generic-library/vpx
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity
092b5bef37
vpx/vp8/encoder/onyx_if.c

5507 lines
174 KiB
C
Raw Normal View History

Initial WebM release
2010-05-18 17:58:33 +02:00
/*
Use WebM in copyright notice for consistency Changes 'The VP8 project' to 'The WebM project', for consistency with other webmproject.org repositories. Fixes issue #97. Change-Id: I37c13ed5fbdb9d334ceef71c6350e9febed9bbba
2010-09-09 14:16:39 +02:00
* Copyright (c) 2010 The WebM project authors. All Rights Reserved.
Initial WebM release
2010-05-18 17:58:33 +02:00
*
cosmetics: trim trailing whitespace When the license headers were updated, they accidentally contained trailing whitespace, so unfortunately we have to touch all the files again. Change-Id: I236c05fade06589e417179c0444cb39b09e4200d
2010-06-18 18:39:21 +02:00
* Use of this source code is governed by a BSD-style license
LICENSE: update with latest text Change-Id: Ieebea089095d9073b3a94932791099f614ce120c
2010-06-04 22:19:40 +02:00
* that can be found in the LICENSE file in the root of the source
* tree. An additional intellectual property rights grant can be found
cosmetics: trim trailing whitespace When the license headers were updated, they accidentally contained trailing whitespace, so unfortunately we have to touch all the files again. Change-Id: I236c05fade06589e417179c0444cb39b09e4200d
2010-06-18 18:39:21 +02:00
* in the file PATENTS. All contributing project authors may
LICENSE: update with latest text Change-Id: Ieebea089095d9073b3a94932791099f614ce120c
2010-06-04 22:19:40 +02:00
* be found in the AUTHORS file in the root of the source tree.
Initial WebM release
2010-05-18 17:58:33 +02:00
*/
#include "onyxc_int.h"
#include "onyx_int.h"
#include "systemdependent.h"
#include "quantize.h"
#include "alloccommon.h"
#include "mcomp.h"
#include "firstpass.h"
#include "psnr.h"
#include "vpx_scale/vpxscale.h"
#include "extend.h"
#include "ratectrl.h"
#include "quant_common.h"
move segmentation_common to encoder vp8_update_gf_useage_maps() is only used by the encoder. This patch fixes the ability to build in decode-only or encode-only configurations. Change-Id: I3a5211428e539886ba998e09e8abd747ac55c9aa
2010-08-13 20:50:51 +02:00
#include "segmentation.h"
Initial WebM release
2010-05-18 17:58:33 +02:00
#include "g_common.h"
#include "vpx_scale/yv12extend.h"
#include "postproc.h"
#include "vpx_mem/vpx_mem.h"
#include "swapyv12buffer.h"
#include "threading.h"
#include "vpx_ports/vpx_timer.h"
Reduced the size of MB_MODE_INFO Moved partition_bmi and partition_count out of MB_MODE_INFO and placed into MACROBLOCK. Also reduced the size of other members of the MB_MODE_INFO struct. For 1080p, the memory was reduced by 1,209,516 bytes. The decoder performance appeared to improve by 3% for the clip used. Note: The main goal for this change is to improve the decoder performance. The encoder will be revisited at a later date for further structure cleanup. Change-Id: I4733621292ee9cc3fffa4046cb3fd4d99bd14613
2010-09-02 22:17:52 +02:00
#include "vpxerrors.h"
isolate new temporal filtering code onyx_if is getting pretty big. split out the temporal code to make it easier to look at. Change-Id: I207c3a94c90e91b32e3ea5e1836a53b7a990fabd
2010-10-20 22:27:33 +02:00
