc_frame(VP8_COMMON *oci, YV12_BUFFER_CONFIG *dest, vp8_ppflags_t
                                 U = B_PREDICTION_MODE_colors[bmi->as_mode][1];
                                 V = B_PREDICTION_MODE_colors[bmi->as_mode][2];
 
-                                POSTPROC_INVOKE(RTCD_VTABLE(oci), blend_b)
+                                vp8_blend_b
                                     (yl+bx, ul+(bx>>1), vl+(bx>>1), Y, U, V, 0xc000, y_stride);
                             }
                             bmi++;
@@ -1045,7 +1108,7 @@ int vp8_post_proc_frame(VP8_COMMON *oci, YV12_BUFFER_CONFIG *dest, vp8_ppflags_t
                     U = MB_PREDICTION_MODE_colors[mi->mbmi.mode][1];
                     V = MB_PREDICTION_MODE_colors[mi->mbmi.mode][2];
 
-                    POSTPROC_INVOKE(RTCD_VTABLE(oci), blend_mb_inner)
+                    vp8_blend_mb_inner
                         (y_ptr+x, u_ptr+(x>>1), v_ptr+(x>>1), Y, U, V, 0xc000, y_stride);
                 }
 
@@ -1084,7 +1147,7 @@ int vp8_post_proc_frame(VP8_COMMON *oci, YV12_BUFFER_CONFIG *dest, vp8_ppflags_t
                     U = MV_REFERENCE_FRAME_colors[mi->mbmi.ref_frame][1];
                     V = MV_REFERENCE_FRAME_colors[mi->mbmi.ref_frame][2];
 
-                    POSTPROC_INVOKE(RTCD_VTABLE(oci), blend_mb_outer)
+                    vp8_blend_mb_outer
                         (y_ptr+x, u_ptr+(x>>1), v_ptr+(x>>1), Y, U, V, 0xc000, y_stride);
                 }
 
@@ -1107,3 +1170,4 @@ int vp8_post_proc_frame(VP8_COMMON *oci, YV12_BUFFER_CONFIG *dest, vp8_ppflags_t
     dest->uv_height = dest->y_height / 2;
     return 0;
 }
+#endif

vp8/common/postproc.h

@@ -12,98 +12,14 @@
 #ifndef POSTPROC_H
 #define POSTPROC_H
 
-#define prototype_postproc_inplace(sym)\
-    void sym (unsigned char *dst, int pitch, int rows, int cols,int flimit)
-
-#define prototype_postproc(sym)\
-    void sym (unsigned char *src, unsigned char *dst, int src_pitch,\
-              int dst_pitch, int rows, int cols, int flimit)
-
-#define prototype_postproc_addnoise(sym) \
-    void sym (unsigned char *s, char *noise, char blackclamp[16],\
-              char whiteclamp[16], char bothclamp[16],\
-              unsigned int w, unsigned int h, int pitch)
-
-#define prototype_postproc_blend_mb_inner(sym)\
-    void sym (unsigned char *y, unsigned char *u, unsigned char *v,\
-              int y1, int u1, int v1, int alpha, int stride)
-
-#define prototype_postproc_blend_mb_outer(sym)\
-    void sym (unsigned char *y, unsigned char *u, unsigned char *v,\
-              int y1, int u1, int v1, int alpha, int stride)
-
-#define prototype_postproc_blend_b(sym)\
-    void sym (unsigned char *y, unsigned char *u, unsigned char *v,\
-              int y1, int u1, int v1, int alpha, int stride)
-
-#if ARCH_X86 || ARCH_X86_64
-#include "x86/postproc_x86.h"
-#endif
-
-#ifndef vp8_postproc_down
-#define vp8_postproc_down vp8_mbpost_proc_down_c
-#endif
-extern prototype_postproc_inplace(vp8_postproc_down);
-
-#ifndef vp8_postproc_across
-#define vp8_postproc_across vp8_mbpost_proc_across_ip_c
-#endif
-extern prototype_postproc_inplace(vp8_postproc_across);
-
-#ifndef vp8_postproc_downacross
-#define vp8_postproc_downacross vp8_post_proc_down_and_across_c
-#endif
-extern prototype_postproc(vp8_postproc_downacross);
-
-#ifndef vp8_postproc_addnoise
-#define vp8_postproc_addnoise vp8_plane_add_noise_c
-#endif
-extern prototype_postproc_addnoise(vp8_postproc_addnoise);
-
-#ifndef vp8_postproc_blend_mb_inner
-#define vp8_postproc_blend_mb_inner vp8_blend_mb_inner_c
-#endif
-extern prototype_postproc_blend_mb_inner(vp8_postproc_blend_mb_inner);
-
-#ifndef vp8_postproc_blend_mb_outer
-#define vp8_postproc_blend_mb_outer vp8_blend_mb_outer_c
-#endif
-extern prototype_postproc_blend_mb_outer(vp8_postproc_blend_mb_outer);
-
-#ifndef vp8_postproc_blend_b
-#define vp8_postproc_blend_b vp8_blend_b_c
-#endif
-extern prototype_postproc_blend_b(vp8_postproc_blend_b);
-
-typedef prototype_postproc((*vp8_postproc_fn_t));
-typedef prototype_postproc_inplace((*vp8_postproc_inplace_fn_t));
-typedef prototype_postproc_addnoise((*vp8_postproc_addnoise_fn_t));
-typedef prototype_postproc_blend_mb_inner((*vp8_postproc_blend_mb_inner_fn_t));
-typedef prototype_postproc_blend_mb_outer((*vp8_postproc_blend_mb_outer_fn_t));
-typedef prototype_postproc_blend_b((*vp8_postproc_blend_b_fn_t));
-typedef struct
-{
-    vp8_postproc_inplace_fn_t           down;
-    vp8_postproc_inplace_fn_t           across;
-    vp8_postproc_fn_t                   downacross;
-    vp8_postproc_addnoise_fn_t          addnoise;
-    vp8_postproc_blend_mb_inner_fn_t    blend_mb_inner;
-    vp8_postproc_blend_mb_outer_fn_t    blend_mb_outer;
-    vp8_postproc_blend_b_fn_t           blend_b;
-} vp8_postproc_rtcd_vtable_t;
-
-#if CONFIG_RUNTIME_CPU_DETECT
-#define POSTPROC_INVOKE(ctx,fn) (ctx)->fn
-#else
-#define POSTPROC_INVOKE(ctx,fn) vp8_postproc_##fn
-#endif
-
 #include "vpx_ports/mem.h"
 struct postproc_state
 {
     int           last_q;
     int           last_noise;
     char          noise[3072];
+    int           last_base_qindex;
+    int           last_frame_valid;
     DECLARE_ALIGNED(16, char, blackclamp[16]);
     DECLARE_ALIGNED(16, char, whiteclamp[16]);
     DECLARE_ALIGNED(16, char, bothclamp[16]);
@@ -118,13 +34,15 @@ void vp8_de_noise(YV12_BUFFER_CONFIG         *source,
                   YV12_BUFFER_CONFIG         *post,
                   int                         q,
                   int                         low_var_thresh,
-                  int                         flag,
-                  vp8_postproc_rtcd_vtable_t *rtcd);
+                  int                         flag);
 
