Added configure option to enable error-concealment. Disabled by default.

Change-Id: I94580a5ecb13520195ea2b8a10ca11bb5a01d2a6
Concealed MBs are always SPLITMV with partition=3. This can be optimized.
2011-04-29 14:08:47 +02:00 · 2011-04-29 13:50:15 +02:00 · 2011-04-29 11:22:09 +02:00 · 2011-04-28 16:28:07 +02:00 · 2011-04-20 12:08:27 +02:00 · 2011-04-19 16:23:05 +02:00
226 changed files with 19662 additions and 3408 deletions
--- a/.mailmap
+++ b/.mailmap
@@ -1,4 +1,2 @@
 Adrian Grange <agrange@google.com>
 Johann Koenig <johannkoenig@google.com>
-Tero Rintaluoma <teror@google.com> <tero.rintaluoma@on2.com>
-Tom Finegan <tomfinegan@google.com>
--- a/9
+++ b/9
@@ -4,18 +4,13 @@
 Aaron Watry <awatry@gmail.com>
 Adrian Grange <agrange@google.com>
 Alex Converse <alex.converse@gmail.com>
-Andoni Morales Alastruey <ylatuya@gmail.com>
 Andres Mejia <mcitadel@gmail.com>
-Attila Nagy <attilanagy@google.com>
 Fabio Pedretti <fabio.ped@libero.it>
 Frank Galligan <fgalligan@google.com>
 Fredrik Söderquist <fs@opera.com>
 Fritz Koenig <frkoenig@google.com>
-Gaute Strokkenes <gaute.strokkenes@broadcom.com>
 Giuseppe Scrivano <gscrivano@gnu.org>
 Guillermo Ballester Valor <gbvalor@gmail.com>
-Henrik Lundin <hlundin@google.com>
-James Berry <jamesberry@google.com>
 James Zern <jzern@google.com>
 Jan Kratochvil <jan.kratochvil@redhat.com>
 Jeff Muizelaar <jmuizelaar@mozilla.com>
@@ -28,14 +23,10 @@ Luca Barbato <lu_zero@gentoo.org>
 Makoto Kato <makoto.kt@gmail.com>
 Martin Ettl <ettl.martin78@googlemail.com>
 Michael Kohler <michaelkohler@live.com>
-Mikhal Shemer <mikhal@google.com>
-Pascal Massimino <pascal.massimino@gmail.com>
-Patrik Westin <patrik.westin@gmail.com>
 Paul Wilkins <paulwilkins@google.com>
 Pavol Rusnak <stick@gk2.sk>
 Philip Jägenstedt <philipj@opera.com>
 Scott LaVarnway <slavarnway@google.com>
-Tero Rintaluoma <teror@google.com>
 Timothy B. Terriberry <tterribe@xiph.org>
 Tom Finegan <tomfinegan@google.com>
 Yaowu Xu <yaowu@google.com>
--- a/77
+++ b/77
@@ -1,80 +1,3 @@
-2011-03-07 v0.9.6 "Bali"
-  Our second named release, focused on a faster, higher quality, encoder.
-
-  - Upgrading:
-    This release is backwards compatible with Aylesbury (v0.9.5). Users
-    of older releases should refer to the Upgrading notes in this
-    document for that release.
-
-  - Enhancements:
-      vpxenc --psnr shows a summary when encode completes
-      --tune=ssim option to enable activity masking
-      improved postproc visualizations for development
-      updated support for Apple iOS to SDK 4.2
-      query decoder to determine which reference frames were updated
-      implemented error tracking in the decoder
-      fix pipe support on windows
-
-  - Speed:
-      Primary focus was on good quality mode, speed 0. Average improvement
-      on x86 about 40%, up to 100% on user-generated content at that speed.
-      Best quality mode speed improved 35%, and realtime speed 10-20%. This
-      release also saw significant improvement in realtime encoding speed
-      on ARM platforms.
-
-        Improved encoder threading
-        Dont pick encoder filter level when loopfilter is disabled.
-        Avoid double copying of key frames into alt and golden buffer
-        FDCT optimizations.
-        x86 sse2 temporal filter
-        SSSE3 version of fast quantizer
-        vp8_rd_pick_best_mbsegmentation code restructure
-        Adjusted breakout RD for SPLITMV
-        Changed segmentation check order
-        Improved rd_pick_intra4x4block
-        Adds armv6 optimized variance calculation
-        ARMv6 optimized sad16x16
-        ARMv6 optimized half pixel variance calculations
-        Full search SAD function optimization in SSE4.1
-        Improve MV prediction accuracy to achieve performance gain
-        Improve MV prediction in vp8_pick_inter_mode() for speed>3
-
-  - Quality:
-      Best quality mode improved PSNR 6.3%, and SSIM 6.1%. This release
-      also includes support for "activity masking," which greatly improves
-      SSIM at the expense of PSNR. For now, this feature is available with
-      the --tune=ssim option. Further experimentation in this area
-      is ongoing. This release also introduces a new rate control mode
-      called "CQ," which changes the allocation of bits within a clip to
-      the sections where they will have the most visual impact.
-
-        Tuning for the more exact quantizer.
-        Relax rate control for last few frames
-        CQ Mode
-        Limit key frame quantizer for forced key frames.
-        KF/GF Pulsing
-        Add simple version of activity masking.
-        make rdmult adaptive for intra in quantizer RDO
-        cap the best quantizer for 2nd order DC
-        change the threshold of DC check for encode breakout
-
-  - Bug Fixes:
-      Fix crash on Sparc Solaris.
-      Fix counter of fixed keyframe distance
-      ARNR filter pointer update bug fix
-      Fixed use of motion percentage in KF/GF group calc
-      Changed condition for using RD in Intra Mode
-      Fix encoder real-time only configuration.
-      Fix ARM encoder crash with multiple token partitions
-      Fixed bug first cluster timecode of webm file is wrong.
-      Fixed various encoder bugs with odd-sized images
-      vp8e_get_preview fixed when spatial resampling enabled
-      quantizer: fix assertion in fast quantizer path
-      Allocate source buffers to be multiples of 16
-      Fix for manual Golden frame frequency
-      Fix drastic undershoot in long form content
-
-
 2010-10-28 v0.9.5 "Aylesbury"
  Our first named release, focused on a faster decoder, and a better encoder.

--- a/4
+++ b/4
@@ -45,14 +45,18 @@ COMPILING THE APPLICATIONS/LIBRARIES:
    armv5te-linux-rvct
    armv5te-linux-gcc
    armv5te-symbian-gcc
+    armv5te-wince-vs8
    armv6-darwin-gcc
    armv6-linux-rvct
    armv6-linux-gcc
    armv6-symbian-gcc
+    armv6-wince-vs8
    iwmmxt-linux-rvct
    iwmmxt-linux-gcc
+    iwmmxt-wince-vs8
    iwmmxt2-linux-rvct
    iwmmxt2-linux-gcc
+    iwmmxt2-wince-vs8
    armv7-linux-rvct
    armv7-linux-gcc
    mips32-linux-gcc
--- a/build/arm-wince-vs8/armasmv5.rules
+++ b/build/arm-wince-vs8/armasmv5.rules
@@ -0,0 +1,20 @@
+<?xml version="1.0" encoding="utf-8"?>
+<VisualStudioToolFile
+	Name="armasm"
+	Version="8.00"
+	>
+	<Rules>
+		<CustomBuildRule
+			Name="ARMASM"
+			DisplayName="Armasm Assembler"
+			CommandLine="armasm -o &quot;$(IntDir)\$(InputName).obj&quot; $(InputPath) -32 -ARCH 5&#x0D;&#x0A;"
+			Outputs="$(IntDir)\$(InputName).obj"
+			FileExtensions="*.asm"
+			ExecutionDescription="Assembling $(InputName).asm"
+			ShowOnlyRuleProperties="false"
+			>
+			<Properties>
+			</Properties>
+		</CustomBuildRule>
+	</Rules>
+</VisualStudioToolFile>
--- a/build/arm-wince-vs8/armasmv6.rules
+++ b/build/arm-wince-vs8/armasmv6.rules
@@ -0,0 +1,20 @@
+<?xml version="1.0" encoding="utf-8"?>
+<VisualStudioToolFile
+	Name="armasm"
+	Version="8.00"
+	>
+	<Rules>
+		<CustomBuildRule
+			Name="ARMASM"
+			DisplayName="Armasm Assembler"
+			CommandLine="armasm -o &quot;$(IntDir)\$(InputName).obj&quot; $(InputPath) -32 -ARCH 6&#x0D;&#x0A;"
+			Outputs="$(IntDir)\$(InputName).obj"
+			FileExtensions="*.asm"
+			ExecutionDescription="Assembling $(InputName).asm"
+			ShowOnlyRuleProperties="false"
+			>
+			<Properties>
+			</Properties>
+		</CustomBuildRule>
+	</Rules>
+</VisualStudioToolFile>
--- a/build/arm-wince-vs8/armasmxscale.rules
+++ b/build/arm-wince-vs8/armasmxscale.rules
@@ -0,0 +1,20 @@
+<?xml version="1.0" encoding="utf-8"?>
+<VisualStudioToolFile
+	Name="armasm"
+	Version="8.00"
+	>
+	<Rules>
+		<CustomBuildRule
+			Name="ARMASM"
+			DisplayName="Armasm Assembler"
+			CommandLine="armasm -o &quot;$(IntDir)\$(InputName).obj&quot; $(InputPath) -32 -cpu XSCALE&#x0D;&#x0A;"
+			Outputs="$(IntDir)\$(InputName).obj"
+			FileExtensions="*.asm"
+			ExecutionDescription="Assembling $(InputName).asm"
+			ShowOnlyRuleProperties="false"
+			>
+			<Properties>
+			</Properties>
+		</CustomBuildRule>
+	</Rules>
+</VisualStudioToolFile>
--- a/build/arm-wince-vs8/obj_int_extract.bat
+++ b/build/arm-wince-vs8/obj_int_extract.bat
@@ -0,0 +1,13 @@
+@echo off
+REM   Copyright (c) 2010 The WebM project authors. All Rights Reserved.
+REM
+REM   Use of this source code is governed by a BSD-style license
+REM   that can be found in the LICENSE file in the root of the source
+REM   tree. An additional intellectual property rights grant can be found
+REM   in the file PATENTS.  All contributing project authors may
+REM   be found in the AUTHORS file in the root of the source tree.
+echo on
+
+
+cl /I ".\\" /I "..\vp6_decoder_sdk" /I "..\vp6_decoder_sdk\vpx_ports" /D "NDEBUG" /D "_WIN32_WCE=0x420" /D "UNDER_CE" /D "WIN32_PLATFORM_PSPC" /D "WINCE" /D "_LIB" /D "ARM" /D "_ARM_" /D "_UNICODE" /D "UNICODE" /FD /EHsc /MT /GS- /fp:fast /GR- /Fo"Pocket_PC_2003__ARMV4_\%1/" /Fd"Pocket_PC_2003__ARMV4_\%1/vc80.pdb" /W3 /nologo /c /TC ..\vp6_decoder_sdk\vp6_decoder\algo\common\arm\dec_asm_offsets_arm.c
+obj_int_extract.exe rvds "Pocket_PC_2003__ARMV4_\%1/dec_asm_offsets_arm.obj"
--- a/build/arm-wince-vs8/vpx.sln
+++ b/build/arm-wince-vs8/vpx.sln
@@ -0,0 +1,88 @@
+Microsoft Visual Studio Solution File, Format Version 9.00
+# Visual Studio 2005
+Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "example", "example.vcproj", "{BA5FE66F-38DD-E034-F542-B1578C5FB950}"
+	ProjectSection(ProjectDependencies) = postProject
+		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74} = {DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}
+		{E1360C65-D375-4335-8057-7ED99CC3F9B2} = {E1360C65-D375-4335-8057-7ED99CC3F9B2}
+	EndProjectSection
+EndProject
+Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "obj_int_extract", "obj_int_extract.vcproj", "{E1360C65-D375-4335-8057-7ED99CC3F9B2}"
+EndProject
+Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "vpx", "vpx.vcproj", "{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}"
+	ProjectSection(ProjectDependencies) = postProject
+		{E1360C65-D375-4335-8057-7ED99CC3F9B2} = {E1360C65-D375-4335-8057-7ED99CC3F9B2}
+	EndProjectSection
+EndProject
+Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "xma", "xma.vcproj", "{A955FC4A-73F1-44F7-135E-30D84D32F022}"
+	ProjectSection(ProjectDependencies) = postProject
+		{E1360C65-D375-4335-8057-7ED99CC3F9B2} = {E1360C65-D375-4335-8057-7ED99CC3F9B2}
+		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74} = {DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}
+	EndProjectSection
+EndProject
+Global
+	GlobalSection(SolutionConfigurationPlatforms) = preSolution
+		Debug|Mixed Platforms = Debug|Mixed Platforms
+		Debug|Pocket PC 2003 (ARMV4) = Debug|Pocket PC 2003 (ARMV4)
+		Debug|Win32 = Debug|Win32
+		Release|Mixed Platforms = Release|Mixed Platforms
+		Release|Pocket PC 2003 (ARMV4) = Release|Pocket PC 2003 (ARMV4)
+		Release|Win32 = Release|Win32
+	EndGlobalSection
+	GlobalSection(ProjectConfigurationPlatforms) = postSolution
+		{BA5FE66F-38DD-E034-F542-B1578C5FB950}.Debug|Mixed Platforms.ActiveCfg = Debug|Pocket PC 2003 (ARMV4)
+		{BA5FE66F-38DD-E034-F542-B1578C5FB950}.Debug|Mixed Platforms.Build.0 = Debug|Pocket PC 2003 (ARMV4)
+		{BA5FE66F-38DD-E034-F542-B1578C5FB950}.Debug|Mixed Platforms.Deploy.0 = Debug|Pocket PC 2003 (ARMV4)
+		{BA5FE66F-38DD-E034-F542-B1578C5FB950}.Debug|Pocket PC 2003 (ARMV4).ActiveCfg = Debug|Pocket PC 2003 (ARMV4)
+		{BA5FE66F-38DD-E034-F542-B1578C5FB950}.Debug|Pocket PC 2003 (ARMV4).Build.0 = Debug|Pocket PC 2003 (ARMV4)
+		{BA5FE66F-38DD-E034-F542-B1578C5FB950}.Debug|Pocket PC 2003 (ARMV4).Deploy.0 = Debug|Pocket PC 2003 (ARMV4)
+		{BA5FE66F-38DD-E034-F542-B1578C5FB950}.Debug|Win32.ActiveCfg = Debug|Pocket PC 2003 (ARMV4)
+		{BA5FE66F-38DD-E034-F542-B1578C5FB950}.Release|Mixed Platforms.ActiveCfg = Release|Pocket PC 2003 (ARMV4)
+		{BA5FE66F-38DD-E034-F542-B1578C5FB950}.Release|Mixed Platforms.Build.0 = Release|Pocket PC 2003 (ARMV4)
+		{BA5FE66F-38DD-E034-F542-B1578C5FB950}.Release|Mixed Platforms.Deploy.0 = Release|Pocket PC 2003 (ARMV4)
+		{BA5FE66F-38DD-E034-F542-B1578C5FB950}.Release|Pocket PC 2003 (ARMV4).ActiveCfg = Release|Pocket PC 2003 (ARMV4)
+		{BA5FE66F-38DD-E034-F542-B1578C5FB950}.Release|Pocket PC 2003 (ARMV4).Build.0 = Release|Pocket PC 2003 (ARMV4)
+		{BA5FE66F-38DD-E034-F542-B1578C5FB950}.Release|Pocket PC 2003 (ARMV4).Deploy.0 = Release|Pocket PC 2003 (ARMV4)
+		{BA5FE66F-38DD-E034-F542-B1578C5FB950}.Release|Win32.ActiveCfg = Release|Pocket PC 2003 (ARMV4)
+		{E1360C65-D375-4335-8057-7ED99CC3F9B2}.Debug|Mixed Platforms.ActiveCfg = Release|Win32
+		{E1360C65-D375-4335-8057-7ED99CC3F9B2}.Debug|Mixed Platforms.Build.0 = Release|Win32
+		{E1360C65-D375-4335-8057-7ED99CC3F9B2}.Debug|Pocket PC 2003 (ARMV4).ActiveCfg = Release|Win32
+		{E1360C65-D375-4335-8057-7ED99CC3F9B2}.Debug|Win32.ActiveCfg = Release|Win32
+		{E1360C65-D375-4335-8057-7ED99CC3F9B2}.Debug|Win32.Build.0 = Release|Win32
+		{E1360C65-D375-4335-8057-7ED99CC3F9B2}.Release|Mixed Platforms.ActiveCfg = Release|Win32
+		{E1360C65-D375-4335-8057-7ED99CC3F9B2}.Release|Mixed Platforms.Build.0 = Release|Win32
+		{E1360C65-D375-4335-8057-7ED99CC3F9B2}.Release|Pocket PC 2003 (ARMV4).ActiveCfg = Release|Win32
+		{E1360C65-D375-4335-8057-7ED99CC3F9B2}.Release|Win32.ActiveCfg = Release|Win32
+		{E1360C65-D375-4335-8057-7ED99CC3F9B2}.Release|Win32.Build.0 = Release|Win32
+		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}.Debug|Mixed Platforms.ActiveCfg = Debug|Pocket PC 2003 (ARMV4)
+		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}.Debug|Mixed Platforms.Build.0 = Debug|Pocket PC 2003 (ARMV4)
+		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}.Debug|Mixed Platforms.Deploy.0 = Debug|Pocket PC 2003 (ARMV4)
+		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}.Debug|Pocket PC 2003 (ARMV4).ActiveCfg = Debug|Pocket PC 2003 (ARMV4)
+		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}.Debug|Pocket PC 2003 (ARMV4).Build.0 = Debug|Pocket PC 2003 (ARMV4)
+		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}.Debug|Pocket PC 2003 (ARMV4).Deploy.0 = Debug|Pocket PC 2003 (ARMV4)
+		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}.Debug|Win32.ActiveCfg = Debug|Pocket PC 2003 (ARMV4)
+		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}.Release|Mixed Platforms.ActiveCfg = Release|Pocket PC 2003 (ARMV4)
+		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}.Release|Mixed Platforms.Build.0 = Release|Pocket PC 2003 (ARMV4)
+		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}.Release|Mixed Platforms.Deploy.0 = Release|Pocket PC 2003 (ARMV4)
+		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}.Release|Pocket PC 2003 (ARMV4).ActiveCfg = Release|Pocket PC 2003 (ARMV4)
+		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}.Release|Pocket PC 2003 (ARMV4).Build.0 = Release|Pocket PC 2003 (ARMV4)
+		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}.Release|Pocket PC 2003 (ARMV4).Deploy.0 = Release|Pocket PC 2003 (ARMV4)
+		{DCE19DAF-69AC-46DB-B14A-39F0FAA5DB74}.Release|Win32.ActiveCfg = Release|Pocket PC 2003 (ARMV4)
+		{A955FC4A-73F1-44F7-135E-30D84D32F022}.Debug|Mixed Platforms.ActiveCfg = Debug|Pocket PC 2003 (ARMV4)
+		{A955FC4A-73F1-44F7-135E-30D84D32F022}.Debug|Mixed Platforms.Build.0 = Debug|Pocket PC 2003 (ARMV4)
+		{A955FC4A-73F1-44F7-135E-30D84D32F022}.Debug|Mixed Platforms.Deploy.0 = Debug|Pocket PC 2003 (ARMV4)
+		{A955FC4A-73F1-44F7-135E-30D84D32F022}.Debug|Pocket PC 2003 (ARMV4).ActiveCfg = Debug|Pocket PC 2003 (ARMV4)
+		{A955FC4A-73F1-44F7-135E-30D84D32F022}.Debug|Pocket PC 2003 (ARMV4).Build.0 = Debug|Pocket PC 2003 (ARMV4)
+		{A955FC4A-73F1-44F7-135E-30D84D32F022}.Debug|Pocket PC 2003 (ARMV4).Deploy.0 = Debug|Pocket PC 2003 (ARMV4)
+		{A955FC4A-73F1-44F7-135E-30D84D32F022}.Debug|Win32.ActiveCfg = Debug|Pocket PC 2003 (ARMV4)
+		{A955FC4A-73F1-44F7-135E-30D84D32F022}.Release|Mixed Platforms.ActiveCfg = Release|Pocket PC 2003 (ARMV4)
+		{A955FC4A-73F1-44F7-135E-30D84D32F022}.Release|Mixed Platforms.Build.0 = Release|Pocket PC 2003 (ARMV4)
+		{A955FC4A-73F1-44F7-135E-30D84D32F022}.Release|Mixed Platforms.Deploy.0 = Release|Pocket PC 2003 (ARMV4)
+		{A955FC4A-73F1-44F7-135E-30D84D32F022}.Release|Pocket PC 2003 (ARMV4).ActiveCfg = Release|Pocket PC 2003 (ARMV4)
+		{A955FC4A-73F1-44F7-135E-30D84D32F022}.Release|Pocket PC 2003 (ARMV4).Build.0 = Release|Pocket PC 2003 (ARMV4)
+		{A955FC4A-73F1-44F7-135E-30D84D32F022}.Release|Pocket PC 2003 (ARMV4).Deploy.0 = Release|Pocket PC 2003 (ARMV4)
+		{A955FC4A-73F1-44F7-135E-30D84D32F022}.Release|Win32.ActiveCfg = Release|Pocket PC 2003 (ARMV4)
+	EndGlobalSection
+	GlobalSection(SolutionProperties) = preSolution
+		HideSolutionNode = FALSE
+	EndGlobalSection
+EndGlobal
--- a/build/make/armlink_adapter.sh
+++ b/build/make/armlink_adapter.sh
@@ -17,17 +17,15 @@ for i; do
        on_of=1
    elif [ "$i" == "-v" ]; then
        verbose=1
-    elif [ "$i" == "-g" ]; then
-        args="${args} --debug"
    elif [ "$on_of" == "1" ]; then
        outfile=$i
-        on_of=0
+    on_of=0
    elif [ -f "$i" ]; then
        infiles="$infiles $i"
    elif [ "${i:0:2}" == "-l" ]; then
        libs="$libs ${i#-l}"
    elif [ "${i:0:2}" == "-L" ]; then
-        libpaths="${libpaths} ${i#-L}"
+    libpaths="${libpaths} ${i#-L}"
    else
        args="${args} ${i}"
    fi
--- a/build/make/configure.sh
+++ b/build/make/configure.sh
@@ -78,7 +78,6 @@ Build options:
  --log=yes|no|FILE           file configure log is written to [config.err]
  --target=TARGET             target platform tuple [generic-gnu]
  --cpu=CPU                   optimize for a specific cpu rather than a family
-  --extra-cflags=ECFLAGS      add ECFLAGS to CFLAGS [$CFLAGS]
  ${toggle_extra_warnings}    emit harmless warnings (always non-fatal)
  ${toggle_werror}            treat warnings as errors, if possible
                              (not available with all compilers)
@@ -443,9 +442,6 @@ process_common_cmdline() {
        ;;
        --cpu=*) tune_cpu="$optval"
        ;;
-        --extra-cflags=*)
-        extra_cflags="${optval}"
-        ;;
        --enable-?*|--disable-?*)
        eval `echo "$opt" | sed 's/--/action=/;s/-/ option=/;s/-/_/g'`
        echo "${CMDLINE_SELECT} ${ARCH_EXT_LIST}" | grep "^ *$option\$" >/dev/null || die_unknown $opt
@@ -624,10 +620,6 @@ process_common_toolchain() {

    # Handle Solaris variants. Solaris 10 needs -lposix4
    case ${toolchain} in
-        sparc-solaris-*)
-            add_extralibs -lposix4
-            add_cflags "-DMUST_BE_ALIGNED"
-            ;;
        *-solaris-*)
            add_extralibs -lposix4
            ;;
@@ -668,12 +660,12 @@ process_common_toolchain() {
            elif enabled armv7
            then
                check_add_cflags -march=armv7-a -mcpu=cortex-a8 -mfpu=neon -mfloat-abi=softfp  #-ftree-vectorize
-                check_add_asflags -mcpu=cortex-a8 -mfpu=neon -mfloat-abi=softfp  #-march=armv7-a
+        check_add_asflags -mcpu=cortex-a8 -mfpu=neon -mfloat-abi=softfp  #-march=armv7-a
            else
                check_add_cflags -march=${tgt_isa}
                check_add_asflags -march=${tgt_isa}
            fi
-            enabled debug && add_asflags -g
+
            asm_conversion_cmd="${source_path}/build/make/ads2gas.pl"
            ;;
        rvct)
@@ -698,24 +690,16 @@ process_common_toolchain() {
            arch_int=${tgt_isa##armv}
            arch_int=${arch_int%%te}
            check_add_asflags --pd "\"ARCHITECTURE SETA ${arch_int}\""
-            enabled debug && add_asflags -g
-            add_cflags --gnu
-            add_cflags --enum_is_int
-            add_cflags --wchar32
        ;;
        esac

        case ${tgt_os} in
-        none*)
-            disable multithread
-            disable os_support
-            ;;
        darwin*)
            SDK_PATH=/Developer/Platforms/iPhoneOS.platform/Developer
            TOOLCHAIN_PATH=${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-darwin9-gcc-4.2.1
            AS=${TOOLCHAIN_PATH}/as
            STRIP=${TOOLCHAIN_PATH}/strip
            NM=${TOOLCHAIN_PATH}/nm
@@ -729,14 +713,14 @@ process_common_toolchain() {
            add_cflags -arch ${tgt_isa}
            add_ldflags -arch_only ${tgt_isa}

-            add_cflags  "-isysroot /Developer/Platforms/iPhoneOS.platform/Developer/SDKs/iPhoneOS4.2.sdk"
+            add_cflags  "-isysroot /Developer/Platforms/iPhoneOS.platform/Developer/SDKs/iPhoneOS3.1.sdk"

            # This should be overridable
-            alt_libc=${SDK_PATH}/SDKs/iPhoneOS4.2.sdk
+            alt_libc=${SDK_PATH}/SDKs/iPhoneOS3.1.sdk

            # Add the paths for the alternate libc
 #            for d in usr/include usr/include/gcc/darwin/4.0/; do
-            for d in usr/include usr/include/gcc/darwin/4.0/ usr/lib/gcc/arm-apple-darwin10/4.2.1/include/; do
+            for d in usr/include usr/include/gcc/darwin/4.0/ usr/lib/gcc/arm-apple-darwin9/4.0.1/include/; do
                try_dir="${alt_libc}/${d}"
                [ -d "${try_dir}" ] && add_cflags -I"${try_dir}"
            done
@@ -758,9 +742,13 @@ process_common_toolchain() {
                    || die "Must supply --libc when targetting *-linux-rvct"

                # Set up compiler
+                add_cflags --gnu
+                add_cflags --enum_is_int
                add_cflags --library_interface=aeabi_glibc
                add_cflags --no_hide_all
+                add_cflags --wchar32
                add_cflags --dwarf2
+                add_cflags --gnu

                # Set up linker
                add_ldflags --sysv --no_startup --no_ref_cpp_init
@@ -867,7 +855,7 @@ process_common_toolchain() {
                setup_gnu_toolchain
                add_cflags -use-msasm -use-asm
                add_ldflags -i-static
-                enabled x86_64 && add_cflags -ipo -no-prec-div -static -xSSE2 -axSSE2
+                enabled x86_64 && add_cflags -ipo -no-prec-div -static -xSSE3 -axSSE3
                enabled x86_64 && AR=xiar
                case ${tune_cpu} in
                    atom*)
@@ -984,12 +972,6 @@ EOF
        add_cflags -D_LARGEFILE_SOURCE
        add_cflags -D_FILE_OFFSET_BITS=64
    fi
-
-    # append any user defined extra cflags
-    if [ -n "${extra_cflags}" ] ; then
-        check_add_cflags ${extra_cflags} || \
-        die "Requested extra CFLAGS '${extra_cflags}' not supported by compiler"
-    fi
 }

 process_toolchain() {
--- a/build/make/gen_msvs_proj.sh
+++ b/build/make/gen_msvs_proj.sh
@@ -32,7 +32,7 @@ Options:
    --name=project_name         Name of the project (required)
    --proj-guid=GUID            GUID to use for the project
    --module-def=filename       File containing export definitions (for DLLs)
-    --ver=version               Version (7,8,9) of visual studio to generate for
+    --ver=version               Version (7,8) of visual studio to generate for
    -Ipath/to/include           Additional include directories
    -DFLAG[=value]              Preprocessor macros to define
    -Lpath/to/lib               Additional library search paths
@@ -132,7 +132,7 @@ generate_filter() {
    open_tag Filter \
        Name=$name \
        Filter=$pats \
-        UniqueIdentifier=`generate_uuid` \
+        UniqueIdentifier=`generate_uuid`

    file_list_sz=${#file_list[@]}
    for i in ${!file_list[@]}; do
@@ -145,21 +145,31 @@ generate_filter() {
                if [ "$pat" == "asm" ] && $asm_use_custom_step; then
                    for plat in "${platforms[@]}"; do
                        for cfg in Debug Release; do
-                            open_tag FileConfiguration \
-                                Name="${cfg}|${plat}" \
-
+                            open_tag  FileConfiguration \
+                            Name="${cfg}|${plat}"
                            tag Tool \
                                Name="VCCustomBuildTool" \
                                Description="Assembling \$(InputFileName)" \
-                                CommandLine="$(eval echo \$asm_${cfg}_cmdline)" \
-                                Outputs="\$(InputName).obj" \
-
+                                CommandLine="$(eval echo \$asm_${cfg}_cmdline)"\
+                                Outputs="\$(InputName).obj"
                            close_tag FileConfiguration
                        done
                    done
                fi

-                close_tag File
+                if [ "${f##*.}" == "cpp" ]; then
+                    for plat in "${platforms[@]}"; do
+                        for cfg in Debug Release; do
+                        open_tag FileConfiguration \
+                            Name="${cfg}|${plat}"
+                        tag Tool \
+                            Name="VCCLCompilerTool" \
+                            CompileAs="2"
+                        close_tag FileConfiguration
+                        done
+                    done
+                fi
+                close_tag  File

                break
            fi
@@ -175,61 +185,57 @@ unset target
 for opt in "$@"; do
    optval="${opt#*=}"
    case "$opt" in
-        --help|-h) show_help
-        ;;
-        --target=*) target="${optval}"
-        ;;
-        --out=*) outfile="$optval"
-        ;;
-        --name=*) name="${optval}"
-        ;;
-        --proj-guid=*) guid="${optval}"
-        ;;
-        --module-def=*) link_opts="${link_opts} ModuleDefinitionFile=${optval}"
-        ;;
-        --exe) proj_kind="exe"
-        ;;
-        --lib) proj_kind="lib"
-        ;;
-        --static-crt) use_static_runtime=true
-        ;;
-        --ver=*)
-            vs_ver="$optval"
-            case "$optval" in
-                [789])
-                ;;
-                *) die Unrecognized Visual Studio Version in $opt
-                ;;
-            esac
-        ;;
-        -I*)
-            opt="${opt%/}"
-            incs="${incs}${incs:+;}&quot;${opt##-I}&quot;"
-            yasmincs="${yasmincs} ${opt}"
-        ;;
-        -D*) defines="${defines}${defines:+;}${opt##-D}"
-        ;;
-        -L*) # fudge . to $(OutDir)
-            if [ "${opt##-L}" == "." ]; then
-                libdirs="${libdirs}${libdirs:+;}&quot;\$(OutDir)&quot;"
-            else
-                 # Also try directories for this platform/configuration
-                 libdirs="${libdirs}${libdirs:+;}&quot;${opt##-L}&quot;"
-                 libdirs="${libdirs}${libdirs:+;}&quot;${opt##-L}/\$(PlatformName)/\$(ConfigurationName)&quot;"
-                 libdirs="${libdirs}${libdirs:+;}&quot;${opt##-L}/\$(PlatformName)&quot;"
-            fi
-        ;;
-        -l*) libs="${libs}${libs:+ }${opt##-l}.lib"
-        ;;
-        -*) die_unknown $opt
-        ;;
-        *)
-            file_list[${#file_list[@]}]="$opt"
-            case "$opt" in
-                 *.asm) uses_asm=true
-                 ;;
-            esac
-        ;;
+    --help|-h) show_help
+    ;;
+    --target=*) target="${optval}"
+    ;;
+    --out=*) outfile="$optval"
+    ;;
+    --name=*) name="${optval}"
+    ;;
+    --proj-guid=*) guid="${optval}"
+    ;;
+    --module-def=*)
+        link_opts="${link_opts} ModuleDefinitionFile=${optval}"
+    ;;
+    --exe) proj_kind="exe"
+    ;;
+    --lib) proj_kind="lib"
+    ;;
+    --static-crt) use_static_runtime=true
+    ;;
+    --ver=*) vs_ver="$optval"
+             case $optval in
+             [789])
+             ;;
+             *) die Unrecognized Visual Studio Version in $opt
+             ;;
+             esac
+    ;;
+    -I*) opt="${opt%/}"
+         incs="${incs}${incs:+;}&quot;${opt##-I}&quot;"
+         yasmincs="${yasmincs} ${opt}"
+    ;;
+    -D*) defines="${defines}${defines:+;}${opt##-D}"
+    ;;
+    -L*) # fudge . to $(OutDir)
+         if [ "${opt##-L}" == "." ]; then
+             libdirs="${libdirs}${libdirs:+;}&quot;\$(OutDir)&quot;"
+         else
+             # Also try directories for this platform/configuration
+             libdirs="${libdirs}${libdirs:+;}&quot;${opt##-L}&quot;"
+             libdirs="${libdirs}${libdirs:+;}&quot;${opt##-L}/\$(PlatformName)/\$(ConfigurationName)&quot;"
+             libdirs="${libdirs}${libdirs:+;}&quot;${opt##-L}/\$(PlatformName)&quot;"
+         fi
+    ;;
+    -l*) libs="${libs}${libs:+ }${opt##-l}.lib"
+    ;;
+    -*) die_unknown $opt
+    ;;
+    *) file_list[${#file_list[@]}]="$opt"
+       case "$opt" in
+       *.asm) uses_asm=true;;
+       esac
    esac
 done
 outfile=${outfile:-/dev/stdout}
@@ -272,7 +278,11 @@ done

 # List Keyword for this target
 case "$target" in
-    x86*) keyword="ManagedCProj"
+    x86*)
+        keyword="ManagedCProj"
+    ;;
+    arm*|iwmmx*)
+        keyword="Win32Proj"
    ;;
    *) die "Unsupported target $target!"
 esac
@@ -288,186 +298,402 @@ case "$target" in
        asm_Debug_cmdline="yasm -Xvc -g cv8 -f \$(PlatformName) ${yasmincs} &quot;\$(InputPath)&quot;"
        asm_Release_cmdline="yasm -Xvc -f \$(PlatformName) ${yasmincs} &quot;\$(InputPath)&quot;"
    ;;
+    arm*|iwmmx*)
+        case "${name}" in
+        obj_int_extract) platforms[0]="Win32"
+        ;;
+        *) platforms[0]="Pocket PC 2003 (ARMV4)"
+        ;;
+        esac
+    ;;
    *) die "Unsupported target $target!"
+esac
+
+# List Command-line Arguments for this target
+case "$target" in
+    arm*|iwmmx*)
+        if [ "$name" == "example" ];then
+            ARGU="--codec vp6 --flipuv --progress _bnd.vp6"
+        fi
+        if [ "$name" == "xma" ];then
+            ARGU="--codec vp6 -h 240 -w 320 -v"
+        fi
    ;;
 esac

 generate_vcproj() {
    case "$proj_kind" in
-        exe) vs_ConfigurationType=1
-        ;;
-        *)   vs_ConfigurationType=4
-        ;;
+    exe) vs_ConfigurationType=1
+    ;;
+    *)   vs_ConfigurationType=4
+    ;;
    esac

    echo "<?xml version=\"1.0\" encoding=\"Windows-1252\"?>"
-    open_tag VisualStudioProject \
-        ProjectType="Visual C++" \
-        Version="${vs_ver_id}" \
-        Name="${name}" \
-        ProjectGUID="{${guid}}" \
-        RootNamespace="${name}" \
-        Keyword="${keyword}" \
+    open_tag  VisualStudioProject \
+                  ProjectType="Visual C++" \
+                  Version="${vs_ver_id}" \
+                  Name="${name}" \
+                  ProjectGUID="{${guid}}" \
+                  RootNamespace="${name}" \
+                  Keyword="${keyword}"