#include "temporal_filter.h"
Add runtime CPU detection support for ARM. The primary goal is to allow a binary to be built which supports NEON, but can fall back to non-NEON routines, since some Android devices do not have NEON, even if they are otherwise ARMv7 (e.g., Tegra). The configure-generated flags HAVE_ARMV7, etc., are used to decide which versions of each function to build, and when CONFIG_RUNTIME_CPU_DETECT is enabled, the correct version is chosen at run time. In order for this to work, the CFLAGS must be set to something appropriate (e.g., without -mfpu=neon for ARMv7, and with appropriate -march and -mcpu for even earlier configurations), or the native C code will not be able to run. The ASFLAGS must remain set for the most advanced instruction set required at build time, since the ARM assembler will refuse to emit them otherwise. I have not attempted to make any changes to configure to do this automatically. Doing so will probably require the addition of new configure options. Many of the hooks for RTCD on ARM were already there, but a lot of the code had bit-rotted, and a good deal of the ARM-specific code is not integrated into the RTCD structs at all. I did not try to resolve the latter, merely to add the minimal amount of protection around them to allow RTCD to work. Those functions that were called based on an ifdef at the calling site were expanded to check the RTCD flags at that site, but they should be added to an RTCD struct somewhere in the future. The functions invoked with global function pointers still are, but these should be moved into an RTCD struct for thread safety (I believe every platform currently supported has atomic pointer stores, but this is not guaranteed). The encoder's boolhuff functions did not even have _c and armv7 suffixes, and the correct version was resolved at link time. The token packing functions did have appropriate suffixes, but the version was selected with a define, with no associated RTCD struct. However, for both of these, the only armv7 instruction they actually used was rbit, and this was completely superfluous, so I reworked them to avoid it. The only non-ARMv4 instruction remaining in them is clz, which is ARMv5 (not even ARMv5TE is required). Considering that there are no ARM-specific configs which are not at least ARMv5TE, I did not try to detect these at runtime, and simply enable them for ARMv5 and above. Finally, the NEON register saving code was completely non-reentrant, since it saved the registers to a global, static variable. I moved the storage for this onto the stack. A single binary built with this code was tested on an ARM11 (ARMv6) and a Cortex A8 (ARMv7 w/NEON), for both the encoder and decoder, and produced identical output, while using the correct accelerated functions on each. I did not test on any earlier processors. Change-Id: I45cbd63a614f4554c3b325c45d46c0806f009eaa
2010-10-21 00:39:11 +02:00
#if ARCH_ARM
#include "vpx_ports/arm.h"
#endif
Reduced the size of MB_MODE_INFO Moved partition_bmi and partition_count out of MB_MODE_INFO and placed into MACROBLOCK. Also reduced the size of other members of the MB_MODE_INFO struct. For 1080p, the memory was reduced by 1,209,516 bytes. The decoder performance appeared to improve by 3% for the clip used. Note: The main goal for this change is to improve the decoder performance. The encoder will be revisited at a later date for further structure cleanup. Change-Id: I4733621292ee9cc3fffa4046cb3fd4d99bd14613