 void vp8_deblock(YV12_BUFFER_CONFIG         *source,
                  YV12_BUFFER_CONFIG         *post,
                  int                         q,
                  int                         low_var_thresh,
-                 int                         flag,
-                 vp8_postproc_rtcd_vtable_t *rtcd);
+                 int                         flag);
+
+#define MFQE_PRECISION 4
+
+void vp8_multiframe_quality_enhance(struct VP8Common *cm);
 #endif

vp8/common/ppc/systemdependent.c

@@ -9,7 +9,6 @@
  */
 
 
-#include "g_common.h"
 #include "subpixel.h"
 #include "loopfilter.h"
 #include "recon.h"

vp8/common/ppc/variance_altivec.asm

@@ -98,7 +98,7 @@
     stw     r4, 0(r7)           ;# sse
 
     mullw   r3, r3, r3          ;# sum*sum
-    srawi   r3, r3, \DS         ;# (sum*sum) >> DS
+    srlwi   r3, r3, \DS         ;# (sum*sum) >> DS
     subf    r3, r3, r4          ;# sse - ((sum*sum) >> DS)
 .endm
 
@@ -142,7 +142,7 @@
     stw     r4, 0(r7)           ;# sse
 
     mullw   r3, r3, r3          ;# sum*sum
-    srawi   r3, r3, \DS         ;# (sum*sum) >> 8
+    srlwi   r3, r3, \DS         ;# (sum*sum) >> 8
     subf    r3, r3, r4          ;# sse - ((sum*sum) >> 8)
 .endm
 
@@ -367,7 +367,7 @@ vp8_variance4x4_ppc:
     stw     r4, 0(r7)           ;# sse
 
     mullw   r3, r3, r3          ;# sum*sum
-    srawi   r3, r3, 4           ;# (sum*sum) >> 4
+    srlwi   r3, r3, 4           ;# (sum*sum) >> 4
     subf    r3, r3, r4          ;# sse - ((sum*sum) >> 4)
 
     epilogue

vp8/common/ppc/variance_subpixel_altivec.asm

@@ -157,7 +157,7 @@
     stw     r4, 0(r9)           ;# sse
 
     mullw   r3, r3, r3          ;# sum*sum
-    srawi   r3, r3, \DS         ;# (sum*sum) >> 8
+    srlwi   r3, r3, \DS         ;# (sum*sum) >> 8
     subf    r3, r3, r4          ;# sse - ((sum*sum) >> 8)
 .endm
 

vp8/common/ppflags.h

@@ -23,7 +23,8 @@ enum
     VP8D_DEBUG_TXT_RATE_INFO    = 1<<6,
     VP8D_DEBUG_DRAW_MV          = 1<<7,
     VP8D_DEBUG_CLR_BLK_MODES    = 1<<8,
-    VP8D_DEBUG_CLR_FRM_REF_BLKS = 1<<9
+    VP8D_DEBUG_CLR_FRM_REF_BLKS = 1<<9,
+    VP8D_MFQE                   = 1<<10
 };
 
 typedef struct

vp8/common/recon.h

@@ -1,111 +0,0 @@
-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-
-#ifndef __INC_RECON_H
-#define __INC_RECON_H
-
-#include "blockd.h"
-
-#define prototype_copy_block(sym) \
-    void sym(unsigned char *src, int src_pitch, unsigned char *dst, int dst_pitch)
-
-#define prototype_recon_block(sym) \
-    void sym(unsigned char *pred, short *diff, int diff_stride, unsigned char *dst, int pitch)
-
-#define prototype_recon_macroblock(sym) \
-    void sym(const struct vp8_recon_rtcd_vtable *rtcd, MACROBLOCKD *x)
-
-#define prototype_build_intra_predictors(sym) \
-    void sym(MACROBLOCKD *x)
-
-#define prototype_intra4x4_predict(sym) \
-    void sym(unsigned char *src, int src_stride, int b_mode, \
-             unsigned char *dst, int dst_stride)
-
-struct vp8_recon_rtcd_vtable;
-
-#if ARCH_X86 || ARCH_X86_64
-#include "x86/recon_x86.h"
-#endif
-
-#if ARCH_ARM
-#include "arm/recon_arm.h"
-#endif
-
-#ifndef vp8_recon_copy16x16
-#define vp8_recon_copy16x16 vp8_copy_mem16x16_c
-#endif
-extern prototype_copy_block(vp8_recon_copy16x16);
-
-#ifndef vp8_recon_copy8x8
-#define vp8_recon_copy8x8 vp8_copy_mem8x8_c
-#endif
-extern prototype_copy_block(vp8_recon_copy8x8);
-
-#ifndef vp8_recon_copy8x4
-#define vp8_recon_copy8x4 vp8_copy_mem8x4_c
-#endif
-extern prototype_copy_block(vp8_recon_copy8x4);
-
-#ifndef vp8_recon_build_intra_predictors_mby
-#define vp8_recon_build_intra_predictors_mby vp8_build_intra_predictors_mby
-#endif
-extern prototype_build_intra_predictors\
-    (vp8_recon_build_intra_predictors_mby);
-
-#ifndef vp8_recon_build_intra_predictors_mby_s
-#define vp8_recon_build_intra_predictors_mby_s vp8_build_intra_predictors_mby_s
-#endif
-extern prototype_build_intra_predictors\
-    (vp8_recon_build_intra_predictors_mby_s);
-
-#ifndef vp8_recon_build_intra_predictors_mbuv
-#define vp8_recon_build_intra_predictors_mbuv vp8_build_intra_predictors_mbuv
-#endif
-extern prototype_build_intra_predictors\
-    (vp8_recon_build_intra_predictors_mbuv);
-
-#ifndef vp8_recon_build_intra_predictors_mbuv_s
-#define vp8_recon_build_intra_predictors_mbuv_s vp8_build_intra_predictors_mbuv_s
-#endif
-extern prototype_build_intra_predictors\
-    (vp8_recon_build_intra_predictors_mbuv_s);
-
-#ifndef vp8_recon_intra4x4_predict
-#define vp8_recon_intra4x4_predict vp8_intra4x4_predict_c
-#endif
-extern prototype_intra4x4_predict\
-    (vp8_recon_intra4x4_predict);
-
-
-typedef prototype_copy_block((*vp8_copy_block_fn_t));
-typedef prototype_build_intra_predictors((*vp8_build_intra_pred_fn_t));
-typedef prototype_intra4x4_predict((*vp8_intra4x4_pred_fn_t));
-typedef struct vp8_recon_rtcd_vtable
-{
-    vp8_copy_block_fn_t  copy16x16;
-    vp8_copy_block_fn_t  copy8x8;
-    vp8_copy_block_fn_t  copy8x4;
-
-    vp8_build_intra_pred_fn_t  build_intra_predictors_mby_s;
-    vp8_build_intra_pred_fn_t  build_intra_predictors_mby;
-    vp8_build_intra_pred_fn_t  build_intra_predictors_mbuv_s;
-    vp8_build_intra_pred_fn_t  build_intra_predictors_mbuv;
-    vp8_intra4x4_pred_fn_t intra4x4_predict;
-} vp8_recon_rtcd_vtable_t;
-
-#if CONFIG_RUNTIME_CPU_DETECT
-#define RECON_INVOKE(ctx,fn) (ctx)->fn
-#else
-#define RECON_INVOKE(ctx,fn) vp8_recon_##fn
-#endif
-
-#endif