-    open_tag Platforms
+    open_tag  Platforms
    for plat in "${platforms[@]}"; do
-        tag Platform Name="$plat"
+        tag   Platform Name="$plat"
    done
    close_tag Platforms

-    open_tag ToolFiles
+    open_tag  ToolFiles
    case "$target" in
        x86*) $uses_asm && tag ToolFile RelativePath="$self_dirname/../x86-msvs/yasm.rules"
        ;;
+        arm*|iwmmx*)
+            if [ "$name" == "vpx" ];then
+            case "$target" in
+                armv5*)
+                    tag ToolFile RelativePath="$self_dirname/../arm-wince-vs8/armasmv5.rules"
+                ;;
+                armv6*)
+                    tag ToolFile RelativePath="$self_dirname/../arm-wince-vs8/armasmv6.rules"
+                ;;
+                iwmmxt*)
+                    tag ToolFile RelativePath="$self_dirname/../arm-wince-vs8/armasmxscale.rules"
+                ;;
+            esac
+            fi
+        ;;
    esac
    close_tag ToolFiles

-    open_tag Configurations
+    open_tag  Configurations
    for plat in "${platforms[@]}"; do
        plat_no_ws=`echo $plat | sed 's/[^A-Za-z0-9_]/_/g'`
-        open_tag Configuration \
-            Name="Debug|$plat" \
-            OutputDirectory="\$(SolutionDir)$plat_no_ws/\$(ConfigurationName)" \
-            IntermediateDirectory="$plat_no_ws/\$(ConfigurationName)/${name}" \
-            ConfigurationType="$vs_ConfigurationType" \
-            CharacterSet="1" \
+        open_tag  Configuration \
+                      Name="Debug|$plat" \
+                      OutputDirectory="\$(SolutionDir)$plat_no_ws/\$(ConfigurationName)" \
+                      IntermediateDirectory="$plat_no_ws/\$(ConfigurationName)/${name}" \
+                      ConfigurationType="$vs_ConfigurationType" \
+                      CharacterSet="1"
+
+        if [ "$target" == "armv6-wince-vs8" ] || [ "$target" == "armv5te-wince-vs8" ] || [ "$target" == "iwmmxt-wince-vs8" ] || [ "$target" == "iwmmxt2-wince-vs8" ];then
+            case "$name" in
+                vpx)         tag Tool \
+                             Name="VCPreBuildEventTool" \
+                             CommandLine="call obj_int_extract.bat \$(ConfigurationName)"
+                             tag Tool \
+                             Name="VCMIDLTool" \
+                             TargetEnvironment="1"
+                             tag Tool \
+                             Name="VCCLCompilerTool" \
+                             ExecutionBucket="7" \
+                             Optimization="0" \
+                             AdditionalIncludeDirectories="$incs" \
+                             PreprocessorDefinitions="_DEBUG;_WIN32_WCE=\$(CEVER);UNDER_CE;\$(PLATFORMDEFINES);WINCE;DEBUG;_LIB;\$(ARCHFAM);\$(_ARCHFAM_);_UNICODE;UNICODE;" \
+                             MinimalRebuild="true" \
+                             RuntimeLibrary="1" \
+                             BufferSecurityCheck="false" \
+                             UsePrecompiledHeader="0" \
+                             WarningLevel="3" \
+                             DebugInformationFormat="1" \
+                             CompileAs="1"
+                             tag Tool \
+                             Name="VCResourceCompilerTool" \
+                             PreprocessorDefinitions="_DEBUG;_WIN32_WCE=\$(CEVER);UNDER_CE;\$(PLATFORMDEFINES)" \
+                             Culture="1033" \
+                             AdditionalIncludeDirectories="\$(IntDir)" \
+                ;;
+                example|xma) tag Tool \
+                             Name="VCCLCompilerTool" \
+                             ExecutionBucket="7" \
+                             Optimization="0" \
+                             AdditionalIncludeDirectories="$incs" \
+                             PreprocessorDefinitions="_DEBUG;_WIN32_WCE=\$(CEVER);UNDER_CE;\$(PLATFORMDEFINES);WINCE;DEBUG;_CONSOLE;\$(ARCHFAM);\$(_ARCHFAM_);_UNICODE;UNICODE;" \
+                             MinimalRebuild="true" \
+                             RuntimeLibrary="1" \
+                             BufferSecurityCheck="false" \
+                             UsePrecompiledHeader="0" \
+                             WarningLevel="3" \
+                             DebugInformationFormat="1" \
+                             CompileAs="1"
+                             tag Tool \
+                             Name="VCResourceCompilerTool" \
+                             PreprocessorDefinitions="_DEBUG;_WIN32_WCE=\$(CEVER);UNDER_CE;\$(PLATFORMDEFINES)" \
+                             Culture="1033" \
+                             AdditionalIncludeDirectories="\$(IntDir)" \
+                ;;
+                obj_int_extract) tag Tool \
+                             Name="VCCLCompilerTool" \
+                             Optimization="0" \
+                             AdditionalIncludeDirectories="$incs" \
+                             PreprocessorDefinitions="WIN32;DEBUG;_CONSOLE" \
+                             RuntimeLibrary="1" \
+                             WarningLevel="3" \
+                             DebugInformationFormat="1" \
+                ;;
+            esac
+        fi

        case "$target" in
-            x86*)
-                case "$name" in
-                    *)
-                        tag Tool \
-                            Name="VCCLCompilerTool" \
-                            Optimization="0" \
-                            AdditionalIncludeDirectories="$incs" \
-                            PreprocessorDefinitions="WIN32;_DEBUG;_CRT_SECURE_NO_WARNINGS;_CRT_SECURE_NO_DEPRECATE;$defines" \
-                            RuntimeLibrary="$debug_runtime" \
-                            UsePrecompiledHeader="0" \
-                            WarningLevel="3" \
-                            DebugInformationFormat="1" \
-                            Detect64BitPortabilityProblems="true" \
+            x86*) tag Tool \
+                Name="VCCLCompilerTool" \
+                Optimization="0" \
+                AdditionalIncludeDirectories="$incs" \
+                PreprocessorDefinitions="WIN32;_DEBUG;_CRT_SECURE_NO_WARNINGS;_CRT_SECURE_NO_DEPRECATE;$defines" \
+                RuntimeLibrary="$debug_runtime" \
+                UsePrecompiledHeader="0" \
+                WarningLevel="3" \
+                DebugInformationFormat="1" \
+                Detect64BitPortabilityProblems="true" \

-                        $uses_asm && tag Tool Name="YASM"  IncludePaths="$incs" Debug="1"
-                    ;;
-                esac
+                $uses_asm && tag Tool Name="YASM"  IncludePaths="$incs" Debug="1"
            ;;
        esac

        case "$proj_kind" in
            exe)
                case "$target" in
-                    x86*)
+                    x86*) tag Tool \
+                          Name="VCLinkerTool" \
+                          AdditionalDependencies="$debug_libs \$(NoInherit)" \
+                          AdditionalLibraryDirectories="$libdirs" \
+                          GenerateDebugInformation="true" \
+                          ProgramDatabaseFile="\$(OutDir)/${name}.pdb" \
+
+                    ;;
+                    arm*|iwmmx*)
                        case "$name" in
-                            *)
-                                tag Tool \
-                                    Name="VCLinkerTool" \
-                                    AdditionalDependencies="$debug_libs \$(NoInherit)" \
-                                    AdditionalLibraryDirectories="$libdirs" \
-                                    GenerateDebugInformation="true" \
-                                    ProgramDatabaseFile="\$(OutDir)/${name}.pdb" \
+                            obj_int_extract) tag Tool \
+                                Name="VCLinkerTool" \
+                                OutputFile="${name}.exe" \
+                                GenerateDebugInformation="true"
+                            ;;
+                            *) tag Tool \
+                                Name="VCLinkerTool" \
+                                AdditionalDependencies="$debug_libs" \
+                                OutputFile="\$(OutDir)/${name}.exe" \
+                                LinkIncremental="2" \
+                                AdditionalLibraryDirectories="${libdirs};&quot;..\lib/$plat_no_ws&quot;" \
+                                DelayLoadDLLs="\$(NOINHERIT)" \
+                                GenerateDebugInformation="true" \
+                                ProgramDatabaseFile="\$(OutDir)/${name}.pdb" \
+                                SubSystem="9" \
+                                StackReserveSize="65536" \
+                                StackCommitSize="4096" \
+                                EntryPointSymbol="mainWCRTStartup" \
+                                TargetMachine="3"
                            ;;
                        esac
-                    ;;
+                     ;;
                 esac
            ;;
            lib)
                case "$target" in
-                    x86*)
-                        tag Tool \
-                            Name="VCLibrarianTool" \
-                            OutputFile="\$(OutDir)/${name}${lib_sfx}d.lib" \
-
-                    ;;
+                      arm*|iwmmx*) tag Tool \
+                                    Name="VCLibrarianTool" \
+                                    AdditionalOptions=" /subsystem:windowsce,4.20 /machine:ARM" \
+                                    OutputFile="\$(OutDir)/${name}.lib" \
+                                ;;
+                                *) tag Tool \
+                                    Name="VCLibrarianTool" \
+                                    OutputFile="\$(OutDir)/${name}${lib_sfx}d.lib" \
+                                ;;
                esac
            ;;
-            dll)
-                tag Tool \
-                    Name="VCLinkerTool" \
-                    AdditionalDependencies="\$(NoInherit)" \
-                    LinkIncremental="2" \
-                    GenerateDebugInformation="true" \
-                    AssemblyDebug="1" \
-                    TargetMachine="1" \
-                    $link_opts \
-
-            ;;
+            dll) tag Tool \
+                 Name="VCLinkerTool" \
+                AdditionalDependencies="\$(NoInherit)" \
+                LinkIncremental="2" \
+                GenerateDebugInformation="true" \
+                AssemblyDebug="1" \
+                TargetMachine="1" \
+                      $link_opts
        esac

+        if [ "$target" == "armv6-wince-vs8" ] || [ "$target" == "armv5te-wince-vs8" ] || [ "$target" == "iwmmxt-wince-vs8" ] || [ "$target" == "iwmmxt2-wince-vs8" ];then
+            case "$name" in
+                vpx)         tag DeploymentTool \
+                             ForceDirty="-1" \
+                             RegisterOutput="0"
+                                ;;
+                example|xma) tag DeploymentTool \
+                             ForceDirty="-1" \
+                             RegisterOutput="0"
+                             tag DebuggerTool \
+                             Arguments="${ARGU}"
+                                ;;
+            esac
+        fi
        close_tag Configuration

-        open_tag Configuration \
-            Name="Release|$plat" \
-            OutputDirectory="\$(SolutionDir)$plat_no_ws/\$(ConfigurationName)" \
-            IntermediateDirectory="$plat_no_ws/\$(ConfigurationName)/${name}" \
-            ConfigurationType="$vs_ConfigurationType" \
-            CharacterSet="1" \
-            WholeProgramOptimization="0" \
+        open_tag  Configuration \
+                      Name="Release|$plat" \
+                      OutputDirectory="\$(SolutionDir)$plat_no_ws/\$(ConfigurationName)" \
+                      IntermediateDirectory="$plat_no_ws/\$(ConfigurationName)/${name}" \
+                      ConfigurationType="$vs_ConfigurationType" \
+                      CharacterSet="1" \
+                      WholeProgramOptimization="0"

-        case "$target" in
-            x86*)
-                case "$name" in
-                    *)
-                        tag Tool \
-                            Name="VCCLCompilerTool" \
-                            AdditionalIncludeDirectories="$incs" \
-                            PreprocessorDefinitions="WIN32;NDEBUG;_CRT_SECURE_NO_WARNINGS;_CRT_SECURE_NO_DEPRECATE;$defines" \
-                            RuntimeLibrary="$release_runtime" \
-                            UsePrecompiledHeader="0" \
-                            WarningLevel="3" \
-                            DebugInformationFormat="0" \
-                            Detect64BitPortabilityProblems="true" \
+        if [ "$target" == "armv6-wince-vs8" ] || [ "$target" == "armv5te-wince-vs8" ] || [ "$target" == "iwmmxt-wince-vs8" ] || [ "$target" == "iwmmxt2-wince-vs8" ];then
+            case "$name" in
+                vpx)         tag Tool \
+                                     Name="VCPreBuildEventTool" \
+                                     CommandLine="call obj_int_extract.bat \$(ConfigurationName)"
+                             tag Tool \
+                                     Name="VCMIDLTool" \
+                                     TargetEnvironment="1"
+                             tag Tool \
+                                             Name="VCCLCompilerTool" \
+                                             ExecutionBucket="7" \
+                                             Optimization="2" \
+                                             FavorSizeOrSpeed="1" \
+                                             AdditionalIncludeDirectories="$incs" \
+                                             PreprocessorDefinitions="NDEBUG;_WIN32_WCE=\$(CEVER);UNDER_CE;\$(PLATFORMDEFINES);WINCE;_LIB;\$(ARCHFAM);\$(_ARCHFAM_);_UNICODE;UNICODE;" \
+                                             RuntimeLibrary="0" \
+                                             BufferSecurityCheck="false" \
+                                             UsePrecompiledHeader="0" \
+                                             WarningLevel="3" \
+                                             DebugInformationFormat="0" \
+                                             CompileAs="1"
+                             tag Tool \
+                                             Name="VCResourceCompilerTool" \
+                                             PreprocessorDefinitions="NDEBUG;_WIN32_WCE=\$(CEVER);UNDER_CE;\$(PLATFORMDEFINES)" \
+                                             Culture="1033" \
+                                             AdditionalIncludeDirectories="\$(IntDir)" \
+                ;;
+                example|xma) tag Tool \
+                             Name="VCCLCompilerTool" \
+                             ExecutionBucket="7" \
+                             Optimization="2" \
+                             FavorSizeOrSpeed="1" \
+                             AdditionalIncludeDirectories="$incs" \
+                             PreprocessorDefinitions="NDEBUG;_WIN32_WCE=\$(CEVER);UNDER_CE;\$(PLATFORMDEFINES);WINCE;_CONSOLE;\$(ARCHFAM);\$(_ARCHFAM_);_UNICODE;UNICODE;" \
+                             RuntimeLibrary="0" \
+                             BufferSecurityCheck="false" \
+                             UsePrecompiledHeader="0" \
+                             WarningLevel="3" \
+                             DebugInformationFormat="0" \
+                             CompileAs="1"
+                             tag Tool \
+                             Name="VCResourceCompilerTool" \
+                             PreprocessorDefinitions="NDEBUG;_WIN32_WCE=\$(CEVER);UNDER_CE;\$(PLATFORMDEFINES)" \
+                             Culture="1033" \
+                             AdditionalIncludeDirectories="\$(IntDir)" \
+                ;;
+                obj_int_extract) tag Tool \
+                             Name="VCCLCompilerTool" \
+                             AdditionalIncludeDirectories="$incs" \
+                             PreprocessorDefinitions="WIN32;NDEBUG;_CONSOLE" \
+                             RuntimeLibrary="0" \
+                             UsePrecompiledHeader="0" \
+                             WarningLevel="3" \
+                             Detect64BitPortabilityProblems="true" \
+                             DebugInformationFormat="0" \
+                ;;
+            esac
+        fi

-                        $uses_asm && tag Tool Name="YASM"  IncludePaths="$incs"
-                    ;;
+    case "$target" in
+        x86*) tag       Tool \
+                      Name="VCCLCompilerTool" \
+                      AdditionalIncludeDirectories="$incs" \
+                      PreprocessorDefinitions="WIN32;NDEBUG;_CRT_SECURE_NO_WARNINGS;_CRT_SECURE_NO_DEPRECATE;$defines" \
+                      RuntimeLibrary="$release_runtime" \
+                      UsePrecompiledHeader="0" \
+                      WarningLevel="3" \
+                      DebugInformationFormat="0" \
+                      Detect64BitPortabilityProblems="true"
+
+                $uses_asm && tag Tool Name="YASM"  IncludePaths="$incs"
+                ;;
                esac
-            ;;
-        esac

        case "$proj_kind" in
            exe)
                case "$target" in
-                    x86*)
+                    x86*) tag Tool \
+                                  Name="VCLinkerTool" \
+                                  AdditionalDependencies="$libs \$(NoInherit)" \
+                                  AdditionalLibraryDirectories="$libdirs" \
+                    ;;
+                    arm*|iwmmx*)
                        case "$name" in
-                            *)
-                                tag Tool \
-                                    Name="VCLinkerTool" \
-                                    AdditionalDependencies="$libs \$(NoInherit)" \
-                                    AdditionalLibraryDirectories="$libdirs" \
-
+                            obj_int_extract) tag Tool \
+                                Name="VCLinkerTool" \
+                                OutputFile="${name}.exe" \
+                                LinkIncremental="1" \
+                                GenerateDebugInformation="false" \
+                                SubSystem="0" \
+                                OptimizeReferences="0" \
+                                EnableCOMDATFolding="0" \
+                                TargetMachine="0"
+                            ;;
+                            *) tag Tool \
+                                Name="VCLinkerTool" \
+                                AdditionalDependencies="$libs" \
+                                OutputFile="\$(OutDir)/${name}.exe" \
+                                LinkIncremental="1" \
+                                AdditionalLibraryDirectories="${libdirs};&quot;..\lib/$plat_no_ws&quot;" \
+                                DelayLoadDLLs="\$(NOINHERIT)" \
+                                GenerateDebugInformation="true" \
+                                ProgramDatabaseFile="\$(OutDir)/${name}.pdb" \
+                                SubSystem="9" \
+                                StackReserveSize="65536" \
+                                StackCommitSize="4096" \
+                                OptimizeReferences="2" \
+                                EnableCOMDATFolding="2" \
+                                EntryPointSymbol="mainWCRTStartup" \
+                                TargetMachine="3"
                            ;;
                        esac
-                    ;;
+                     ;;
                 esac
            ;;
-            lib)
+        lib)
                case "$target" in
-                    x86*)
-                        tag Tool \
-                            Name="VCLibrarianTool" \
-                            OutputFile="\$(OutDir)/${name}${lib_sfx}.lib" \
-
-                    ;;
+                      arm*|iwmmx*) tag Tool \
+                                    Name="VCLibrarianTool" \
+                                    AdditionalOptions=" /subsystem:windowsce,4.20 /machine:ARM" \
+                                    OutputFile="\$(OutDir)/${name}.lib" \
+                                ;;
+                                *) tag Tool \
+                                    Name="VCLibrarianTool" \
+                                    OutputFile="\$(OutDir)/${name}${lib_sfx}.lib" \
+                                ;;
                esac
-            ;;
-            dll) # note differences to debug version: LinkIncremental, AssemblyDebug
-                tag Tool \
-                    Name="VCLinkerTool" \
-                    AdditionalDependencies="\$(NoInherit)" \
-                    LinkIncremental="1" \
-                    GenerateDebugInformation="true" \
-                    TargetMachine="1" \
-                    $link_opts \
-
-            ;;
+        ;;
+        dll) # note differences to debug version: LinkIncremental, AssemblyDebug
+             tag Tool \
+                      Name="VCLinkerTool" \
+                      AdditionalDependencies="\$(NoInherit)" \
+                      LinkIncremental="1" \
+                      GenerateDebugInformation="true" \
+                      TargetMachine="1" \
+                      $link_opts
        esac

+        if [ "$target" == "armv6-wince-vs8" ] || [ "$target" == "armv5te-wince-vs8" ] || [ "$target" == "iwmmxt-wince-vs8" ] || [ "$target" == "iwmmxt2-wince-vs8" ];then
+            case "$name" in
+                vpx)         tag DeploymentTool \
+                             ForceDirty="-1" \
+                             RegisterOutput="0"
+                ;;
+                example|xma) tag DeploymentTool \
+                             ForceDirty="-1" \
+                             RegisterOutput="0"
+                             tag DebuggerTool \
+                             Arguments="${ARGU}"
+                                ;;
+            esac
+        fi
+
        close_tag Configuration
    done
    close_tag Configurations

-    open_tag Files
-    generate_filter srcs   "Source Files"   "c;def;odl;idl;hpj;bat;asm;asmx"
-    generate_filter hdrs   "Header Files"   "h;hm;inl;inc;xsd"
+    open_tag  Files
+    generate_filter srcs   "Source Files"   "cpp;c;cc;cxx;def;odl;idl;hpj;bat;asm;asmx"
+    generate_filter hdrs   "Header Files"   "h;hpp;hxx;hm;inl;inc;xsd"
    generate_filter resrcs "Resource Files" "rc;ico;cur;bmp;dlg;rc2;rct;bin;rgs;gif;jpg;jpeg;jpe;resx;tiff;tif;png;wav"
    generate_filter resrcs "Build Files"    "mk"
    close_tag Files
--- a/build/make/gen_msvs_sln.sh
+++ b/build/make/gen_msvs_sln.sh
@@ -139,6 +139,9 @@ process_global() {
            echo "${indent}${proj_guid}.${config}.ActiveCfg = ${config}"
            echo "${indent}${proj_guid}.${config}.Build.0 = ${config}"

+            if [ "$target" == "armv6-wince-vs8" ] || [ "$target" == "armv5te-wince-vs8" ] || [ "$target" == "iwmmxt-wince-vs8" ] || [ "$target" == "iwmmxt2-wince-vs8" ];then
+                echo "${indent}${proj_guid}.${config}.Deploy.0 = ${config}"
+            fi
        done
        IFS=${IFS_bak}
    done
--- a/build/make/obj_int_extract.c
+++ b/build/make/obj_int_extract.c
@@ -590,7 +590,7 @@ int parse_coff(unsigned __int8 *buf, size_t sz)
    //log_msg("COFF: Symbol table at offset %u\n", symtab_ptr);
    //log_msg("COFF: raw data pointer ofset for section .data is %u\n", sectionrawdata_ptr);

-    fp = fopen("assembly_offsets.asm", "w");
+    fp = fopen("vpx_asm_offsets.asm", "w");

    if (fp == NULL)
    {
--- a/17
+++ b/17
@@ -37,9 +37,11 @@ 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_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
  ${toggle_small}                 favor smaller size over speed
+  ${toggle_arm_asm_detok}         assembly version of the detokenizer (ARM platforms only)
  ${toggle_postproc_visualizer}   macro block / block level visualizers

 Codecs:
@@ -78,21 +80,22 @@ EOF
 # alphabetically by architecture, generic-gnu last.
 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="${all_platforms} armv5te-wince-vs8"
 all_platforms="${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="${all_platforms} armv6-wince-vs8"
 all_platforms="${all_platforms} iwmmxt-linux-rvct"
 all_platforms="${all_platforms} iwmmxt-linux-gcc"
+all_platforms="${all_platforms} iwmmxt-wince-vs8"
 all_platforms="${all_platforms} iwmmxt2-linux-rvct"
 all_platforms="${all_platforms} iwmmxt2-linux-gcc"
+all_platforms="${all_platforms} iwmmxt2-wince-vs8"
 all_platforms="${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
-all_platforms="${all_platforms} armv7-none-rvct"     #neon Cortex-A8
 all_platforms="${all_platforms} mips32-linux-gcc"
 all_platforms="${all_platforms} ppc32-darwin8-gcc"
 all_platforms="${all_platforms} ppc32-darwin9-gcc"
@@ -157,7 +160,6 @@ enable fast_unaligned #allow unaligned accesses, if supported by hw
 enable md5
 enable spatial_resampling
 enable multithread
-enable os_support

 [ -d ${source_path}/../include ] && enable alt_tree_layout
 for d in vp8; do
@@ -248,10 +250,11 @@ CONFIG_LIST="
    static_msvcrt
    spatial_resampling
    realtime_only
+    error_concealment
    shared
    small
+    arm_asm_detok
    postproc_visualizer
-    os_support
 "
 CMDLINE_SELECT="
    extra_warnings
@@ -288,8 +291,10 @@ CMDLINE_SELECT="
    mem_tracker
    spatial_resampling
    realtime_only
+    error_concealment
    shared
    small
+    arm_asm_detok
    postproc_visualizer
 "

@@ -298,7 +303,7 @@ process_cmdline() {
        optval="${opt#*=}"
        case "$opt" in
        --disable-codecs) for c in ${CODECS}; do disable $c; done ;;
-        *) process_common_cmdline "$opt"
+        *) process_common_cmdline $opt
        ;;
        esac
    done
--- a/docs.mk
+++ b/docs.mk
@@ -34,8 +34,7 @@ TXT_DOX = $(call enabled,TXT_DOX)

 EXAMPLE_PATH += $(SRC_PATH_BARE) #for CHANGELOG, README, etc

-doxyfile: $(if $(findstring examples, $(ALL_TARGETS)),examples.doxy)
-doxyfile: libs.doxy_template libs.doxy
+doxyfile: libs.doxy_template libs.doxy examples.doxy
 	@echo "    [CREATE] $@"
 	@cat $^ > $@
 	@echo "STRIP_FROM_PATH += $(SRC_PATH_BARE) $(BUILD_ROOT)" >> $@
--- a/examples.mk
+++ b/examples.mk
@@ -77,6 +77,11 @@ GEN_EXAMPLES-$(CONFIG_ENCODERS) += decode_with_drops.c
 endif
 decode_with_drops.GUID           = CE5C53C4-8DDA-438A-86ED-0DDD3CDB8D26
 decode_with_drops.DESCRIPTION    = Drops frames while decoding
+ifeq ($(CONFIG_DECODERS),yes)
+GEN_EXAMPLES-$(CONFIG_ENCODERS) += decode_with_partial_drops.c
+endif
+decode_partial_with_drops.GUID           = CE5C53C4-8DDA-438A-86ED-0DDD3CDB8D27
+decode_partial_with_drops.DESCRIPTION    = Drops parts of frames while decoding
 GEN_EXAMPLES-$(CONFIG_ENCODERS) += error_resilient.c
 error_resilient.GUID             = DF5837B9-4145-4F92-A031-44E4F832E00C
 error_resilient.DESCRIPTION      = Error Resiliency Feature
@@ -93,16 +98,8 @@ vp8cx_set_ref.DESCRIPTION           = VP8 set encoder reference frame


 # Handle extra library flags depending on codec configuration
-
-# We should not link to math library (libm) on RVCT
-# when building for bare-metal targets
-ifeq ($(CONFIG_OS_SUPPORT), yes)
 CODEC_EXTRA_LIBS-$(CONFIG_VP8)         += m
-else
-    ifeq ($(CONFIG_GCC), yes)
-    CODEC_EXTRA_LIBS-$(CONFIG_VP8)         += m
-    endif
-endif
+
 #
 # End of specified files. The rest of the build rules should happen
 # automagically from here.
--- a/examples/decode_with_partial_drops.txt
+++ b/examples/decode_with_partial_drops.txt
@@ -0,0 +1,213 @@
+@TEMPLATE decoder_tmpl.c
+Decode With Drops Example
+=========================
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ INTRODUCTION
+This is an example utility which drops a series of frames, as specified
+on the command line. This is useful for observing the error recovery
+features of the codec.
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ INTRODUCTION
+
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ EXTRA_INCLUDES
+#include <time.h>
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ EXTRA_INCLUDES
+
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ HELPERS
+struct parsed_header
+{
+    char key_frame;
+    int version;
+    char show_frame;
+    int first_part_size;
+};
+
+int next_packet(struct parsed_header* hdr, int pos, int length, int mtu)
+{
+    int size = 0;
+    int remaining = length - pos;
+    /* Uncompressed part is 3 bytes for P frames and 10 bytes for I frames */
+    int uncomp_part_size = (hdr->key_frame ? 10 : 3);
+    /* number of bytes yet to send from header and the first partition */
+    int remainFirst = uncomp_part_size + hdr->first_part_size - pos;
+    if (remainFirst > 0)
+    {
+        if (remainFirst <= mtu)
+        {
+            size = remainFirst;
+        }
+        else
+        {
+            size = mtu;
+        }
+
+        return size;
+    }
+
+    /* second partition; just slot it up according to MTU */
+    if (remaining <= mtu)
+    {
+        size = remaining;
+        return size;
+    }
+    return mtu;
+}
+
+void throw_packets(unsigned char* frame, int* size, int loss_rate, int* thrown, int* kept)
+{
+    unsigned char loss_frame[256*1024];
+    int pkg_size = 1;
+    int count = 0;
+    int pos = 0;
+    int loss_pos = 0;
+    struct parsed_header hdr;
+    unsigned int tmp;
+    int mtu = 100;
+
+    if (*size < 3)
+    {
+        return;
+    }
+    putc('|', stdout);
+    /* parse uncompressed 3 bytes */
+    tmp = (frame[2] << 16) | (frame[1] << 8) | frame[0];
+    hdr.key_frame = !(tmp & 0x1); /* inverse logic */
+    hdr.version = (tmp >> 1) & 0x7;
+    hdr.show_frame = (tmp >> 4) & 0x1;
+    hdr.first_part_size = (tmp >> 5) & 0x7FFFF;
+
+    /* don't drop key frames */
+    if (hdr.key_frame)
+    {
+        int i;
+        *kept = *size/mtu + ((*size % mtu > 0) ? 1 : 0); /* approximate */
+        for (i=0; i < *kept; i++)
+            putc('.', stdout);
+        return;
+    }
+
+    while ((pkg_size = next_packet(&hdr, pos, *size, mtu)) > 0)
+    {
+        int loss_event = ((rand() + 1.0)/(RAND_MAX + 1.0) < loss_rate/100.0);
+        if (*thrown == 0 && !loss_event)
+        {
+            memcpy(loss_frame + loss_pos, frame + pos, pkg_size);
+            loss_pos += pkg_size;
+            (*kept)++;
+            putc('.', stdout);
+        }
+        else
+        {
+            (*thrown)++;
+            putc('X', stdout);
+        }
+        pos += pkg_size;
+    }
+    memcpy(frame, loss_frame, loss_pos);
+    memset(frame + loss_pos, 0, *size - loss_pos);
+    *size = loss_pos;
+}
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ HELPERS
+
+Usage
+-----
+This example adds a single argument to the `simple_decoder` example,
+which specifies the range or pattern of frames to drop. The parameter is
+parsed as follows:
+
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ USAGE
+if(argc!=4 && argc != 5)
+    die("Usage: %s <infile> <outfile> <N-M|N/M|L,S>\n", argv[0]);
+{
+    char *nptr;
+    n = strtol(argv[3], &nptr, 0);
+    mode = (*nptr == '\0' || *nptr == ',') ? 2 : (*nptr == '-') ? 1 : 0;
+
+    m = strtol(nptr+1, NULL, 0);
+    if((!n && !m) || (*nptr != '-' && *nptr != '/' &&
+        *nptr != '\0' && *nptr != ','))
+        die("Couldn't parse pattern %s\n", argv[3]);
+}
+seed = (m > 0) ? m : (unsigned int)time(NULL);
+srand(seed);thrown_frame = 0;
+printf("Seed: %u\n", seed);
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ USAGE
+
+
+Dropping A Range Of Frames
+--------------------------
+To drop a range of frames, specify the starting frame and the ending
+frame to drop, separated by a dash. The following command will drop
+frames 5 through 10 (base 1).
+
+  $ ./decode_with_drops in.ivf out.i420 5-10
+
+
+Dropping A Pattern Of Frames
+----------------------------
+To drop a pattern of frames, specify the number of frames to drop and
+the number of frames after which to repeat the pattern, separated by
+a forward-slash. The following command will drop 3 of 7 frames.
+Specifically, it will decode 4 frames, then drop 3 frames, and then
+repeat.
+
+  $ ./decode_with_drops in.ivf out.i420 3/7
+
+
+Extra Variables
+---------------
+This example maintains the pattern passed on the command line in the
+`n`, `m`, and `is_range` variables:
+
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ EXTRA_VARS
+int              n, m, mode;                                        //
+unsigned int     seed;
+int              thrown=0, kept=0;
+int              thrown_frame=0, kept_frame=0;
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ EXTRA_VARS
+
+
+Making The Drop Decision
+------------------------
+The example decides whether to drop the frame based on the current
+frame number, immediately before decoding the frame.
+
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ PRE_DECODE
+/* Decide whether to throw parts of the frame or the whole frame
+   depending on the drop mode */
+thrown_frame = 0;
+kept_frame = 0;
+switch (mode)
+{
+case 0:
+    if (m - (frame_cnt-1)%m <= n)
+    {
+        frame_sz = 0;
+    }
+    break;
+case 1:
+    if (frame_cnt >= n && frame_cnt <= m)
+    {
+        frame_sz = 0;
+    }
+    break;
+case 2:
+    throw_packets(frame, &frame_sz, n, &thrown_frame, &kept_frame);
+    break;
+default: break;
+}
+if (mode < 2)
+{
+    if (frame_sz == 0)
+    {
+        putc('X', stdout);
+        thrown_frame++;
+    }
+    else
+    {
+        putc('.', stdout);
+        kept_frame++;
+    }
+}
+thrown += thrown_frame;
+kept += kept_frame;
+fflush(stdout);
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ PRE_DECODE
--- a/examples/decoder_tmpl.c
+++ b/examples/decoder_tmpl.c
@@ -19,7 +19,7 @@
 #define VPX_CODEC_DISABLE_COMPAT 1
 #include "vpx/vpx_decoder.h"
 #include "vpx/vp8dx.h"
-#define interface (vpx_codec_vp8_dx())
+#define interface (&vpx_codec_vp8_dx_algo)
@EXTRA_INCLUDES


@@ -42,6 +42,8 @@ static void die(const char *fmt, ...) {

@DIE_CODEC

+@HELPERS
+
 int main(int argc, char **argv) {
    FILE            *infile, *outfile;
    vpx_codec_ctx_t  codec;
--- a/examples/decoder_tmpl.txt
+++ b/examples/decoder_tmpl.txt
@@ -2,7 +2,7 @@
 #define VPX_CODEC_DISABLE_COMPAT 1
 #include "vpx/vpx_decoder.h"
 #include "vpx/vp8dx.h"
-#define interface (vpx_codec_vp8_dx())
+#define interface (&vpx_codec_vp8_dx_algo)
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ DEC_INCLUDES


--- a/examples/encoder_tmpl.c
+++ b/examples/encoder_tmpl.c
@@ -19,7 +19,7 @@
 #define VPX_CODEC_DISABLE_COMPAT 1
 #include "vpx/vpx_encoder.h"
 #include "vpx/vp8cx.h"
-#define interface (vpx_codec_vp8_cx())
+#define interface (&vpx_codec_vp8_cx_algo)
 #define fourcc    0x30385056
@EXTRA_INCLUDES