2010-09-02 22:17:52 +02:00
Initial WebM release
2010-05-18 17:58:33 +02:00
#include <math.h>
#include <stdio.h>
#include <limits.h>
#if CONFIG_RUNTIME_CPU_DETECT
#define IF_RTCD(x) (x)
#define RTCD(x) &cpi->common.rtcd.x
#else
#define IF_RTCD(x) NULL
#define RTCD(x) NULL
#endif
extern void vp8cx_init_mv_bits_sadcost();
extern void vp8cx_pick_filter_level_fast(YV12_BUFFER_CONFIG *sd, VP8_COMP *cpi);
extern void vp8cx_set_alt_lf_level(VP8_COMP *cpi, int filt_val);
extern void vp8cx_pick_filter_level(YV12_BUFFER_CONFIG *sd, VP8_COMP *cpi);
extern void vp8_init_loop_filter(VP8_COMMON *cm);
extern void vp8_loop_filter_frame(VP8_COMMON *cm, MACROBLOCKD *mbd, int filt_val);
extern void vp8_loop_filter_frame_yonly(VP8_COMMON *cm, MACROBLOCKD *mbd, int filt_val, int sharpness_lvl);
extern void vp8_dmachine_specific_config(VP8_COMP *cpi);
extern void vp8_cmachine_specific_config(VP8_COMP *cpi);
extern void vp8_calc_auto_iframe_target_size(VP8_COMP *cpi);
extern void vp8_deblock_frame(YV12_BUFFER_CONFIG *source, YV12_BUFFER_CONFIG *post, int filt_lvl, int low_var_thresh, int flag);
extern void print_parms(VP8_CONFIG *ocf, char *filenam);
extern unsigned int vp8_get_processor_freq();
extern void print_tree_update_probs();
extern void vp8cx_create_encoder_threads(VP8_COMP *cpi);
extern void vp8cx_remove_encoder_threads(VP8_COMP *cpi);
#if HAVE_ARMV7
extern void vp8_yv12_copy_frame_func_neon(YV12_BUFFER_CONFIG *src_ybc, YV12_BUFFER_CONFIG *dst_ybc);
extern void vp8_yv12_copy_src_frame_func_neon(YV12_BUFFER_CONFIG *src_ybc, YV12_BUFFER_CONFIG *dst_ybc);
#endif
int vp8_estimate_entropy_savings(VP8_COMP *cpi);
int vp8_calc_ss_err(YV12_BUFFER_CONFIG *source, YV12_BUFFER_CONFIG *dest, const vp8_variance_rtcd_vtable_t *rtcd);
int vp8_calc_low_ss_err(YV12_BUFFER_CONFIG *source, YV12_BUFFER_CONFIG *dest, const vp8_variance_rtcd_vtable_t *rtcd);
abstract apply_temporal_filter allow for optimized versions of apply_temporal_filter (now vp8_apply_temporal_filter_c) the function was previously declared as static and appears to have been inlined. with this change, that's no longer possible. performance takes a small hit. the declaration for vp8_cx_temp_filter_c was moved to onyx_if.c because of a circular dependency. for rtcd, temporal_filter.h holds the definition for the rtcd table, so it needs to be included by onyx_int.h. however, onyx_int.h holds the definition for VP8_COMP which is needed for the function prototype. blah. Change-Id: I499c055fdc652ac4659c21c5a55fe10ceb7e95e3
2010-11-29 20:21:11 +01:00
extern void vp8cx_temp_filter_c(VP8_COMP *cpi);
Initial WebM release
2010-05-18 17:58:33 +02:00
Rename mode_ref_lf_test_function This function graduated from being a test func to something that's on by default. Rename it and remove some spurious comments that confuse its status. Change-Id: I689695a3ad29c35e9a72a43ec93766733ac6c20b
2010-09-29 19:53:08 +02:00
static void set_default_lf_deltas(VP8_COMP *cpi);
Initial WebM release
2010-05-18 17:58:33 +02:00
extern const int vp8_gf_interval_table[101];
#if CONFIG_PSNR
#include "math.h"