vp8/common/reconinter.c

@@ -9,10 +9,10 @@
  */
 
 
+#include <limits.h>
 #include "vpx_config.h"
+#include "vpx_rtcd.h"
 #include "vpx/vpx_integer.h"
-#include "recon.h"
-#include "subpixel.h"
 #include "blockd.h"
 #include "reconinter.h"
 #if CONFIG_RUNTIME_CPU_DETECT
@@ -123,25 +123,19 @@ void vp8_copy_mem8x4_c(
 }
 
 
-void vp8_build_inter_predictors_b(BLOCKD *d, int pitch, vp8_subpix_fn_t sppf)
+void vp8_build_inter_predictors_b(BLOCKD *d, int pitch, unsigned char *base_pre, int pre_stride, vp8_subpix_fn_t sppf)
 {
     int r;
-    unsigned char *ptr_base;
-    unsigned char *ptr;
     unsigned char *pred_ptr = d->predictor;
-
-    ptr_base = *(d->base_pre);
+    unsigned char *ptr;
+    ptr = base_pre + d->offset + (d->bmi.mv.as_mv.row >> 3) * pre_stride + (d->bmi.mv.as_mv.col >> 3);
 
     if (d->bmi.mv.as_mv.row & 7 || d->bmi.mv.as_mv.col & 7)
     {
-        ptr = ptr_base + d->pre + (d->bmi.mv.as_mv.row >> 3) * d->pre_stride + (d->bmi.mv.as_mv.col >> 3);
-        sppf(ptr, d->pre_stride, d->bmi.mv.as_mv.col & 7, d->bmi.mv.as_mv.row & 7, pred_ptr, pitch);
+        sppf(ptr, pre_stride, d->bmi.mv.as_mv.col & 7, d->bmi.mv.as_mv.row & 7, pred_ptr, pitch);
     }
     else
     {
-        ptr_base += d->pre + (d->bmi.mv.as_mv.row >> 3) * d->pre_stride + (d->bmi.mv.as_mv.col >> 3);
-        ptr = ptr_base;
-
         for (r = 0; r < 4; r++)
         {
 #if !(CONFIG_FAST_UNALIGNED)
@@ -153,65 +147,53 @@ void vp8_build_inter_predictors_b(BLOCKD *d, int pitch, vp8_subpix_fn_t sppf)
             *(uint32_t *)pred_ptr = *(uint32_t *)ptr ;
 #endif
             pred_ptr     += pitch;
-            ptr         += d->pre_stride;
+            ptr         += pre_stride;
         }
     }
 }
 
-static void build_inter_predictors4b(MACROBLOCKD *x, BLOCKD *d, unsigned char *dst, int dst_stride)
+static void build_inter_predictors4b(MACROBLOCKD *x, BLOCKD *d, unsigned char *dst, int dst_stride, unsigned char *base_pre, int pre_stride)
 {
-    unsigned char *ptr_base;
     unsigned char *ptr;
-
-    ptr_base = *(d->base_pre);
-    ptr = ptr_base + d->pre + (d->bmi.mv.as_mv.row >> 3) * d->pre_stride + (d->bmi.mv.as_mv.col >> 3);
+    ptr = base_pre + d->offset + (d->bmi.mv.as_mv.row >> 3) * pre_stride + (d->bmi.mv.as_mv.col >> 3);
 
     if (d->bmi.mv.as_mv.row & 7 || d->bmi.mv.as_mv.col & 7)
     {
-        x->subpixel_predict8x8(ptr, d->pre_stride, d->bmi.mv.as_mv.col & 7, d->bmi.mv.as_mv.row & 7, dst, dst_stride);
+        x->subpixel_predict8x8(ptr, pre_stride, d->bmi.mv.as_mv.col & 7, d->bmi.mv.as_mv.row & 7, dst, dst_stride);
     }
     else
     {
-        RECON_INVOKE(&x->rtcd->recon, copy8x8)(ptr, d->pre_stride, dst, dst_stride);
+        vp8_copy_mem8x8(ptr, pre_stride, dst, dst_stride);
     }
 }
 
-static void build_inter_predictors2b(MACROBLOCKD *x, BLOCKD *d, unsigned char *dst, int dst_stride)
+static void build_inter_predictors2b(MACROBLOCKD *x, BLOCKD *d, unsigned char *dst, int dst_stride, unsigned char *base_pre, int pre_stride)
 {
-    unsigned char *ptr_base;
     unsigned char *ptr;
-
-    ptr_base = *(d->base_pre);
-    ptr = ptr_base + d->pre + (d->bmi.mv.as_mv.row >> 3) * d->pre_stride + (d->bmi.mv.as_mv.col >> 3);
+    ptr = base_pre + d->offset + (d->bmi.mv.as_mv.row >> 3) * pre_stride + (d->bmi.mv.as_mv.col >> 3);
 
     if (d->bmi.mv.as_mv.row & 7 || d->bmi.mv.as_mv.col & 7)
     {
-        x->subpixel_predict8x4(ptr, d->pre_stride, d->bmi.mv.as_mv.col & 7, d->bmi.mv.as_mv.row & 7, dst, dst_stride);
+        x->subpixel_predict8x4(ptr, pre_stride, d->bmi.mv.as_mv.col & 7, d->bmi.mv.as_mv.row & 7, dst, dst_stride);
     }
     else
     {
-        RECON_INVOKE(&x->rtcd->recon, copy8x4)(ptr, d->pre_stride, dst, dst_stride);
+        vp8_copy_mem8x4(ptr, pre_stride, dst, dst_stride);
     }
 }
 