--- a/examples/encoder_tmpl.txt
+++ b/examples/encoder_tmpl.txt
@@ -2,7 +2,7 @@
 #define VPX_CODEC_DISABLE_COMPAT 1
 #include "vpx/vpx_encoder.h"
 #include "vpx/vp8cx.h"
-#define interface (vpx_codec_vp8_cx())
+#define interface (&vpx_codec_vp8_cx_algo)
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ENC_INCLUDES


--- a/examples/simple_decoder.txt
+++ b/examples/simple_decoder.txt
@@ -33,7 +33,7 @@ Initializing The Codec
 ----------------------
 The decoder is initialized by the following code. This is an example for
 the VP8 decoder, but the code is analogous for all algorithms. Replace
-`vpx_codec_vp8_dx()` with a pointer to the interface exposed by the
+`&vpx_codec_vp8_dx_algo` with a pointer to the interface exposed by the
 algorithm you want to use. The `cfg` argument is left as NULL in this
 example, because we want the algorithm to determine the stream
 configuration (width/height) and allocate memory automatically. This
--- a/libs.mk
+++ b/libs.mk
@@ -126,6 +126,29 @@ INSTALL-SRCS-$(CONFIG_CODEC_SRCS) += $(call enabled,CODEC_EXPORTS)
 ifeq ($(CONFIG_EXTERNAL_BUILD),yes)
 ifeq ($(CONFIG_MSVS),yes)

+ifeq ($(ARCH_ARM),yes)
+ifeq ($(HAVE_ARMV5TE),yes)
+ARM_ARCH=v5
+endif
+ifeq ($(HAVE_ARMV6),yes)
+ARM_ARCH=v6
+endif
+obj_int_extract.vcproj: $(SRC_PATH_BARE)/build/make/obj_int_extract.c
+	@cp $(SRC_PATH_BARE)/build/arm-wince-vs8/obj_int_extract.bat .
+	@echo "    [CREATE] $@"
+	$(SRC_PATH_BARE)/build/make/gen_msvs_proj.sh\
+			--exe\
+			--target=$(TOOLCHAIN)\
+            $(if $(CONFIG_STATIC_MSVCRT),--static-crt) \
+            --name=obj_int_extract\
+            --proj-guid=E1360C65-D375-4335-8057-7ED99CC3F9B2\
+            --out=$@ $^\
+            -I".&quot;;&quot;$(SRC_PATH_BARE)"
+
+PROJECTS-$(BUILD_LIBVPX) += obj_int_extract.vcproj
+PROJECTS-$(BUILD_LIBVPX) += obj_int_extract.bat
+endif
+
 vpx.def: $(call enabled,CODEC_EXPORTS)
 	@echo "    [CREATE] $@"
 	$(SRC_PATH_BARE)/build/make/gen_msvs_def.sh\
@@ -207,39 +230,10 @@ endif
 #
 # Add assembler dependencies for configuration and offsets
 #
+#$(filter %$(ASM).o,$(OBJS-yes)): $(BUILD_PFX)vpx_config.asm $(BUILD_PFX)vpx_asm_offsets.asm
 $(filter %.s.o,$(OBJS-yes)):   $(BUILD_PFX)vpx_config.asm
 $(filter %.asm.o,$(OBJS-yes)): $(BUILD_PFX)vpx_config.asm

-#
-# Calculate platform- and compiler-specific offsets for hand coded assembly
-#
-ifeq ($(ARCH_ARM), yes)
-  asm_com_offsets.asm: obj_int_extract
-  asm_com_offsets.asm: $(VP8_PREFIX)common/asm_com_offsets.c.o
-	./obj_int_extract rvds $< $(ADS2GAS) > $@
-  OBJS-yes += $(VP8_PREFIX)common/asm_com_offsets.c.o
-  CLEAN-OBJS += asm_com_offsets.asm
-  $(filter %$(ASM).o,$(OBJS-yes)): $(BUILD_PFX)asm_com_offsets.asm
-
-  ifeq ($(CONFIG_VP8_ENCODER), yes)
-    asm_enc_offsets.asm: obj_int_extract
-    asm_enc_offsets.asm: $(VP8_PREFIX)encoder/asm_enc_offsets.c.o
-	./obj_int_extract rvds $< $(ADS2GAS) > $@
-    OBJS-yes += $(VP8_PREFIX)encoder/asm_enc_offsets.c.o
-    CLEAN-OBJS += asm_enc_offsets.asm
-    $(filter %$(ASM).o,$(OBJS-yes)): $(BUILD_PFX)asm_enc_offsets.asm
-  endif
-
-  ifeq ($(CONFIG_VP8_DECODER), yes)
-    asm_dec_offsets.asm: obj_int_extract
-    asm_dec_offsets.asm: $(VP8_PREFIX)decoder/asm_dec_offsets.c.o
-	./obj_int_extract rvds $< $(ADS2GAS) > $@
-    OBJS-yes += $(VP8_PREFIX)decoder/asm_dec_offsets.c.o
-    CLEAN-OBJS += asm_dec_offsets.asm
-    $(filter %$(ASM).o,$(OBJS-yes)): $(BUILD_PFX)asm_dec_offsets.asm
-  endif
-endif
-
 $(shell $(SRC_PATH_BARE)/build/make/version.sh "$(SRC_PATH_BARE)" $(BUILD_PFX)vpx_version.h)
 CLEAN-OBJS += $(BUILD_PFX)vpx_version.h

--- a/mainpage.dox
+++ b/mainpage.dox
@@ -31,7 +31,7 @@
  The WebM project is an open source project supported by its community. For
  questions about this SDK, please mail the apps-devel@webmproject.org list.
  To contribute, see http://www.webmproject.org/code/contribute and mail
-  codec-devel@webmproject.org.
+  vpx-devel@webmproject.org.
 */

 /*!\page changelog CHANGELOG
--- a/md5_utils.c
+++ b/md5_utils.c
@@ -20,6 +20,8 @@
 * Still in the public domain.
 */

+#include <sys/types.h>    /* for stupid systems */
+
 #include <string.h>   /* for memcpy() */

 #include "md5_utils.h"
--- a/solution.mk
+++ b/solution.mk
@@ -9,6 +9,32 @@
 ##


+ifeq ($(ARCH_ARM),yes)
+ARM_DEVELOP=no
+ARM_DEVELOP:=$(if $(filter %vpx.vcproj,$(wildcard *.vcproj)),yes)
+
+ifeq ($(ARM_DEVELOP),yes)
+vpx.sln:
+	@echo "    [COPY] $@"
+	@cp $(SRC_PATH_BARE)/build/arm-wince-vs8/vpx.sln .
+PROJECTS-yes += vpx.sln
+else
+vpx.sln: $(wildcard *.vcproj)
+	@echo "    [CREATE] $@"
+	$(SRC_PATH_BARE)/build/make/gen_msvs_sln.sh \
+            $(if $(filter %vpx.vcproj,$^),--dep=vpxdec:vpx) \
+            $(if $(filter %vpx.vcproj,$^),--dep=xma:vpx) \
+            --ver=$(CONFIG_VS_VERSION)\
+            --target=$(TOOLCHAIN)\
+            --out=$@ $^
+vpx.sln.mk: vpx.sln
+	@true
+
+PROJECTS-yes += vpx.sln vpx.sln.mk
+-include vpx.sln.mk
+endif
+
+else
 vpx.sln: $(wildcard *.vcproj)
 	@echo "    [CREATE] $@"
 	$(SRC_PATH_BARE)/build/make/gen_msvs_sln.sh \
@@ -22,6 +48,7 @@ vpx.sln.mk: vpx.sln

 PROJECTS-yes += vpx.sln vpx.sln.mk
 -include vpx.sln.mk
+endif

 # Always install this file, as it is an unconditional post-build rule.
 INSTALL_MAPS += src/%     $(SRC_PATH_BARE)/%
--- a/usage.dox
+++ b/usage.dox
@@ -25,7 +25,7 @@
    codec may write into to store details about a single instance of that codec.
    Most of the context is implementation specific, and thus opaque to the
    application. The context structure as seen by the application is of fixed
-    size, and thus can be allocated with automatic storage or dynamically
+    size, and thus can be allocated eith with automatic storage or dynamically
    on the heap.

    Most operations require an initialized codec context. Codec context
@@ -74,7 +74,7 @@
    the ABI is versioned. The ABI version number must be passed at
    initialization time to ensure the application is using a header file that
    matches the library. The current ABI version number is stored in the
-    preprocessor macros #VPX_CODEC_ABI_VERSION, #VPX_ENCODER_ABI_VERSION, and
+    prepropcessor macros #VPX_CODEC_ABI_VERSION, #VPX_ENCODER_ABI_VERSION, and
    #VPX_DECODER_ABI_VERSION. For convenience, each initialization function has
    a wrapper macro that inserts the correct version number. These macros are
    named like the initialization methods, but without the _ver suffix.
@@ -125,7 +125,7 @@

    The special value <code>0</code> is reserved to represent an infinite
    deadline. In this case, the codec will perform as much processing as
-    possible to yield the highest quality frame.
+    possible to yeild the highest quality frame.

    By convention, the value <code>1</code> is used to mean "return as fast as
    possible."
@@ -135,7 +135,7 @@

 /*! \page usage_xma External Memory Allocation
    Applications that wish to have fine grained control over how and where
-    decoders allocate memory \ref MAY make use of the eXternal Memory Allocation
+    decoders allocate memory \ref MAY make use of the e_xternal Memory Allocation
    (XMA) interface. Not all codecs support the XMA \ref usage_features.

    To use a decoder in XMA mode, the decoder \ref MUST be initialized with the
@@ -143,7 +143,7 @@
    allocate is heavily dependent on the size of the encoded video frames. The
    size of the video must be known before requesting the decoder's memory map.
    This stream information can be obtained with the vpx_codec_peek_stream_info()
-    function, which does not require a constructed decoder context. If the exact
+    function, which does not require a contructed decoder context. If the exact
    stream is not known, a stream info structure can be created that reflects
    the maximum size that the decoder instance is required to support.

@@ -175,7 +175,7 @@
    \section usage_xma_seg_szalign Segment Size and Alignment
    The sz (size) and align (alignment) parameters describe the required size
    and alignment of the requested segment. Alignment will always be a power of
-    two. Applications \ref MUST honor the alignment requested. Failure to do so
+    two. Applications \ref MUST honor the aligment requested. Failure to do so
    could result in program crashes or may incur a speed penalty.

    \section usage_xma_seg_flags Segment Flags
--- a/vp8/common/alloccommon.c
+++ b/vp8/common/alloccommon.c
@@ -12,10 +12,12 @@
 #include "vpx_ports/config.h"
 #include "blockd.h"
 #include "vpx_mem/vpx_mem.h"
+#include "error_concealment.h"
 #include "onyxc_int.h"
 #include "findnearmv.h"
 #include "entropymode.h"
 #include "systemdependent.h"
+#include "vpxerrors.h"


 extern  void vp8_init_scan_order_mask();
@@ -27,6 +29,9 @@ void vp8_update_mode_info_border(MODE_INFO *mi, int rows, int cols)

    for (i = 0; i < rows; i++)
    {
+        /* TODO(holmer): Bug? This updates the last element of each row
+         * rather than the border element!
+         */
        vpx_memset(&mi[i*cols-1], 0, sizeof(MODE_INFO));
    }
 }
@@ -43,9 +48,11 @@ void vp8_de_alloc_frame_buffers(VP8_COMMON *oci)

    vpx_free(oci->above_context);
    vpx_free(oci->mip);
+    vpx_free(oci->prev_mip);

    oci->above_context = 0;
    oci->mip = 0;
+    oci->prev_mip = 0;

 }

@@ -70,7 +77,7 @@ int vp8_alloc_frame_buffers(VP8_COMMON *oci, int width, int height)
      if (vp8_yv12_alloc_frame_buffer(&oci->yv12_fb[i],  width, height, VP8BORDERINPIXELS) < 0)
        {
            vp8_de_alloc_frame_buffers(oci);
-            return 1;
+            return ALLOC_FAILURE;
        }
    }

@@ -87,13 +94,13 @@ int vp8_alloc_frame_buffers(VP8_COMMON *oci, int width, int height)
    if (vp8_yv12_alloc_frame_buffer(&oci->temp_scale_frame,   width, 16, VP8BORDERINPIXELS) < 0)
    {
        vp8_de_alloc_frame_buffers(oci);
-        return 1;
+        return ALLOC_FAILURE;
    }

    if (vp8_yv12_alloc_frame_buffer(&oci->post_proc_buffer, width, height, VP8BORDERINPIXELS) < 0)
    {
        vp8_de_alloc_frame_buffers(oci);
-        return 1;
+        return ALLOC_FAILURE;
    }

    oci->mb_rows = height >> 4;
@@ -105,18 +112,28 @@ int vp8_alloc_frame_buffers(VP8_COMMON *oci, int width, int height)
    if (!oci->mip)
    {
        vp8_de_alloc_frame_buffers(oci);
-        return 1;
+        return ALLOC_FAILURE;
    }

    oci->mi = oci->mip + oci->mode_info_stride + 1;

+    /* allocate memory for last frame MODE_INFO array */
+    oci->prev_mip = vpx_calloc((oci->mb_cols + 1) * (oci->mb_rows + 1), sizeof(MODE_INFO));
+
+    if (!oci->prev_mip)
+    {
+        vp8_de_alloc_frame_buffers(oci);
+        return ALLOC_FAILURE;
+    }
+
+    oci->prev_mi = oci->prev_mip + oci->mode_info_stride + 1;

    oci->above_context = vpx_calloc(sizeof(ENTROPY_CONTEXT_PLANES) * oci->mb_cols, 1);

    if (!oci->above_context)
    {
        vp8_de_alloc_frame_buffers(oci);
-        return 1;
+        return ALLOC_FAILURE;
    }

    vp8_update_mode_info_border(oci->mi, oci->mb_rows, oci->mb_cols);
--- a/vp8/common/arm/arm_systemdependent.c
+++ b/vp8/common/arm/arm_systemdependent.c
@@ -11,13 +11,13 @@

 #include "vpx_ports/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"
+#include "g_common.h"
+#include "pragmas.h"
+#include "subpixel.h"
+#include "loopfilter.h"
+#include "recon.h"
+#include "idct.h"
+#include "onyxc_int.h"

 extern void (*vp8_build_intra_predictors_mby_ptr)(MACROBLOCKD *x);
 extern void vp8_build_intra_predictors_mby(MACROBLOCKD *x);
--- a/vp8/common/arm/armv6/bilinearfilter_v6.asm
+++ b/vp8/common/arm/armv6/bilinearfilter_v6.asm
@@ -15,19 +15,19 @@
    AREA    |.text|, CODE, READONLY  ; name this block of code

 ;-------------------------------------
-; r0    unsigned char  *src_ptr,
-; r1    unsigned short *dst_ptr,
-; r2    unsigned int    src_pitch,
-; r3    unsigned int    height,
-; stack unsigned int    width,
-; stack const short    *vp8_filter
+; r0    unsigned char *src_ptr,
+; r1    unsigned short *output_ptr,
+; r2    unsigned int src_pixels_per_line,
+; r3    unsigned int output_height,
+; stack    unsigned int output_width,
+; stack    const short *vp8_filter
 ;-------------------------------------
 ; The output is transposed stroed in output array to make it easy for second pass filtering.
 |vp8_filter_block2d_bil_first_pass_armv6| PROC
    stmdb   sp!, {r4 - r11, lr}

    ldr     r11, [sp, #40]                  ; vp8_filter address
-    ldr     r4, [sp, #36]                   ; width
+    ldr     r4, [sp, #36]                   ; output width

    mov     r12, r3                         ; outer-loop counter
    sub     r2, r2, r4                      ; src increment for height loop
@@ -38,10 +38,10 @@

    ldr     r5, [r11]                       ; load up filter coefficients

-    mov     r3, r3, lsl #1                  ; height*2
+    mov     r3, r3, lsl #1                  ; output_height*2
    add     r3, r3, #2                      ; plus 2 to make output buffer 4-bit aligned since height is actually (height+1)

-    mov     r11, r1                         ; save dst_ptr for each row
+    mov     r11, r1                         ; save output_ptr for each row

    cmp     r5, #128                        ; if filter coef = 128, then skip the filter
    beq     bil_null_1st_filter
@@ -140,17 +140,17 @@

 ;---------------------------------
 ; r0    unsigned short *src_ptr,
-; r1    unsigned char  *dst_ptr,
-; r2    int             dst_pitch,
-; r3    unsigned int    height,
-; stack unsigned int    width,
-; stack const short    *vp8_filter
+; r1    unsigned char *output_ptr,
+; r2    int output_pitch,
+; r3    unsigned int  output_height,
+; stack unsigned int  output_width,
+; stack const short *vp8_filter
 ;---------------------------------
 |vp8_filter_block2d_bil_second_pass_armv6| PROC
    stmdb   sp!, {r4 - r11, lr}

    ldr     r11, [sp, #40]                  ; vp8_filter address
-    ldr     r4, [sp, #36]                   ; width
+    ldr     r4, [sp, #36]                   ; output width

    ldr     r5, [r11]                       ; load up filter coefficients
    mov     r12, r4                         ; outer-loop counter = width, since we work on transposed data matrix
--- a/vp8/common/arm/armv6/sixtappredict8x4_v6.asm
+++ b/vp8/common/arm/armv6/sixtappredict8x4_v6.asm
@@ -243,6 +243,8 @@ skip_secondpass_hloop
    ENDP

 ;-----------------
+    AREA    subpelfilters8_dat, DATA, READWRITE         ;read/write by default
+;Data section with name data_area is specified. DCD reserves space in memory for 48 data.
 ;One word each is reserved. Label filter_coeff can be used to access the data.
 ;Data address: filter_coeff, filter_coeff+4, filter_coeff+8 ...
 _filter8_coeff_
--- a/vp8/common/arm/bilinearfilter_arm.c
+++ b/vp8/common/arm/bilinearfilter_arm.c
@@ -10,29 +10,128 @@


 #include <math.h>
-#include "vp8/common/filter.h"
-#include "vp8/common/subpixel.h"
-#include "bilinearfilter_arm.h"
+#include "subpixel.h"
+
+#define BLOCK_HEIGHT_WIDTH 4
+#define VP8_FILTER_WEIGHT 128
+#define VP8_FILTER_SHIFT  7
+
+static const short bilinear_filters[8][2] =
+{
+    { 128,   0 },
+    { 112,  16 },
+    {  96,  32 },
+    {  80,  48 },
+    {  64,  64 },
+    {  48,  80 },
+    {  32,  96 },
+    {  16, 112 }
+};
+
+
+extern void vp8_filter_block2d_bil_first_pass_armv6
+(
+    unsigned char *src_ptr,
+    unsigned short *output_ptr,
+    unsigned int src_pixels_per_line,
+    unsigned int output_height,
+    unsigned int output_width,
+    const short *vp8_filter
+);
+
+extern void vp8_filter_block2d_bil_second_pass_armv6
+(
+    unsigned short *src_ptr,
+    unsigned char  *output_ptr,
+    int output_pitch,
+    unsigned int  output_height,
+    unsigned int  output_width,
+    const short *vp8_filter
+);
+
+#if 0
+void vp8_filter_block2d_bil_first_pass_6
+(
+    unsigned char *src_ptr,
+    unsigned short *output_ptr,
+    unsigned int src_pixels_per_line,
+    unsigned int output_height,
+    unsigned int output_width,
+    const short *vp8_filter
+)
+{
+    unsigned int i, j;
+
+    for ( i=0; i<output_height; i++ )
+    {
+        for ( j=0; j<output_width; j++ )
+        {
+            /* Apply bilinear filter */
+            output_ptr[j] = ( ( (int)src_ptr[0]          * vp8_filter[0]) +
+                               ((int)src_ptr[1] * vp8_filter[1]) +
+                                (VP8_FILTER_WEIGHT/2) ) >> VP8_FILTER_SHIFT;
+            src_ptr++;
+        }
+
+        /* Next row... */
+        src_ptr    += src_pixels_per_line - output_width;
+        output_ptr += output_width;
+    }
+}
+
+void vp8_filter_block2d_bil_second_pass_6
+(
+    unsigned short *src_ptr,
+    unsigned char  *output_ptr,
+    int output_pitch,
+    unsigned int  output_height,
+    unsigned int  output_width,
+    const short *vp8_filter
+)
+{
+    unsigned int  i,j;
+    int  Temp;
+
+    for ( i=0; i<output_height; i++ )
+    {
+        for ( j=0; j<output_width; j++ )
+        {
+            /* Apply filter */
+            Temp =  ((int)src_ptr[0]         * vp8_filter[0]) +
+                    ((int)src_ptr[output_width] * vp8_filter[1]) +
+                    (VP8_FILTER_WEIGHT/2);
+            output_ptr[j] = (unsigned int)(Temp >> VP8_FILTER_SHIFT);
+            src_ptr++;
+        }
+
+        /* Next row... */
+        /*src_ptr    += src_pixels_per_line - output_width;*/
+        output_ptr += output_pitch;
+    }
+}
+#endif

 void vp8_filter_block2d_bil_armv6
 (
    unsigned char *src_ptr,
-    unsigned char *dst_ptr,
-    unsigned int   src_pitch,
+    unsigned char *output_ptr,
+    unsigned int   src_pixels_per_line,
    unsigned int   dst_pitch,
-    const short   *HFilter,
-    const short   *VFilter,
+    const short      *HFilter,
+    const short      *VFilter,
    int            Width,
    int            Height
 )
 {
-    unsigned short FData[36*16]; /* Temp data buffer used in filtering */
+
+    unsigned short FData[36*16]; /* Temp data bufffer used in filtering */

    /* First filter 1-D horizontally... */
-    vp8_filter_block2d_bil_first_pass_armv6(src_ptr, FData, src_pitch, Height + 1, Width, HFilter);
+    /* pixel_step = 1; */
+    vp8_filter_block2d_bil_first_pass_armv6(src_ptr, FData, src_pixels_per_line, Height + 1, Width, HFilter);

    /* then 1-D vertically... */
-    vp8_filter_block2d_bil_second_pass_armv6(FData, dst_ptr, dst_pitch, Height, Width, VFilter);
+    vp8_filter_block2d_bil_second_pass_armv6(FData, output_ptr, dst_pitch, Height, Width, VFilter);
 }


@@ -49,8 +148,8 @@ void vp8_bilinear_predict4x4_armv6
    const short  *HFilter;
    const short  *VFilter;

-    HFilter = vp8_bilinear_filters[xoffset];
-    VFilter = vp8_bilinear_filters[yoffset];
+    HFilter = bilinear_filters[xoffset];
+    VFilter = bilinear_filters[yoffset];

    vp8_filter_block2d_bil_armv6(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter, VFilter, 4, 4);
 }
@@ -68,8 +167,8 @@ void vp8_bilinear_predict8x8_armv6
    const short  *HFilter;
    const short  *VFilter;

-    HFilter = vp8_bilinear_filters[xoffset];
-    VFilter = vp8_bilinear_filters[yoffset];
+    HFilter = bilinear_filters[xoffset];
+    VFilter = bilinear_filters[yoffset];

    vp8_filter_block2d_bil_armv6(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter, VFilter, 8, 8);
 }
@@ -87,8 +186,8 @@ void vp8_bilinear_predict8x4_armv6
    const short  *HFilter;
    const short  *VFilter;

-    HFilter = vp8_bilinear_filters[xoffset];
-    VFilter = vp8_bilinear_filters[yoffset];
+    HFilter = bilinear_filters[xoffset];
+    VFilter = bilinear_filters[yoffset];

    vp8_filter_block2d_bil_armv6(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter, VFilter, 8, 4);
 }
@@ -106,8 +205,8 @@ void vp8_bilinear_predict16x16_armv6
    const short  *HFilter;
    const short  *VFilter;

-    HFilter = vp8_bilinear_filters[xoffset];
-    VFilter = vp8_bilinear_filters[yoffset];
+    HFilter = bilinear_filters[xoffset];
+    VFilter = bilinear_filters[yoffset];

    vp8_filter_block2d_bil_armv6(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter, VFilter, 16, 16);
 }
--- a/vp8/common/arm/bilinearfilter_arm.h
+++ b/vp8/common/arm/bilinearfilter_arm.h
@@ -1,35 +0,0 @@
-/*
- *  Copyright (c) 2011 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 BILINEARFILTER_ARM_H
-#define BILINEARFILTER_ARM_H
-
-extern void vp8_filter_block2d_bil_first_pass_armv6
-(
-    const unsigned char  *src_ptr,
-    unsigned short       *dst_ptr,
-    unsigned int          src_pitch,
-    unsigned int          height,
-    unsigned int          width,
-    const short          *vp8_filter
-);
-
-extern void vp8_filter_block2d_bil_second_pass_armv6
-(
-    const unsigned short *src_ptr,
-    unsigned char        *dst_ptr,
-    int                   dst_pitch,
-    unsigned int          height,
-    unsigned int          width,
-    const short         *vp8_filter
-);
-
-#endif /* BILINEARFILTER_ARM_H */
--- a/vp8/common/arm/filter_arm.c
+++ b/vp8/common/arm/filter_arm.c
@@ -11,10 +11,26 @@

 #include "vpx_ports/config.h"
 #include <math.h>
-#include "vp8/common/filter.h"
-#include "vp8/common/subpixel.h"
+#include "subpixel.h"
 #include "vpx_ports/mem.h"

+#define BLOCK_HEIGHT_WIDTH 4
+#define VP8_FILTER_WEIGHT 128
+#define VP8_FILTER_SHIFT  7
+
+DECLARE_ALIGNED(16, static const short, sub_pel_filters[8][6]) =
+{
+    { 0,  0,  128,    0,   0,  0 },         /* note that 1/8 pel positions are just as per alpha -0.5 bicubic */
+    { 0, -6,  123,   12,  -1,  0 },
+    { 2, -11, 108,   36,  -8,  1 },         /* New 1/4 pel 6 tap filter */
+    { 0, -9,   93,   50,  -6,  0 },
+    { 3, -16,  77,   77, -16,  3 },         /* New 1/2 pel 6 tap filter */
+    { 0, -6,   50,   93,  -9,  0 },
+    { 1, -8,   36,  108, -11,  2 },         /* New 1/4 pel 6 tap filter */
+    { 0, -1,   12,  123,  -6,  0 },
+};
+
+
 extern void vp8_filter_block2d_first_pass_armv6
 (
    unsigned char *src_ptr,
@@ -77,11 +93,11 @@ void vp8_sixtap_predict_armv6
 {
    const short  *HFilter;
    const short  *VFilter;
-    DECLARE_ALIGNED_ARRAY(4, short, FData, 12*4); /* Temp data buffer used in filtering */
+    DECLARE_ALIGNED_ARRAY(4, short, FData, 12*4); /* Temp data bufffer used in filtering */


-    HFilter = vp8_sub_pel_filters[xoffset];   /* 6 tap */
-    VFilter = vp8_sub_pel_filters[yoffset];   /* 6 tap */
+    HFilter = sub_pel_filters[xoffset];   /* 6 tap */
+    VFilter = sub_pel_filters[yoffset];       /* 6 tap */

    /* Vfilter is null. First pass only */
    if (xoffset && !yoffset)
@@ -113,6 +129,47 @@ void vp8_sixtap_predict_armv6
    }
 }

+#if 0
+void vp8_sixtap_predict8x4_armv6
+(
+    unsigned char  *src_ptr,
+    int  src_pixels_per_line,
+    int  xoffset,
+    int  yoffset,
+    unsigned char *dst_ptr,
+    int  dst_pitch
+)
+{
+    const short  *HFilter;
+    const short  *VFilter;
+    DECLARE_ALIGNED_ARRAY(4, short, FData, 16*8); /* Temp data bufffer used in filtering */
+
+    HFilter = sub_pel_filters[xoffset];   /* 6 tap */
+    VFilter = sub_pel_filters[yoffset];       /* 6 tap */
+
+
+    /*if (xoffset && !yoffset)
+    {
+        vp8_filter_block2d_first_pass_only_armv6 (  src_ptr, dst_ptr, src_pixels_per_line, 8, dst_pitch, HFilter );
+    }*/
+    /* Hfilter is null. Second pass only */
+    /*else if (!xoffset && yoffset)
+    {
+        vp8_filter_block2d_second_pass_only_armv6 ( src_ptr, dst_ptr, src_pixels_per_line, 8, dst_pitch, VFilter );
+    }
+    else
+    {
+        if (yoffset & 0x1)
+            vp8_filter_block2d_first_pass_armv6 ( src_ptr-src_pixels_per_line, FData+1, src_pixels_per_line, 8, 7, HFilter );
+        else*/
+
+        vp8_filter_block2d_first_pass_armv6 ( src_ptr-(2*src_pixels_per_line), FData, src_pixels_per_line, 8, 9, HFilter );
+
+        vp8_filter_block2d_second_pass_armv6 ( FData+2, dst_ptr, dst_pitch, 4, 8, VFilter );
+    /*}*/
+}
+#endif
+
 void vp8_sixtap_predict8x8_armv6
 (
    unsigned char  *src_ptr,
@@ -125,10 +182,10 @@ void vp8_sixtap_predict8x8_armv6
 {
    const short  *HFilter;
    const short  *VFilter;
-    DECLARE_ALIGNED_ARRAY(4, short, FData, 16*8); /* Temp data buffer used in filtering */
+    DECLARE_ALIGNED_ARRAY(4, short, FData, 16*8); /* Temp data bufffer used in filtering */

-    HFilter = vp8_sub_pel_filters[xoffset];   /* 6 tap */
-    VFilter = vp8_sub_pel_filters[yoffset];   /* 6 tap */
+    HFilter = sub_pel_filters[xoffset];   /* 6 tap */
+    VFilter = sub_pel_filters[yoffset];       /* 6 tap */

    if (xoffset && !yoffset)
    {
@@ -167,10 +224,10 @@ void vp8_sixtap_predict16x16_armv6
 {
    const short  *HFilter;
    const short  *VFilter;
-    DECLARE_ALIGNED_ARRAY(4, short, FData, 24*16);    /* Temp data buffer used in filtering */
+    DECLARE_ALIGNED_ARRAY(4, short, FData, 24*16);    /* Temp data bufffer used in filtering */

-    HFilter = vp8_sub_pel_filters[xoffset];   /* 6 tap */
-    VFilter = vp8_sub_pel_filters[yoffset];   /* 6 tap */
+    HFilter = sub_pel_filters[xoffset];   /* 6 tap */
+    VFilter = sub_pel_filters[yoffset];       /* 6 tap */

    if (xoffset && !yoffset)
    {
--- a/vp8/common/arm/loopfilter_arm.c
+++ b/vp8/common/arm/loopfilter_arm.c
@@ -11,8 +11,8 @@

 #include "vpx_ports/config.h"
 #include <math.h>
-#include "vp8/common/loopfilter.h"
-#include "vp8/common/onyxc_int.h"
+#include "loopfilter.h"
+#include "onyxc_int.h"

 extern prototype_loopfilter(vp8_loop_filter_horizontal_edge_armv6);
 extern prototype_loopfilter(vp8_loop_filter_vertical_edge_armv6);
@@ -41,13 +41,13 @@ void vp8_loop_filter_mbh_armv6(unsigned char *y_ptr, unsigned char *u_ptr, unsig
                               int y_stride, int uv_stride, loop_filter_info *lfi, int simpler_lpf)
 {
    (void) simpler_lpf;
-    vp8_mbloop_filter_horizontal_edge_armv6(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->thr, 2);
+    vp8_mbloop_filter_horizontal_edge_armv6(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->mbthr, 2);

    if (u_ptr)
-        vp8_mbloop_filter_horizontal_edge_armv6(u_ptr, uv_stride, lfi->mbflim, lfi->lim, lfi->thr, 1);
+        vp8_mbloop_filter_horizontal_edge_armv6(u_ptr, uv_stride, lfi->uvmbflim, lfi->uvlim, lfi->uvmbthr, 1);

    if (v_ptr)
-        vp8_mbloop_filter_horizontal_edge_armv6(v_ptr, uv_stride, lfi->mbflim, lfi->lim, lfi->thr, 1);
+        vp8_mbloop_filter_horizontal_edge_armv6(v_ptr, uv_stride, lfi->uvmbflim, lfi->uvlim, lfi->uvmbthr, 1);
 }

 void vp8_loop_filter_mbhs_armv6(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
@@ -57,7 +57,7 @@ void vp8_loop_filter_mbhs_armv6(unsigned char *y_ptr, unsigned char *u_ptr, unsi
    (void) v_ptr;
    (void) uv_stride;
    (void) simpler_lpf;
-    vp8_loop_filter_simple_horizontal_edge_armv6(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->thr, 2);
+    vp8_loop_filter_simple_horizontal_edge_armv6(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->mbthr, 2);
 }

 /* Vertical MB Filtering */
@@ -65,13 +65,13 @@ void vp8_loop_filter_mbv_armv6(unsigned char *y_ptr, unsigned char *u_ptr, unsig
                               int y_stride, int uv_stride, loop_filter_info *lfi, int simpler_lpf)
 {
    (void) simpler_lpf;
-    vp8_mbloop_filter_vertical_edge_armv6(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->thr, 2);
+    vp8_mbloop_filter_vertical_edge_armv6(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->mbthr, 2);

    if (u_ptr)
-        vp8_mbloop_filter_vertical_edge_armv6(u_ptr, uv_stride, lfi->mbflim, lfi->lim, lfi->thr, 1);
+        vp8_mbloop_filter_vertical_edge_armv6(u_ptr, uv_stride, lfi->uvmbflim, lfi->uvlim, lfi->uvmbthr, 1);

    if (v_ptr)
-        vp8_mbloop_filter_vertical_edge_armv6(v_ptr, uv_stride, lfi->mbflim, lfi->lim, lfi->thr, 1);
+        vp8_mbloop_filter_vertical_edge_armv6(v_ptr, uv_stride, lfi->uvmbflim, lfi->uvlim, lfi->uvmbthr, 1);
 }

 void vp8_loop_filter_mbvs_armv6(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
@@ -81,7 +81,7 @@ void vp8_loop_filter_mbvs_armv6(unsigned char *y_ptr, unsigned char *u_ptr, unsi
    (void) v_ptr;
    (void) uv_stride;
    (void) simpler_lpf;
-    vp8_loop_filter_simple_vertical_edge_armv6(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->thr, 2);
+    vp8_loop_filter_simple_vertical_edge_armv6(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->mbthr, 2);
 }