extern double vp8_calc_ssim
(
YV12_BUFFER_CONFIG *source,
YV12_BUFFER_CONFIG *dest,
int lumamask,
double *weight
);
extern double vp8_calc_ssimg
(
YV12_BUFFER_CONFIG *source,
YV12_BUFFER_CONFIG *dest,
double *ssim_y,
double *ssim_u,
double *ssim_v
);
#endif
#ifdef OUTPUT_YUV_SRC
FILE *yuv_file;
#endif
#if 0
FILE *framepsnr;
FILE *kf_list;
FILE *keyfile;
#endif
#if 0
extern int skip_true_count;
extern int skip_false_count;
#endif
#ifdef ENTROPY_STATS
extern int intra_mode_stats[10][10][10];
#endif
#ifdef SPEEDSTATS
unsigned int frames_at_speed[16] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0};
unsigned int tot_pm = 0;
unsigned int cnt_pm = 0;
unsigned int tot_ef = 0;
unsigned int cnt_ef = 0;
#endif
#ifdef MODE_STATS
extern unsigned __int64 Sectionbits[50];
extern int y_modes[5] ;
extern int uv_modes[4] ;
extern int b_modes[10] ;
extern int inter_y_modes[10] ;
extern int inter_uv_modes[4] ;
extern unsigned int inter_b_modes[15];
#endif
extern void (*vp8_short_fdct4x4)(short *input, short *output, int pitch);
extern void (*vp8_short_fdct8x4)(short *input, short *output, int pitch);
extern const int vp8_bits_per_mb[2][QINDEX_RANGE];
extern const int qrounding_factors[129];
extern const int qzbin_factors[129];
extern void vp8cx_init_quantizer(VP8_COMP *cpi);
extern const int vp8cx_base_skip_false_prob[128];
Control of active min quantizer for two pass. Create look up tables for controlling the active quantizer range. Some initial tuning to improve quality circa 0.5% on test set. Clean up of some stats output code Change-Id: Ia698a8525f8b8129a503cadace3ee73fe888f543
2010-09-29 13:03:19 +02:00
// Tables relating active max Q to active min Q
static const int kf_low_motion_minq[QINDEX_RANGE] =
{
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 4, 4,
5, 5, 5, 6, 6, 6, 7, 7, 7, 8, 8, 8, 9, 9, 10,10,
11,11,12,12,13,13,14,14,15,15,16,16,17,17,18,18,
19,19,20,20,21,21,22,22,23,23,24,24,25,25,26,26,
27,27,28,28,29,29,30,30,31,32,33,34,35,36,37,38,
};
static const int kf_high_motion_minq[QINDEX_RANGE] =
{
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1,
2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 4, 4, 5, 5, 5, 5,
6, 6, 6, 6, 7, 7, 7, 7, 8, 8, 8, 9, 9, 9, 10,10,
11,11,12,12,13,13,14,14,15,15,16,16,17,17,18,18,
19,19,20,20,21,21,22,22,23,23,24,24,25,25,26,26,
27,27,28,28,29,29,30,30,31,31,32,32,33,33,34,34,
35,35,36,36,37,38,39,40,41,42,43,44,45,46,47,48,
};
static const int gf_low_motion_minq[QINDEX_RANGE] =
{
0,0,0,0,1,1,1,1,1,1,1,1,2,2,2,2,
3,3,3,3,4,4,4,4,5,5,5,5,6,6,6,6,
7,7,7,7,8,8,8,8,9,9,9,9,10,10,10,10,
11,11,12,12,13,13,14,14,15,15,16,16,17,17,18,18,
19,19,20,20,21,21,22,22,23,23,24,24,25,25,26,26,
27,27,28,28,29,29,30,30,31,31,32,32,33,33,34,34,
35,35,36,36,37,37,38,38,39,39,40,40,41,41,42,42,
43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58
};
static const int gf_mid_motion_minq[QINDEX_RANGE] =
{
0,0,0,0,1,1,1,1,1,1,2,2,3,3,3,4,
4,4,5,5,5,6,6,6,7,7,7,8,8,8,9,9,
9,10,10,10,10,11,11,11,12,12,12,12,13,13,13,14,
14,14,15,15,16,16,17,17,18,18,19,19,20,20,21,21,
22,22,23,23,24,24,25,25,26,26,27,27,28,28,29,29,
30,30,31,31,32,32,33,33,34,34,35,35,36,36,37,37,
38,39,39,40,40,41,41,42,42,43,43,44,45,46,47,48,
49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,
};