-static void build_inter_predictors_b(BLOCKD *d, unsigned char *dst, int dst_stride, vp8_subpix_fn_t sppf)
+static void build_inter_predictors_b(BLOCKD *d, unsigned char *dst, int dst_stride, unsigned char *base_pre, int pre_stride, vp8_subpix_fn_t sppf)
 {
     int r;
-    unsigned char *ptr_base;
     unsigned char *ptr;
-
-    ptr_base = *(d->base_pre);
+    ptr = base_pre + d->offset + (d->bmi.mv.as_mv.row >> 3) * pre_stride + (d->bmi.mv.as_mv.col >> 3);
 
     if (d->bmi.mv.as_mv.row & 7 || d->bmi.mv.as_mv.col & 7)
     {
-        ptr = ptr_base + d->pre + (d->bmi.mv.as_mv.row >> 3) * d->pre_stride + (d->bmi.mv.as_mv.col >> 3);
-        sppf(ptr, d->pre_stride, d->bmi.mv.as_mv.col & 7, d->bmi.mv.as_mv.row & 7, dst, dst_stride);
+        sppf(ptr, pre_stride, d->bmi.mv.as_mv.col & 7, d->bmi.mv.as_mv.row & 7, dst, dst_stride);
     }
     else
     {
-        ptr_base += d->pre + (d->bmi.mv.as_mv.row >> 3) * d->pre_stride + (d->bmi.mv.as_mv.col >> 3);
-        ptr = ptr_base;
-
         for (r = 0; r < 4; r++)
         {
 #if !(CONFIG_FAST_UNALIGNED)
@@ -223,7 +205,7 @@ static void build_inter_predictors_b(BLOCKD *d, unsigned char *dst, int dst_stri
             *(uint32_t *)dst = *(uint32_t *)ptr ;
 #endif
             dst     += dst_stride;
-            ptr         += d->pre_stride;
+            ptr     += pre_stride;
         }
     }
 }
@@ -239,22 +221,13 @@ void vp8_build_inter16x16_predictors_mbuv(MACROBLOCKD *x)
     int mv_row = x->mode_info_context->mbmi.mv.as_mv.row;
     int mv_col = x->mode_info_context->mbmi.mv.as_mv.col;
     int offset;
-    int pre_stride = x->block[16].pre_stride;
+    int pre_stride = x->pre.uv_stride;
 
     /* calc uv motion vectors */
-    if (mv_row < 0)
-        mv_row -= 1;
-    else
-        mv_row += 1;
-
-    if (mv_col < 0)
-        mv_col -= 1;
-    else
-        mv_col += 1;
-
+    mv_row += 1 | (mv_row >> (sizeof(int) * CHAR_BIT - 1));
+    mv_col += 1 | (mv_col >> (sizeof(int) * CHAR_BIT - 1));
     mv_row /= 2;
     mv_col /= 2;
-
     mv_row &= x->fullpixel_mask;
     mv_col &= x->fullpixel_mask;
 
@@ -269,8 +242,8 @@ void vp8_build_inter16x16_predictors_mbuv(MACROBLOCKD *x)
     }
     else
     {
-        RECON_INVOKE(&x->rtcd->recon, copy8x8)(uptr, pre_stride, upred_ptr, 8);
-        RECON_INVOKE(&x->rtcd->recon, copy8x8)(vptr, pre_stride, vpred_ptr, 8);
+        vp8_copy_mem8x8(uptr, pre_stride, upred_ptr, 8);
+        vp8_copy_mem8x8(vptr, pre_stride, vpred_ptr, 8);
     }
 }
 
@@ -278,6 +251,8 @@ void vp8_build_inter16x16_predictors_mbuv(MACROBLOCKD *x)
 void vp8_build_inter4x4_predictors_mbuv(MACROBLOCKD *x)
 {
     int i, j;
+    int pre_stride = x->pre.uv_stride;
+    unsigned char *base_pre;
 
     /* build uv mvs */
     for (i = 0; i < 2; i++)
@@ -295,8 +270,7 @@ void vp8_build_inter4x4_predictors_mbuv(MACROBLOCKD *x)
                    + x->block[yoffset+4].bmi.mv.as_mv.row
                    + x->block[yoffset+5].bmi.mv.as_mv.row;
 
-            if (temp < 0) temp -= 4;
-            else temp += 4;
+            temp += 4 + ((temp >> (sizeof(int) * CHAR_BIT - 1)) << 3);
 
             x->block[uoffset].bmi.mv.as_mv.row = (temp / 8) & x->fullpixel_mask;
 
@@ -305,54 +279,69 @@ void vp8_build_inter4x4_predictors_mbuv(MACROBLOCKD *x)
                    + x->block[yoffset+4].bmi.mv.as_mv.col
                    + x->block[yoffset+5].bmi.mv.as_mv.col;
 
-            if (temp < 0) temp -= 4;
-            else temp += 4;
+            temp += 4 + ((temp >> (sizeof(int) * CHAR_BIT - 1)) << 3);
 
             x->block[uoffset].bmi.mv.as_mv.col = (temp / 8) & x->fullpixel_mask;
 
-            x->block[voffset].bmi.mv.as_mv.row =
-                x->block[uoffset].bmi.mv.as_mv.row ;
-            x->block[voffset].bmi.mv.as_mv.col =
-                x->block[uoffset].bmi.mv.as_mv.col ;
+            x->block[voffset].bmi.mv.as_int = x->block[uoffset].bmi.mv.as_int;
         }
     }
 
-    for (i = 16; i < 24; i += 2)
+    base_pre = x->pre.u_buffer;
+    for (i = 16; i < 20; i += 2)
     {
         BLOCKD *d0 = &x->block[i];
         BLOCKD *d1 = &x->block[i+1];
 
         if (d0->bmi.mv.as_int == d1->bmi.mv.as_int)
-            build_inter_predictors2b(x, d0, d0->predictor, 8);
+            build_inter_predictors2b(x, d0, d0->predictor, 8, base_pre, pre_stride);
         else
         {
-            vp8_build_inter_predictors_b(d0, 8, x->subpixel_predict);
-            vp8_build_inter_predictors_b(d1, 8, x->subpixel_predict);
+            vp8_build_inter_predictors_b(d0, 8, base_pre, pre_stride, x->subpixel_predict);
+            vp8_build_inter_predictors_b(d1, 8, base_pre, pre_stride, x->subpixel_predict);
+        }
+    }
+
+    base_pre = x->pre.v_buffer;
+    for (i = 20; i < 24; i += 2)
+    {
+        BLOCKD *d0 = &x->block[i];
+        BLOCKD *d1 = &x->block[i+1];
+
+        if (d0->bmi.mv.as_int == d1->bmi.mv.as_int)
+            build_inter_predictors2b(x, d0, d0->predictor, 8, base_pre, pre_stride);
+        else
+        {
+            vp8_build_inter_predictors_b(d0, 8, base_pre, pre_stride, x->subpixel_predict);
+            vp8_build_inter_predictors_b(d1, 8, base_pre, pre_stride, x->subpixel_predict);
         }
     }
 }
 