 /* Horizontal B Filtering */
@@ -94,10 +94,10 @@ void vp8_loop_filter_bh_armv6(unsigned char *y_ptr, unsigned char *u_ptr, unsign
    vp8_loop_filter_horizontal_edge_armv6(y_ptr + 12 * y_stride, y_stride, lfi->flim, lfi->lim, lfi->thr, 2);

    if (u_ptr)
-        vp8_loop_filter_horizontal_edge_armv6(u_ptr + 4 * uv_stride, uv_stride, lfi->flim, lfi->lim, lfi->thr, 1);
+        vp8_loop_filter_horizontal_edge_armv6(u_ptr + 4 * uv_stride, uv_stride, lfi->uvflim, lfi->uvlim, lfi->uvthr, 1);

    if (v_ptr)
-        vp8_loop_filter_horizontal_edge_armv6(v_ptr + 4 * uv_stride, uv_stride, lfi->flim, lfi->lim, lfi->thr, 1);
+        vp8_loop_filter_horizontal_edge_armv6(v_ptr + 4 * uv_stride, uv_stride, lfi->uvflim, lfi->uvlim, lfi->uvthr, 1);
 }

 void vp8_loop_filter_bhs_armv6(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
@@ -122,10 +122,10 @@ void vp8_loop_filter_bv_armv6(unsigned char *y_ptr, unsigned char *u_ptr, unsign
    vp8_loop_filter_vertical_edge_armv6(y_ptr + 12, y_stride, lfi->flim, lfi->lim, lfi->thr, 2);

    if (u_ptr)
-        vp8_loop_filter_vertical_edge_armv6(u_ptr + 4, uv_stride, lfi->flim, lfi->lim, lfi->thr, 1);
+        vp8_loop_filter_vertical_edge_armv6(u_ptr + 4, uv_stride, lfi->uvflim, lfi->uvlim, lfi->uvthr, 1);

    if (v_ptr)
-        vp8_loop_filter_vertical_edge_armv6(v_ptr + 4, uv_stride, lfi->flim, lfi->lim, lfi->thr, 1);
+        vp8_loop_filter_vertical_edge_armv6(v_ptr + 4, uv_stride, lfi->uvflim, lfi->uvlim, lfi->uvthr, 1);
 }

 void vp8_loop_filter_bvs_armv6(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
@@ -148,10 +148,10 @@ void vp8_loop_filter_mbh_neon(unsigned char *y_ptr, unsigned char *u_ptr, unsign
                              int y_stride, int uv_stride, loop_filter_info *lfi, int simpler_lpf)
 {
    (void) simpler_lpf;
-    vp8_mbloop_filter_horizontal_edge_y_neon(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->thr, 2);
+    vp8_mbloop_filter_horizontal_edge_y_neon(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->mbthr, 2);

    if (u_ptr)
-        vp8_mbloop_filter_horizontal_edge_uv_neon(u_ptr, uv_stride, lfi->mbflim, lfi->lim, lfi->thr, v_ptr);
+        vp8_mbloop_filter_horizontal_edge_uv_neon(u_ptr, uv_stride, lfi->uvmbflim, lfi->uvlim, lfi->uvmbthr, v_ptr);
 }

 void vp8_loop_filter_mbhs_neon(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
@@ -161,7 +161,7 @@ void vp8_loop_filter_mbhs_neon(unsigned char *y_ptr, unsigned char *u_ptr, unsig
    (void) v_ptr;
    (void) uv_stride;
    (void) simpler_lpf;
-    vp8_loop_filter_simple_horizontal_edge_neon(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->thr, 2);
+    vp8_loop_filter_simple_horizontal_edge_neon(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->mbthr, 2);
 }

 /* Vertical MB Filtering */
@@ -169,10 +169,10 @@ void vp8_loop_filter_mbv_neon(unsigned char *y_ptr, unsigned char *u_ptr, unsign
                              int y_stride, int uv_stride, loop_filter_info *lfi, int simpler_lpf)
 {
    (void) simpler_lpf;
-    vp8_mbloop_filter_vertical_edge_y_neon(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->thr, 2);
+    vp8_mbloop_filter_vertical_edge_y_neon(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->mbthr, 2);

    if (u_ptr)
-        vp8_mbloop_filter_vertical_edge_uv_neon(u_ptr, uv_stride, lfi->mbflim, lfi->lim, lfi->thr, v_ptr);
+        vp8_mbloop_filter_vertical_edge_uv_neon(u_ptr, uv_stride, lfi->uvmbflim, lfi->uvlim, lfi->uvmbthr, v_ptr);
 }

 void vp8_loop_filter_mbvs_neon(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
@@ -182,7 +182,7 @@ void vp8_loop_filter_mbvs_neon(unsigned char *y_ptr, unsigned char *u_ptr, unsig
    (void) v_ptr;
    (void) uv_stride;
    (void) simpler_lpf;
-    vp8_loop_filter_simple_vertical_edge_neon(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->thr, 2);
+    vp8_loop_filter_simple_vertical_edge_neon(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->mbthr, 2);
 }

 /* Horizontal B Filtering */
@@ -195,7 +195,7 @@ void vp8_loop_filter_bh_neon(unsigned char *y_ptr, unsigned char *u_ptr, unsigne
    vp8_loop_filter_horizontal_edge_y_neon(y_ptr + 12 * y_stride, y_stride, lfi->flim, lfi->lim, lfi->thr, 2);

    if (u_ptr)
-        vp8_loop_filter_horizontal_edge_uv_neon(u_ptr + 4 * uv_stride, uv_stride, lfi->flim, lfi->lim, lfi->thr, v_ptr + 4 * uv_stride);
+        vp8_loop_filter_horizontal_edge_uv_neon(u_ptr + 4 * uv_stride, uv_stride, lfi->uvflim, lfi->uvlim, lfi->uvthr, v_ptr + 4 * uv_stride);
 }

 void vp8_loop_filter_bhs_neon(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
@@ -220,7 +220,7 @@ void vp8_loop_filter_bv_neon(unsigned char *y_ptr, unsigned char *u_ptr, unsigne
    vp8_loop_filter_vertical_edge_y_neon(y_ptr + 12, y_stride, lfi->flim, lfi->lim, lfi->thr, 2);

    if (u_ptr)
-        vp8_loop_filter_vertical_edge_uv_neon(u_ptr + 4, uv_stride, lfi->flim, lfi->lim, lfi->thr, v_ptr + 4);
+        vp8_loop_filter_vertical_edge_uv_neon(u_ptr + 4, uv_stride, lfi->uvflim, lfi->uvlim, lfi->uvthr, v_ptr + 4);
 }

 void vp8_loop_filter_bvs_neon(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
--- a/vp8/common/arm/neon/bilinearpredict16x16_neon.asm
+++ b/vp8/common/arm/neon/bilinearpredict16x16_neon.asm
@@ -350,7 +350,10 @@ filt_blk2d_spo16x16_loop_neon
    ENDP

 ;-----------------
-
+    AREA    bifilters16_dat, DATA, READWRITE            ;read/write by default
+;Data section with name data_area is specified. DCD reserves space in memory for 48 data.
+;One word each is reserved. Label filter_coeff can be used to access the data.
+;Data address: filter_coeff, filter_coeff+4, filter_coeff+8 ...
 _bifilter16_coeff_
    DCD     bifilter16_coeff
 bifilter16_coeff
--- a/vp8/common/arm/neon/bilinearpredict4x4_neon.asm
+++ b/vp8/common/arm/neon/bilinearpredict4x4_neon.asm
@@ -123,7 +123,10 @@ skip_secondpass_filter
    ENDP

 ;-----------------
-
+    AREA    bilinearfilters4_dat, DATA, READWRITE           ;read/write by default
+;Data section with name data_area is specified. DCD reserves space in memory for 48 data.
+;One word each is reserved. Label filter_coeff can be used to access the data.
+;Data address: filter_coeff, filter_coeff+4, filter_coeff+8 ...
 _bifilter4_coeff_
    DCD     bifilter4_coeff
 bifilter4_coeff
--- a/vp8/common/arm/neon/bilinearpredict8x4_neon.asm
+++ b/vp8/common/arm/neon/bilinearpredict8x4_neon.asm
@@ -128,7 +128,10 @@ skip_secondpass_filter
    ENDP

 ;-----------------
-
+    AREA    bifilters8x4_dat, DATA, READWRITE           ;read/write by default
+;Data section with name data_area is specified. DCD reserves space in memory for 48 data.
+;One word each is reserved. Label filter_coeff can be used to access the data.
+;Data address: filter_coeff, filter_coeff+4, filter_coeff+8 ...
 _bifilter8x4_coeff_
    DCD     bifilter8x4_coeff
 bifilter8x4_coeff
--- a/vp8/common/arm/neon/bilinearpredict8x8_neon.asm
+++ b/vp8/common/arm/neon/bilinearpredict8x8_neon.asm
@@ -176,7 +176,10 @@ skip_secondpass_filter
    ENDP

 ;-----------------
-
+    AREA    bifilters8_dat, DATA, READWRITE         ;read/write by default
+;Data section with name data_area is specified. DCD reserves space in memory for 48 data.
+;One word each is reserved. Label filter_coeff can be used to access the data.
+;Data address: filter_coeff, filter_coeff+4, filter_coeff+8 ...
 _bifilter8_coeff_
    DCD     bifilter8_coeff
 bifilter8_coeff
--- a/vp8/common/arm/neon/loopfilter_neon.asm
+++ b/vp8/common/arm/neon/loopfilter_neon.asm
@@ -397,8 +397,7 @@
    bx          lr
    ENDP        ; |vp8_loop_filter_horizontal_edge_y_neon|

-;-----------------
-
+    AREA    loopfilter_dat, DATA, READONLY
 _lf_coeff_
    DCD     lf_coeff
 lf_coeff
--- a/vp8/common/arm/neon/loopfiltersimplehorizontaledge_neon.asm
+++ b/vp8/common/arm/neon/loopfiltersimplehorizontaledge_neon.asm
@@ -104,7 +104,10 @@
    ENDP        ; |vp8_loop_filter_simple_horizontal_edge_neon|

 ;-----------------
-
+    AREA    hloopfiltery_dat, DATA, READWRITE           ;read/write by default
+;Data section with name data_area is specified. DCD reserves space in memory for 16 data.
+;One word each is reserved. Label filter_coeff can be used to access the data.
+;Data address: filter_coeff, filter_coeff+4, filter_coeff+8 ...
 _lfhy_coeff_
    DCD     lfhy_coeff
 lfhy_coeff
--- a/vp8/common/arm/neon/loopfiltersimpleverticaledge_neon.asm
+++ b/vp8/common/arm/neon/loopfiltersimpleverticaledge_neon.asm
@@ -145,7 +145,10 @@
    ENDP        ; |vp8_loop_filter_simple_vertical_edge_neon|

 ;-----------------
-
+    AREA    vloopfiltery_dat, DATA, READWRITE           ;read/write by default
+;Data section with name data_area is specified. DCD reserves space in memory for 16 data.
+;One word each is reserved. Label filter_coeff can be used to access the data.
+;Data address: filter_coeff, filter_coeff+4, filter_coeff+8 ...
 _vlfy_coeff_
    DCD     vlfy_coeff
 vlfy_coeff
--- a/vp8/common/arm/neon/mbloopfilter_neon.asm
+++ b/vp8/common/arm/neon/mbloopfilter_neon.asm
@@ -505,8 +505,7 @@
    bx          lr
    ENDP        ; |vp8_mbloop_filter_neon|

-;-----------------
-
+    AREA    mbloopfilter_dat, DATA, READONLY
 _mblf_coeff_
    DCD     mblf_coeff
 mblf_coeff
--- a/vp8/common/arm/neon/recon_neon.c
+++ b/vp8/common/arm/neon/recon_neon.c
@@ -10,8 +10,8 @@


 #include "vpx_ports/config.h"
-#include "vp8/common/recon.h"
-#include "vp8/common/blockd.h"
+#include "recon.h"
+#include "blockd.h"

 extern void vp8_recon16x16mb_neon(unsigned char *pred_ptr, short *diff_ptr, unsigned char *dst_ptr, int ystride, unsigned char *udst_ptr, unsigned char *vdst_ptr);

--- a/vp8/common/arm/neon/shortidct4x4llm_neon.asm
+++ b/vp8/common/arm/neon/shortidct4x4llm_neon.asm
@@ -113,7 +113,10 @@
    ENDP

 ;-----------------
-
+    AREA    idct4x4_dat, DATA, READWRITE            ;read/write by default
+;Data section with name data_area is specified. DCD reserves space in memory for 48 data.
+;One word each is reserved. Label filter_coeff can be used to access the data.
+;Data address: filter_coeff, filter_coeff+4, filter_coeff+8 ...
 _idct_coeff_
    DCD     idct_coeff
 idct_coeff
--- a/vp8/common/arm/neon/sixtappredict16x16_neon.asm
+++ b/vp8/common/arm/neon/sixtappredict16x16_neon.asm
@@ -476,7 +476,10 @@ secondpass_only_inner_loop_neon
    ENDP

 ;-----------------
-
+    AREA    subpelfilters16_dat, DATA, READWRITE            ;read/write by default
+;Data section with name data_area is specified. DCD reserves space in memory for 48 data.
+;One word each is reserved. Label filter_coeff can be used to access the data.
+;Data address: filter_coeff, filter_coeff+4, filter_coeff+8 ...
 _filter16_coeff_
    DCD     filter16_coeff
 filter16_coeff
--- a/vp8/common/arm/neon/sixtappredict4x4_neon.asm
+++ b/vp8/common/arm/neon/sixtappredict4x4_neon.asm
@@ -407,7 +407,10 @@ secondpass_filter4x4_only
    ENDP

 ;-----------------
-
+    AREA    subpelfilters4_dat, DATA, READWRITE         ;read/write by default
+;Data section with name data_area is specified. DCD reserves space in memory for 48 data.
+;One word each is reserved. Label filter_coeff can be used to access the data.
+;Data address: filter_coeff, filter_coeff+4, filter_coeff+8 ...
 _filter4_coeff_
    DCD     filter4_coeff
 filter4_coeff
--- a/vp8/common/arm/neon/sixtappredict8x4_neon.asm
+++ b/vp8/common/arm/neon/sixtappredict8x4_neon.asm
@@ -458,7 +458,10 @@ secondpass_filter8x4_only
    ENDP

 ;-----------------
-
+    AREA    subpelfilters8_dat, DATA, READWRITE         ;read/write by default
+;Data section with name data_area is specified. DCD reserves space in memory for 48 data.
+;One word each is reserved. Label filter_coeff can be used to access the data.
+;Data address: filter_coeff, filter_coeff+4, filter_coeff+8 ...
 _filter8_coeff_
    DCD     filter8_coeff
 filter8_coeff
--- a/vp8/common/arm/neon/sixtappredict8x8_neon.asm
+++ b/vp8/common/arm/neon/sixtappredict8x8_neon.asm
@@ -509,7 +509,10 @@ filt_blk2d_spo8x8_loop_neon
    ENDP

 ;-----------------
-
+    AREA    subpelfilters8_dat, DATA, READWRITE         ;read/write by default
+;Data section with name data_area is specified. DCD reserves space in memory for 48 data.
+;One word each is reserved. Label filter_coeff can be used to access the data.
+;Data address: filter_coeff, filter_coeff+4, filter_coeff+8 ...
 _filter8_coeff_
    DCD     filter8_coeff
 filter8_coeff
--- a/vp8/common/arm/reconintra_arm.c
+++ b/vp8/common/arm/reconintra_arm.c
@@ -10,10 +10,10 @@


 #include "vpx_ports/config.h"
-#include "vp8/common/blockd.h"
-#include "vp8/common/reconintra.h"
+#include "blockd.h"
+#include "reconintra.h"
 #include "vpx_mem/vpx_mem.h"
-#include "vp8/common/recon.h"
+#include "recon.h"

 #if HAVE_ARMV7
 extern void vp8_build_intra_predictors_mby_neon_func(
--- a/vp8/common/arm/vpx_asm_offsets.c
+++ b/vp8/common/arm/vpx_asm_offsets.c
@@ -1,5 +1,5 @@
 /*
- *  Copyright (c) 2011 The WebM project authors. All Rights Reserved.
+ *  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
@@ -12,7 +12,13 @@
 #include "vpx_ports/config.h"
 #include <stddef.h>

+#if CONFIG_VP8_ENCODER
+#include "vpx_scale/yv12config.h"
+#endif
+
+#if CONFIG_VP8_DECODER
 #include "onyxd_int.h"
+#endif

 #define DEFINE(sym, val) int sym = val;

@@ -25,6 +31,29 @@
 * {
 */

+#if CONFIG_VP8_DECODER || CONFIG_VP8_ENCODER
+DEFINE(yv12_buffer_config_y_width,              offsetof(YV12_BUFFER_CONFIG, y_width));
+DEFINE(yv12_buffer_config_y_height,             offsetof(YV12_BUFFER_CONFIG, y_height));
+DEFINE(yv12_buffer_config_y_stride,             offsetof(YV12_BUFFER_CONFIG, y_stride));
+DEFINE(yv12_buffer_config_uv_width,             offsetof(YV12_BUFFER_CONFIG, uv_width));
+DEFINE(yv12_buffer_config_uv_height,            offsetof(YV12_BUFFER_CONFIG, uv_height));
+DEFINE(yv12_buffer_config_uv_stride,            offsetof(YV12_BUFFER_CONFIG, uv_stride));
+DEFINE(yv12_buffer_config_y_buffer,             offsetof(YV12_BUFFER_CONFIG, y_buffer));
+DEFINE(yv12_buffer_config_u_buffer,             offsetof(YV12_BUFFER_CONFIG, u_buffer));
+DEFINE(yv12_buffer_config_v_buffer,             offsetof(YV12_BUFFER_CONFIG, v_buffer));
+DEFINE(yv12_buffer_config_border,               offsetof(YV12_BUFFER_CONFIG, border));
+#endif
+
+#if CONFIG_VP8_DECODER
+DEFINE(mb_diff,                                 offsetof(MACROBLOCKD, diff));
+DEFINE(mb_predictor,                            offsetof(MACROBLOCKD, predictor));
+DEFINE(mb_dst_y_stride,                         offsetof(MACROBLOCKD, dst.y_stride));
+DEFINE(mb_dst_y_buffer,                         offsetof(MACROBLOCKD, dst.y_buffer));
+DEFINE(mb_dst_u_buffer,                         offsetof(MACROBLOCKD, dst.u_buffer));
+DEFINE(mb_dst_v_buffer,                         offsetof(MACROBLOCKD, dst.v_buffer));
+DEFINE(mb_up_available,                         offsetof(MACROBLOCKD, up_available));
+DEFINE(mb_left_available,                       offsetof(MACROBLOCKD, left_available));
+
 DEFINE(detok_scan,                              offsetof(DETOK, scan));
 DEFINE(detok_ptr_block2leftabove,               offsetof(DETOK, ptr_block2leftabove));
 DEFINE(detok_coef_tree_ptr,                     offsetof(DETOK, vp8_coef_tree_ptr));
@@ -48,6 +77,7 @@ DEFINE(bool_decoder_range,                      offsetof(BOOL_DECODER, range));

 DEFINE(tokenextrabits_min_val,                  offsetof(TOKENEXTRABITS, min_val));
 DEFINE(tokenextrabits_length,                   offsetof(TOKENEXTRABITS, Length));
+#endif

 //add asserts for any offset that is not supported by assembly code
 //add asserts for any size that is not supported by assembly code
--- a/vp8/common/asm_com_offsets.c
+++ b/vp8/common/asm_com_offsets.c
@@ -1,49 +0,0 @@
-/*
- *  Copyright (c) 2011 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_ports/config.h"
-#include <stddef.h>
-
-#include "vpx_scale/yv12config.h"
-
-#define ct_assert(name,cond) \
-    static void assert_##name(void) UNUSED;\
-    static void assert_##name(void) {switch(0){case 0:case !!(cond):;}}
-
-#define DEFINE(sym, val) int sym = val;
-
-/*
-#define BLANK() asm volatile("\n->" : : )
-*/
-
-/*
- * int main(void)
- * {
- */
-
-//vpx_scale
-DEFINE(yv12_buffer_config_y_width,              offsetof(YV12_BUFFER_CONFIG, y_width));
-DEFINE(yv12_buffer_config_y_height,             offsetof(YV12_BUFFER_CONFIG, y_height));
-DEFINE(yv12_buffer_config_y_stride,             offsetof(YV12_BUFFER_CONFIG, y_stride));
-DEFINE(yv12_buffer_config_uv_width,             offsetof(YV12_BUFFER_CONFIG, uv_width));
-DEFINE(yv12_buffer_config_uv_height,            offsetof(YV12_BUFFER_CONFIG, uv_height));
-DEFINE(yv12_buffer_config_uv_stride,            offsetof(YV12_BUFFER_CONFIG, uv_stride));
-DEFINE(yv12_buffer_config_y_buffer,             offsetof(YV12_BUFFER_CONFIG, y_buffer));
-DEFINE(yv12_buffer_config_u_buffer,             offsetof(YV12_BUFFER_CONFIG, u_buffer));
-DEFINE(yv12_buffer_config_v_buffer,             offsetof(YV12_BUFFER_CONFIG, v_buffer));
-DEFINE(yv12_buffer_config_border,               offsetof(YV12_BUFFER_CONFIG, border));
-
-//add asserts for any offset that is not supported by assembly code
-//add asserts for any size that is not supported by assembly code
-/*
- * return 0;
- * }
- */
--- a/vp8/common/blockd.c
+++ b/vp8/common/blockd.c
@@ -12,6 +12,8 @@
 #include "blockd.h"
 #include "vpx_mem/vpx_mem.h"

+const int vp8_block2type[25] = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2, 2, 2, 2, 2, 1};
+
 const unsigned char vp8_block2left[25] =
 {
    0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 5, 5, 6, 6, 7, 7, 8
--- a/vp8/common/blockd.h
+++ b/vp8/common/blockd.h
@@ -28,6 +28,11 @@ void vpx_log(const char *format, ...);
 #define DCPREDSIMTHRESH 0
 #define DCPREDCNTTHRESH 3

+#define Y1CONTEXT 0
+#define UCONTEXT 1
+#define VCONTEXT 2
+#define Y2CONTEXT 3
+
 #define MB_FEATURE_TREE_PROBS   3
 #define MAX_MB_SEGMENTS         4

@@ -43,11 +48,6 @@ typedef struct
    int r, c;
 } POS;

-#define PLANE_TYPE_Y_NO_DC    0
-#define PLANE_TYPE_Y2         1
-#define PLANE_TYPE_UV         2
-#define PLANE_TYPE_Y_WITH_DC  3
-

 typedef char ENTROPY_CONTEXT;
 typedef struct
@@ -58,6 +58,8 @@ typedef struct
    ENTROPY_CONTEXT y2;
 } ENTROPY_CONTEXT_PLANES;

+extern const int vp8_block2type[25];
+
 extern const unsigned char vp8_block2left[25];
 extern const unsigned char vp8_block2above[25];

--- a/vp8/common/boolcoder.h
+++ b/vp8/common/boolcoder.h
@@ -0,0 +1,570 @@
+/*
+ *  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 bool_coder_h
+#define bool_coder_h 1
+
+/* Arithmetic bool coder with largish probability range.
+   Timothy S Murphy  6 August 2004 */
+
+/* So as not to force users to drag in too much of my idiosyncratic C++ world,
+   I avoid fancy storage management. */
+
+#include <assert.h>
+
+#include <stddef.h>
+#include <stdio.h>
+
+typedef unsigned char vp8bc_index_t; // probability index
+
+/* There are a couple of slight variants in the details of finite-precision
+   arithmetic coding.  May be safely ignored by most users. */
+
+enum vp8bc_rounding
+{
+    vp8bc_down = 0,     // just like VP8
+    vp8bc_down_full = 1, // handles minimum probability correctly
+    vp8bc_up = 2
+};
+
+#if _MSC_VER
+
+/* Note that msvc by default does not inline _anything_ (regardless of the
+   setting of inline_depth) and that a command-line option (-Ob1 or -Ob2)
+   is required to inline even the smallest functions. */
+
+#   pragma inline_depth( 255)           // I mean it when I inline something
+#   pragma warning( disable : 4099)     // No class vs. struct harassment
+#   pragma warning( disable : 4250)     // dominance complaints
+#   pragma warning( disable : 4284)     // operator-> in templates
+#   pragma warning( disable : 4800)     // bool conversion
+
+// don't let prefix ++,-- stand in for postfix, disaster would ensue
+
+#   pragma warning( error : 4620 4621)
+
+#endif  // _MSC_VER
+
+
+#if __cplusplus
+
+// Sometimes one wishes to be definite about integer lengths.
+
+struct int_types
+{
+    typedef const bool cbool;
+    typedef const signed char cchar;
+    typedef const short cshort;
+    typedef const int cint;
+    typedef const int clong;
+
+    typedef const double cdouble;
+    typedef const size_t csize_t;
+
+    typedef unsigned char uchar;    // 8 bits
+    typedef const uchar cuchar;
+
+    typedef short int16;
+    typedef unsigned short uint16;
+    typedef const int16 cint16;
+    typedef const uint16 cuint16;
+
+    typedef int int32;
+    typedef unsigned int uint32;
+    typedef const int32 cint32;
+    typedef const uint32 cuint32;
+
+    typedef unsigned int uint;
+    typedef unsigned int ulong;
+    typedef const uint cuint;
+    typedef const ulong culong;
+
+
+    // All structs consume space, may as well have a vptr.
+
+    virtual ~int_types();
+};
+
+
+struct bool_coder_spec;
+struct bool_coder;
+struct bool_writer;
+struct bool_reader;
+
+
+struct bool_coder_namespace : int_types
+{
+    typedef vp8bc_index_t Index;
+    typedef bool_coder_spec Spec;
+    typedef const Spec c_spec;
+
+    enum Rounding
+    {
+        Down = vp8bc_down,
+        down_full = vp8bc_down_full,
+        Up = vp8bc_up
+    };
+};
+
+
+// Archivable specification of a bool coder includes rounding spec
+// and probability mapping table.  The latter replaces a uchar j
+// (0 <= j < 256) with an arbitrary uint16 tbl[j] = p.
+// p/65536 is then the probability of a zero.
+
+struct bool_coder_spec : bool_coder_namespace
+{
+    friend struct bool_coder;
+    friend struct bool_writer;
+    friend struct bool_reader;
+    friend struct bool_coder_spec_float;
+    friend struct bool_coder_spec_explicit_table;
+    friend struct bool_coder_spec_exponential_table;
+    friend struct BPsrc;
+private:
+    uint w;                 // precision
+    Rounding r;
+
+    uint ebits, mbits, ebias;
+    uint32 mmask;
+
+    Index max_index, half_index;
+
+    uint32 mantissa(Index i) const
+    {
+        assert(i < half_index);
+        return (1 << mbits) + (i & mmask);
+    }
+    uint exponent(Index i) const
+    {
+        assert(i < half_index);
+        return ebias - (i >> mbits);
+    }
+
+    uint16 Ptbl[256];       // kinda clunky, but so is storage management.
+
+    /* Cost in bits of encoding a zero at every probability, scaled by 2^20.
+       Assumes that index is at most 8 bits wide. */
+
+    uint32 Ctbl[256];
+
+    uint32 split(Index i, uint32 R) const    // 1 <= split <= max( 1, R-1)
+    {
+        if (!ebias)
+            return 1 + (((R - 1) * Ptbl[i]) >> 16);
+
+        if (i >= half_index)
+            return R - split(max_index - i, R);
+
+        return 1 + (((R - 1) * mantissa(i)) >> exponent(i));
+    }
+
+    uint32 max_range() const
+    {
+        return (1 << w) - (r == down_full ? 0 : 1);
+    }
+    uint32 min_range() const
+    {
+        return (1 << (w - 1)) + (r == down_full ? 1 : 0);
+    }
+    uint32 Rinc() const
+    {
+        return r == Up ? 1 : 0;
+    }
+
+    void check_prec() const;
+
+    bool float_init(uint Ebits, uint Mbits);
+
+    void cost_init();
+
+    bool_coder_spec(
+        uint prec, Rounding rr, uint Ebits = 0, uint Mbits = 0
+    )
+        : w(prec), r(rr)
+    {
+        float_init(Ebits, Mbits);
+    }
+public:
+    // Read complete spec from file.
+    bool_coder_spec(FILE *);
+
+    // Write spec to file.
+    void dump(FILE *) const;
+
+    // return probability index best approximating prob.
+    Index operator()(double prob) const;
+
+    // probability corresponding to index
+    double operator()(Index i) const;
+
+    Index complement(Index i) const
+    {
+        return max_index - i;
+    }
+
+    Index max_index() const
+    {
+        return max_index;
+    }
+    Index half_index() const
+    {
+        return half_index;
+    }
+
+    uint32 cost_zero(Index i) const
+    {
+        return Ctbl[i];
+    }
+    uint32 cost_one(Index i) const
+    {
+        return Ctbl[ max_index - i];
+    }
+    uint32 cost_bit(Index i, bool b) const
+    {
+        return Ctbl[b? max_index-i:i];
+    }
+};
+
+
+/* Pseudo floating-point probability specification.
+
+   At least one of Ebits and Mbits must be nonzero.
+
+   Since all arithmetic is done at 32 bits, Ebits is at most 5.
+
+   Total significant bits in index is Ebits + Mbits + 1.
+
+   Below the halfway point (i.e. when the top significant bit is 0),
+   the index is (e << Mbits) + m.
+
+   The exponent e is between 0 and (2**Ebits) - 1,
+   the mantissa m is between 0 and (2**Mbits) - 1.
+
+   Prepending an implicit 1 to the mantissa, the probability is then
+
+        (2**Mbits + m) >> (e - 2**Ebits - 1 - Mbits),
+
+   which has (1/2)**(2**Ebits + 1) as a minimum
+   and (1/2) * [1 - 2**(Mbits + 1)] as a maximum.
+
+   When the index is above the halfway point, the probability is the
+   complement of the probability associated to the complement of the index.
+
+   Note that the probability increases with the index and that, because of
+   the symmetry, we cannot encode probability exactly 1/2; though we
+   can get as close to 1/2 as we like, provided we have enough Mbits.
+
+   The latter is of course not a problem in practice, one never has
+   exact probabilities and entropy errors are second order, that is, the
+   "overcoding" of a zero will be largely compensated for by the
+   "undercoding" of a one (or vice-versa).
+
+   Compared to arithmetic probability specs (a la VP8), this will do better
+   at very high and low probabilities and worse at probabilities near 1/2,
+   as well as facilitating the usage of wider or narrower probability indices.
+*/
+
+struct bool_coder_spec_float : bool_coder_spec
+{
+    bool_coder_spec_float(
+        uint Ebits = 3, uint Mbits = 4, Rounding rr = down_full, uint prec = 12
+    )
+        : bool_coder_spec(prec, rr, Ebits, Mbits)
+    {
+        cost_init();
+    }
+};
+
+
+struct bool_coder_spec_explicit_table : bool_coder_spec
+{
+    bool_coder_spec_explicit_table(
+        cuint16 probability_table[256] = 0,  // default is tbl[i] = i << 8.
+        Rounding = down_full,
+        uint precision = 16
+    );
+};
+
+// Contruct table via multiplicative interpolation between
+// p[128] = 1/2  and p[0] = (1/2)^x.
+// Since we are working with 16-bit precision, x is at most 16.
+// For probabilities to increase with i, we must have x > 1.
+// For 0 <= i <= 128, p[i] = (1/2)^{ 1 + [1 - (i/128)]*[x-1] }.
+// Finally, p[128+i] = 1 - p[128 - i].
+
+struct bool_coder_spec_exponential_table : bool_coder_spec
+{
+    bool_coder_spec_exponential_table(uint x, Rounding = down_full, uint prec = 16);
+};
+
+
+// Commonalities between writer and reader.
+
+struct bool_coder : bool_coder_namespace
+{
+    friend struct bool_writer;
+    friend struct bool_reader;
+    friend struct BPsrc;
+private:
+    uint32 Low, Range;
+    cuint32 min_range;
+    cuint32 rinc;
+    c_spec spec;
+
+    void _reset()
+    {
+        Low = 0;
+        Range = spec.max_range();
+    }
+
+    bool_coder(c_spec &s)
+        :  min_range(s.min_range()),
+           rinc(s.Rinc()),
+           spec(s)
+    {
+        _reset();
+    }
+
+    uint32 half() const
+    {
+        return 1 + ((Range - 1) >> 1);
+    }
+public:
+    c_spec &Spec() const
+    {
+        return spec;
+    }
+};
+
+
+struct bool_writer : bool_coder
+{
+    friend struct BPsrc;
+private:
+    uchar *Bstart, *Bend, *B;
+    int bit_lag;
+    bool is_toast;
+    void carry();
+    void reset()
+    {
+        _reset();
+        bit_lag = 32 - spec.w;
+        is_toast = 0;
+    }
+    void raw(bool value, uint32 split);
+public:
+    bool_writer(c_spec &, uchar *Dest, size_t Len);
+    virtual ~bool_writer();
+
+    void operator()(Index p, bool v)
+    {
+        raw(v, spec.split(p, Range));
+    }
+
+    uchar *buf() const
+    {
+        return Bstart;
+    }
+    size_t bytes_written() const
+    {
+        return B - Bstart;
+    }
+
+    // Call when done with input, flushes internal state.
+    // DO NOT write any more data after calling this.
+
+    bool_writer &flush();
+
+    void write_bits(int n, uint val)
+    {
+        if (n)
+        {
+            uint m = 1 << (n - 1);
+
+            do
+            {
+                raw((bool)(val & m), half());
+            }
+            while (m >>= 1);
+        }
+    }
+
+#   if 0
+    // We are agnostic about storage management.
+    // By default, overflows throw an assert but user can
+    // override to provide an expanding buffer using ...
+
+    virtual void overflow(uint Len) const;
+
+    // ... this function copies already-written data into new buffer
+    // and retains new buffer location.
+
+    void new_buffer(uchar *dest, uint Len);
+
+    // Note that storage management is the user's responsibility.
+#   endif
+};
+
+
+// This could be adjusted to use a little less lookahead.
+
+struct bool_reader : bool_coder
+{
+    friend struct BPsrc;
+private:
+    cuchar *const Bstart;   // for debugging
+    cuchar *B;
+    cuchar *const Bend;
+    cuint shf;
+    uint bct;
+    bool raw(uint32 split);
+public:
+    bool_reader(c_spec &s, cuchar *src, size_t Len);
+
+    bool operator()(Index p)
+    {
+        return raw(spec.split(p, Range));
+    }
+
+    uint read_bits(int num_bits)
+    {
+        uint v = 0;
+
+        while (--num_bits >= 0)
+            v += v + (raw(half()) ? 1 : 0);
+
+        return v;
+    }
+};
+
+extern "C" {
+
+#endif /*  __cplusplus */
+
+
+    /* C interface */
+
+    typedef struct bool_coder_spec bool_coder_spec;
+    typedef struct bool_writer bool_writer;
+    typedef struct bool_reader bool_reader;
+
+    typedef const bool_coder_spec c_bool_coder_spec;
+    typedef const bool_writer c_bool_writer;
+    typedef const bool_reader c_bool_reader;
+
+
+    /* Optionally override default precision when constructing coder_specs.
+       Just pass a zero pointer if you don't care.
+       Precision is at most 16 bits for table specs, at most 23 otherwise. */
+
+    struct vp8bc_prec
+    {
+        enum vp8bc_rounding r;      /* see top header file for def */
+        unsigned int prec;          /* range precision in bits */
+    };
+
+    typedef const struct vp8bc_prec vp8bc_c_prec;
+
+    /* bool_coder_spec contains mapping of uchars to actual probabilities
+       (16 bit uints) as well as (usually immaterial) selection of
+       exact finite-precision algorithm used (for now, the latter can only
+       be overridden using the C++ interface).
+       See comments above the corresponding C++ constructors for discussion,
+       especially of exponential probability table generation. */
+
+    bool_coder_spec *vp8bc_vp8spec(); // just like vp8
+
+    bool_coder_spec *vp8bc_literal_spec(
+        const unsigned short prob_map[256],  // 0 is like vp8 w/more precision
+        vp8bc_c_prec*
+    );
+
+    bool_coder_spec *vp8bc_float_spec(
+        unsigned int exponent_bits, unsigned int mantissa_bits, vp8bc_c_prec*
+    );
+
+    bool_coder_spec *vp8bc_exponential_spec(unsigned int min_exp, vp8bc_c_prec *);
+
+    bool_coder_spec *vp8bc_spec_from_file(FILE *);
+
+
+    void vp8bc_destroy_spec(c_bool_coder_spec *);
+
+    void vp8bc_spec_to_file(c_bool_coder_spec *, FILE *);
+
+
+    /* Nearest index to supplied probability of zero, 0 <= prob <= 1. */
+
+    vp8bc_index_t vp8bc_index(c_bool_coder_spec *, double prob);
+
+    vp8bc_index_t vp8bc_index_from_counts(
+        c_bool_coder_spec *p, unsigned int zero_ct, unsigned int one_ct
+    );
+
+    /* In case you want to look */
+
+    double vp8bc_probability(c_bool_coder_spec *, vp8bc_index_t);
+
+    /* Opposite index */
+
+    vp8bc_index_t vp8bc_complement(c_bool_coder_spec *, vp8bc_index_t);
+
+    /* Cost in bits of encoding a zero at given probability, scaled by 2^20.
+       (assumes that an int holds at least 32 bits). */
+
+    unsigned int vp8bc_cost_zero(c_bool_coder_spec *, vp8bc_index_t);
+
+    unsigned int vp8bc_cost_one(c_bool_coder_spec *, vp8bc_index_t);
+    unsigned int vp8bc_cost_bit(c_bool_coder_spec *, vp8bc_index_t, int);
+
+
+    /* bool_writer interface */
+
+    /* Length = 0 disables checking for writes beyond buffer end. */
+
+    bool_writer *vp8bc_create_writer(
+        c_bool_coder_spec *, unsigned char *Destination, size_t Length
+    );
+
+    /* Flushes out any buffered data and returns total # of bytes written. */
+
+    size_t vp8bc_destroy_writer(bool_writer *);
+
+    void vp8bc_write_bool(bool_writer *, int boolean_val, vp8bc_index_t false_prob);
+
+    void vp8bc_write_bits(
+        bool_writer *, unsigned int integer_value, int number_of_bits
+    );
+
+    c_bool_coder_spec *vp8bc_writer_spec(c_bool_writer *);
+
+
+    /* bool_reader interface */
+
+    /* Length = 0 disables checking for reads beyond buffer end. */
+
+    bool_reader *vp8bc_create_reader(
+        c_bool_coder_spec *, const unsigned char *Source, size_t Length
+    );
+    void vp8bc_destroy_reader(bool_reader *);
+
+    int vp8bc_read_bool(bool_reader *, vp8bc_index_t false_prob);
+
+    unsigned int vp8bc_read_bits(bool_reader *, int number_of_bits);
+
+    c_bool_coder_spec *vp8bc_reader_spec(c_bool_reader *);
+
+#if __cplusplus
+}
+#endif
+
+#endif  /* bool_coder_h */
--- a/vp8/common/codec_common_interface.h
+++ b/vp8/common/codec_common_interface.h
@@ -0,0 +1,93 @@
+/*
+ *  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 CODEC_COMMON_INTERFACE_H
+#define CODEC_COMMON_INTERFACE_H
+
+#define __export
+#define _export
+#define dll_export   __declspec( dllexport )
+#define dll_import   __declspec( dllimport )
+
+// Playback ERROR Codes.
+#define NO_DECODER_ERROR            0
+#define REMOTE_DECODER_ERROR        -1
+
+#define DFR_BAD_DCT_COEFF           -100
+#define DFR_ZERO_LENGTH_FRAME       -101
+#define DFR_FRAME_SIZE_INVALID      -102
+#define DFR_OUTPUT_BUFFER_OVERFLOW  -103
+#define DFR_INVALID_FRAME_HEADER    -104
+#define FR_INVALID_MODE_TOKEN       -110
+#define ETR_ALLOCATION_ERROR        -200
+#define ETR_INVALID_ROOT_PTR        -201
+#define SYNCH_ERROR                 -400
+#define BUFFER_UNDERFLOW_ERROR      -500
+#define PB_IB_OVERFLOW_ERROR        -501
+
+// External error triggers
+#define PB_HEADER_CHECKSUM_ERROR    -601
+#define PB_DATA_CHECKSUM_ERROR      -602
+
+// DCT Error Codes
+#define DDCT_EXPANSION_ERROR        -700
+#define DDCT_INVALID_TOKEN_ERROR    -701
+
+// exception_errors
+#define GEN_EXCEPTIONS              -800
+#define EX_UNQUAL_ERROR             -801
+
+// Unrecoverable error codes
+#define FATAL_PLAYBACK_ERROR        -1000
+#define GEN_ERROR_CREATING_CDC      -1001
+#define GEN_THREAD_CREATION_ERROR   -1002
+#define DFR_CREATE_BMP_FAILED       -1003
+
+// YUV buffer configuration structure
+typedef struct
+{
+    int     y_width;
+    int     y_height;
+    int     y_stride;
+
+    int     uv_width;
+    int     uv_height;
+    int     uv_stride;
+
+    unsigned char   *y_buffer;
+    unsigned char   *u_buffer;
+    unsigned char   *v_buffer;
+
+} YUV_BUFFER_CONFIG;
+typedef enum
+{
+    C_SET_KEY_FRAME,
+    C_SET_FIXED_Q,
+    C_SET_FIRSTPASS_FILE,
+    C_SET_EXPERIMENTAL_MIN,
+    C_SET_EXPERIMENTAL_MAX = C_SET_EXPERIMENTAL_MIN + 255,
+    C_SET_CHECKPROTECT,
+    C_SET_TESTMODE,
+    C_SET_INTERNAL_SIZE,
+    C_SET_RECOVERY_FRAME,
+    C_SET_REFERENCEFRAME,
+    C_SET_GOLDENFRAME
+
+#ifndef VP50_COMP_INTERFACE
+    // Specialist test facilities.
+//    C_VCAP_PARAMS,              // DO NOT USE FOR NOW WITH VFW CODEC
+#endif
+
+} C_SETTING;
+
+typedef unsigned long C_SET_VALUE;
+
+
+#endif
--- a/vp8/common/filter_c.c
+++ b/vp8/common/filter_c.c
@@ -10,10 +10,13 @@