static const int gf_high_motion_minq[QINDEX_RANGE] =
{
0,0,0,0,1,1,1,1,1,2,2,2,3,3,3,4,
4,4,5,5,5,6,6,6,7,7,7,8,8,8,9,9,
9,10,10,10,11,11,12,12,13,13,14,14,15,15,16,16,
17,17,18,18,19,19,20,20,21,21,22,22,23,23,24,24,
25,25,26,26,27,27,28,28,29,29,30,30,31,31,32,32,
33,33,34,34,35,35,36,36,37,37,38,38,39,39,40,40,
41,41,42,42,43,44,45,46,47,48,49,50,51,52,53,54,
55,56,57,58,59,60,62,64,66,68,70,72,74,76,78,80,
};
static const int inter_minq[QINDEX_RANGE] =
{
Change to inter_minq table. The inter_minq table controls the range of quantizers available for a particular frame in two pass relative to a max Q value. The changes reduces the range somewhat. The effect of this was a small increase (0.3% average) in psnr for the test set but it should also help encode speed somewhat for higher quality modes as it will reduce the number of iterations in the recode loop. The change damps the range of quantizers available locally within a section of a clip and should therefore help keep quality more uniform. If there is systematic overshoot or undershoot the range can shift gradually to accommodate. However, there is some increased risk of overshoot or undershoot against the target bit rate in VBR mode and this risk will be more pronounced for short clips. The change damps the range of quantizers available locally within a section of a clip and should therefore help keep quality more uniform. If there is systematic overshoot or undershoot the range can shift gradually to accommodate. However, there is some increased risk of overshoot or undershoot against the target bit rate in VBR mode and this risk will be more pronounced for short clips. Change-Id: I84465567d49ae767c6c73ff2a2aac30c895adb52
2010-12-04 11:04:12 +01:00
0,0,1,1,2,3,3,4,4,5,6,6,7,8,8,9,
9,10,11,11,12,13,13,14,15,15,16,17,17,18,19,20,
20,21,22,22,23,24,24,25,26,27,27,28,29,30,30,31,
32,33,33,34,35,36,36,37,38,39,39,40,41,42,42,43,
44,45,46,46,47,48,49,50,50,51,52,53,54,55,55,56,
57,58,59,60,60,61,62,63,64,65,66,67,67,68,69,70,
71,72,73,74,75,75,76,77,78,79,80,81,82,83,84,85,
86,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100
Control of active min quantizer for two pass. Create look up tables for controlling the active quantizer range. Some initial tuning to improve quality circa 0.5% on test set. Clean up of some stats output code Change-Id: Ia698a8525f8b8129a503cadace3ee73fe888f543
2010-09-29 13:03:19 +02:00
};
Initial WebM release
2010-05-18 17:58:33 +02:00
void vp8_initialize()
{
static int init_done = 0;
if (!init_done)
{
vp8_scale_machine_specific_config();
vp8_initialize_common();
//vp8_dmachine_specific_config();
vp8_tokenize_initialize();
vp8cx_init_mv_bits_sadcost();
init_done = 1;
}
}
#ifdef PACKET_TESTING
extern FILE *vpxlogc;
#endif
static void setup_features(VP8_COMP *cpi)
{
// Set up default state for MB feature flags
cpi->mb.e_mbd.segmentation_enabled = 0;
cpi->mb.e_mbd.update_mb_segmentation_map = 0;
cpi->mb.e_mbd.update_mb_segmentation_data = 0;
vpx_memset(cpi->mb.e_mbd.mb_segment_tree_probs, 255, sizeof(cpi->mb.e_mbd.mb_segment_tree_probs));
vpx_memset(cpi->mb.e_mbd.segment_feature_data, 0, sizeof(cpi->mb.e_mbd.segment_feature_data));
cpi->mb.e_mbd.mode_ref_lf_delta_enabled = 0;
cpi->mb.e_mbd.mode_ref_lf_delta_update = 0;