 
 /*encoder only*/
-void vp8_build_inter16x16_predictors_mby(MACROBLOCKD *x)
+void vp8_build_inter16x16_predictors_mby(MACROBLOCKD *x,
+                                         unsigned char *dst_y,
+                                         int dst_ystride)
 {
     unsigned char *ptr_base;
     unsigned char *ptr;
-    unsigned char *pred_ptr = x->predictor;
     int mv_row = x->mode_info_context->mbmi.mv.as_mv.row;
     int mv_col = x->mode_info_context->mbmi.mv.as_mv.col;
-    int pre_stride = x->block[0].pre_stride;
+    int pre_stride = x->pre.y_stride;
 
     ptr_base = x->pre.y_buffer;
     ptr = ptr_base + (mv_row >> 3) * pre_stride + (mv_col >> 3);
 
     if ((mv_row | mv_col) & 7)
     {
-        x->subpixel_predict16x16(ptr, pre_stride, mv_col & 7, mv_row & 7, pred_ptr, 16);
+        x->subpixel_predict16x16(ptr, pre_stride, mv_col & 7, mv_row & 7,
+                                 dst_y, dst_ystride);
     }
     else
     {
-        RECON_INVOKE(&x->rtcd->recon, copy16x16)(ptr, pre_stride, pred_ptr, 16);
+        vp8_copy_mem16x16(ptr, pre_stride, dst_y,
+            dst_ystride);
     }
 }
 
@@ -406,7 +395,7 @@ void vp8_build_inter16x16_predictors_mb(MACROBLOCKD *x,
     int_mv _16x16mv;
 
     unsigned char *ptr_base = x->pre.y_buffer;
-    int pre_stride = x->block[0].pre_stride;
+    int pre_stride = x->pre.y_stride;
 
     _16x16mv.as_int = x->mode_info_context->mbmi.mv.as_int;
 
@@ -423,23 +412,14 @@ void vp8_build_inter16x16_predictors_mb(MACROBLOCKD *x,
     }
     else
     {
-        RECON_INVOKE(&x->rtcd->recon, copy16x16)(ptr, pre_stride, dst_y, dst_ystride);
+        vp8_copy_mem16x16(ptr, pre_stride, dst_y, dst_ystride);
     }
 
     /* calc uv motion vectors */
-    if ( _16x16mv.as_mv.row < 0)
-      _16x16mv.as_mv.row -= 1;
-    else
-      _16x16mv.as_mv.row += 1;
-
-    if (_16x16mv.as_mv.col < 0)
-        _16x16mv.as_mv.col -= 1;
-    else
-        _16x16mv.as_mv.col += 1;
-
+    _16x16mv.as_mv.row += 1 | (_16x16mv.as_mv.row >> (sizeof(int) * CHAR_BIT - 1));
+    _16x16mv.as_mv.col += 1 | (_16x16mv.as_mv.col >> (sizeof(int) * CHAR_BIT - 1));
     _16x16mv.as_mv.row /= 2;
     _16x16mv.as_mv.col /= 2;
-
     _16x16mv.as_mv.row &= x->fullpixel_mask;
     _16x16mv.as_mv.col &= x->fullpixel_mask;
 
@@ -455,19 +435,21 @@ void vp8_build_inter16x16_predictors_mb(MACROBLOCKD *x,
     }
     else
     {
-        RECON_INVOKE(&x->rtcd->recon, copy8x8)(uptr, pre_stride, dst_u, dst_uvstride);
-        RECON_INVOKE(&x->rtcd->recon, copy8x8)(vptr, pre_stride, dst_v, dst_uvstride);
+        vp8_copy_mem8x8(uptr, pre_stride, dst_u, dst_uvstride);
+        vp8_copy_mem8x8(vptr, pre_stride, dst_v, dst_uvstride);
     }
 }
 
 static void build_inter4x4_predictors_mb(MACROBLOCKD *x)
 {
     int i;
+    unsigned char *base_dst = x->dst.y_buffer;
+    unsigned char *base_pre = x->pre.y_buffer;
 
     if (x->mode_info_context->mbmi.partitioning < 3)
     {
         BLOCKD *b;
-        int dst_stride = x->block[ 0].dst_stride;
+        int dst_stride = x->dst.y_stride;
 
         x->block[ 0].bmi = x->mode_info_context->bmi[ 0];
         x->block[ 2].bmi = x->mode_info_context->bmi[ 2];
@@ -482,13 +464,13 @@ static void build_inter4x4_predictors_mb(MACROBLOCKD *x)
         }
 
         b = &x->block[ 0];
-        build_inter_predictors4b(x, b, *(b->base_dst) + b->dst, dst_stride);
+        build_inter_predictors4b(x, b, base_dst + b->offset, dst_stride, base_pre, dst_stride);
         b = &x->block[ 2];
-        build_inter_predictors4b(x, b, *(b->base_dst) + b->dst, dst_stride);
+        build_inter_predictors4b(x, b, base_dst + b->offset, dst_stride, base_pre, dst_stride);
         b = &x->block[ 8];
-        build_inter_predictors4b(x, b, *(b->base_dst) + b->dst, dst_stride);
+        build_inter_predictors4b(x, b, base_dst + b->offset, dst_stride, base_pre, dst_stride);
         b = &x->block[10];
-        build_inter_predictors4b(x, b, *(b->base_dst) + b->dst, dst_stride);
+        build_inter_predictors4b(x, b, base_dst + b->offset, dst_stride, base_pre, dst_stride);
     }
     else
     {
@@ -496,7 +478,7 @@ static void build_inter4x4_predictors_mb(MACROBLOCKD *x)
         {
             BLOCKD *d0 = &x->block[i];
             BLOCKD *d1 = &x->block[i+1];
-            int dst_stride = x->block[ 0].dst_stride;
+            int dst_stride = x->dst.y_stride;
 
             x->block[i+0].bmi = x->mode_info_context->bmi[i+0];
             x->block[i+1].bmi = x->mode_info_context->bmi[i+1];
@@ -507,31 +489,51 @@ static void build_inter4x4_predictors_mb(MACROBLOCKD *x)
             }
 
             if (d0->bmi.mv.as_int == d1->bmi.mv.as_int)
-                build_inter_predictors2b(x, d0, *(d0->base_dst) + d0->dst, dst_stride);
+                build_inter_predictors2b(x, d0, base_dst + d0->offset, dst_stride, base_pre, dst_stride);
             else
             {
-                build_inter_predictors_b(d0, *(d0->base_dst) + d0->dst, dst_stride, x->subpixel_predict);
-                build_inter_predictors_b(d1, *(d1->base_dst) + d1->dst, dst_stride, x->subpixel_predict);
+                build_inter_predictors_b(d0, base_dst + d0->offset, dst_stride, base_pre, dst_stride, x->subpixel_predict);
+                build_inter_predictors_b(d1, base_dst + d1->offset, dst_stride, base_pre, dst_stride, x->subpixel_predict);
             }
 
         }
 
     }
-
-    for (i = 16; i < 24; i += 2)
+    base_dst = x->dst.u_buffer;
+    base_pre = x->pre.u_buffer;
+    for (i = 16; i < 20; i += 2)
     {
         BLOCKD *d0 = &x->block[i];
         BLOCKD *d1 = &x->block[i+1];
-        int dst_stride = x->block[ 16].dst_stride;
+        int dst_stride = x->dst.uv_stride;
 