 #include <stdlib.h>
-#include "filter.h"
-#include "vpx_ports/mem.h"

-DECLARE_ALIGNED(16, const short, vp8_bilinear_filters[8][2]) =
+#define BLOCK_HEIGHT_WIDTH 4
+#define VP8_FILTER_WEIGHT 128
+#define VP8_FILTER_SHIFT  7
+
+
+static const int bilinear_filters[8][2] =
 {
    { 128,   0 },
    { 112,  16 },
@@ -25,7 +28,8 @@ DECLARE_ALIGNED(16, const short, vp8_bilinear_filters[8][2]) =
    {  16, 112 }
 };

-DECLARE_ALIGNED(16, const short, vp8_sub_pel_filters[8][6]) =
+
+static const short sub_pel_filters[8][6] =
 {

    { 0,  0,  128,    0,   0,  0 },         /* note that 1/8 pel positions are just as per alpha -0.5 bicubic */
@@ -36,6 +40,9 @@ DECLARE_ALIGNED(16, const short, vp8_sub_pel_filters[8][6]) =
    { 0, -6,   50,   93,  -9,  0 },
    { 1, -8,   36,  108, -11,  2 },         /* New 1/4 pel 6 tap filter */
    { 0, -1,   12,  123,  -6,  0 },
+
+
+
 };

 void vp8_filter_block2d_first_pass
@@ -139,7 +146,7 @@ void vp8_filter_block2d
    const short  *VFilter
 )
 {
-    int FData[9*4]; /* Temp data buffer used in filtering */
+    int FData[9*4]; /* Temp data bufffer used in filtering */

    /* First filter 1-D horizontally... */
    vp8_filter_block2d_first_pass(src_ptr - (2 * src_pixels_per_line), FData, src_pixels_per_line, 1, 9, 4, HFilter);
@@ -188,8 +195,8 @@ void vp8_sixtap_predict_c
    const short  *HFilter;
    const short  *VFilter;

-    HFilter = vp8_sub_pel_filters[xoffset];   /* 6 tap */
-    VFilter = vp8_sub_pel_filters[yoffset];   /* 6 tap */
+    HFilter = sub_pel_filters[xoffset];   /* 6 tap */
+    VFilter = sub_pel_filters[yoffset];   /* 6 tap */

    vp8_filter_block2d(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter, VFilter);
 }
@@ -205,10 +212,10 @@ void vp8_sixtap_predict8x8_c
 {
    const short  *HFilter;
    const short  *VFilter;
-    int FData[13*16];   /* Temp data buffer used in filtering */
+    int FData[13*16];   /* Temp data bufffer used in filtering */

-    HFilter = vp8_sub_pel_filters[xoffset];   /* 6 tap */
-    VFilter = vp8_sub_pel_filters[yoffset];   /* 6 tap */
+    HFilter = sub_pel_filters[xoffset];   /* 6 tap */
+    VFilter = sub_pel_filters[yoffset];   /* 6 tap */

    /* First filter 1-D horizontally... */
    vp8_filter_block2d_first_pass(src_ptr - (2 * src_pixels_per_line), FData, src_pixels_per_line, 1, 13, 8, HFilter);
@@ -231,10 +238,10 @@ void vp8_sixtap_predict8x4_c
 {
    const short  *HFilter;
    const short  *VFilter;
-    int FData[13*16];   /* Temp data buffer used in filtering */
+    int FData[13*16];   /* Temp data bufffer used in filtering */

-    HFilter = vp8_sub_pel_filters[xoffset];   /* 6 tap */
-    VFilter = vp8_sub_pel_filters[yoffset];   /* 6 tap */
+    HFilter = sub_pel_filters[xoffset];   /* 6 tap */
+    VFilter = sub_pel_filters[yoffset];   /* 6 tap */

    /* First filter 1-D horizontally... */
    vp8_filter_block2d_first_pass(src_ptr - (2 * src_pixels_per_line), FData, src_pixels_per_line, 1, 9, 8, HFilter);
@@ -257,11 +264,11 @@ void vp8_sixtap_predict16x16_c
 {
    const short  *HFilter;
    const short  *VFilter;
-    int FData[21*24];   /* Temp data buffer used in filtering */
+    int FData[21*24];   /* Temp data bufffer used in filtering */


-    HFilter = vp8_sub_pel_filters[xoffset];   /* 6 tap */
-    VFilter = vp8_sub_pel_filters[yoffset];   /* 6 tap */
+    HFilter = sub_pel_filters[xoffset];   /* 6 tap */
+    VFilter = sub_pel_filters[yoffset];   /* 6 tap */

    /* First filter 1-D horizontally... */
    vp8_filter_block2d_first_pass(src_ptr - (2 * src_pixels_per_line), FData, src_pixels_per_line, 1, 21, 16, HFilter);
@@ -276,50 +283,57 @@ void vp8_sixtap_predict16x16_c
 *
 *  ROUTINE       : filter_block2d_bil_first_pass
 *
- *  INPUTS        : UINT8  *src_ptr    : Pointer to source block.
- *                  UINT32  src_stride : Stride of source block.
- *                  UINT32  height     : Block height.
- *                  UINT32  width      : Block width.
- *                  INT32  *vp8_filter : Array of 2 bi-linear filter taps.
+ *  INPUTS        : UINT8  *src_ptr          : Pointer to source block.
+ *                  UINT32 src_pixels_per_line : Stride of input block.
+ *                  UINT32 pixel_step        : Offset between filter input samples (see notes).
+ *                  UINT32 output_height     : Input block height.
+ *                  UINT32 output_width      : Input block width.
+ *                  INT32  *vp8_filter          : Array of 2 bi-linear filter taps.
 *
- *  OUTPUTS       : INT32  *dst_ptr    : Pointer to filtered block.
+ *  OUTPUTS       : INT32 *output_ptr        : Pointer to filtered block.
 *
 *  RETURNS       : void
 *
- *  FUNCTION      : Applies a 1-D 2-tap bi-linear filter to the source block
- *                  in the horizontal direction to produce the filtered output
- *                  block. Used to implement first-pass of 2-D separable filter.
+ *  FUNCTION      : Applies a 1-D 2-tap bi-linear filter to the source block in
+ *                  either horizontal or vertical direction to produce the
+ *                  filtered output block. Used to implement first-pass
+ *                  of 2-D separable filter.
 *
 *  SPECIAL NOTES : Produces INT32 output to retain precision for next pass.
 *                  Two filter taps should sum to VP8_FILTER_WEIGHT.
+ *                  pixel_step defines whether the filter is applied
+ *                  horizontally (pixel_step=1) or vertically (pixel_step=stride).
+ *                  It defines the offset required to move from one input
+ *                  to the next.
 *
 ****************************************************************************/
 void vp8_filter_block2d_bil_first_pass
 (
-    unsigned char  *src_ptr,
-    unsigned short *dst_ptr,
-    unsigned int    src_stride,
-    unsigned int    height,
-    unsigned int    width,
-    const short    *vp8_filter
+    unsigned char *src_ptr,
+    unsigned short *output_ptr,
+    unsigned int src_pixels_per_line,
+    int pixel_step,
+    unsigned int output_height,
+    unsigned int output_width,
+    const int *vp8_filter
 )
 {
    unsigned int i, j;

-    for (i = 0; i < height; i++)
+    for (i = 0; i < output_height; i++)
    {
-        for (j = 0; j < width; j++)
+        for (j = 0; j < output_width; j++)
        {
            /* Apply bilinear filter */
-            dst_ptr[j] = (((int)src_ptr[0] * vp8_filter[0]) +
-                          ((int)src_ptr[1] * vp8_filter[1]) +
-                          (VP8_FILTER_WEIGHT / 2)) >> VP8_FILTER_SHIFT;
+            output_ptr[j] = (((int)src_ptr[0]          * vp8_filter[0]) +
+                             ((int)src_ptr[pixel_step] * vp8_filter[1]) +
+                             (VP8_FILTER_WEIGHT / 2)) >> VP8_FILTER_SHIFT;
            src_ptr++;
        }

        /* Next row... */
-        src_ptr += src_stride - width;
-        dst_ptr += width;
+        src_ptr    += src_pixels_per_line - output_width;
+        output_ptr += output_width;
    }
 }

@@ -327,51 +341,60 @@ void vp8_filter_block2d_bil_first_pass
 *
 *  ROUTINE       : filter_block2d_bil_second_pass
 *
- *  INPUTS        : INT32  *src_ptr    : Pointer to source block.
- *                  UINT32  dst_pitch  : Destination block pitch.
- *                  UINT32  height     : Block height.
- *                  UINT32  width      : Block width.
- *                  INT32  *vp8_filter : Array of 2 bi-linear filter taps.
+ *  INPUTS        : INT32  *src_ptr          : Pointer to source block.
+ *                  UINT32 src_pixels_per_line : Stride of input block.
+ *                  UINT32 pixel_step        : Offset between filter input samples (see notes).
+ *                  UINT32 output_height     : Input block height.
+ *                  UINT32 output_width      : Input block width.
+ *                  INT32  *vp8_filter          : Array of 2 bi-linear filter taps.
 *
- *  OUTPUTS       : UINT16 *dst_ptr    : Pointer to filtered block.
+ *  OUTPUTS       : UINT16 *output_ptr       : Pointer to filtered block.
 *
 *  RETURNS       : void
 *
- *  FUNCTION      : Applies a 1-D 2-tap bi-linear filter to the source block
- *                  in the vertical direction to produce the filtered output
- *                  block. Used to implement second-pass of 2-D separable filter.
+ *  FUNCTION      : Applies a 1-D 2-tap bi-linear filter to the source block in
+ *                  either horizontal or vertical direction to produce the
+ *                  filtered output block. Used to implement second-pass
+ *                  of 2-D separable filter.
 *
 *  SPECIAL NOTES : Requires 32-bit input as produced by filter_block2d_bil_first_pass.
 *                  Two filter taps should sum to VP8_FILTER_WEIGHT.
+ *                  pixel_step defines whether the filter is applied
+ *                  horizontally (pixel_step=1) or vertically (pixel_step=stride).
+ *                  It defines the offset required to move from one input
+ *                  to the next.
 *
 ****************************************************************************/
 void vp8_filter_block2d_bil_second_pass
 (
    unsigned short *src_ptr,
-    unsigned char  *dst_ptr,
-    int             dst_pitch,
-    unsigned int    height,
-    unsigned int    width,
-    const short    *vp8_filter
+    unsigned char  *output_ptr,
+    int output_pitch,
+    unsigned int  src_pixels_per_line,
+    unsigned int  pixel_step,
+    unsigned int  output_height,
+    unsigned int  output_width,
+    const int *vp8_filter
 )
 {
    unsigned int  i, j;
    int  Temp;

-    for (i = 0; i < height; i++)
+    for (i = 0; i < output_height; i++)
    {
-        for (j = 0; j < width; j++)
+        for (j = 0; j < output_width; j++)
        {
            /* Apply filter */
-            Temp = ((int)src_ptr[0]     * vp8_filter[0]) +
-                   ((int)src_ptr[width] * vp8_filter[1]) +
+            Temp = ((int)src_ptr[0]         * vp8_filter[0]) +
+                   ((int)src_ptr[pixel_step] * vp8_filter[1]) +
                   (VP8_FILTER_WEIGHT / 2);
-            dst_ptr[j] = (unsigned int)(Temp >> VP8_FILTER_SHIFT);
+            output_ptr[j] = (unsigned int)(Temp >> VP8_FILTER_SHIFT);
            src_ptr++;
        }

        /* Next row... */
-        dst_ptr += dst_pitch;
+        src_ptr    += src_pixels_per_line - output_width;
+        output_ptr += output_pitch;
    }
 }

@@ -381,14 +404,11 @@ void vp8_filter_block2d_bil_second_pass
 *  ROUTINE       : filter_block2d_bil
 *
 *  INPUTS        : UINT8  *src_ptr          : Pointer to source block.
- *                  UINT32  src_pitch        : Stride of source block.
- *                  UINT32  dst_pitch        : Stride of destination block.
- *                  INT32  *HFilter          : Array of 2 horizontal filter taps.
- *                  INT32  *VFilter          : Array of 2 vertical filter taps.
- *                  INT32  Width             : Block width
- *                  INT32  Height            : Block height
+ *                  UINT32 src_pixels_per_line : Stride of input block.
+ *                  INT32  *HFilter         : Array of 2 horizontal filter taps.
+ *                  INT32  *VFilter         : Array of 2 vertical filter taps.
 *
- *  OUTPUTS       : UINT16 *dst_ptr       : Pointer to filtered block.
+ *  OUTPUTS       : UINT16 *output_ptr       : Pointer to filtered block.
 *
 *  RETURNS       : void
 *
@@ -402,23 +422,23 @@ void vp8_filter_block2d_bil_second_pass
 void vp8_filter_block2d_bil
 (
    unsigned char *src_ptr,
-    unsigned char *dst_ptr,
-    unsigned int   src_pitch,
+    unsigned char *output_ptr,
+    unsigned int   src_pixels_per_line,
    unsigned int   dst_pitch,
-    const short   *HFilter,
-    const short   *VFilter,
+    const int      *HFilter,
+    const int      *VFilter,
    int            Width,
    int            Height
 )
 {

-    unsigned short FData[17*16];    /* Temp data buffer used in filtering */
+    unsigned short FData[17*16];    /* Temp data bufffer used in filtering */

    /* First filter 1-D horizontally... */
-    vp8_filter_block2d_bil_first_pass(src_ptr, FData, src_pitch, Height + 1, Width, HFilter);
+    vp8_filter_block2d_bil_first_pass(src_ptr, FData, src_pixels_per_line, 1, Height + 1, Width, HFilter);

    /* then 1-D vertically... */
-    vp8_filter_block2d_bil_second_pass(FData, dst_ptr, dst_pitch, Height, Width, VFilter);
+    vp8_filter_block2d_bil_second_pass(FData, output_ptr, dst_pitch, Width, Width, Height, Width, VFilter);
 }


@@ -432,11 +452,11 @@ void vp8_bilinear_predict4x4_c
    int dst_pitch
 )
 {
-    const short *HFilter;
-    const short *VFilter;
+    const int  *HFilter;
+    const int  *VFilter;

-    HFilter = vp8_bilinear_filters[xoffset];
-    VFilter = vp8_bilinear_filters[yoffset];
+    HFilter = bilinear_filters[xoffset];
+    VFilter = bilinear_filters[yoffset];
 #if 0
    {
        int i;
@@ -470,11 +490,11 @@ void vp8_bilinear_predict8x8_c
    int  dst_pitch
 )
 {
-    const short *HFilter;
-    const short *VFilter;
+    const int  *HFilter;
+    const int  *VFilter;

-    HFilter = vp8_bilinear_filters[xoffset];
-    VFilter = vp8_bilinear_filters[yoffset];
+    HFilter = bilinear_filters[xoffset];
+    VFilter = bilinear_filters[yoffset];

    vp8_filter_block2d_bil(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter, VFilter, 8, 8);

@@ -490,11 +510,11 @@ void vp8_bilinear_predict8x4_c
    int  dst_pitch
 )
 {
-    const short *HFilter;
-    const short *VFilter;
+    const int  *HFilter;
+    const int  *VFilter;

-    HFilter = vp8_bilinear_filters[xoffset];
-    VFilter = vp8_bilinear_filters[yoffset];
+    HFilter = bilinear_filters[xoffset];
+    VFilter = bilinear_filters[yoffset];

    vp8_filter_block2d_bil(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter, VFilter, 8, 4);

@@ -510,11 +530,11 @@ void vp8_bilinear_predict16x16_c
    int  dst_pitch
 )
 {
-    const short *HFilter;
-    const short *VFilter;
+    const int  *HFilter;
+    const int  *VFilter;

-    HFilter = vp8_bilinear_filters[xoffset];
-    VFilter = vp8_bilinear_filters[yoffset];
+    HFilter = bilinear_filters[xoffset];
+    VFilter = bilinear_filters[yoffset];

    vp8_filter_block2d_bil(src_ptr, dst_ptr, src_pixels_per_line, dst_pitch, HFilter, VFilter, 16, 16);
 }
--- a/vp8/common/findnearmv.c
+++ b/vp8/common/findnearmv.c
@@ -11,9 +11,47 @@

 #include "findnearmv.h"

+#define FINDNEAR_SEARCH_SITES   3
+
 /* Predict motion vectors using those from already-decoded nearby blocks.
   Note that we only consider one 4x4 subblock from each candidate 16x16
   macroblock.   */
+
+typedef union
+{
+    unsigned int as_int;
+    MV           as_mv;
+} int_mv;        /* facilitates rapid equality tests */
+
+static void mv_bias(const MODE_INFO *x, int refframe, int_mv *mvp, const int *ref_frame_sign_bias)
+{
+    MV xmv;
+    xmv = x->mbmi.mv.as_mv;
+
+    if (ref_frame_sign_bias[x->mbmi.ref_frame] != ref_frame_sign_bias[refframe])
+    {
+        xmv.row *= -1;
+        xmv.col *= -1;
+    }
+
+    mvp->as_mv = xmv;
+}
+
+
+void vp8_clamp_mv(MV *mv, const MACROBLOCKD *xd)
+{
+    if (mv->col < (xd->mb_to_left_edge - LEFT_TOP_MARGIN))
+        mv->col = xd->mb_to_left_edge - LEFT_TOP_MARGIN;
+    else if (mv->col > xd->mb_to_right_edge + RIGHT_BOTTOM_MARGIN)
+        mv->col = xd->mb_to_right_edge + RIGHT_BOTTOM_MARGIN;
+
+    if (mv->row < (xd->mb_to_top_edge - LEFT_TOP_MARGIN))
+        mv->row = xd->mb_to_top_edge - LEFT_TOP_MARGIN;
+    else if (mv->row > xd->mb_to_bottom_edge + RIGHT_BOTTOM_MARGIN)
+        mv->row = xd->mb_to_bottom_edge + RIGHT_BOTTOM_MARGIN;
+}
+
+
 void vp8_find_near_mvs
 (
    MACROBLOCKD *xd,
@@ -44,7 +82,7 @@ void vp8_find_near_mvs
        if (above->mbmi.mv.as_int)
        {
            (++mv)->as_int = above->mbmi.mv.as_int;
-            mv_bias(ref_frame_sign_bias[above->mbmi.ref_frame], refframe, mv, ref_frame_sign_bias);
+            mv_bias(above, refframe, mv, ref_frame_sign_bias);
            ++cntx;
        }

@@ -59,7 +97,7 @@ void vp8_find_near_mvs
            int_mv this_mv;

            this_mv.as_int = left->mbmi.mv.as_int;
-            mv_bias(ref_frame_sign_bias[left->mbmi.ref_frame], refframe, &this_mv, ref_frame_sign_bias);
+            mv_bias(left, refframe, &this_mv, ref_frame_sign_bias);

            if (this_mv.as_int != mv->as_int)
            {
@@ -81,7 +119,7 @@ void vp8_find_near_mvs
            int_mv this_mv;

            this_mv.as_int = aboveleft->mbmi.mv.as_int;
-            mv_bias(ref_frame_sign_bias[aboveleft->mbmi.ref_frame], refframe, &this_mv, ref_frame_sign_bias);
+            mv_bias(aboveleft, refframe, &this_mv, ref_frame_sign_bias);

            if (this_mv.as_int != mv->as_int)
            {
--- a/vp8/common/findnearmv.h
+++ b/vp8/common/findnearmv.h
@@ -17,41 +17,6 @@
 #include "modecont.h"
 #include "treecoder.h"

-typedef union
-{
-    unsigned int as_int;
-    MV           as_mv;
-} int_mv;        /* facilitates rapid equality tests */
-
-static void mv_bias(int refmb_ref_frame_sign_bias, int refframe, int_mv *mvp, const int *ref_frame_sign_bias)
-{
-    MV xmv;
-    xmv = mvp->as_mv;
-
-    if (refmb_ref_frame_sign_bias != ref_frame_sign_bias[refframe])
-    {
-        xmv.row *= -1;
-        xmv.col *= -1;
-    }
-
-    mvp->as_mv = xmv;
-}
-
-#define LEFT_TOP_MARGIN (16 << 3)
-#define RIGHT_BOTTOM_MARGIN (16 << 3)
-static void vp8_clamp_mv(MV *mv, const MACROBLOCKD *xd)
-{
-    if (mv->col < (xd->mb_to_left_edge - LEFT_TOP_MARGIN))
-        mv->col = xd->mb_to_left_edge - LEFT_TOP_MARGIN;
-    else if (mv->col > xd->mb_to_right_edge + RIGHT_BOTTOM_MARGIN)
-        mv->col = xd->mb_to_right_edge + RIGHT_BOTTOM_MARGIN;
-
-    if (mv->row < (xd->mb_to_top_edge - LEFT_TOP_MARGIN))
-        mv->row = xd->mb_to_top_edge - LEFT_TOP_MARGIN;
-    else if (mv->row > xd->mb_to_bottom_edge + RIGHT_BOTTOM_MARGIN)
-        mv->row = xd->mb_to_bottom_edge + RIGHT_BOTTOM_MARGIN;
-}
-
 void vp8_find_near_mvs
 (
    MACROBLOCKD *xd,
@@ -70,4 +35,8 @@ const B_MODE_INFO *vp8_left_bmi(const MODE_INFO *cur_mb, int b);

 const B_MODE_INFO *vp8_above_bmi(const MODE_INFO *cur_mb, int b, int mi_stride);

+#define LEFT_TOP_MARGIN (16 << 3)
+#define RIGHT_BOTTOM_MARGIN (16 << 3)
+
+
 #endif
--- a/vp8/common/fourcc.hpp
+++ b/vp8/common/fourcc.hpp
@@ -0,0 +1,121 @@
+/*
+ *  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 FOURCC_HPP
+#define FOURCC_HPP
+
+#include <iosfwd>
+#include <cstring>
+
+
+#if defined(__POWERPC__) || defined(__APPLE__) || defined(__MERKS__)
+using namespace std;
+#endif
+
+class four_cc
+{
+public:
+
+    four_cc();
+    four_cc(const char*);
+    explicit four_cc(unsigned long);
+
+    bool operator==(const four_cc&) const;
+    bool operator!=(const four_cc&) const;
+
+    bool operator==(const char*) const;
+    bool operator!=(const char*) const;
+
+    operator unsigned long() const;
+    unsigned long as_long() const;
+
+    four_cc& operator=(unsigned long);
+
+    char operator[](int) const;
+
+    std::ostream& put(std::ostream&) const;
+
+    bool printable() const;
+
+private:
+
+    union
+    {
+        char code[4];
+        unsigned long code_as_long;
+    };
+
+};
+
+
+inline four_cc::four_cc()
+{
+}
+
+inline four_cc::four_cc(unsigned long x)
+    : code_as_long(x)
+{
+}
+
+inline four_cc::four_cc(const char* str)
+{
+    memcpy(code, str, 4);
+}
+
+
+inline bool four_cc::operator==(const four_cc& rhs) const
+{
+    return code_as_long == rhs.code_as_long;
+}
+
+inline bool four_cc::operator!=(const four_cc& rhs) const
+{
+    return !operator==(rhs);
+}
+
+inline bool four_cc::operator==(const char* rhs) const
+{
+    return (memcmp(code, rhs, 4) == 0);
+}
+
+inline bool four_cc::operator!=(const char* rhs) const
+{
+    return !operator==(rhs);
+}
+
+
+inline four_cc::operator unsigned long() const
+{
+    return code_as_long;
+}
+
+inline unsigned long four_cc::as_long() const
+{
+    return code_as_long;
+}
+
+inline char four_cc::operator[](int i) const
+{
+    return code[i];
+}
+
+inline four_cc& four_cc::operator=(unsigned long val)
+{
+    code_as_long = val;
+    return *this;
+}
+
+inline std::ostream& operator<<(std::ostream& os, const four_cc& rhs)
+{
+    return rhs.put(os);
+}
+
+#endif
--- a/vp8/common/generic/systemdependent.c
+++ b/vp8/common/generic/systemdependent.c
@@ -10,12 +10,12 @@


 #include "vpx_ports/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 "vp8/common/onyxc_int.h"
+#include "g_common.h"
+#include "subpixel.h"
+#include "loopfilter.h"
+#include "recon.h"
+#include "idct.h"
+#include "onyxc_int.h"

 extern void vp8_arch_x86_common_init(VP8_COMMON *ctx);
 extern void vp8_arch_arm_common_init(VP8_COMMON *ctx);
--- a/vp8/common/loopfilter.c
+++ b/vp8/common/loopfilter.c
@@ -28,13 +28,13 @@ void vp8_loop_filter_mbh_c(unsigned char *y_ptr, unsigned char *u_ptr, unsigned
                           int y_stride, int uv_stride, loop_filter_info *lfi, int simpler_lpf)
 {
    (void) simpler_lpf;
-    vp8_mbloop_filter_horizontal_edge_c(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->thr, 2);
+    vp8_mbloop_filter_horizontal_edge_c(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->mbthr, 2);

    if (u_ptr)
-        vp8_mbloop_filter_horizontal_edge_c(u_ptr, uv_stride, lfi->mbflim, lfi->lim, lfi->thr, 1);
+        vp8_mbloop_filter_horizontal_edge_c(u_ptr, uv_stride, lfi->uvmbflim, lfi->uvlim, lfi->uvmbthr, 1);

    if (v_ptr)
-        vp8_mbloop_filter_horizontal_edge_c(v_ptr, uv_stride, lfi->mbflim, lfi->lim, lfi->thr, 1);
+        vp8_mbloop_filter_horizontal_edge_c(v_ptr, uv_stride, lfi->uvmbflim, lfi->uvlim, lfi->uvmbthr, 1);
 }

 void vp8_loop_filter_mbhs_c(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
@@ -44,7 +44,7 @@ void vp8_loop_filter_mbhs_c(unsigned char *y_ptr, unsigned char *u_ptr, unsigned
    (void) v_ptr;
    (void) uv_stride;
    (void) simpler_lpf;
-    vp8_loop_filter_simple_horizontal_edge_c(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->thr, 2);
+    vp8_loop_filter_simple_horizontal_edge_c(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->mbthr, 2);
 }

 /* Vertical MB Filtering */
@@ -52,13 +52,13 @@ void vp8_loop_filter_mbv_c(unsigned char *y_ptr, unsigned char *u_ptr, unsigned
                           int y_stride, int uv_stride, loop_filter_info *lfi, int simpler_lpf)
 {
    (void) simpler_lpf;
-    vp8_mbloop_filter_vertical_edge_c(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->thr, 2);
+    vp8_mbloop_filter_vertical_edge_c(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->mbthr, 2);

    if (u_ptr)
-        vp8_mbloop_filter_vertical_edge_c(u_ptr, uv_stride, lfi->mbflim, lfi->lim, lfi->thr, 1);
+        vp8_mbloop_filter_vertical_edge_c(u_ptr, uv_stride, lfi->uvmbflim, lfi->uvlim, lfi->uvmbthr, 1);

    if (v_ptr)
-        vp8_mbloop_filter_vertical_edge_c(v_ptr, uv_stride, lfi->mbflim, lfi->lim, lfi->thr, 1);
+        vp8_mbloop_filter_vertical_edge_c(v_ptr, uv_stride, lfi->uvmbflim, lfi->uvlim, lfi->uvmbthr, 1);
 }

 void vp8_loop_filter_mbvs_c(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
@@ -68,7 +68,7 @@ void vp8_loop_filter_mbvs_c(unsigned char *y_ptr, unsigned char *u_ptr, unsigned
    (void) v_ptr;
    (void) uv_stride;
    (void) simpler_lpf;
-    vp8_loop_filter_simple_vertical_edge_c(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->thr, 2);
+    vp8_loop_filter_simple_vertical_edge_c(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->mbthr, 2);
 }

 /* Horizontal B Filtering */
@@ -81,10 +81,10 @@ void vp8_loop_filter_bh_c(unsigned char *y_ptr, unsigned char *u_ptr, unsigned c
    vp8_loop_filter_horizontal_edge_c(y_ptr + 12 * y_stride, y_stride, lfi->flim, lfi->lim, lfi->thr, 2);

    if (u_ptr)
-        vp8_loop_filter_horizontal_edge_c(u_ptr + 4 * uv_stride, uv_stride, lfi->flim, lfi->lim, lfi->thr, 1);
+        vp8_loop_filter_horizontal_edge_c(u_ptr + 4 * uv_stride, uv_stride, lfi->uvflim, lfi->uvlim, lfi->uvthr, 1);

    if (v_ptr)
-        vp8_loop_filter_horizontal_edge_c(v_ptr + 4 * uv_stride, uv_stride, lfi->flim, lfi->lim, lfi->thr, 1);
+        vp8_loop_filter_horizontal_edge_c(v_ptr + 4 * uv_stride, uv_stride, lfi->uvflim, lfi->uvlim, lfi->uvthr, 1);
 }

 void vp8_loop_filter_bhs_c(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
@@ -109,10 +109,10 @@ void vp8_loop_filter_bv_c(unsigned char *y_ptr, unsigned char *u_ptr, unsigned c
    vp8_loop_filter_vertical_edge_c(y_ptr + 12, y_stride, lfi->flim, lfi->lim, lfi->thr, 2);

    if (u_ptr)
-        vp8_loop_filter_vertical_edge_c(u_ptr + 4, uv_stride, lfi->flim, lfi->lim, lfi->thr, 1);
+        vp8_loop_filter_vertical_edge_c(u_ptr + 4, uv_stride, lfi->uvflim, lfi->uvlim, lfi->uvthr, 1);

    if (v_ptr)
-        vp8_loop_filter_vertical_edge_c(v_ptr + 4, uv_stride, lfi->flim, lfi->lim, lfi->thr, 1);
+        vp8_loop_filter_vertical_edge_c(v_ptr + 4, uv_stride, lfi->uvflim, lfi->uvlim, lfi->uvthr, 1);
 }

 void vp8_loop_filter_bvs_c(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
@@ -137,6 +137,8 @@ void vp8_init_loop_filter(VP8_COMMON *cm)

    int block_inside_limit = 0;
    int HEVThresh;
+    const int yhedge_boost  = 2;
+    const int uvhedge_boost = 2;