vpx_memset(cpi->mb.e_mbd.ref_lf_deltas, 0, sizeof(cpi->mb.e_mbd.ref_lf_deltas));
vpx_memset(cpi->mb.e_mbd.mode_lf_deltas, 0, sizeof(cpi->mb.e_mbd.mode_lf_deltas));
Fix loopfilter delta zero transitions Loopfilter deltas are initialized to zero on keyframes in the decoder. The values then persist from the previous frame unless an update bit is set in the bitstream. This data is not included in the entropy data saved by the 'refresh entropy' bit in the bitstream, so it is effectively an additional contextual element beyond the 3 ref-frames and the entropy data. The encoder was treating this delta update bit as update-if-nonzero, meaning that the value would be refreshed even if it hadn't changed, and more significantly, if the correct value for the delta changed to zero, the update wouldn't be sent, and the decoder would preserve the last (presumably non-zero) value. This patch updates the encoder to send an update only if the value has changed from the previously transmitted value. It also forces the value to be transmitted in error resilient mode, to account for lost context in the event of lost frames. Change-Id: I56671d5b42965d0166ac226765dbfce3e5301868
2010-09-29 19:04:04 +02:00
vpx_memset(cpi->mb.e_mbd.last_ref_lf_deltas, 0, sizeof(cpi->mb.e_mbd.ref_lf_deltas));
vpx_memset(cpi->mb.e_mbd.last_mode_lf_deltas, 0, sizeof(cpi->mb.e_mbd.mode_lf_deltas));
Initial WebM release
2010-05-18 17:58:33 +02:00
Rename mode_ref_lf_test_function This function graduated from being a test func to something that's on by default. Rename it and remove some spurious comments that confuse its status. Change-Id: I689695a3ad29c35e9a72a43ec93766733ac6c20b
2010-09-29 19:53:08 +02:00
set_default_lf_deltas(cpi);
Initial WebM release
2010-05-18 17:58:33 +02:00
}
void vp8_dealloc_compressor_data(VP8_COMP *cpi)
{
Improve MV prediction accuracy to achieve performance gain Add vp8_mv_pred() to better predict starting MV for NEWMV mode in vp8_rd_pick_inter_mode(). Set different search ranges according to MV prediction accuracy, which improves encoder performance without hurting the quality. Also, as Yaowu suggested, using diamond search result as full search starting point and therefore adjusting(reducing) full search range helps the performance. Change-Id: Ie4a3c8df87e697c1f4f6e2ddb693766bba1b77b6
2010-12-03 17:26:21 +01:00
// Delete last frame MV storage buffers
if (cpi->lfmv != 0)
vpx_free(cpi->lfmv);
cpi->lfmv = 0;
if (cpi->lf_ref_frame_sign_bias != 0)
vpx_free(cpi->lf_ref_frame_sign_bias);
cpi->lf_ref_frame_sign_bias = 0;
if (cpi->lf_ref_frame != 0)
vpx_free(cpi->lf_ref_frame);
cpi->lf_ref_frame = 0;
Initial WebM release
2010-05-18 17:58:33 +02:00
// Delete sementation map
if (cpi->segmentation_map != 0)
vpx_free(cpi->segmentation_map);
cpi->segmentation_map = 0;
if (cpi->active_map != 0)
vpx_free(cpi->active_map);
cpi->active_map = 0;
// Delete first pass motion map
if (cpi->fp_motion_map != 0)
vpx_free(cpi->fp_motion_map);
cpi->fp_motion_map = 0;
vp8_de_alloc_frame_buffers(&cpi->common);
vp8_yv12_de_alloc_frame_buffer(&cpi->last_frame_uf);
vp8_yv12_de_alloc_frame_buffer(&cpi->scaled_source);
#if VP8_TEMPORAL_ALT_REF
vp8_yv12_de_alloc_frame_buffer(&cpi->alt_ref_buffer.source_buffer);
#endif
{
int i;
for (i = 0; i < MAX_LAG_BUFFERS; i++)