         /* Note: uv mvs already clamped in build_4x4uvmvs() */
 
         if (d0->bmi.mv.as_int == d1->bmi.mv.as_int)
-            build_inter_predictors2b(x, d0, *(d0->base_dst) + d0->dst, dst_stride);
+            build_inter_predictors2b(x, d0, base_dst + d0->offset, dst_stride, base_pre, dst_stride);
         else
         {
-            build_inter_predictors_b(d0, *(d0->base_dst) + d0->dst, dst_stride, x->subpixel_predict);
-            build_inter_predictors_b(d1, *(d1->base_dst) + d1->dst, dst_stride, x->subpixel_predict);
+            build_inter_predictors_b(d0, base_dst + d0->offset, dst_stride, base_pre, dst_stride, x->subpixel_predict);
+            build_inter_predictors_b(d1, base_dst + d1->offset, dst_stride, base_pre, dst_stride, x->subpixel_predict);
+        }
+    }
+
+    base_dst = x->dst.v_buffer;
+    base_pre = x->pre.v_buffer;
+    for (i = 20; i < 24; i += 2)
+    {
+        BLOCKD *d0 = &x->block[i];
+        BLOCKD *d1 = &x->block[i+1];
+        int dst_stride = x->dst.uv_stride;
+
+        /* Note: uv mvs already clamped in build_4x4uvmvs() */
+
+        if (d0->bmi.mv.as_int == d1->bmi.mv.as_int)
+            build_inter_predictors2b(x, d0, base_dst + d0->offset, dst_stride, base_pre, dst_stride);
+        else
+        {
+            build_inter_predictors_b(d0, base_dst + d0->offset, dst_stride, base_pre, dst_stride, x->subpixel_predict);
+            build_inter_predictors_b(d1, base_dst + d1->offset, dst_stride, base_pre, dst_stride, x->subpixel_predict);
         }
     }
 }
@@ -556,8 +558,7 @@ void build_4x4uvmvs(MACROBLOCKD *x)
                  + x->mode_info_context->bmi[yoffset + 4].mv.as_mv.row
                  + x->mode_info_context->bmi[yoffset + 5].mv.as_mv.row;
 
-            if (temp < 0) temp -= 4;
-            else temp += 4;
+            temp += 4 + ((temp >> (sizeof(int) * CHAR_BIT - 1)) << 3);
 
             x->block[uoffset].bmi.mv.as_mv.row = (temp / 8) & x->fullpixel_mask;
 
@@ -566,18 +567,14 @@ void build_4x4uvmvs(MACROBLOCKD *x)
                  + x->mode_info_context->bmi[yoffset + 4].mv.as_mv.col
                  + x->mode_info_context->bmi[yoffset + 5].mv.as_mv.col;
 
-            if (temp < 0) temp -= 4;
-            else temp += 4;
+            temp += 4 + ((temp >> (sizeof(int) * CHAR_BIT - 1)) << 3);
 
             x->block[uoffset].bmi.mv.as_mv.col = (temp / 8) & x->fullpixel_mask;
 
             if (x->mode_info_context->mbmi.need_to_clamp_mvs)
                 clamp_uvmv_to_umv_border(&x->block[uoffset].bmi.mv.as_mv, x);
 
-            x->block[voffset].bmi.mv.as_mv.row =
-                x->block[uoffset].bmi.mv.as_mv.row ;
-            x->block[voffset].bmi.mv.as_mv.col =
-                x->block[uoffset].bmi.mv.as_mv.col ;
+            x->block[voffset].bmi.mv.as_int = x->block[uoffset].bmi.mv.as_int;
         }
     }
 }
@@ -596,69 +593,3 @@ void vp8_build_inter_predictors_mb(MACROBLOCKD *xd)
         build_inter4x4_predictors_mb(xd);
     }
 }
-/* encoder only*/
-static void build_inter4x4_predictors_mb_e(MACROBLOCKD *x)
-{
-    int i;
-
-    if (x->mode_info_context->mbmi.partitioning < 3)
-    {
-        x->block[ 0].bmi = x->mode_info_context->bmi[ 0];
-        x->block[ 2].bmi = x->mode_info_context->bmi[ 2];
-        x->block[ 8].bmi = x->mode_info_context->bmi[ 8];
-        x->block[10].bmi = x->mode_info_context->bmi[10];
-
-        build_inter_predictors4b(x, &x->block[ 0], x->block[ 0].predictor, 16);
-        build_inter_predictors4b(x, &x->block[ 2], x->block[ 2].predictor, 16);
-        build_inter_predictors4b(x, &x->block[ 8], x->block[ 8].predictor, 16);
-        build_inter_predictors4b(x, &x->block[10], x->block[10].predictor, 16);
-    }
-    else
-    {
-        for (i = 0; i < 16; i += 2)
-        {
-            BLOCKD *d0 = &x->block[i];
-            BLOCKD *d1 = &x->block[i+1];
-
-            x->block[i+0].bmi = x->mode_info_context->bmi[i+0];
-            x->block[i+1].bmi = x->mode_info_context->bmi[i+1];
-
-            if (d0->bmi.mv.as_int == d1->bmi.mv.as_int)
-                build_inter_predictors2b(x, d0, d0->predictor, 16);
-            else
-            {
-                build_inter_predictors_b(d0, d0->predictor, 16, x->subpixel_predict);
-                build_inter_predictors_b(d1, d1->predictor, 16, x->subpixel_predict);
-            }
-
-        }
-
-    }
-
-    for (i = 16; i < 24; i += 2)
-    {
-        BLOCKD *d0 = &x->block[i];
-        BLOCKD *d1 = &x->block[i+1];
-
-        if (d0->bmi.mv.as_int == d1->bmi.mv.as_int)
-            build_inter_predictors2b(x, d0, d0->predictor, 8);
-        else
-        {
-            build_inter_predictors_b(d0, d0->predictor, 8, x->subpixel_predict);
-            build_inter_predictors_b(d1, d1->predictor, 8, x->subpixel_predict);
-        }
-    }
-}
-void vp8_build_inter_predictors_mb_e(MACROBLOCKD *xd)
-{
-    if (xd->mode_info_context->mbmi.mode != SPLITMV)
-    {
-        vp8_build_inter16x16_predictors_mb(xd, xd->predictor, &xd->predictor[256],
-                                           &xd->predictor[320], 16, 8);
-    }
-    else
-    {
-        build_4x4uvmvs(xd);
-        build_inter4x4_predictors_mb_e(xd);
-    }
-}

vp8/common/reconinter.h

@@ -21,11 +21,15 @@ extern void vp8_build_inter16x16_predictors_mb(MACROBLOCKD *x,
                                                int dst_uvstride);
 