    /* For each possible value for the loop filter fill out a "loop_filter_info" entry. */
    for (i = 0; i <= MAX_LOOP_FILTER; i++)
@@ -180,9 +182,15 @@ void vp8_init_loop_filter(VP8_COMMON *cm)
        for (j = 0; j < 16; j++)
        {
            lfi[i].lim[j] = block_inside_limit;
-            lfi[i].mbflim[j] = filt_lvl + 2;
+            lfi[i].mbflim[j] = filt_lvl + yhedge_boost;
+            lfi[i].mbthr[j] = HEVThresh;
            lfi[i].flim[j] = filt_lvl;
            lfi[i].thr[j] = HEVThresh;
+            lfi[i].uvlim[j] = block_inside_limit;
+            lfi[i].uvmbflim[j] = filt_lvl + uvhedge_boost;
+            lfi[i].uvmbthr[j] = HEVThresh;
+            lfi[i].uvflim[j] = filt_lvl;
+            lfi[i].uvthr[j] = HEVThresh;
        }

    }
@@ -241,52 +249,57 @@ void vp8_frame_init_loop_filter(loop_filter_info *lfi, int frame_type)
        for (j = 0; j < 16; j++)
        {
            /*lfi[i].lim[j] = block_inside_limit;
-            lfi[i].mbflim[j] = filt_lvl+2;*/
+            lfi[i].mbflim[j] = filt_lvl+yhedge_boost;*/
+            lfi[i].mbthr[j] = HEVThresh;
            /*lfi[i].flim[j] = filt_lvl;*/
            lfi[i].thr[j] = HEVThresh;
+            /*lfi[i].uvlim[j] = block_inside_limit;
+            lfi[i].uvmbflim[j] = filt_lvl+uvhedge_boost;*/
+            lfi[i].uvmbthr[j] = HEVThresh;
+            /*lfi[i].uvflim[j] = filt_lvl;*/
+            lfi[i].uvthr[j] = HEVThresh;
        }
    }
 }


-int vp8_adjust_mb_lf_value(MACROBLOCKD *mbd, int filter_level)
+void vp8_adjust_mb_lf_value(MACROBLOCKD *mbd, int *filter_level)
 {
    MB_MODE_INFO *mbmi = &mbd->mode_info_context->mbmi;

    if (mbd->mode_ref_lf_delta_enabled)
    {
        /* Apply delta for reference frame */
-        filter_level += mbd->ref_lf_deltas[mbmi->ref_frame];
+        *filter_level += mbd->ref_lf_deltas[mbmi->ref_frame];

        /* Apply delta for mode */
        if (mbmi->ref_frame == INTRA_FRAME)
        {
            /* Only the split mode BPRED has a further special case */
            if (mbmi->mode == B_PRED)
-                filter_level +=  mbd->mode_lf_deltas[0];
+                *filter_level +=  mbd->mode_lf_deltas[0];
        }
        else
        {
            /* Zero motion mode */
            if (mbmi->mode == ZEROMV)
-                filter_level +=  mbd->mode_lf_deltas[1];
+                *filter_level +=  mbd->mode_lf_deltas[1];

            /* Split MB motion mode */
            else if (mbmi->mode == SPLITMV)
-                filter_level +=  mbd->mode_lf_deltas[3];
+                *filter_level +=  mbd->mode_lf_deltas[3];

            /* All other inter motion modes (Nearest, Near, New) */
            else
-                filter_level +=  mbd->mode_lf_deltas[2];
+                *filter_level +=  mbd->mode_lf_deltas[2];
        }

        /* Range check */
-        if (filter_level > MAX_LOOP_FILTER)
-            filter_level = MAX_LOOP_FILTER;
-        else if (filter_level < 0)
-            filter_level = 0;
+        if (*filter_level > MAX_LOOP_FILTER)
+            *filter_level = MAX_LOOP_FILTER;
+        else if (*filter_level < 0)
+            *filter_level = 0;
    }
-    return filter_level;
 }


@@ -360,7 +373,7 @@ void vp8_loop_filter_frame
             * These specified to 8th pel as they are always compared to values that are in 1/8th pel units
             * Apply any context driven MB level adjustment
             */
-            filter_level = vp8_adjust_mb_lf_value(mbd, filter_level);
+            vp8_adjust_mb_lf_value(mbd, &filter_level);

            if (filter_level)
            {
@@ -460,7 +473,7 @@ void vp8_loop_filter_frame_yonly
            filter_level = baseline_filter_level[Segment];

            /* Apply any context driven MB level adjustment */
-            filter_level = vp8_adjust_mb_lf_value(mbd, filter_level);
+            vp8_adjust_mb_lf_value(mbd, &filter_level);

            if (filter_level)
            {
--- a/vp8/common/loopfilter.h
+++ b/vp8/common/loopfilter.h
@@ -32,6 +32,12 @@ typedef struct
    DECLARE_ALIGNED(16, signed char, flim[16]);
    DECLARE_ALIGNED(16, signed char, thr[16]);
    DECLARE_ALIGNED(16, signed char, mbflim[16]);
+    DECLARE_ALIGNED(16, signed char, mbthr[16]);
+    DECLARE_ALIGNED(16, signed char, uvlim[16]);
+    DECLARE_ALIGNED(16, signed char, uvflim[16]);
+    DECLARE_ALIGNED(16, signed char, uvthr[16]);
+    DECLARE_ALIGNED(16, signed char, uvmbflim[16]);
+    DECLARE_ALIGNED(16, signed char, uvmbthr[16]);
 } loop_filter_info;


--- a/vp8/common/mac_specs.h
+++ b/vp8/common/mac_specs.h
@@ -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.
+ */
+
+
+#if !defined(_mac_specs_h)
+#define _mac_specs_h
+
+
+#if defined(__cplusplus)
+extern "C" {
+#endif
+
+    extern unsigned int vp8_read_tsc();
+
+    extern unsigned int vp8_get_processor_freq();
+
+    extern unsigned int vpx_has_altivec();
+
+#if defined(__cplusplus)
+}
+#endif
+
+
+#endif
--- a/vp8/common/onyxc_int.h
+++ b/vp8/common/onyxc_int.h
@@ -140,6 +140,8 @@ typedef struct VP8Common

    MODE_INFO *mip; /* Base of allocated array */
    MODE_INFO *mi;  /* Corresponds to upper left visible macroblock */
+    MODE_INFO *prev_mip; /* MODE_INFO array 'mip' from last decoded frame */
+    MODE_INFO *prev_mi;  /* 'mi' from last frame (points into prev_mip) */


    INTERPOLATIONFILTERTYPE mcomp_filter_type;
@@ -200,7 +202,7 @@ typedef struct VP8Common
 } VP8_COMMON;


-int vp8_adjust_mb_lf_value(MACROBLOCKD *mbd, int filter_level);
+void vp8_adjust_mb_lf_value(MACROBLOCKD *mbd, int *filter_level);
 void vp8_init_loop_filter(VP8_COMMON *cm);
 void vp8_frame_init_loop_filter(loop_filter_info *lfi, int frame_type);
 extern void vp8_loop_filter_frame(VP8_COMMON *cm,    MACROBLOCKD *mbd,  int filt_val);
--- a/vp8/common/partialgfupdate.h
+++ b/vp8/common/partialgfupdate.h
@@ -1,5 +1,5 @@
 /*
- *  Copyright (c) 2011 The WebM project authors. All Rights Reserved.
+ *  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
@@ -9,14 +9,11 @@
 */


-#ifndef FILTER_H
-#define FILTER_H
+#ifndef __INC_PARTIALGFUPDATE_H
+#define __INC_PARTIALGFUPDATE_H

-#define BLOCK_HEIGHT_WIDTH 4
-#define VP8_FILTER_WEIGHT 128
-#define VP8_FILTER_SHIFT  7
+#include "onyxc_int.h"

-extern const short vp8_bilinear_filters[8][2];
-extern const short vp8_sub_pel_filters[8][6];
+extern void update_gf_selective(ONYX_COMMON *cm, MACROBLOCKD *x);

-#endif //FILTER_H
+#endif
--- a/vp8/common/ppc/loopfilter_altivec.c
+++ b/vp8/common/ppc/loopfilter_altivec.c
@@ -56,10 +56,10 @@ void loop_filter_mbh_ppc(unsigned char *y_ptr, unsigned char *u_ptr, unsigned ch
                         int y_stride, int uv_stride, loop_filter_info *lfi, int simpler_lpf)
 {
    (void)simpler_lpf;
-    mbloop_filter_horizontal_edge_y_ppc(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->thr);
+    mbloop_filter_horizontal_edge_y_ppc(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->mbthr);

    if (u_ptr)
-        mbloop_filter_horizontal_edge_uv_ppc(u_ptr, v_ptr, uv_stride, lfi->mbflim, lfi->lim, lfi->thr);
+        mbloop_filter_horizontal_edge_uv_ppc(u_ptr, v_ptr, uv_stride, lfi->uvmbflim, lfi->uvlim, lfi->uvmbthr);
 }

 void loop_filter_mbhs_ppc(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
@@ -77,10 +77,10 @@ void loop_filter_mbv_ppc(unsigned char *y_ptr, unsigned char *u_ptr, unsigned ch
                         int y_stride, int uv_stride, loop_filter_info *lfi, int simpler_lpf)
 {
    (void)simpler_lpf;
-    mbloop_filter_vertical_edge_y_ppc(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->thr);
+    mbloop_filter_vertical_edge_y_ppc(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->mbthr);

    if (u_ptr)
-        mbloop_filter_vertical_edge_uv_ppc(u_ptr, v_ptr, uv_stride, lfi->mbflim, lfi->lim, lfi->thr);
+        mbloop_filter_vertical_edge_uv_ppc(u_ptr, v_ptr, uv_stride, lfi->uvmbflim, lfi->uvlim, lfi->uvmbthr);
 }

 void loop_filter_mbvs_ppc(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
@@ -104,7 +104,7 @@ void loop_filter_bh_ppc(unsigned char *y_ptr, unsigned char *u_ptr, unsigned cha
    loop_filter_horizontal_edge_y_ppc(y_ptr + 12 * y_stride, y_stride, lfi->flim, lfi->lim, lfi->thr);

    if (u_ptr)
-        loop_filter_horizontal_edge_uv_ppc(u_ptr + 4 * uv_stride, v_ptr + 4 * uv_stride, uv_stride, lfi->flim, lfi->lim, lfi->thr);
+        loop_filter_horizontal_edge_uv_ppc(u_ptr + 4 * uv_stride, v_ptr + 4 * uv_stride, uv_stride, lfi->uvflim, lfi->uvlim, lfi->uvthr);
 }

 void loop_filter_bhs_ppc(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
@@ -127,7 +127,7 @@ void loop_filter_bv_ppc(unsigned char *y_ptr, unsigned char *u_ptr, unsigned cha
    loop_filter_vertical_edge_y_ppc(y_ptr, y_stride, lfi->flim, lfi->lim, lfi->thr);

    if (u_ptr)
-        loop_filter_vertical_edge_uv_ppc(u_ptr + 4, v_ptr + 4, uv_stride, lfi->flim, lfi->lim, lfi->thr);
+        loop_filter_vertical_edge_uv_ppc(u_ptr + 4, v_ptr + 4, uv_stride, lfi->uvflim, lfi->uvlim, lfi->uvthr);
 }

 void loop_filter_bvs_ppc(unsigned char *y_ptr, unsigned char *u_ptr, unsigned char *v_ptr,
--- a/vp8/common/predictdc.c
+++ b/vp8/common/predictdc.c
@@ -0,0 +1,44 @@
+/*
+ *  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 <stdlib.h>
+#include "blockd.h"
+
+
+void vp8_predict_dc(short *lastdc, short *thisdc, short quant, short *cons)
+{
+    int diff;
+    int sign;
+    int last_dc = *lastdc;
+    int this_dc = *thisdc;
+
+    if (*cons  > DCPREDCNTTHRESH)
+    {
+        this_dc += last_dc;
+    }
+
+    diff = abs(last_dc - this_dc);
+    sign  = (last_dc >> 31) ^(this_dc >> 31);
+    sign |= (!last_dc | !this_dc);
+
+    if (sign)
+    {
+        *cons = 0;
+    }
+    else
+    {
+        if (diff <= DCPREDSIMTHRESH * quant)
+            (*cons)++ ;
+    }
+
+    *thisdc = this_dc;
+    *lastdc = this_dc;
+}
--- a/vp8/common/predictdc.h
+++ b/vp8/common/predictdc.h
@@ -0,0 +1,18 @@
+/*
+ *  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 __PREDICTDC_H
+#define __PREDICTDC_H
+
+void uvvp8_predict_dc(short *lastdc, short *thisdc, short quant, short *cons);
+void vp8_predict_dc(short *lastdc, short *thisdc, short quant, short *cons);
+
+#endif
--- a/vp8/common/proposed.h
+++ b/vp8/common/proposed.h
@@ -0,0 +1,71 @@
+/*
+ *  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.
+ */
+
+
+typedef struct core_codec *codec_ptr;
+typedef struct interface_table *interface_ptr;
+
+typedef struct
+{
+    void (*Initialize)();
+    void (*Shutdown)();
+    codec_ptr(*Create)();
+    int (*compress_frame)(codec_ptr, unsigned int *frame_flags, YV12_BUFFER_CONFIG *sd, unsigned long *size, char *dest, INT64 time_stamp);
+    int (*show_frame)(codec_ptr , YV12_BUFFER_CONFIG *dest, int deblock_level, int noise_level, int flags);
+    void (*Remove)(codec_ptr *comp);
+    interface_ptr(*get_interface)(unsigned int id);
+
+} core_codec;
+
+typedef struct
+{
+    int (*set_bitrate)(codec_ptr, END_USAGE usage, int Datarate);
+    int (*get_bitrate)(codec_ptr, END_USAGE *usage, int *Datarate);
+    int (*set_mode)(codec_ptr, MODE mode, int Speed, char *File);
+    int (*get_mode)(codec_ptr, MODE *mode, int *Speed, char **File);
+} codec_settings_basic;
+
+typedef struct
+{
+    int (*set_bitrate)(codec_ptr, END_USAGE usage, int Datarate);
+    int (*get_bitrate)(codec_ptr, END_USAGE *usage, int *Datarate);
+    int (*set_mode)(codec_ptr, MODE  mode, int  Speed, char   *File);
+    int (*get_mode)(codec_ptr, MODE *mode, int *Speed, char **File);
+    int (*set_denoise)(codec_ptr, int  Level);
+    int (*get_denoise)(codec_ptr, int *Level);
+    int (*set_sharpness)(codec_ptr, int  sharpness);
+    int (*get_sharpness)(codec_ptr, int *sharpness);
+    int (*set_keyframing)(codec_ptr, int  Auto, int  max_distance);
+    int (*get_keyframing)(codec_ptr, int *Auto, int *max_distance);
+    int (*set_buffering)(codec_ptr, int  buffer_level, int  max_buffer_level);
+    int (*get_buffering)(codec_ptr, int *buffer_level, int *max_buffer_level);
+    int (*set_adjust_frame_rate)(codec_ptr, int Allowed, int at_buffer_level_pct);
+    int (*get_adjust_frame_rate)(codec_ptr, int *Allowed, int *at_buffer_level_pct);
+    int (*set_adjust_frame_size)(codec_ptr, int Allowed, int down_at_buffer_level_pct, int up_at_buffer_level_pct);
+    int (*get_adjust_frame_size)(codec_ptr, int *Allowed, int *down_at_buffer_level_pct, int *up_at_buffer_level_pct);
+    int (*set_adjust_quality)(codec_ptr, int Allowed, int min_quantizer, int max_quantizer);
+    int (*get_adjust_quality)(codec_ptr, int *Allowed, int *min_quantizer, int *max_quantizer);
+    int (*set_vbrparms)(codec_ptr, int Bias, int Min, int Max);
+    int (*get_vbrparms)(codec_ptr, int *Bias, int *Min, int *Max);
+
+} codec_settings_v1;
+
+typedef struct
+{
+    int (*request_recovery)(codec_ptr);
+    int (*request_droppable)(codec_ptr);
+    int (*internal_size)(codec_ptr, VPX_SCALING Vertical, VPX_SCALING Horizontal);
+    int (*update_last)(codec_ptr);
+    int (*update_gold)(codec_ptr);
+    int (*use_only_last)(codec_ptr);
+    int (*use_only_gold)(codec_ptr);
+    int (*update_entropy)(codec_ptr);
+
+} codec_realtime_requests;
--- a/vp8/common/recon.h
+++ b/vp8/common/recon.h
@@ -94,5 +94,6 @@ typedef struct vp8_recon_rtcd_vtable
 #define RECON_INVOKE(ctx,fn) vp8_recon_##fn
 #endif

+void vp8_recon_intra4x4mb(const vp8_recon_rtcd_vtable_t *rtcd, MACROBLOCKD *x);
 void vp8_recon_intra_mbuv(const vp8_recon_rtcd_vtable_t *rtcd, MACROBLOCKD *x);
 #endif
--- a/vp8/common/reconintra4x4.c
+++ b/vp8/common/reconintra4x4.c
@@ -313,3 +313,89 @@ void vp8_intra_prediction_down_copy(MACROBLOCKD *x)
 }


+void vp8_recon_intra4x4mb(const vp8_recon_rtcd_vtable_t *rtcd, MACROBLOCKD *x)
+{
+    int i;
+
+    vp8_intra_prediction_down_copy(x);
+
+#if ARCH_ARM
+    {
+        BLOCKD *b = &x->block[0];
+
+        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
+        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
+        b += 1;
+
+        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
+        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
+        b += 1;
+
+        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
+        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
+        b += 1;
+
+        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
+        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
+        b += 1;
+
+        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
+        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
+        b += 1;
+
+        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
+        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
+        b += 1;
+
+        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
+        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
+        b += 1;
+
+        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
+        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
+        b += 1;
+
+        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
+        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
+        b += 1;
+
+        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
+        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
+        b += 1;
+
+        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
+        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
+        b += 1;
+
+        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
+        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
+        b += 1;
+
+        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
+        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
+        b += 1;
+
+        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
+        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
+        b += 1;
+
+        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
+        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
+        b += 1;
+
+        vp8_predict_intra4x4(b, b->bmi.mode, b->predictor);
+        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
+    }
+#else
+    for (i = 0; i < 16; i++)
+    {
+        BLOCKD *b = &x->block[i];
+
+        vp8_predict_intra4x4(b, x->block[i].bmi.mode, x->block[i].predictor);
+        RECON_INVOKE(rtcd, recon)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
+    }
+#endif
+
+    vp8_recon_intra_mbuv(rtcd, x);
+
+}
--- a/vp8/common/threading.h
+++ b/vp8/common/threading.h
@@ -14,8 +14,6 @@

 #define VPXINFINITE 10000       /* 10second. */

-#if CONFIG_OS_SUPPORT && CONFIG_MULTITHREAD
-
 /* Thread management macros */
 #ifdef _WIN32
 /* Win32 */
@@ -90,6 +88,4 @@
 #define x86_pause_hint()
 #endif

-#endif /* CONFIG_OS_SUPPORT && CONFIG_MULTITHREAD */
-
 #endif
--- a/vp8/common/vfwsetting.hpp
+++ b/vp8/common/vfwsetting.hpp
@@ -0,0 +1,76 @@
+/*
+ *  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.
+ */
+
+
+#if !defined(VFWSETTING_HPP)
+#define VFWSETTING_HPP
+//______________________________________________________________________________
+//
+//  VFWSetting.hpp
+//
+
+#include "four_cc.hpp"
+#include <iosfwd>
+
+namespace vpxvp
+{
+
+    //--------------------------------------
+    class VFWSetting
+    {
+        friend std::ostream& operator<<(std::ostream& os, const VFWSetting& vfws);
+
+    public:
+
+        enum Mode
+        {
+            m_setting,
+            m_config
+        };
+
+        enum
+        {
+            header_size = 8,
+            Size = 16
+        };
+
+        VFWSetting(four_cc fcc);
+        ~VFWSetting();
+
+        four_cc fcc() const;
+        Mode mode() const;
+
+        int setting() const;
+        int value() const;
+        void setting_value(int i_setting, int i_value);  //  Sets mode to m_setting
+
+        long size() const;
+        const void* data() const;
+        int data(const void* p_data, unsigned long ul_size);
+
+    private:
+
+        VFWSetting(const VFWSetting& vfws);  //  Not implemented
+        VFWSetting& operator=(const VFWSetting& vfws);  //  Not implemented
+
+        int extract_(const void* p_data, unsigned long ul_size);
+        void update_() const;
+
+        four_cc m_fcc;
+        Mode m_mode;
+        int m_i_setting;
+        int m_i_value;
+
+        mutable unsigned char m_p_data[Size];
+    };
+
+}  //  namespace vpxvp
+
+#endif  //  VFWSETTING_HPP
--- a/vp8/common/vpx_ref_build_prefix.h
+++ b/vp8/common/vpx_ref_build_prefix.h
@@ -0,0 +1,24 @@
+/*
+ *  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 _VPX_REF_BUILD_PREFIX_h
+#define _VPX_REF_BUILD_PREFIX_h
+
+#if defined(__cplusplus)
+extern "C" {
+#endif
+
+
+#if defined(__cplusplus)
+}
+#endif
+
+#endif /* include guards */
--- a/vp8/common/vpxblit.h
+++ b/vp8/common/vpxblit.h
@@ -0,0 +1,112 @@
+/*
+ *  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 VPXBLIT_H_INCL
+#define VPXBLIT_H_INCL
+/*==============================================================================
+                              Includes
+==============================================================================*/
+
+/*==============================================================================
+                              Defines
+==============================================================================*/
+
+
+#ifdef VPX_BIG_ENDIAN
+#define BYTE_ZERO(X)    ((X & 0xFF000000) >> (24 - 2)   )
+#define BYTE_ONE(X)     ((X & 0x00FF0000) >> (16 - 2)   )
+#define BYTE_TWO(X)     ((X & 0x0000FF00) >> (8 - 2)    )
+#define BYTE_THREE(X)   ((X & 0x000000FF) << (0 + 2)    )
+
+#define BYTE_ZERO_UV(X) ((X & 0x0000FF00) >> (8 - 2)    )
+#define BYTE_ONE_UV(X)  ((X & 0x000000FF) << (0 + 2)    )
+
+#define REREFERENCE(X) (*((int *) &(X)))
+
+#else
+
+#define BYTE_THREE(X)   ((X & 0xFF000000) >> (24 - 2)   )
+#define BYTE_TWO(X)     ((X & 0x00FF0000) >> (16 - 2)   )
+#define BYTE_ONE(X)     ((X & 0x0000FF00) >> (8 - 2)    )
+#define BYTE_ZERO(X)    ((X & 0x000000FF) << (0 + 2)    )
+
+#define BYTE_ONE_UV(X) ((X & 0x0000FF00) >> (8 - 2) )
+#define BYTE_ZERO_UV(X)     ((X & 0x000000FF) << (0 + 2)    )
+
+#define REREFERENCE(X) (*((int *) &(X)))
+
+#endif
+
+
+/*==============================================================================
+                            Type Definitions
+==============================================================================*/
+typedef struct  // YUV buffer configuration structure
+{
+    int   y_width;
+    int   y_height;
+    int   y_stride;
+
+    int   uv_width;
+    int   uv_height;
+    int   uv_stride;
+
+    char *y_buffer;
+    char *u_buffer;
+    char *v_buffer;
+
+    char *uv_start;
+    int   uv_dst_area;
+    int   uv_used_area;
+
+} VPX_BLIT_CONFIG;
+
+typedef struct tx86_params
+{
+    unsigned int pushed_registers[6];
+    unsigned int return_address;
+    unsigned int dst;
+    unsigned int scrn_pitch;
+    VPX_BLIT_CONFIG *buff_config;
+} x86_params;
+
+/*=============================================================================
+                                Enums
+==============================================================================*/
+
+
+/*==============================================================================
+                              Structures
+==============================================================================*/
+
+/*==============================================================================
+                             Constants
+==============================================================================*/
+
+
+/*==============================================================================
+                               Variables
+==============================================================================*/
+
+
+
+
+/*==============================================================================
+                            Function Protoypes/MICROS
+==============================================================================*/
+int vpx_get_size_of_pixel(unsigned int bd);
+void *vpx_get_blitter(unsigned int bd);
+void vpx_set_blit(void);
+void vpx_destroy_blit(void);
+
+
+
+#endif //VPXBLIT_H_INCL
--- a/vp8/common/vpxblit_c64.h
+++ b/vp8/common/vpxblit_c64.h
@@ -0,0 +1,48 @@
+/*
+ *  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 _VPX_BLIT_C64_h
+#define _VPX_BLIT_C64_h
+
+/****************************************************************************
+*  Typedefs
+****************************************************************************/
+
+typedef struct  // YUV buffer configuration structure
+{
+    int     y_width;
+    int     y_height;
+    int     y_stride;
+
+    int     uv_width;
+    int     uv_height;
+    int     uv_stride;
+
+    unsigned char *y_buffer;
+    unsigned char *u_buffer;
+    unsigned char *v_buffer;
+
+    unsigned char *y_ptr_scrn;
+    unsigned char *u_ptr_scrn;
+    unsigned char *v_ptr_scrn;
+
+} DXV_YUV_BUFFER_CONFIG;
+
+typedef struct
+{
+    unsigned char *rgbptr_scrn;
+    unsigned char *y_ptr_scrn;
+    unsigned char *u_ptr_scrn;
+    unsigned char *v_ptr_scrn;
+    unsigned char *rgbptr_scrn2;
+} DXV_FINAL_VIDEO;
+
+#endif /* include guards */
--- a/vp8/common/vpxerrors.h
+++ b/vp8/common/vpxerrors.h
@@ -0,0 +1,13 @@
+/*
+ *  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.
+ */
+
+
+
+#define ALLOC_FAILURE -2
--- a/vp8/common/x86/loopfilter_x86.c
+++ b/vp8/common/x86/loopfilter_x86.c
@@ -10,7 +10,7 @@


 #include "vpx_ports/config.h"
-#include "vp8/common/loopfilter.h"
+#include "loopfilter.h"

 prototype_loopfilter(vp8_loop_filter_horizontal_edge_c);
 prototype_loopfilter(vp8_loop_filter_vertical_edge_c);
@@ -45,13 +45,13 @@ void vp8_loop_filter_mbh_mmx(unsigned char *y_ptr, unsigned char *u_ptr, unsigne
                             int y_stride, int uv_stride, loop_filter_info *lfi, int simpler_lpf)
 {
    (void) simpler_lpf;
-    vp8_mbloop_filter_horizontal_edge_mmx(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->thr, 2);
+    vp8_mbloop_filter_horizontal_edge_mmx(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->mbthr, 2);

    if (u_ptr)
-        vp8_mbloop_filter_horizontal_edge_mmx(u_ptr, uv_stride, lfi->mbflim, lfi->lim, lfi->thr, 1);
+        vp8_mbloop_filter_horizontal_edge_mmx(u_ptr, uv_stride, lfi->uvmbflim, lfi->uvlim, lfi->uvmbthr, 1);

    if (v_ptr)
-        vp8_mbloop_filter_horizontal_edge_mmx(v_ptr, uv_stride, lfi->mbflim, lfi->lim, lfi->thr, 1);
+        vp8_mbloop_filter_horizontal_edge_mmx(v_ptr, uv_stride, lfi->uvmbflim, lfi->uvlim, lfi->uvmbthr, 1);
 }


@@ -62,7 +62,7 @@ void vp8_loop_filter_mbhs_mmx(unsigned char *y_ptr, unsigned char *u_ptr, unsign
    (void) v_ptr;
    (void) uv_stride;
    (void) simpler_lpf;
-    vp8_loop_filter_simple_horizontal_edge_mmx(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->thr, 2);
+    vp8_loop_filter_simple_horizontal_edge_mmx(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->mbthr, 2);
 }


@@ -71,13 +71,13 @@ void vp8_loop_filter_mbv_mmx(unsigned char *y_ptr, unsigned char *u_ptr, unsigne
                             int y_stride, int uv_stride, loop_filter_info *lfi, int simpler_lpf)
 {
    (void) simpler_lpf;
-    vp8_mbloop_filter_vertical_edge_mmx(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->thr, 2);
+    vp8_mbloop_filter_vertical_edge_mmx(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->mbthr, 2);

    if (u_ptr)
-        vp8_mbloop_filter_vertical_edge_mmx(u_ptr, uv_stride, lfi->mbflim, lfi->lim, lfi->thr, 1);
+        vp8_mbloop_filter_vertical_edge_mmx(u_ptr, uv_stride, lfi->uvmbflim, lfi->uvlim, lfi->uvmbthr, 1);

    if (v_ptr)
-        vp8_mbloop_filter_vertical_edge_mmx(v_ptr, uv_stride, lfi->mbflim, lfi->lim, lfi->thr, 1);
+        vp8_mbloop_filter_vertical_edge_mmx(v_ptr, uv_stride, lfi->uvmbflim, lfi->uvlim, lfi->uvmbthr, 1);
 }


@@ -88,7 +88,7 @@ void vp8_loop_filter_mbvs_mmx(unsigned char *y_ptr, unsigned char *u_ptr, unsign
    (void) v_ptr;
    (void) uv_stride;
    (void) simpler_lpf;
-    vp8_loop_filter_simple_vertical_edge_mmx(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->thr, 2);
+    vp8_loop_filter_simple_vertical_edge_mmx(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->mbthr, 2);
 }


@@ -102,10 +102,10 @@ void vp8_loop_filter_bh_mmx(unsigned char *y_ptr, unsigned char *u_ptr, unsigned
    vp8_loop_filter_horizontal_edge_mmx(y_ptr + 12 * y_stride, y_stride, lfi->flim, lfi->lim, lfi->thr, 2);

    if (u_ptr)
-        vp8_loop_filter_horizontal_edge_mmx(u_ptr + 4 * uv_stride, uv_stride, lfi->flim, lfi->lim, lfi->thr, 1);
+        vp8_loop_filter_horizontal_edge_mmx(u_ptr + 4 * uv_stride, uv_stride, lfi->uvflim, lfi->uvlim, lfi->uvthr, 1);

    if (v_ptr)
-        vp8_loop_filter_horizontal_edge_mmx(v_ptr + 4 * uv_stride, uv_stride, lfi->flim, lfi->lim, lfi->thr, 1);
+        vp8_loop_filter_horizontal_edge_mmx(v_ptr + 4 * uv_stride, uv_stride, lfi->uvflim, lfi->uvlim, lfi->uvthr, 1);
 }


@@ -132,10 +132,10 @@ void vp8_loop_filter_bv_mmx(unsigned char *y_ptr, unsigned char *u_ptr, unsigned
    vp8_loop_filter_vertical_edge_mmx(y_ptr + 12, y_stride, lfi->flim, lfi->lim, lfi->thr, 2);

    if (u_ptr)
-        vp8_loop_filter_vertical_edge_mmx(u_ptr + 4, uv_stride, lfi->flim, lfi->lim, lfi->thr, 1);
+        vp8_loop_filter_vertical_edge_mmx(u_ptr + 4, uv_stride, lfi->uvflim, lfi->uvlim, lfi->uvthr, 1);

    if (v_ptr)
-        vp8_loop_filter_vertical_edge_mmx(v_ptr + 4, uv_stride, lfi->flim, lfi->lim, lfi->thr, 1);
+        vp8_loop_filter_vertical_edge_mmx(v_ptr + 4, uv_stride, lfi->uvflim, lfi->uvlim, lfi->uvthr, 1);
 }


@@ -159,10 +159,10 @@ void vp8_loop_filter_mbh_sse2(unsigned char *y_ptr, unsigned char *u_ptr, unsign
                              int y_stride, int uv_stride, loop_filter_info *lfi, int simpler_lpf)
 {
    (void) simpler_lpf;
-    vp8_mbloop_filter_horizontal_edge_sse2(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->thr, 2);
+    vp8_mbloop_filter_horizontal_edge_sse2(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->mbthr, 2);

    if (u_ptr)
-        vp8_mbloop_filter_horizontal_edge_uv_sse2(u_ptr, uv_stride, lfi->mbflim, lfi->lim, lfi->thr, v_ptr);
+        vp8_mbloop_filter_horizontal_edge_uv_sse2(u_ptr, uv_stride, lfi->uvmbflim, lfi->uvlim, lfi->uvmbthr, v_ptr);
 }


@@ -173,7 +173,7 @@ void vp8_loop_filter_mbhs_sse2(unsigned char *y_ptr, unsigned char *u_ptr, unsig
    (void) v_ptr;
    (void) uv_stride;
    (void) simpler_lpf;
-    vp8_loop_filter_simple_horizontal_edge_sse2(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->thr, 2);
+    vp8_loop_filter_simple_horizontal_edge_sse2(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->mbthr, 2);
 }


@@ -182,10 +182,10 @@ void vp8_loop_filter_mbv_sse2(unsigned char *y_ptr, unsigned char *u_ptr, unsign
                              int y_stride, int uv_stride, loop_filter_info *lfi, int simpler_lpf)
 {
    (void) simpler_lpf;
-    vp8_mbloop_filter_vertical_edge_sse2(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->thr, 2);
+    vp8_mbloop_filter_vertical_edge_sse2(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->mbthr, 2);

    if (u_ptr)
-        vp8_mbloop_filter_vertical_edge_uv_sse2(u_ptr, uv_stride, lfi->mbflim, lfi->lim, lfi->thr, v_ptr);
+        vp8_mbloop_filter_vertical_edge_uv_sse2(u_ptr, uv_stride, lfi->uvmbflim, lfi->uvlim, lfi->uvmbthr, v_ptr);
 }


@@ -196,7 +196,7 @@ void vp8_loop_filter_mbvs_sse2(unsigned char *y_ptr, unsigned char *u_ptr, unsig
    (void) v_ptr;
    (void) uv_stride;
    (void) simpler_lpf;
-    vp8_loop_filter_simple_vertical_edge_sse2(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->thr, 2);
+    vp8_loop_filter_simple_vertical_edge_sse2(y_ptr, y_stride, lfi->mbflim, lfi->lim, lfi->mbthr, 2);
 }


@@ -210,7 +210,7 @@ void vp8_loop_filter_bh_sse2(unsigned char *y_ptr, unsigned char *u_ptr, unsigne
    vp8_loop_filter_horizontal_edge_sse2(y_ptr + 12 * y_stride, y_stride, lfi->flim, lfi->lim, lfi->thr, 2);

    if (u_ptr)
-        vp8_loop_filter_horizontal_edge_uv_sse2(u_ptr + 4 * uv_stride, uv_stride, lfi->flim, lfi->lim, lfi->thr, v_ptr + 4 * uv_stride);
+        vp8_loop_filter_horizontal_edge_uv_sse2(u_ptr + 4 * uv_stride, uv_stride, lfi->uvflim, lfi->uvlim, lfi->uvthr, v_ptr + 4 * uv_stride);
 }