vp8_yv12_de_alloc_frame_buffer(&cpi->src_buffer[i].source_buffer);
cpi->source_buffer_count = 0;
}
vpx_free(cpi->tok);
cpi->tok = 0;
Moved gf_active code to encoder only The gf_active code is only used by the encoder, so it was moved from common and decoder. Change-Id: Iada15acd5b2b33ff70c34668ca87d4cfd0d05025
2010-08-11 17:02:31 +02:00
// Structure used to minitor GF useage
if (cpi->gf_active_flags != 0)
vpx_free(cpi->gf_active_flags);
cpi->gf_active_flags = 0;
Reduced the size of MB_MODE_INFO Moved partition_bmi and partition_count out of MB_MODE_INFO and placed into MACROBLOCK. Also reduced the size of other members of the MB_MODE_INFO struct. For 1080p, the memory was reduced by 1,209,516 bytes. The decoder performance appeared to improve by 3% for the clip used. Note: The main goal for this change is to improve the decoder performance. The encoder will be revisited at a later date for further structure cleanup. Change-Id: I4733621292ee9cc3fffa4046cb3fd4d99bd14613
2010-09-02 22:17:52 +02:00
if(cpi->mb.pip)
vpx_free(cpi->mb.pip);
cpi->mb.pip = 0;
Fix a memory leak problem in encoder Deallocating the buffers before re-allocating them. The fix passed James Berry's test program for memory leak check. Change-Id: I18c3cf665412c0e313a523e3d435106c03ca438d
2010-12-06 23:21:37 +01:00
if(cpi->total_stats)
vpx_free(cpi->total_stats);
cpi->total_stats = 0;
if(cpi->this_frame_stats)
vpx_free(cpi->this_frame_stats);
cpi->this_frame_stats = 0;
Initial WebM release
2010-05-18 17:58:33 +02:00
}
static void enable_segmentation(VP8_PTR ptr)
{
VP8_COMP *cpi = (VP8_COMP *)(ptr);
// Set the appropriate feature bit
cpi->mb.e_mbd.segmentation_enabled = 1;
cpi->mb.e_mbd.update_mb_segmentation_map = 1;
cpi->mb.e_mbd.update_mb_segmentation_data = 1;
}
static void disable_segmentation(VP8_PTR ptr)
{
VP8_COMP *cpi = (VP8_COMP *)(ptr);
// Clear the appropriate feature bit
cpi->mb.e_mbd.segmentation_enabled = 0;
}
// Valid values for a segment are 0 to 3
// Segmentation map is arrange as [Rows][Columns]
static void set_segmentation_map(VP8_PTR ptr, unsigned char *segmentation_map)
{
VP8_COMP *cpi = (VP8_COMP *)(ptr);
// Copy in the new segmentation map
vpx_memcpy(cpi->segmentation_map, segmentation_map, (cpi->common.mb_rows * cpi->common.mb_cols));
// Signal that the map should be updated.
cpi->mb.e_mbd.update_mb_segmentation_map = 1;
cpi->mb.e_mbd.update_mb_segmentation_data = 1;
}
// The values given for each segment can be either deltas (from the default value chosen for the frame) or absolute values.
//
// Valid range for abs values is (0-127 for MB_LVL_ALT_Q) , (0-63 for SEGMENT_ALT_LF)
// Valid range for delta values are (+/-127 for MB_LVL_ALT_Q) , (+/-63 for SEGMENT_ALT_LF)
//
// abs_delta = SEGMENT_DELTADATA (deltas) abs_delta = SEGMENT_ABSDATA (use the absolute values given).
//
//
static void set_segment_data(VP8_PTR ptr, signed char *feature_data, unsigned char abs_delta)
{
VP8_COMP *cpi = (VP8_COMP *)(ptr);
cpi->mb.e_mbd.mb_segement_abs_delta = abs_delta;
vpx_memcpy(cpi->segment_feature_data, feature_data, sizeof(cpi->segment_feature_data));
}
static void segmentation_test_function(VP8_PTR ptr)
{
VP8_COMP *cpi = (VP8_COMP *)(ptr);
unsigned char *seg_map;
signed char feature_data[MB_LVL_MAX][MAX_MB_SEGMENTS];
// Create a temporary map for segmentation data.