 
-extern void vp8_build_inter16x16_predictors_mby(MACROBLOCKD *x);
-extern void vp8_build_inter_predictors_b(BLOCKD *d, int pitch, vp8_subpix_fn_t sppf);
+extern void vp8_build_inter16x16_predictors_mby(MACROBLOCKD *x,
+                                                unsigned char *dst_y,
+                                                int dst_ystride);
+extern void vp8_build_inter_predictors_b(BLOCKD *d, int pitch,
+                                         unsigned char *base_pre,
+                                         int pre_stride,
+                                         vp8_subpix_fn_t sppf);
 
 extern void vp8_build_inter16x16_predictors_mbuv(MACROBLOCKD *x);
 extern void vp8_build_inter4x4_predictors_mbuv(MACROBLOCKD *x);
-extern void vp8_build_inter_predictors_mb_e(MACROBLOCKD *xd);
 
 #endif

vp8/common/reconintra.c

@@ -10,147 +10,24 @@
 
 
 #include "vpx_config.h"
-#include "recon.h"
-#include "reconintra.h"
+#include "vpx_rtcd.h"
 #include "vpx_mem/vpx_mem.h"
+#include "blockd.h"
 
-/* For skip_recon_mb(), add vp8_build_intra_predictors_mby_s(MACROBLOCKD *x) and
- * vp8_build_intra_predictors_mbuv_s(MACROBLOCKD *x).
- */
-
-void vp8_build_intra_predictors_mby(MACROBLOCKD *x)
+void vp8_build_intra_predictors_mby_s_c(MACROBLOCKD *x,
+                                          unsigned char * yabove_row,
+                                          unsigned char * yleft,
+                                          int left_stride,
+                                          unsigned char * ypred_ptr,
+                                          int y_stride)
 {
-
-    unsigned char *yabove_row = x->dst.y_buffer - x->dst.y_stride;
     unsigned char yleft_col[16];
     unsigned char ytop_left = yabove_row[-1];
-    unsigned char *ypred_ptr = x->predictor;
     int r, c, i;
 
     for (i = 0; i < 16; i++)
     {
-        yleft_col[i] = x->dst.y_buffer [i* x->dst.y_stride -1];
-    }
-
-    /* for Y */
-    switch (x->mode_info_context->mbmi.mode)
-    {
-    case DC_PRED:
-    {
-        int expected_dc;
-        int i;
-        int shift;
-        int average = 0;
-
-
-        if (x->up_available || x->left_available)
-        {
-            if (x->up_available)
-            {
-                for (i = 0; i < 16; i++)
-                {
-                    average += yabove_row[i];
-                }
-            }
-
-            if (x->left_available)
-            {
-
-                for (i = 0; i < 16; i++)
-                {
-                    average += yleft_col[i];
-                }
-
-            }
-
-
-
-            shift = 3 + x->up_available + x->left_available;
-            expected_dc = (average + (1 << (shift - 1))) >> shift;
-        }
-        else
-        {
-            expected_dc = 128;
-        }
-
-        vpx_memset(ypred_ptr, expected_dc, 256);
-    }
-    break;
-    case V_PRED:
-    {
-
-        for (r = 0; r < 16; r++)
-        {
-
-            ((int *)ypred_ptr)[0] = ((int *)yabove_row)[0];
-            ((int *)ypred_ptr)[1] = ((int *)yabove_row)[1];
-            ((int *)ypred_ptr)[2] = ((int *)yabove_row)[2];
-            ((int *)ypred_ptr)[3] = ((int *)yabove_row)[3];
-            ypred_ptr += 16;
-        }
-    }
-    break;
-    case H_PRED:
-    {
-
-        for (r = 0; r < 16; r++)
-        {
-
-            vpx_memset(ypred_ptr, yleft_col[r], 16);
-            ypred_ptr += 16;
-        }
-
-    }
-    break;
-    case TM_PRED:
-    {
-
-        for (r = 0; r < 16; r++)
-        {
-            for (c = 0; c < 16; c++)
-            {
-                int pred =  yleft_col[r] + yabove_row[ c] - ytop_left;
-
-                if (pred < 0)
-                    pred = 0;
-
-                if (pred > 255)
-                    pred = 255;
-
-                ypred_ptr[c] = pred;
-            }
-
-            ypred_ptr += 16;
-        }
-
-    }
-    break;
-    case B_PRED:
-    case NEARESTMV:
-    case NEARMV:
-    case ZEROMV:
-    case NEWMV:
-    case SPLITMV:
-    case MB_MODE_COUNT:
-        break;
-    }
-}
-
-void vp8_build_intra_predictors_mby_s(MACROBLOCKD *x)
-{
-
-    unsigned char *yabove_row = x->dst.y_buffer - x->dst.y_stride;
-    unsigned char yleft_col[16];
-    unsigned char ytop_left = yabove_row[-1];
-    unsigned char *ypred_ptr = x->predictor;
-    int r, c, i;
-
-    int y_stride = x->dst.y_stride;
-    ypred_ptr = x->dst.y_buffer; /*x->predictor;*/
-
-    for (i = 0; i < 16; i++)
-    {
-        yleft_col[i] = x->dst.y_buffer [i* x->dst.y_stride -1];
+        yleft_col[i] = yleft[i* left_stride];
     }
 
     /* for Y */
@@ -198,7 +75,7 @@ void vp8_build_intra_predictors_mby_s(MACROBLOCKD *x)
         for (r = 0; r < 16; r++)
         {
             vpx_memset(ypred_ptr, expected_dc, 16);
-            ypred_ptr += y_stride; /*16;*/
+            ypred_ptr += y_stride;
         }
     }
     break;
@@ -212,7 +89,7 @@ void vp8_build_intra_predictors_mby_s(MACROBLOCKD *x)
             ((int *)ypred_ptr)[1] = ((int *)yabove_row)[1];
             ((int *)ypred_ptr)[2] = ((int *)yabove_row)[2];
             ((int *)ypred_ptr)[3] = ((int *)yabove_row)[3];
-            ypred_ptr += y_stride; /*16;*/
+            ypred_ptr += y_stride;
         }
     }
     break;
@@ -223,7 +100,7 @@ void vp8_build_intra_predictors_mby_s(MACROBLOCKD *x)
         {
 
             vpx_memset(ypred_ptr, yleft_col[r], 16);
-            ypred_ptr += y_stride;  /*16;*/
+            ypred_ptr += y_stride;
         }
 
     }
@@ -246,7 +123,7 @@ void vp8_build_intra_predictors_mby_s(MACROBLOCKD *x)
                 ypred_ptr[c] = pred;
             }
 
-            ypred_ptr += y_stride;  /*16;*/
+            ypred_ptr += y_stride;
         }
 
     }
@@ -262,162 +139,27 @@ void vp8_build_intra_predictors_mby_s(MACROBLOCKD *x)
     }
 }
 