@@ -237,7 +237,7 @@ void vp8_loop_filter_bv_sse2(unsigned char *y_ptr, unsigned char *u_ptr, unsigne
    vp8_loop_filter_vertical_edge_sse2(y_ptr + 12, y_stride, lfi->flim, lfi->lim, lfi->thr, 2);

    if (u_ptr)
-        vp8_loop_filter_vertical_edge_uv_sse2(u_ptr + 4, uv_stride, lfi->flim, lfi->lim, lfi->thr, v_ptr + 4);
+        vp8_loop_filter_vertical_edge_uv_sse2(u_ptr + 4, uv_stride, lfi->uvflim, lfi->uvlim, lfi->uvthr, v_ptr + 4);
 }


--- a/vp8/common/x86/vp8_asm_stubs.c
+++ b/vp8/common/x86/vp8_asm_stubs.c
@@ -11,7 +11,7 @@

 #include "vpx_ports/config.h"
 #include "vpx_ports/mem.h"
-#include "vp8/common/subpixel.h"
+#include "subpixel.h"

 extern const short vp8_six_tap_mmx[8][6*8];
 extern const short vp8_bilinear_filters_mmx[8][2*8];
--- a/vp8/common/x86/x86_systemdependent.c
+++ b/vp8/common/x86/x86_systemdependent.c
@@ -11,13 +11,13 @@

 #include "vpx_ports/config.h"
 #include "vpx_ports/x86.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 "vp8/common/pragmas.h"
-#include "vp8/common/onyxc_int.h"
+#include "g_common.h"
+#include "subpixel.h"
+#include "loopfilter.h"
+#include "recon.h"
+#include "idct.h"
+#include "pragmas.h"
+#include "onyxc_int.h"

 void vp8_arch_x86_common_init(VP8_COMMON *ctx)
 {
--- a/vp8/decoder/arm/arm_dsystemdependent.c
+++ b/vp8/decoder/arm/arm_dsystemdependent.c
@@ -11,11 +11,12 @@

 #include "vpx_ports/config.h"
 #include "vpx_ports/arm.h"
-#include "vp8/common/blockd.h"
-#include "vp8/common/pragmas.h"
-#include "vp8/common/postproc.h"
-#include "vp8/decoder/dequantize.h"
-#include "vp8/decoder/onyxd_int.h"
+#include "blockd.h"
+#include "pragmas.h"
+#include "postproc.h"
+#include "dboolhuff.h"
+#include "dequantize.h"
+#include "onyxd_int.h"

 void vp8_arch_arm_decode_init(VP8D_COMP *pbi)
 {
@@ -34,6 +35,12 @@ void vp8_arch_arm_decode_init(VP8D_COMP *pbi)
        pbi->dequant.dc_idct_add_y_block = vp8_dequant_dc_idct_add_y_block_v6;
        pbi->dequant.idct_add_y_block    = vp8_dequant_idct_add_y_block_v6;
        pbi->dequant.idct_add_uv_block   = vp8_dequant_idct_add_uv_block_v6;
+#if 0 /*For use with RTCD, when implemented*/
+        pbi->dboolhuff.start             = vp8dx_start_decode_c;
+        pbi->dboolhuff.fill              = vp8dx_bool_decoder_fill_c;
+        pbi->dboolhuff.debool            = vp8dx_decode_bool_c;
+        pbi->dboolhuff.devalue           = vp8dx_decode_value_c;
+#endif
    }
 #endif

@@ -47,6 +54,12 @@ void vp8_arch_arm_decode_init(VP8D_COMP *pbi)
        pbi->dequant.dc_idct_add_y_block = vp8_dequant_dc_idct_add_y_block_neon;
        pbi->dequant.idct_add_y_block    = vp8_dequant_idct_add_y_block_neon;
        pbi->dequant.idct_add_uv_block   = vp8_dequant_idct_add_uv_block_neon;
+#if 0 /*For use with RTCD, when implemented*/
+        pbi->dboolhuff.start             = vp8dx_start_decode_c;
+        pbi->dboolhuff.fill              = vp8dx_bool_decoder_fill_c;
+        pbi->dboolhuff.debool            = vp8dx_decode_bool_c;
+        pbi->dboolhuff.devalue           = vp8dx_decode_value_c;
+#endif
    }
 #endif
 #endif
--- a/vp8/decoder/arm/armv6/dboolhuff_v6.asm
+++ b/vp8/decoder/arm/armv6/dboolhuff_v6.asm
@@ -0,0 +1,163 @@
+;
+;  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.
+;
+
+
+    EXPORT  |vp8_decode_value_v6|
+    EXPORT  |vp8dx_start_decode_v6|
+    EXPORT  |vp8dx_stop_decode_v6|
+    EXPORT  |vp8dx_decode_bool_v6|
+
+    ARM
+    REQUIRE8
+    PRESERVE8
+
+    INCLUDE vpx_asm_offsets.asm
+
+br      RN  r0
+prob    RN  r1
+bits    RN  r1
+    AREA    |.text|, CODE, READONLY  ; name this block of code
+
+;   int z = 0;
+;   int bit;
+;   for ( bit=bits-1; bit>=0; bit-- )
+;   {
+;       z |= (vp8dx_decode_bool(br, 0x80)<<bit);
+;   }
+;   return z;
+
+;int vp8_decode_value_v6 ( BOOL_DECODER *br, int bits )
+|vp8_decode_value_v6| PROC
+    stmdb   sp!, {r4 - r6, lr}
+    mov     r4, br
+    mov     r5, bits
+    mov     r6, #0
+
+    subs    r5, r5, #1
+    bmi     decode_value_exit
+
+decode_value_loop
+    mov     prob, #0x80
+    mov     br, r4
+    bl      vp8dx_decode_bool_v6_internal     ; needed for conversion to s file
+    orr     r6, r6, r0, lsl r5
+    subs    r5, r5, #1
+    bpl     decode_value_loop
+
+decode_value_exit
+    mov     r0, r6
+    ldmia   sp!, {r4 - r6, pc}
+    ENDP    ; |vp8_decode_value_v6|
+
+
+;void vp8dx_start_decode_v6 ( BOOL_DECODER *br, unsigned char *source )
+|vp8dx_start_decode_v6| PROC
+    stmdb   sp!, {r4 - r5, lr}
+    mov     r2, #0
+    mov     r3, #255
+
+    str     r2, [br, #bool_decoder_lowvalue]
+    str     r3, [br, #bool_decoder_range]
+    str     r1, [br, #bool_decoder_buffer]
+
+    mov     r3, #8
+    mov     r2, #4
+    str     r3, [br, #bool_decoder_count]
+    str     r2, [br, #bool_decoder_pos]
+
+    ldrb    r2, [r1, #3]
+    ldrb    r3, [r1, #2]
+    ldrb    r4, [r1, #1]
+    ldrb    r5, [r1]
+
+    orr     r1, r2, r3, lsl #8
+    orr     r1, r1, r4, lsl #16
+    orr     r1, r1, r5, lsl #24
+
+    str     r1, [br, #bool_decoder_value]
+
+    ldmia   sp!, {r4 - r5, pc}
+    ENDP    ; |vp8dx_start_decode_v6|
+
+
+;void vp8dx_stop_decode_v6 ( BOOL_DECODER *bc );
+|vp8dx_stop_decode_v6| PROC
+    mov     pc, lr
+    ENDP    ; |vp8dx_stop_decode_v6|
+
+
+; bigsplit  RN  r1
+; buffer_v  RN  r1
+; count_v       RN  r4
+; range_v       RN  r2
+; value_v       RN  r3
+; pos_v     RN  r5
+; split     RN  r6
+; bit           RN  lr
+;int vp8dx_decode_bool_v6 ( BOOL_DECODER *br, int probability )
+|vp8dx_decode_bool_v6| PROC
+vp8dx_decode_bool_v6_internal
+    stmdb   sp!, {r4 - r6, lr}
+
+    ldr     r2, [br, #bool_decoder_range]
+    ldr     r3, [br, #bool_decoder_value]
+
+    mov     r6, r2, lsl #8
+    sub     r6, r6, #256                ;   split = 1 +  (((range-1) * probability) >> 8)
+    mov     r12, #1
+    smlawb  r6, r6, prob, r12
+
+    mov     lr, #0
+    subs    r5, r3, r6, lsl #24
+
+    ;cmp        r3, r1
+    movhs   lr, #1
+    movhs   r3, r5
+    subhs   r2, r2, r6
+    movlo   r2, r6
+
+    cmp     r2, #0x80
+    blt     range_less_0x80
+    ;strd   r2, r3, [br, #bool_decoder_range]
+    str     r2, [br, #bool_decoder_range]
+    str     r3, [br, #bool_decoder_value]
+    mov     r0, lr
+    ldmia   sp!, {r4 - r6, pc}
+
+range_less_0x80
+    ldr     r5, [br, #bool_decoder_pos]
+    ldr     r1, [br, #bool_decoder_buffer]
+    ldr     r4, [br, #bool_decoder_count]
+    add     r1, r1, r5
+
+    clz       r12, r2
+    sub       r12, r12, #24
+    subs      r4, r4, r12
+    ldrleb    r6, [r1], #1
+    mov       r2, r2, lsl r12
+    mov       r3, r3, lsl r12
+    addle     r4, r4, #8
+    rsble     r12, r4, #8
+    addle     r5, r5, #1
+    orrle     r3, r3, r6, lsl r12
+
+    ;strd       r2, r3, [br, #bool_decoder_range]
+    ;strd       r4, r5, [br, #bool_decoder_count]
+    str         r2, [br, #bool_decoder_range]
+    str         r3, [br, #bool_decoder_value]
+    str         r4, [br, #bool_decoder_count]
+    str         r5, [br, #bool_decoder_pos]
+
+    mov     r0, lr
+
+    ldmia   sp!, {r4 - r6, pc}
+    ENDP    ; |vp8dx_decode_bool_v6|
+
+    END
--- a/vp8/decoder/arm/armv6/idct_blk_v6.c
+++ b/vp8/decoder/arm/armv6/idct_blk_v6.c
@@ -9,8 +9,8 @@
 */

 #include "vpx_ports/config.h"
-#include "vp8/common/idct.h"
-#include "vp8/decoder/dequantize.h"
+#include "idct.h"
+#include "dequantize.h"

 void vp8_dequant_dc_idct_add_y_block_v6
            (short *q, short *dq, unsigned char *pre,
--- a/vp8/decoder/arm/dboolhuff_arm.h
+++ b/vp8/decoder/arm/dboolhuff_arm.h
@@ -0,0 +1,43 @@
+#ifndef DBOOLHUFF_ARM_H
+#define DBOOLHUFF_ARM_H
+
+/* JLK
+ * There are currently no arm-optimized versions of
+ * these functions. As they are implemented, they
+ * can be uncommented below and added to
+ * arm/dsystemdependent.c
+ *
+ * The existing asm code is likely so different as
+ * to be useless. However, its been left (for now)
+ * for reference.
+ */
+#if 0
+#if HAVE_ARMV6
+#undef vp8_dbool_start
+#define vp8_dbool_start vp8dx_start_decode_v6
+
+#undef vp8_dbool_fill
+#define vp8_dbool_fill vp8_bool_decoder_fill_v6
+
+#undef vp8_dbool_debool
+#define vp8_dbool_debool vp8_decode_bool_v6
+
+#undef vp8_dbool_devalue
+#define vp8_dbool_devalue vp8_decode_value_v6
+#endif /* HAVE_ARMV6 */
+
+#if HAVE_ARMV7
+#undef vp8_dbool_start
+#define vp8_dbool_start vp8dx_start_decode_neon
+
+#undef vp8_dbool_fill
+#define vp8_dbool_fill vp8_bool_decoder_fill_neon
+
+#undef vp8_dbool_debool
+#define vp8_dbool_debool vp8_decode_bool_neon
+
+#undef vp8_dbool_devalue
+#define vp8_dbool_devalue vp8_decode_value_neon
+#endif /* HAVE_ARMV7 */
+#endif
+#endif /* DBOOLHUFF_ARM_H */
--- a/vp8/decoder/arm/dequantize_arm.c
+++ b/vp8/decoder/arm/dequantize_arm.c
@@ -10,8 +10,9 @@


 #include "vpx_ports/config.h"
-#include "vp8/decoder/dequantize.h"
-#include "vp8/common/idct.h"
+#include "dequantize.h"
+#include "predictdc.h"
+#include "idct.h"
 #include "vpx_mem/vpx_mem.h"

 #if HAVE_ARMV7
--- a/vp8/decoder/arm/detokenize.asm
+++ b/vp8/decoder/arm/detokenize.asm
@@ -0,0 +1,320 @@
+;
+;  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.
+;
+
+
+    EXPORT  |vp8_decode_mb_tokens_v6|
+
+    AREA    |.text|, CODE, READONLY  ; name this block of code
+
+    INCLUDE vpx_asm_offsets.asm
+
+l_qcoeff    EQU     0
+l_i         EQU     4
+l_type      EQU     8
+l_stop      EQU     12
+l_c         EQU     16
+l_l_ptr     EQU     20
+l_a_ptr     EQU     24
+l_bc        EQU     28
+l_coef_ptr  EQU     32
+l_stacksize EQU     64
+
+
+;; constant offsets -- these should be created at build time
+c_block2above_offset         EQU 25
+c_entropy_nodes              EQU 11
+c_dct_eob_token              EQU 11
+
+|vp8_decode_mb_tokens_v6| PROC
+    stmdb       sp!, {r4 - r11, lr}
+    sub         sp, sp, #l_stacksize
+    mov         r7, r1                      ; type
+    mov         r9, r0                      ; detoken
+
+    ldr         r1, [r9, #detok_current_bc]
+    ldr         r0, [r9, #detok_qcoeff_start_ptr]
+    mov         r11, #0                     ; i
+    mov         r3, #16                     ; stop
+
+    cmp         r7, #1                      ; type ?= 1
+    addeq       r11, r11, #24               ; i = 24
+    addeq       r3, r3, #8                  ; stop = 24
+    addeq       r0, r0, #3, 24              ; qcoefptr += 24*16
+
+    str         r0, [sp, #l_qcoeff]
+    str         r11, [sp, #l_i]
+    str         r7, [sp, #l_type]
+    str         r3, [sp, #l_stop]
+    str         r1, [sp, #l_bc]
+
+    add         lr, r9, r7, lsl #2          ; detoken + type*4
+
+    ldr         r8, [r1, #bool_decoder_user_buffer]
+
+    ldr         r10, [lr, #detok_coef_probs]
+    ldr         r5, [r1, #bool_decoder_count]
+    ldr         r6, [r1, #bool_decoder_range]
+    ldr         r4, [r1, #bool_decoder_value]
+
+    str         r10, [sp, #l_coef_ptr]
+
+BLOCK_LOOP
+    ldr         r3, [r9, #detok_ptr_block2leftabove]
+    ldr         r1, [r9, #detok_L]
+    ldr         r2, [r9, #detok_A]
+    ldrb        r12, [r3, r11]!             ; block2left[i]
+    ldrb        r3, [r3, #c_block2above_offset]; block2above[i]
+
+    cmp         r7, #0                      ; c = !type
+    moveq       r7, #1
+    movne       r7, #0
+
+    ldrb        r0, [r1, r12]!              ; *(L += block2left[i])
+    ldrb        r3, [r2, r3]!               ; *(A += block2above[i])
+    mov         lr, #c_entropy_nodes        ; ENTROPY_NODES = 11
+
+; VP8_COMBINEENTROPYCONTETEXTS(t, *a, *l) => t = ((*a) != 0) + ((*l) !=0)
+    cmp         r0, #0                      ; *l ?= 0
+    movne       r0, #1
+    cmp         r3, #0                      ; *a ?= 0
+    addne       r0, r0, #1                  ; t
+
+    str         r1, [sp, #l_l_ptr]          ; save &l
+    str         r2, [sp, #l_a_ptr]          ; save &a
+    smlabb      r0, r0, lr, r10             ; Prob = coef_probs + (t * ENTROPY_NODES)
+    mov         r1, #0                      ; t = 0
+    str         r7, [sp, #l_c]
+
+    ;align 4
+COEFF_LOOP
+    ldr         r3, [r9, #detok_ptr_coef_bands_x]
+    ldr         lr, [r9, #detok_coef_tree_ptr]
+    ;STALL
+    ldrb        r3, [r3, r7]                ; coef_bands_x[c]
+    ;STALL
+    ;STALL
+    add         r0, r0, r3                  ; Prob += coef_bands_x[c]
+
+get_token_loop
+    ldrb        r2, [r0, +r1, asr #1]       ; Prob[t >> 1]
+    mov         r3, r6, lsl #8              ; range << 8
+    sub         r3, r3, #256                ; (range << 8) - (1 << 8)
+    mov         r10, #1                     ; 1
+
+    smlawb      r2, r3, r2, r10             ; split = 1 + (((range-1) * probability) >> 8)
+
+    ldrb        r12, [r8]                   ; load cx data byte in stall slot : r8 = bufptr
+    ;++
+
+    subs        r3, r4, r2, lsl #24         ; value-(split<<24): used later to calculate shift for NORMALIZE
+    addhs       r1, r1, #1                  ; t += 1
+    movhs       r4, r3                      ; value -= bigsplit (split << 24)
+    subhs       r2, r6, r2                  ; range -= split
+ ;   movlo       r6, r2                      ; range = split
+
+    ldrsb     r1, [lr, r1]                  ; t = onyx_coef_tree_ptr[t]
+
+; NORMALIZE
+    clz         r3, r2                      ; vp8dx_bitreader_norm[range] + 24
+    sub         r3, r3, #24                 ; vp8dx_bitreader_norm[range]
+    subs        r5, r5, r3                  ; count -= shift
+    mov         r6, r2, lsl r3              ; range <<= shift
+    mov         r4, r4, lsl r3              ; value <<= shift
+
+; if count <= 0, += BR_COUNT; value |= *bufptr++ << (BR_COUNT-count); BR_COUNT = 8, but need to upshift values by +16
+    addle         r5, r5, #8                ; count += 8
+    rsble         r3, r5, #24               ; 24 - count
+    addle         r8, r8, #1                ; bufptr++
+    orrle         r4, r4, r12, lsl r3       ; value |= *bufptr << shift + 16
+
+    cmp         r1, #0                      ; t ?= 0
+    bgt         get_token_loop              ; while (t > 0)
+
+    cmn         r1, #c_dct_eob_token        ; if(t == -DCT_EOB_TOKEN)
+    beq         END_OF_BLOCK                ; break
+
+    rsb         lr, r1, #0                  ; v = -t;
+
+    cmp         lr, #4                      ; if(v > FOUR_TOKEN)
+    ble         SKIP_EXTRABITS
+
+    ldr         r3, [r9, #detok_teb_base_ptr]
+    mov         r11, #1                     ; 1 in split = 1 + ... nope, v+= 1 << bits_count
+    add         r7, r3, lr, lsl #4          ; detok_teb_base_ptr + (v << 4)
+
+    ldrsh       lr, [r7, #tokenextrabits_min_val] ; v = teb_ptr->min_val
+    ldrsh       r0, [r7, #tokenextrabits_length] ; bits_count = teb_ptr->Length
+
+extrabits_loop
+    add         r3, r0, r7                  ; &teb_ptr->Probs[bits_count]
+
+    ldrb        r2, [r3, #4]                ; probability. why +4?
+    mov         r3, r6, lsl #8              ; range << 8
+    sub         r3, r3, #256                ; range << 8 + 1 << 8
+
+    smlawb      r2, r3, r2, r11             ; split = 1 +  (((range-1) * probability) >> 8)
+
+    ldrb        r12, [r8]                   ; *bufptr
+    ;++
+
+    subs        r10, r4, r2, lsl #24        ; value - (split<<24)
+    movhs       r4, r10                     ; value = value - (split << 24)
+    subhs       r2, r6, r2                  ; range = range - split
+    addhs       lr, lr, r11, lsl r0         ; v += ((UINT16)1<<bits_count)
+
+; NORMALIZE
+    clz         r3, r2                      ; shift - leading zeros in split
+    sub         r3, r3, #24                 ; don't count first 3 bytes
+    subs        r5, r5, r3                  ; count -= shift
+    mov         r6, r2, lsl r3              ; range = range << shift
+    mov         r4, r4, lsl r3              ; value <<= shift
+
+    addle       r5, r5, #8                  ; count += BR_COUNT
+    addle       r8, r8, #1                  ; bufptr++
+    rsble       r3, r5, #24                 ; BR_COUNT - count
+    orrle       r4, r4, r12, lsl r3         ; value |= *bufptr << (BR_COUNT - count)
+
+    subs        r0, r0, #1                  ; bits_count --
+    bpl         extrabits_loop
+
+
+SKIP_EXTRABITS
+    ldr         r11, [sp, #l_qcoeff]
+    ldr         r0, [sp, #l_coef_ptr]       ; Prob = coef_probs
+
+    cmp         r1, #0                      ; check for nonzero token - if (t)
+    beq         SKIP_EOB_CHECK              ; if t is zero, we will skip the eob table chec
+
+    add         r3, r6, #1                  ; range + 1
+    mov         r2, r3, lsr #1              ; split = (range + 1) >> 1
+
+    subs        r3, r4, r2, lsl #24         ; value - (split<<24)
+    movhs       r4, r3                      ; value -= (split << 24)
+    subhs       r2, r6, r2                  ; range -= split
+    mvnhs       r3, lr                      ; -v
+    addhs       lr, r3, #1                  ; v = (v ^ -1) + 1
+
+; NORMALIZE
+    clz         r3, r2                      ; leading 0s in split
+    sub         r3, r3, #24                 ; shift
+    subs        r5, r5, r3                  ; count -= shift
+    mov         r6, r2, lsl r3              ; range <<= shift
+    mov         r4, r4, lsl r3              ; value <<= shift
+    ldrleb      r2, [r8], #1                ; *(bufptr++)
+    addle       r5, r5, #8                  ; count += 8
+    rsble       r3, r5, #24                 ; BR_COUNT - count
+    orrle       r4, r4, r2, lsl r3          ; value |= *bufptr << (BR_COUNT - count)
+
+    add         r0, r0, #11                 ; Prob += ENTROPY_NODES (11)
+
+    cmn         r1, #1                      ; t < -ONE_TOKEN
+
+    addlt       r0, r0, #11                 ; Prob += ENTROPY_NODES (11)
+
+    mvn         r1, #1                      ; t = -1 ???? C is -2
+
+SKIP_EOB_CHECK
+    ldr         r7, [sp, #l_c]              ; c
+    ldr         r3, [r9, #detok_scan]
+    add         r1, r1, #2                  ; t+= 2
+    cmp         r7, #15                     ; c should will be one higher
+
+    ldr         r3, [r3, +r7, lsl #2]       ; scan[c] this needs pre-inc c value
+    add         r7, r7, #1                  ; c++
+    add         r3, r11, r3, lsl #1         ; qcoeff + scan[c]
+
+    str         r7, [sp, #l_c]              ; store c
+    strh        lr, [r3]                    ; qcoef_ptr[scan[c]] = v
+
+    blt         COEFF_LOOP
+
+    sub         r7, r7, #1                  ; if(t != -DCT_EOB_TOKEN) --c
+
+END_OF_BLOCK
+    ldr         r3, [sp, #l_type]           ; type
+    ldr         r10, [sp, #l_coef_ptr]      ; coef_ptr
+    ldr         r0, [sp, #l_qcoeff]         ; qcoeff
+    ldr         r11, [sp, #l_i]             ; i
+    ldr         r12, [sp, #l_stop]          ; stop
+
+    cmp         r3, #0                      ; type ?= 0
+    moveq       r1, #1
+    movne       r1, #0
+    add         r3, r11, r9                 ; detok + i
+
+    cmp         r7, r1                      ; c ?= !type
+    strb        r7, [r3, #detok_eob]        ; eob[i] = c
+
+    ldr         r7, [sp, #l_l_ptr]          ; l
+    ldr         r2, [sp, #l_a_ptr]          ; a
+    movne       r3, #1                      ; t
+    moveq       r3, #0
+
+    add         r0, r0, #32                 ; qcoeff += 32 (16 * 2?)
+    add         r11, r11, #1                ; i++
+    strb        r3, [r7]                    ; *l = t
+    strb        r3, [r2]                    ; *a = t
+    str         r0, [sp, #l_qcoeff]         ; qcoeff
+    str         r11, [sp, #l_i]             ; i
+
+    cmp         r11, r12                    ; i < stop
+    ldr         r7, [sp, #l_type]           ; type
+
+    blt         BLOCK_LOOP
+
+    cmp         r11, #25                    ; i ?= 25
+    bne         ln2_decode_mb_to
+
+    ldr         r12, [r9, #detok_qcoeff_start_ptr]
+    ldr         r10, [r9, #detok_coef_probs]
+    mov         r7, #0                      ; type/i = 0
+    mov         r3, #16                     ; stop = 16
+    str         r12, [sp, #l_qcoeff]        ; qcoeff_ptr = qcoeff_start_ptr
+    str         r7, [sp, #l_i]
+    str         r7, [sp, #l_type]
+    str         r3, [sp, #l_stop]
+
+    str         r10, [sp, #l_coef_ptr]      ; coef_probs = coef_probs[type=0]
+
+    b           BLOCK_LOOP
+
+ln2_decode_mb_to
+    cmp         r11, #16                    ; i ?= 16
+    bne         ln1_decode_mb_to
+
+    mov         r10, #detok_coef_probs
+    add         r10, r10, #2*4              ; coef_probs[type]
+    ldr         r10, [r9, r10]              ; detok + detok_coef_probs[type]
+
+    mov         r7, #2                      ; type = 2
+    mov         r3, #24                     ; stop = 24
+
+    str         r7, [sp, #l_type]
+    str         r3, [sp, #l_stop]
+
+    str         r10, [sp, #l_coef_ptr]      ; coef_probs = coef_probs[type]
+    b           BLOCK_LOOP
+
+ln1_decode_mb_to
+    ldr         r2, [sp, #l_bc]
+    mov         r0, #0
+    nop
+
+    str         r8, [r2, #bool_decoder_user_buffer]
+    str         r5, [r2, #bool_decoder_count]
+    str         r4, [r2, #bool_decoder_value]
+    str         r6, [r2, #bool_decoder_range]
+
+    add         sp, sp, #l_stacksize
+    ldmia       sp!, {r4 - r11, pc}
+
+    ENDP  ; |vp8_decode_mb_tokens_v6|
+
+    END
--- a/vp8/decoder/arm/detokenize_arm.h
+++ b/vp8/decoder/arm/detokenize_arm.h
@@ -0,0 +1,22 @@
+/*
+ *  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 DETOKENIZE_ARM_H
+#define DETOKENIZE_ARM_H
+
+#if HAVE_ARMV6
+#if CONFIG_ARM_ASM_DETOK
+void vp8_init_detokenizer(VP8D_COMP *dx);
+void vp8_decode_mb_tokens_v6(DETOK *detoken, int type);
+#endif
+#endif
+
+#endif
--- a/vp8/decoder/arm/neon/dboolhuff_neon.asm
+++ b/vp8/decoder/arm/neon/dboolhuff_neon.asm
@@ -0,0 +1,160 @@
+;
+;  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.
+;
+
+
+    EXPORT  |vp8_decode_value_neon|
+    EXPORT  |vp8dx_start_decode_neon|
+    EXPORT  |vp8dx_stop_decode_neon|
+    EXPORT  |vp8dx_decode_bool_neon|
+
+    ARM
+    REQUIRE8
+    PRESERVE8
+
+    INCLUDE vpx_asm_offsets.asm
+
+    AREA    |.text|, CODE, READONLY  ; name this block of code
+
+;   int z = 0;
+;   int bit;
+;   for ( bit=bits-1; bit>=0; bit-- )
+;   {
+;       z |= (vp8dx_decode_bool(br, 0x80)<<bit);
+;   }
+;   return z;
+
+;int vp8_decode_value_neon ( BOOL_DECODER *br, int bits )
+|vp8_decode_value_neon| PROC
+    stmdb   sp!, {r4 - r6, lr}
+    mov     r4, r0
+    mov     r5, r1
+    mov     r6, #0
+
+    subs    r5, r5, #1
+    bmi     decode_value_exit
+
+decode_value_loop
+    mov     r1, #0x80
+    mov     r0, r4
+    bl      vp8dx_decode_bool_neon_internal       ; needed for conversion to s file
+    orr     r6, r6, r0, lsl r5
+    subs    r5, r5, #1
+    bpl     decode_value_loop
+
+decode_value_exit
+    mov     r0, r6
+    ldmia   sp!, {r4 - r6, pc}
+    ENDP    ; |vp8_decode_value_neon|
+
+
+;void vp8dx_start_decode_neon ( BOOL_DECODER *br, unsigned char *source )
+|vp8dx_start_decode_neon| PROC
+    stmdb   sp!, {r4 - r5, lr}
+    mov     r2, #0
+    mov     r3, #255
+
+    str     r2, [r0, #bool_decoder_lowvalue]
+    str     r3, [r0, #bool_decoder_range]
+    str     r1, [r0, #bool_decoder_buffer]
+
+    mov     r3, #8
+    mov     r2, #4
+    str     r3, [r0, #bool_decoder_count]
+    str     r2, [r0, #bool_decoder_pos]
+
+    ldrb    r2, [r1, #3]
+    ldrb    r3, [r1, #2]
+    ldrb    r4, [r1, #1]
+    ldrb    r5, [r1]
+
+    orr     r1, r2, r3, lsl #8
+    orr     r1, r1, r4, lsl #16
+    orr     r1, r1, r5, lsl #24
+
+    str     r1, [r0, #bool_decoder_value]
+
+    ldmia   sp!, {r4 - r5, pc}
+    ENDP    ; |vp8dx_start_decode_neon|
+
+
+;void vp8dx_stop_decode_neon ( BOOL_DECODER *bc );
+|vp8dx_stop_decode_neon| PROC
+    mov     pc, lr
+    ENDP    ; |vp8dx_stop_decode_neon|
+
+
+; bigsplit  RN  r1
+; buffer_v  RN  r1
+; count_v       RN  r4
+; range_v       RN  r2
+; value_v       RN  r3
+; pos_v     RN  r5
+; split     RN  r6
+; bit           RN  lr
+;int vp8dx_decode_bool_neon ( BOOL_DECODER *br, int probability )
+|vp8dx_decode_bool_neon| PROC
+vp8dx_decode_bool_neon_internal
+;LDRD and STRD doubleword data transfers must be eight-byte aligned. Use ALIGN 8
+;before memory allocation
+    stmdb   sp!, {r4 - r5, lr}
+
+    ldr     r2, [r0, #bool_decoder_range]       ;load range (r2), value(r3)
+    ldr     r3, [r0, #bool_decoder_value]
+    ;ldrd   r2, r3, [r0, #bool_decoder_range]   ;ldrd costs 2 cycles
+    ;
+
+    mov     r4, r2, lsl #8
+    sub     r4, r4, #256
+    mov     r12, #1
+
+    smlawb  r4, r4, r1, r12         ;split = 1 +  (((range-1) * probability) >> 8)
+
+    mov     lr, r0
+    mov     r0, #0                  ;bit = 0
+    ;
+    subs    r5, r3, r4, lsl #24
+
+    subhs   r2, r2, r4              ;range = br->range-split
+    movlo   r2, r4                  ;range = split
+    movhs   r0, #1                  ;bit = 1
+    movhs   r3, r5                  ;value = value-bigsplit
+
+    cmp     r2, #0x80
+    blt     range_less_0x80
+    strd    r2, r3, [lr, #bool_decoder_range]   ;store result
+
+    ldmia   sp!, {r4 - r5, pc}
+
+range_less_0x80
+
+    ldrd    r4, r5, [lr, #bool_decoder_count]   ;load count, pos, buffer
+    ldr     r1, [lr, #bool_decoder_buffer]
+
+    clz     r12, r2
+    add     r1, r1, r5
+
+    sub     r12, r12, #24
+    subs    r4, r4, r12             ;count -= shift
+    mov     r2, r2, lsl r12         ;range <<= shift
+    mov     r3, r3, lsl r12         ;value <<= shift
+    addle   r4, r4, #8              ;count += 8
+    ldrleb  r12, [r1], #1           ;br->buffer[br->pos]
+
+    rsble   r1, r4, #8              ;-count
+    addle   r5, r5, #1              ;br->pos++
+    orrle   r3, r3, r12, lsl r1     ;value |= (br->buffer[br->pos]) << (-count)
+
+    strd    r2, r3, [lr, #bool_decoder_range]   ;store result
+    strd    r4, r5, [lr, #bool_decoder_count]
+
+    ldmia   sp!, {r4 - r5, pc}
+    ENDP    ; |vp8dx_decode_bool_neon|
+
+    END
--- a/vp8/decoder/arm/neon/idct_blk_neon.c
+++ b/vp8/decoder/arm/neon/idct_blk_neon.c
@@ -9,8 +9,8 @@
 */

 #include "vpx_ports/config.h"
-#include "vp8/common/idct.h"
-#include "vp8/decoder/dequantize.h"
+#include "idct.h"
+#include "dequantize.h"

 /* place these declarations here because we don't want to maintain them
 * outside of this scope
--- a/vp8/decoder/dboolhuff.c
+++ b/vp8/decoder/dboolhuff.c
@@ -26,9 +26,8 @@ DECLARE_ALIGNED(16, const unsigned char, vp8dx_bitreader_norm[256]) =
 };


-int vp8dx_start_decode(BOOL_DECODER *br,
-                       const unsigned char *source,
-                       unsigned int source_sz)
+int vp8dx_start_decode_c(BOOL_DECODER *br, const unsigned char *source,
+                        unsigned int source_sz)
 {
    br->user_buffer_end = source+source_sz;
    br->user_buffer     = source;
@@ -40,13 +39,13 @@ int vp8dx_start_decode(BOOL_DECODER *br,
        return 1;

    /* Populate the buffer */
-    vp8dx_bool_decoder_fill(br);
+    vp8dx_bool_decoder_fill_c(br);

    return 0;
 }


-void vp8dx_bool_decoder_fill(BOOL_DECODER *br)
+void vp8dx_bool_decoder_fill_c(BOOL_DECODER *br)
 {
    const unsigned char *bufptr;
    const unsigned char *bufend;
@@ -63,3 +62,69 @@ void vp8dx_bool_decoder_fill(BOOL_DECODER *br)
    br->value = value;
    br->count = count;
 }
+
+#if 0
+/*
+ * Until optimized versions of these functions are available, we
+ * keep the implementation in the header to allow inlining.
+ *
+ * The RTCD-style invocations are still in place so this can
+ * be switched by just uncommenting these functions here and
+ * the DBOOLHUFF_INVOKE calls in the header.
+ */
+int vp8dx_decode_bool_c(BOOL_DECODER *br, int probability)
+{
+    unsigned int bit=0;
+    VP8_BD_VALUE value;
+    unsigned int split;
+    VP8_BD_VALUE bigsplit;
+    int count;
+    unsigned int range;
+
+    value = br->value;
+    count = br->count;
+    range = br->range;
+
+    split = 1 + (((range-1) * probability) >> 8);
+    bigsplit = (VP8_BD_VALUE)split << (VP8_BD_VALUE_SIZE - 8);
+
+    range = split;
+    if(value >= bigsplit)
+    {
+        range = br->range-split;
+        value = value-bigsplit;
+        bit = 1;
+    }
+
+    /*if(range>=0x80)
+    {
+        br->value = value;
+        br->range = range;
+        return bit;
+    }*/
+
+    {
+        register unsigned int shift = vp8dx_bitreader_norm[range];
+        range <<= shift;
+        value <<= shift;
+        count -= shift;
+    }
+    br->value = value;
+    br->count = count;
+    br->range = range;
+    if (count < 0)
+        vp8dx_bool_decoder_fill_c(br);
+    return bit;
+}
+
+int vp8dx_decode_value_c(BOOL_DECODER *br, int bits)
+{
+    int z = 0;
+    int bit;
+    for ( bit=bits-1; bit>=0; bit-- )
+    {
+        z |= (vp8dx_decode_bool(br, 0x80)<<bit);
+    }
+    return z;
+}
+#endif
--- a/vp8/decoder/dboolhuff.h
+++ b/vp8/decoder/dboolhuff.h
@@ -25,6 +25,10 @@ typedef size_t VP8_BD_VALUE;
  Even relatively modest values like 100 would work fine.*/
 # define VP8_LOTS_OF_BITS (0x40000000)

+
+
+struct vp8_dboolhuff_rtcd_vtable;
+
 typedef struct
 {
    const unsigned char *user_buffer_end;
@@ -32,15 +36,82 @@ typedef struct
    VP8_BD_VALUE         value;
    int                  count;
    unsigned int         range;
+#if CONFIG_RUNTIME_CPU_DETECT
+    struct vp8_dboolhuff_rtcd_vtable *rtcd;
+#endif
 } BOOL_DECODER;

+#define prototype_dbool_start(sym) int sym(BOOL_DECODER *br, \
+    const unsigned char *source, unsigned int source_sz)
+#define prototype_dbool_fill(sym) void sym(BOOL_DECODER *br)
+#define prototype_dbool_debool(sym) int sym(BOOL_DECODER *br, int probability)
+#define prototype_dbool_devalue(sym) int sym(BOOL_DECODER *br, int bits)
+
+#if ARCH_ARM
+#include "arm/dboolhuff_arm.h"
+#endif
+
+#ifndef vp8_dbool_start
+#define vp8_dbool_start vp8dx_start_decode_c
+#endif
+
+#ifndef vp8_dbool_fill
+#define vp8_dbool_fill vp8dx_bool_decoder_fill_c
+#endif
+
+#ifndef vp8_dbool_debool
+#define vp8_dbool_debool vp8dx_decode_bool_c
+#endif
+
+#ifndef vp8_dbool_devalue
+#define vp8_dbool_devalue vp8dx_decode_value_c
+#endif
+
+extern prototype_dbool_start(vp8_dbool_start);
+extern prototype_dbool_fill(vp8_dbool_fill);
+extern prototype_dbool_debool(vp8_dbool_debool);
+extern prototype_dbool_devalue(vp8_dbool_devalue);
+
+typedef prototype_dbool_start((*vp8_dbool_start_fn_t));
+typedef prototype_dbool_fill((*vp8_dbool_fill_fn_t));
+typedef prototype_dbool_debool((*vp8_dbool_debool_fn_t));
+typedef prototype_dbool_devalue((*vp8_dbool_devalue_fn_t));
+
+typedef struct vp8_dboolhuff_rtcd_vtable {
+    vp8_dbool_start_fn_t   start;
+    vp8_dbool_fill_fn_t    fill;
+    vp8_dbool_debool_fn_t  debool;
+    vp8_dbool_devalue_fn_t devalue;
+} vp8_dboolhuff_rtcd_vtable_t;
+
+/* There are no processor-specific versions of these
+ * functions right now. Disable RTCD to avoid using
+ * function pointers which gives a speed boost
+ */
+/*#ifdef ENABLE_RUNTIME_CPU_DETECT
+#define DBOOLHUFF_INVOKE(ctx,fn) (ctx)->fn
+#define IF_RTCD(x) (x)
+#else*/
+#define DBOOLHUFF_INVOKE(ctx,fn) vp8_dbool_##fn
+#define IF_RTCD(x) NULL
+/*#endif*/
+
 DECLARE_ALIGNED(16, extern const unsigned char, vp8dx_bitreader_norm[256]);

-int vp8dx_start_decode(BOOL_DECODER *br,
-                       const unsigned char *source,
-                       unsigned int source_sz);
-
-void vp8dx_bool_decoder_fill(BOOL_DECODER *br);
+/* wrapper functions to hide RTCD. static means inline means hopefully no
+ * penalty
+ */
+static int vp8dx_start_decode(BOOL_DECODER *br,
+        struct vp8_dboolhuff_rtcd_vtable *rtcd,
+        const unsigned char *source, unsigned int source_sz) {
+#if CONFIG_RUNTIME_CPU_DETECT
+    br->rtcd = rtcd;
+#endif
+    return DBOOLHUFF_INVOKE(rtcd, start)(br, source, source_sz);
+}
+static void vp8dx_bool_decoder_fill(BOOL_DECODER *br) {
+    DBOOLHUFF_INVOKE(br->rtcd, fill)(br);
+}

 /*The refill loop is used in several places, so define it in a macro to make
   sure they're all consistent.
@@ -55,7 +126,7 @@ void vp8dx_bool_decoder_fill(BOOL_DECODER *br);
        for(shift = VP8_BD_VALUE_SIZE - 8 - ((_count) + 8); shift >= 0; ) \
        { \
            if((_bufptr) >= (_bufend)) { \
-                (_count) = VP8_LOTS_OF_BITS; \
+                (_count) += VP8_LOTS_OF_BITS; \
                break; \
            } \
            (_count) += 8; \
@@ -67,6 +138,12 @@ void vp8dx_bool_decoder_fill(BOOL_DECODER *br);


 static int vp8dx_decode_bool(BOOL_DECODER *br, int probability) {
+  /*
+   * Until optimized versions of this function are available, we
+   * keep the implementation in the header to allow inlining.
+   *
+   *return DBOOLHUFF_INVOKE(br->rtcd, debool)(br, probability);
+   */
    unsigned int bit = 0;
    VP8_BD_VALUE value;
    unsigned int split;
@@ -90,6 +167,13 @@ static int vp8dx_decode_bool(BOOL_DECODER *br, int probability) {
        bit = 1;
    }

+    /*if(range>=0x80)
+    {
+        br->value = value;
+        br->range = range;
+        return bit
+    }*/
+
    {
        register unsigned int shift = vp8dx_bitreader_norm[range];
        range <<= shift;
@@ -106,6 +190,12 @@ static int vp8dx_decode_bool(BOOL_DECODER *br, int probability) {

 static int vp8_decode_value(BOOL_DECODER *br, int bits)
 {
+  /*
+   * Until optimized versions of this function are available, we
+   * keep the implementation in the header to allow inlining.
+   *
+   *return DBOOLHUFF_INVOKE(br->rtcd, devalue)(br, bits);
+   */
    int z = 0;
    int bit;

@@ -119,18 +209,19 @@ static int vp8_decode_value(BOOL_DECODER *br, int bits)

 static int vp8dx_bool_error(BOOL_DECODER *br)
 {
-  /* Check if we have reached the end of the buffer.
-   *
-   * Variable 'count' stores the number of bits in the 'value' buffer,
-   * minus 8. So if count == 8, there are 16 bits available to be read.
-   * Normally, count is filled with 8 and one byte is filled into the
-   * value buffer. When we reach the end of the buffer, count is instead
-   * filled with VP8_LOTS_OF_BITS, 8 of which represent the last 8 real
-   * bits from the bitstream. So the last bit in the bitstream will be
-   * represented by count == VP8_LOTS_OF_BITS - 16.
-   */
-    if ((br->count > VP8_BD_VALUE_SIZE)
-        && (br->count <= VP8_LOTS_OF_BITS - 16))
+    /* Check if we have reached the end of the buffer.
+     *
+     * Variable 'count' stores the number of bits in the 'value' buffer, minus
+     * 8. The top byte is part of the algorithm, and the remainder is buffered
+     * to be shifted into it. So if count == 8, the top 16 bits of 'value' are
+     * occupied, 8 for the algorithm and 8 in the buffer.
+     *
+     * When reading a byte from the user's buffer, count is filled with 8 and
+     * one byte is filled into the value buffer. When we reach the end of the
+     * data, count is additionally filled with VP8_LOTS_OF_BITS. So when
+     * count == VP8_LOTS_OF_BITS - 1, the user's data has been exhausted.
+     */
+    if ((br->count > VP8_BD_VALUE_SIZE) && (br->count < VP8_LOTS_OF_BITS))
    {
       /* We have tried to decode bits after the end of
        * stream was encountered.
--- a/vp8/decoder/decodemv.c
+++ b/vp8/decoder/decodemv.c
@@ -10,10 +10,10 @@


 #include "treereader.h"
-#include "vp8/common/entropymv.h"
-#include "vp8/common/entropymode.h"
+#include "entropymv.h"
+#include "entropymode.h"
 #include "onyxd_int.h"
-#include "vp8/common/findnearmv.h"
+#include "findnearmv.h"

 #if CONFIG_DEBUG
 #include <assert.h>
@@ -251,6 +251,7 @@ void vp8_mb_mode_mv_init(VP8D_COMP *pbi)
    vp8_reader *const bc = & pbi->bc;
    MV_CONTEXT *const mvc = pbi->common.fc.mvc;

+    pbi->mvs_corrupt_from_mb = -1;
    pbi->prob_skip_false = 0;
    if (pbi->common.mb_no_coeff_skip)
        pbi->prob_skip_false = (vp8_prob)vp8_read_literal(bc, 8);
@@ -412,7 +413,7 @@ void vp8_read_mb_modes_mv(VP8D_COMP *pbi, MODE_INFO *mi, MB_MODE_INFO *mbmi,

                    do {
                        mi->bmi[ *fill_offset] = bmi;
-                      fill_offset++;
+                        fill_offset++;

                    }while (--fill_count);
                }
@@ -543,12 +544,18 @@ void vp8_decode_mode_mvs(VP8D_COMP *pbi)

        while (++mb_col < pbi->common.mb_cols)
        {
+            int mb_num = mb_row * pbi->common.mb_cols + mb_col;
            /*vp8_read_mb_modes_mv(pbi, xd->mode_info_context, &xd->mode_info_context->mbmi, mb_row, mb_col);*/
            if(pbi->common.frame_type == KEY_FRAME)
                vp8_kfread_modes(pbi, mi, mb_row, mb_col);
            else
                vp8_read_mb_modes_mv(pbi, mi, &mi->mbmi, mb_row, mb_col);

+            /* look for corruption */
+            if (vp8dx_bool_error(&pbi->bc) && mb_num < pbi->mvs_corrupt_from_mb)
+                pbi->mvs_corrupt_from_mb = mb_num;
+
+
            mi++;       /* next macroblock */
        }

--- a/vp8/decoder/decodframe.c
+++ b/vp8/decoder/decodframe.c
@@ -10,27 +10,29 @@


 #include "onyxd_int.h"
-#include "vp8/common/header.h"
-#include "vp8/common/reconintra.h"
-#include "vp8/common/reconintra4x4.h"
-#include "vp8/common/recon.h"
-#include "vp8/common/reconinter.h"
+#include "header.h"
+#include "reconintra.h"
+#include "reconintra4x4.h"
+#include "recon.h"
+#include "reconinter.h"
 #include "dequantize.h"
 #include "detokenize.h"
-#include "vp8/common/invtrans.h"
-#include "vp8/common/alloccommon.h"
-#include "vp8/common/entropymode.h"
-#include "vp8/common/quant_common.h"
+#include "invtrans.h"
+#include "alloccommon.h"
+#include "entropymode.h"
+#include "quant_common.h"
 #include "vpx_scale/vpxscale.h"
 #include "vpx_scale/yv12extend.h"
-#include "vp8/common/setupintrarecon.h"
+#include "setupintrarecon.h"

 #include "decodemv.h"
-#include "vp8/common/extend.h"
+#include "extend.h"
+#include "error_concealment.h"
 #include "vpx_mem/vpx_mem.h"
-#include "vp8/common/idct.h"
+#include "idct.h"
 #include "dequantize.h"
-#include "vp8/common/threading.h"
+#include "predictdc.h"
+#include "threading.h"
 #include "decoderthreading.h"
 #include "dboolhuff.h"

@@ -175,11 +177,21 @@ void clamp_mvs(MACROBLOCKD *xd)

 }

-void vp8_decode_macroblock(VP8D_COMP *pbi, MACROBLOCKD *xd)
+void vp8_decode_macroblock(VP8D_COMP *pbi, MACROBLOCKD *xd, unsigned int mb_idx)
 {
    int eobtotal = 0;
    int i, do_clamp = xd->mode_info_context->mbmi.need_to_clamp_mvs;

+    /* TODO(holmer): change when we have MB level error tracking
+     * The residuals may not match the predicted signal when a macroblock is
+     * corrupted due to previous losses. Should we try to add the residual
+     * anyway, or just throw it away? Should test this on a couple of files.
+     */
+    if (pbi->ec_enabled && mb_idx >= pbi->mvs_corrupt_from_mb)
+    {
+        xd->mode_info_context->mbmi.mb_skip_coeff = 1;
+    }
+
    if (xd->mode_info_context->mbmi.mb_skip_coeff)
    {
        vp8_reset_mb_tokens_context(xd);
@@ -332,6 +344,7 @@ void vp8_decode_mb_row(VP8D_COMP *pbi,
    int dst_fb_idx = pc->new_fb_idx;
    int recon_y_stride = pc->yv12_fb[ref_fb_idx].y_stride;
    int recon_uv_stride = pc->yv12_fb[ref_fb_idx].uv_stride;
+    int corrupt_partition = 0;

    vpx_memset(&pc->left_context, 0, sizeof(pc->left_context));
    recon_yoffset = mb_row * recon_y_stride * 16;
@@ -346,6 +359,28 @@ void vp8_decode_mb_row(VP8D_COMP *pbi,

    for (mb_col = 0; mb_col < pc->mb_cols; mb_col++)
    {
+        /* Distance of Mb to the various image edges.
+         * These are specified to 8th pel as they are always compared to values that are in 1/8th pel units
+         */
+        xd->mb_to_left_edge = -((mb_col * 16) << 3);
+        xd->mb_to_right_edge = ((pc->mb_cols - 1 - mb_col) * 16) << 3;
+
+        if (xd->mode_info_context->mbmi.ref_frame == INTRA_FRAME &&
+            corrupt_partition)
+        {
+            /* We have an intra block with corrupt coefficients,
+             * better to conceal with an inter block. Interpolate MVs
+             * from neighboring MBs.
+             */
+            /* TODO(holmer): We will fail to conceal an intra block with missing
+             * coefficients if it's the first block with missing coefficients,
+             * since the bool dec error detection is done after reconstruction.
+             */
+            vp8_interpolate_motion(xd,
+                                   mb_row, mb_col,
+                                   pc->mb_rows, pc->mb_cols,
+                                   pc->mode_info_stride);
+        }

        if (xd->mode_info_context->mbmi.mode == SPLITMV || xd->mode_info_context->mbmi.mode == B_PRED)
        {
@@ -356,12 +391,6 @@ void vp8_decode_mb_row(VP8D_COMP *pbi,
            }
        }

-        /* Distance of Mb to the various image edges.
-         * These are specified to 8th pel as they are always compared to values that are in 1/8th pel units
-         */
-        xd->mb_to_left_edge = -((mb_col * 16) << 3);
-        xd->mb_to_right_edge = ((pc->mb_cols - 1 - mb_col) * 16) << 3;
-
        xd->dst.y_buffer = pc->yv12_fb[dst_fb_idx].y_buffer + recon_yoffset;
        xd->dst.u_buffer = pc->yv12_fb[dst_fb_idx].u_buffer + recon_uvoffset;
        xd->dst.v_buffer = pc->yv12_fb[dst_fb_idx].v_buffer + recon_uvoffset;
@@ -394,10 +423,11 @@ void vp8_decode_mb_row(VP8D_COMP *pbi,
        else
        pbi->debugoutput =0;
        */
-        vp8_decode_macroblock(pbi, xd);
+        vp8_decode_macroblock(pbi, xd, mb_row * pc->mb_cols + mb_col);

        /* check if the boolean decoder has suffered an error */
-        xd->corrupted |= vp8dx_bool_error(xd->current_bc);
+        corrupt_partition = vp8dx_bool_error(xd->current_bc);
+        xd->corrupted |= corrupt_partition;

        recon_yoffset += 16;
        recon_uvoffset += 8;
@@ -468,13 +498,14 @@ static void setup_token_decoder(VP8D_COMP *pbi,
            partition_size = user_data_end - partition;
        }

-        if (partition + partition_size > user_data_end
-            || partition + partition_size < partition)
+        if (!pbi->ec_enabled && (partition + partition_size > user_data_end
+            || partition + partition_size < partition))
            vpx_internal_error(&pc->error, VPX_CODEC_CORRUPT_FRAME,
                               "Truncated packet or corrupt partition "
                               "%d length", i + 1);

-        if (vp8dx_start_decode(bool_decoder, partition, partition_size))
+        if (vp8dx_start_decode(bool_decoder, IF_RTCD(&pbi->dboolhuff),
+                               partition, partition_size))
            vpx_internal_error(&pc->error, VPX_CODEC_MEM_ERROR,
                               "Failed to allocate bool decoder %d", i + 1);

@@ -483,11 +514,9 @@ static void setup_token_decoder(VP8D_COMP *pbi,
        bool_decoder++;
    }

-#if CONFIG_MULTITHREAD
    /* Clamp number of decoder threads */
    if (pbi->decoding_thread_count > num_part - 1)
        pbi->decoding_thread_count = num_part - 1;
-#endif
 }


@@ -584,63 +613,87 @@ int vp8_decode_frame(VP8D_COMP *pbi)
    pc->yv12_fb[pc->new_fb_idx].corrupted = 0;

    if (data_end - data < 3)
-        vpx_internal_error(&pc->error, VPX_CODEC_CORRUPT_FRAME,
-                           "Truncated packet");
-    pc->frame_type = (FRAME_TYPE)(data[0] & 1);
-    pc->version = (data[0] >> 1) & 7;
-    pc->show_frame = (data[0] >> 4) & 1;
-    first_partition_length_in_bytes =
-        (data[0] | (data[1] << 8) | (data[2] << 16)) >> 5;
-    data += 3;
-
-    if (data + first_partition_length_in_bytes > data_end
-        || data + first_partition_length_in_bytes < data)
-        vpx_internal_error(&pc->error, VPX_CODEC_CORRUPT_FRAME,
-                           "Truncated packet or corrupt partition 0 length");
-    vp8_setup_version(pc);
-
-    if (pc->frame_type == KEY_FRAME)
    {
-        const int Width = pc->Width;
-        const int Height = pc->Height;
-
-        /* vet via sync code */
-        if (data[0] != 0x9d || data[1] != 0x01 || data[2] != 0x2a)
-            vpx_internal_error(&pc->error, VPX_CODEC_UNSUP_BITSTREAM,
-                               "Invalid frame sync code");
-
-        pc->Width = (data[3] | (data[4] << 8)) & 0x3fff;
-        pc->horiz_scale = data[4] >> 6;
-        pc->Height = (data[5] | (data[6] << 8)) & 0x3fff;
-        pc->vert_scale = data[6] >> 6;
-        data += 7;
-
-        if (Width != pc->Width  ||  Height != pc->Height)
+        if (pbi->ec_enabled)
        {
-            int prev_mb_rows = pc->mb_rows;
+            /* Declare the missing frame as an inter frame since it will
+               be handled as an inter frame when we have estimated its
+               motion vectors. */
+            pc->frame_type = INTER_FRAME;
+            pc->version = 0;
+            pc->show_frame = 1;
+            first_partition_length_in_bytes = 0;
+        }
+        else
+        {
+            vpx_internal_error(&pc->error, VPX_CODEC_CORRUPT_FRAME,
+                               "Truncated packet");
+        }
+    }
+    else
+    {
+        pc->frame_type = (FRAME_TYPE)(data[0] & 1);
+        pc->version = (data[0] >> 1) & 7;
+        pc->show_frame = (data[0] >> 4) & 1;
+        first_partition_length_in_bytes =
+            (data[0] | (data[1] << 8) | (data[2] << 16)) >> 5;
+        data += 3;

-            if (pc->Width <= 0)
+        if (!pbi->ec_enabled && (data + first_partition_length_in_bytes > data_end
+            || data + first_partition_length_in_bytes < data))
+            vpx_internal_error(&pc->error, VPX_CODEC_CORRUPT_FRAME,
+                               "Truncated packet or corrupt partition 0 length");
+        vp8_setup_version(pc);
+
+        if (pc->frame_type == KEY_FRAME)
+        {
+            const int Width = pc->Width;
+            const int Height = pc->Height;
+
+            /* vet via sync code */
+            if (!pbi->ec_enabled &&
+                (data[0] != 0x9d || data[1] != 0x01 || data[2] != 0x2a))
+                vpx_internal_error(&pc->error, VPX_CODEC_UNSUP_BITSTREAM,
+                                   "Invalid frame sync code");
+
+            pc->Width = (data[3] | (data[4] << 8)) & 0x3fff;
+            pc->horiz_scale = data[4] >> 6;
+            pc->Height = (data[5] | (data[6] << 8)) & 0x3fff;
+            pc->vert_scale = data[6] >> 6;
+            data += 7;
+
+            if (Width != pc->Width  ||  Height != pc->Height)
            {
-                pc->Width = Width;
-                vpx_internal_error(&pc->error, VPX_CODEC_CORRUPT_FRAME,
-                                   "Invalid frame width");
-            }
+                int prev_mb_rows = pc->mb_rows;

-            if (pc->Height <= 0)
-            {
-                pc->Height = Height;
-                vpx_internal_error(&pc->error, VPX_CODEC_CORRUPT_FRAME,
-                                   "Invalid frame height");
-            }
+                if (pc->Width <= 0)
+                {
+                    pc->Width = Width;
+                    vpx_internal_error(&pc->error, VPX_CODEC_CORRUPT_FRAME,
+                                       "Invalid frame width");
+                }

-            if (vp8_alloc_frame_buffers(pc, pc->Width, pc->Height))
-                vpx_internal_error(&pc->error, VPX_CODEC_MEM_ERROR,
-                                   "Failed to allocate frame buffers");
+                if (pc->Height <= 0)
+                {
+                    pc->Height = Height;
+                    vpx_internal_error(&pc->error, VPX_CODEC_CORRUPT_FRAME,
+                                       "Invalid frame height");
+                }
+
+                if (vp8_alloc_frame_buffers(pc, pc->Width, pc->Height))
+                    vpx_internal_error(&pc->error, VPX_CODEC_MEM_ERROR,
+                                       "Failed to allocate frame buffers");
+
+                if (vp8_alloc_overlap_lists(pbi))
+                    vpx_internal_error(&pc->error, VPX_CODEC_MEM_ERROR,
+                                       "Failed to allocate overlap lists for "
+                                       "error concealment");

 #if CONFIG_MULTITHREAD
-            if (pbi->b_multithreaded_rd)
-                vp8mt_alloc_temp_buffers(pbi, pc->Width, prev_mb_rows);
+                if (pbi->b_multithreaded_rd)
+                    vp8mt_alloc_temp_buffers(pbi, pc->Width, prev_mb_rows);
 #endif
+            }
        }
    }

@@ -651,7 +704,8 @@ int vp8_decode_frame(VP8D_COMP *pbi)

    init_frame(pbi);

-    if (vp8dx_start_decode(bc, data, data_end - data))
+    if (vp8dx_start_decode(bc, IF_RTCD(&pbi->dboolhuff),
+                           data, data_end - data))
        vpx_internal_error(&pc->error, VPX_CODEC_MEM_ERROR,
                           "Failed to allocate bool decoder 0");
    if (pc->frame_type == KEY_FRAME) {
@@ -845,9 +899,7 @@ int vp8_decode_frame(VP8D_COMP *pbi)
    vpx_memcpy(&xd->dst, &pc->yv12_fb[pc->new_fb_idx], sizeof(YV12_BUFFER_CONFIG));

    /* set up frame new frame for intra coded blocks */
-#if CONFIG_MULTITHREAD
    if (!(pbi->b_multithreaded_rd) || pc->multi_token_partition == ONE_PARTITION || !(pc->filter_level))
-#endif
        vp8_setup_intra_recon(&pc->yv12_fb[pc->new_fb_idx]);

    vp8_setup_block_dptrs(xd);
@@ -863,11 +915,16 @@ int vp8_decode_frame(VP8D_COMP *pbi)

    vp8_decode_mode_mvs(pbi);

+    if (pbi->ec_enabled &&
+            pbi->mvs_corrupt_from_mb < (unsigned int)pc->mb_cols * pc->mb_rows)
+    {
+        vp8_estimate_missing_mvs(pbi);
+    }
+
    vpx_memset(pc->above_context, 0, sizeof(ENTROPY_CONTEXT_PLANES) * pc->mb_cols);

    vpx_memcpy(&xd->block[0].bmi, &xd->mode_info_context->bmi[0], sizeof(B_MODE_INFO));

-#if CONFIG_MULTITHREAD
    if (pbi->b_multithreaded_rd && pc->multi_token_partition != ONE_PARTITION)
    {
        vp8mt_decode_mb_rows(pbi, xd);
@@ -882,7 +939,6 @@ int vp8_decode_frame(VP8D_COMP *pbi)
        vp8_yv12_extend_frame_borders_ptr(&pc->yv12_fb[pc->new_fb_idx]);    /*cm->frame_to_show);*/
    }
    else
-#endif
    {
        int ibc = 0;
        int num_part = 1 << pc->multi_token_partition;
--- a/vp8/decoder/dequantize.c
+++ b/vp8/decoder/dequantize.c
@@ -11,7 +11,8 @@

 #include "vpx_ports/config.h"
 #include "dequantize.h"
-#include "vp8/common/idct.h"
+#include "predictdc.h"
+#include "idct.h"
 #include "vpx_mem/vpx_mem.h"

 extern void vp8_short_idct4x4llm_c(short *input, short *output, int pitch) ;
--- a/vp8/decoder/dequantize.h
+++ b/vp8/decoder/dequantize.h
@@ -11,7 +11,7 @@

 #ifndef DEQUANTIZE_H
 #define DEQUANTIZE_H
-#include "vp8/common/blockd.h"
+#include "blockd.h"

 #define prototype_dequant_block(sym) \
    void sym(BLOCKD *x)
--- a/vp8/decoder/detokenize.c
+++ b/vp8/decoder/detokenize.c
@@ -9,8 +9,8 @@
 */


-#include "vp8/common/type_aliases.h"
-#include "vp8/common/blockd.h"
+#include "type_aliases.h"
+#include "blockd.h"
 #include "onyxd_int.h"
 #include "vpx_mem/vpx_mem.h"
 #include "vpx_ports/mem.h"
@@ -74,6 +74,37 @@ void vp8_reset_mb_tokens_context(MACROBLOCKD *x)
    }
 }

+#if CONFIG_ARM_ASM_DETOK
+/* mashup of vp8_block2left and vp8_block2above so we only need one pointer
+ * for the assembly version.
+ */
+DECLARE_ALIGNED(16, const UINT8, vp8_block2leftabove[25*2]) =
+{
+    /* vp8_block2left */
+    0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 5, 5, 6, 6, 7, 7, 8,
+    /* vp8_block2above */
+    0, 1, 2, 3, 0, 1, 2, 3, 0, 1, 2, 3, 0, 1, 2, 3, 4, 5, 4, 5, 6, 7, 6, 7, 8
+};
+
+void vp8_init_detokenizer(VP8D_COMP *dx)
+{
+    const VP8_COMMON *const oc = & dx->common;
+    MACROBLOCKD *x = & dx->mb;
+
+    dx->detoken.vp8_coef_tree_ptr = vp8_coef_tree;
+    dx->detoken.ptr_block2leftabove = vp8_block2leftabove;
+    dx->detoken.ptr_coef_bands_x = vp8_coef_bands_x;
+    dx->detoken.scan = vp8_default_zig_zag1d;
+    dx->detoken.teb_base_ptr = vp8d_token_extra_bits2;
+    dx->detoken.qcoeff_start_ptr = &x->qcoeff[0];
+
+    dx->detoken.coef_probs[0] = (oc->fc.coef_probs [0] [ 0 ] [0]);
+    dx->detoken.coef_probs[1] = (oc->fc.coef_probs [1] [ 0 ] [0]);
+    dx->detoken.coef_probs[2] = (oc->fc.coef_probs [2] [ 0 ] [0]);
+    dx->detoken.coef_probs[3] = (oc->fc.coef_probs [3] [ 0 ] [0]);
+}
+#endif
+
 DECLARE_ALIGNED(16, extern const unsigned char, vp8dx_bitreader_norm[256]);
 #define FILL \
    if(count < 0) \
@@ -171,6 +202,35 @@ DECLARE_ALIGNED(16, extern const unsigned char, vp8dx_bitreader_norm[256]);
    }\
    NORMALIZE

+#if CONFIG_ARM_ASM_DETOK
+int vp8_decode_mb_tokens(VP8D_COMP *dx, MACROBLOCKD *x)
+{
+    int eobtotal = 0;
+    int i, type;
+
+    dx->detoken.current_bc = x->current_bc;
+    dx->detoken.A = x->above_context;
+    dx->detoken.L = x->left_context;
+
+    type = 3;
+
+    if (x->mode_info_context->mbmi.mode != B_PRED && x->mode_info_context->mbmi.mode != SPLITMV)
+    {
+        type = 1;
+        eobtotal -= 16;
+    }
+
+    vp8_decode_mb_tokens_v6(&dx->detoken, type);
+
+    for (i = 0; i < 25; i++)
+    {
+        x->eobs[i] = dx->detoken.eob[i];
+        eobtotal += dx->detoken.eob[i];
+    }
+
+    return eobtotal;
+}
+#else
 int vp8_decode_mb_tokens(VP8D_COMP *dx, MACROBLOCKD *x)
 {
    ENTROPY_CONTEXT *A = (ENTROPY_CONTEXT *)x->above_context;
@@ -363,3 +423,4 @@ BLOCK_FINISHED:
    return eobtotal;

 }
+#endif /*!CONFIG_ASM_DETOK*/
--- a/vp8/decoder/detokenize.h
+++ b/vp8/decoder/detokenize.h
@@ -14,6 +14,10 @@

 #include "onyxd_int.h"

+#if ARCH_ARM
+#include "arm/detokenize_arm.h"
+#endif
+
 void vp8_reset_mb_tokens_context(MACROBLOCKD *x);
 int vp8_decode_mb_tokens(VP8D_COMP *, MACROBLOCKD *);

--- a/vp8/decoder/ec_types.h
+++ b/vp8/decoder/ec_types.h
@@ -0,0 +1,40 @@
+/*
+ *  Copyright (c) 2011 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 VP8_DEC_EC_TYPES_H
+#define VP8_DEC_EC_TYPES_H
+
+#define MAX_OVERLAPS 16
+
+typedef struct
+{
+    int overlap;
+    B_MODE_INFO *bmi;
+    MV_REFERENCE_FRAME ref_frame;
+} OVERLAP_NODE;
+
+typedef struct
+{
+    /* TODO(holmer): This array should be exchanged for a linked list */
+    OVERLAP_NODE overlaps[MAX_OVERLAPS];
+} B_OVERLAP;
+
+typedef struct
+{
+    B_OVERLAP overlaps[16];
+} MB_OVERLAP;
+
+typedef struct
+{
+    MV mv;
+    MV_REFERENCE_FRAME ref_frame;
+} EC_BLOCK;
+
+#endif /* VP8_DEC_EC_TYPES_H */
--- a/Show More
+++ b/Show More
Author	SHA1	Message	Date
Stefan Holmer	3cf0ef4593	Added configure option to enable error-concealment. Disabled by default. Change-Id: I94580a5ecb13520195ea2b8a10ca11bb5a01d2a6	2011-04-29 14:08:47 +02:00
Stefan Holmer	0909b83427	Concealed MBs are always SPLITMV with partition=3. This can be optimized. Also changed the criterion for when to skip decoding the residual, now only skipping for blocks which actually is missing residual. Now using mvs_corrupt_from_mb for this decision since asking the bool decoder doesn't work (it has already finished decoding). Change-Id: I3175f11c84ae701fc2935ebe22e1d75297072eae	2011-04-29 13:50:15 +02:00
John Koleszar	62da6700dc	Update VP8DX_BOOL_DECODER_FILL to better detect EOS Allow more reliable detection of truncated bitstreams by being more precise with the count of "virtual" bits in the value buffer. Specifically, the VP8_LOTS_OF_BITS value is accumulated into count, rather than being assigned, which was losing the prior value, increasing the required tolerance when testing for the error condition. Change-Id: Ib5172eaa57323b939c439fff8a8ab5fa38da9b69	2011-04-29 11:22:09 +02:00
Stefan Holmer	98ea0d71a4	Added more descriptive comments and did some smaller refactoring. Also changed to setting the mb_skip_coeff flag when a macroblock needs to be concealed. Change-Id: I0bbf6de899f5b27f4a8ca0454da7e928e8b23919	2011-04-28 16:28:07 +02:00
Stefan Holmer	8d49ea12c2	Added correct handling of motion vectors outside frame boundaries. Change-Id: Ibf81e1d188d8dd6de877e1c52761fa212e848865	2011-04-20 12:08:27 +02:00
Stefan Holmer	766ad7edb6	Reverting some of the changes done in a64b37..., moving back the bool dec error check to vp8_decode_mb_row. Change-Id: I717ee57efc29b8e0619d6f00d1c64d0d20114a8b	2011-04-19 16:23:05 +02:00
Stefan Holmer	20431c1354	Forgot to remove two lines in previous submit Change-Id: Idbc0bc328cf2f99071008fd4a54ea00bac7beb94	2011-04-19 15:38:39 +02:00
Stefan Holmer	1b913c1f78	Refactored find_neighboring_blocks() and moved the test for corrupt stream and intra concealment inside vp8_decode_macroblock to be able tocapture and conceal errors in the residual before reconstruction. Change-Id: Id0f0bd87945a9bb1db0c20bb5467e2ff9aae5d28	2011-04-19 15:33:46 +02:00
Stefan Holmer	a64b37fdbc	Added spatial motion vector interpolation. Used for intra blocks with missing residual coefficients. Change-Id: I3e765b5dee251362d1330ebbcf9fa22d852377a1	2011-04-19 12:45:51 +02:00
Stefan Holmer	a2951d8deb	Implemented a first version of the motion vector extrapolation error concealment algorithm. Tested on foreman_cif.yuv only. Some special cases are still not handled in a good way, for instance when receiving intra blocks without coefficients. Change-Id: Ie7bb41855860923b313645dacb3cf70f1e350549	2011-04-01 11:55:30 +02:00
Stefan Holmer	83a2b4e114	Added a first simple version of error-concealment Added a first very simple version of error-concealment which simply repeats the last decoded motion vector for corrupt MBs. Change-Id: Ia83e111649afe11870c3c66065977bd0610c4fa1	2011-02-01 17:30:51 +01:00
Henrik Lundin	1422ce5cff	Error concealment in decoder Implementing an error concealment in the VP8 decoder. Change-Id: I63934df71191ad0b1e65c89725d9e021e1d8d93d	2011-01-20 11:22:50 +01:00