-void vp8_build_intra_predictors_mbuv(MACROBLOCKD *x)
+void vp8_build_intra_predictors_mbuv_s_c(MACROBLOCKD *x,
+                                         unsigned char * uabove_row,
+                                         unsigned char * vabove_row,
+                                         unsigned char * uleft,
+                                         unsigned char * vleft,
+                                         int left_stride,
+                                         unsigned char * upred_ptr,
+                                         unsigned char * vpred_ptr,
+                                         int pred_stride)
 {
-    unsigned char *uabove_row = x->dst.u_buffer - x->dst.uv_stride;
-    unsigned char uleft_col[16];
+    unsigned char uleft_col[8];
     unsigned char utop_left = uabove_row[-1];
-    unsigned char *vabove_row = x->dst.v_buffer - x->dst.uv_stride;
-    unsigned char vleft_col[20];
+    unsigned char vleft_col[8];
     unsigned char vtop_left = vabove_row[-1];
-    unsigned char *upred_ptr = &x->predictor[256];
-    unsigned char *vpred_ptr = &x->predictor[320];
-    int i, j;
-
-    for (i = 0; i < 8; i++)
-    {
-        uleft_col[i] = x->dst.u_buffer [i* x->dst.uv_stride -1];
-        vleft_col[i] = x->dst.v_buffer [i* x->dst.uv_stride -1];
-    }
-
-    switch (x->mode_info_context->mbmi.uv_mode)
-    {
-    case DC_PRED:
-    {
-        int expected_udc;
-        int expected_vdc;
-        int i;
-        int shift;
-        int Uaverage = 0;
-        int Vaverage = 0;
-
-        if (x->up_available)
-        {
-            for (i = 0; i < 8; i++)
-            {
-                Uaverage += uabove_row[i];
-                Vaverage += vabove_row[i];
-            }
-        }
-
-        if (x->left_available)
-        {
-            for (i = 0; i < 8; i++)
-            {
-                Uaverage += uleft_col[i];
-                Vaverage += vleft_col[i];
-            }
-        }
-
-        if (!x->up_available && !x->left_available)
-        {
-            expected_udc = 128;
-            expected_vdc = 128;
-        }
-        else
-        {
-            shift = 2 + x->up_available + x->left_available;
-            expected_udc = (Uaverage + (1 << (shift - 1))) >> shift;
-            expected_vdc = (Vaverage + (1 << (shift - 1))) >> shift;
-        }
-
-
-        vpx_memset(upred_ptr, expected_udc, 64);
-        vpx_memset(vpred_ptr, expected_vdc, 64);
-
-
-    }
-    break;
-    case V_PRED:
-    {
-        int i;
-
-        for (i = 0; i < 8; i++)
-        {
-            vpx_memcpy(upred_ptr, uabove_row, 8);
-            vpx_memcpy(vpred_ptr, vabove_row, 8);
-            upred_ptr += 8;
-            vpred_ptr += 8;
-        }
-
-    }
-    break;
-    case H_PRED:
-    {
-        int i;
-
-        for (i = 0; i < 8; i++)
-        {
-            vpx_memset(upred_ptr, uleft_col[i], 8);
-            vpx_memset(vpred_ptr, vleft_col[i], 8);
-            upred_ptr += 8;
-            vpred_ptr += 8;
-        }
-    }
-
-    break;
-    case TM_PRED:
-    {
-        int i;
-
-        for (i = 0; i < 8; i++)
-        {
-            for (j = 0; j < 8; j++)
-            {
-                int predu = uleft_col[i] + uabove_row[j] - utop_left;
-                int predv = vleft_col[i] + vabove_row[j] - vtop_left;
-
-                if (predu < 0)
-                    predu = 0;
-
-                if (predu > 255)
-                    predu = 255;
-
-                if (predv < 0)
-                    predv = 0;
-
-                if (predv > 255)
-                    predv = 255;
-
-                upred_ptr[j] = predu;
-                vpred_ptr[j] = predv;
-            }
-
-            upred_ptr += 8;
-            vpred_ptr += 8;
-        }
-
-    }
-    break;
-    case B_PRED:
-    case NEARESTMV:
-    case NEARMV:
-    case ZEROMV:
-    case NEWMV:
-    case SPLITMV:
-    case MB_MODE_COUNT:
-        break;
-    }
-}
-
-void vp8_build_intra_predictors_mbuv_s(MACROBLOCKD *x)
-{
-    unsigned char *uabove_row = x->dst.u_buffer - x->dst.uv_stride;
-    unsigned char uleft_col[16];
-    unsigned char utop_left = uabove_row[-1];
-    unsigned char *vabove_row = x->dst.v_buffer - x->dst.uv_stride;
-    unsigned char vleft_col[20];
-    unsigned char vtop_left = vabove_row[-1];
-    unsigned char *upred_ptr = x->dst.u_buffer; /*&x->predictor[256];*/
-    unsigned char *vpred_ptr = x->dst.v_buffer; /*&x->predictor[320];*/
-    int uv_stride = x->dst.uv_stride;
 
     int i, j;
 
     for (i = 0; i < 8; i++)
     {
-        uleft_col[i] = x->dst.u_buffer [i* x->dst.uv_stride -1];
-        vleft_col[i] = x->dst.v_buffer [i* x->dst.uv_stride -1];
+        uleft_col[i] = uleft [i* left_stride];
+        vleft_col[i] = vleft [i* left_stride];
     }
 
     switch (x->mode_info_context->mbmi.uv_mode)
@@ -468,8 +210,8 @@ void vp8_build_intra_predictors_mbuv_s(MACROBLOCKD *x)
         {
             vpx_memset(upred_ptr, expected_udc, 8);
             vpx_memset(vpred_ptr, expected_vdc, 8);
-            upred_ptr += uv_stride; /*8;*/
-            vpred_ptr += uv_stride; /*8;*/
+            upred_ptr += pred_stride;
+            vpred_ptr += pred_stride;
         }
     }
     break;
@@ -481,8 +223,8 @@ void vp8_build_intra_predictors_mbuv_s(MACROBLOCKD *x)
         {
             vpx_memcpy(upred_ptr, uabove_row, 8);
             vpx_memcpy(vpred_ptr, vabove_row, 8);
-            upred_ptr += uv_stride; /*8;*/
-            vpred_ptr += uv_stride; /*8;*/
+            upred_ptr += pred_stride;
+            vpred_ptr += pred_stride;
         }
 
     }
@@ -495,8 +237,8 @@ void vp8_build_intra_predictors_mbuv_s(MACROBLOCKD *x)
         {
             vpx_memset(upred_ptr, uleft_col[i], 8);
             vpx_memset(vpred_ptr, vleft_col[i], 8);
-            upred_ptr += uv_stride; /*8;*/
-            vpred_ptr += uv_stride; /*8;*/
+            upred_ptr += pred_stride;
+            vpred_ptr += pred_stride;
         }
     }
 
@@ -528,8 +270,8 @@ void vp8_build_intra_predictors_mbuv_s(MACROBLOCKD *x)
                 vpred_ptr[j] = predv;
             }
 
-            upred_ptr += uv_stride; /*8;*/
-            vpred_ptr += uv_stride; /*8;*/
+            upred_ptr += pred_stride;
+            vpred_ptr += pred_stride;
         }
 
     }