CHECK_MEM_ERROR(seg_map, vpx_calloc(cpi->common.mb_rows * cpi->common.mb_cols, 1));
// MB loop to set local segmentation map
/*for ( i = 0; i < cpi->common.mb_rows; i++ )
{
for ( j = 0; j < cpi->common.mb_cols; j++ )
{
//seg_map[(i*cpi->common.mb_cols) + j] = (j % 2) + ((i%2)* 2);
//if ( j < cpi->common.mb_cols/2 )
// Segment 1 around the edge else 0
if ( (i == 0) || (j == 0) || (i == (cpi->common.mb_rows-1)) || (j == (cpi->common.mb_cols-1)) )
seg_map[(i*cpi->common.mb_cols) + j] = 1;
//else if ( (i < 2) || (j < 2) || (i > (cpi->common.mb_rows-3)) || (j > (cpi->common.mb_cols-3)) )
// seg_map[(i*cpi->common.mb_cols) + j] = 2;
//else if ( (i < 5) || (j < 5) || (i > (cpi->common.mb_rows-6)) || (j > (cpi->common.mb_cols-6)) )
// seg_map[(i*cpi->common.mb_cols) + j] = 3;
else
seg_map[(i*cpi->common.mb_cols) + j] = 0;
}
}*/
// Set the segmentation Map
set_segmentation_map(ptr, seg_map);
// Activate segmentation.
enable_segmentation(ptr);
// Set up the quant segment data
feature_data[MB_LVL_ALT_Q][0] = 0;
feature_data[MB_LVL_ALT_Q][1] = 4;
feature_data[MB_LVL_ALT_Q][2] = 0;
feature_data[MB_LVL_ALT_Q][3] = 0;
// Set up the loop segment data
feature_data[MB_LVL_ALT_LF][0] = 0;
feature_data[MB_LVL_ALT_LF][1] = 0;
feature_data[MB_LVL_ALT_LF][2] = 0;
feature_data[MB_LVL_ALT_LF][3] = 0;
// Initialise the feature data structure
// SEGMENT_DELTADATA 0, SEGMENT_ABSDATA 1
set_segment_data(ptr, &feature_data[0][0], SEGMENT_DELTADATA);
// Delete sementation map
if (seg_map != 0)
vpx_free(seg_map);
seg_map = 0;
}
// A simple function to cyclically refresh the background at a lower Q
static void cyclic_background_refresh(VP8_COMP *cpi, int Q, int lf_adjustment)
{
unsigned char *seg_map;
signed char feature_data[MB_LVL_MAX][MAX_MB_SEGMENTS];
int i;
int block_count = cpi->cyclic_refresh_mode_max_mbs_perframe;
int mbs_in_frame = cpi->common.mb_rows * cpi->common.mb_cols;
// Create a temporary map for segmentation data.
CHECK_MEM_ERROR(seg_map, vpx_calloc(cpi->common.mb_rows * cpi->common.mb_cols, 1));
cpi->cyclic_refresh_q = Q;
for (i = Q; i > 0; i--)
{
if (vp8_bits_per_mb[cpi->common.frame_type][i] >= ((vp8_bits_per_mb[cpi->common.frame_type][Q]*(Q + 128)) / 64))
//if ( vp8_bits_per_mb[cpi->common.frame_type][i] >= ((vp8_bits_per_mb[cpi->common.frame_type][Q]*((2*Q)+96))/64) )
{
break;
}
}
cpi->cyclic_refresh_q = i;
// Only update for inter frames
if (cpi->common.frame_type != KEY_FRAME)
{
// Cycle through the macro_block rows
// MB loop to set local segmentation map
for (i = cpi->cyclic_refresh_mode_index; i < mbs_in_frame; i++)
{
// If the MB is as a candidate for clean up then mark it for possible boost/refresh (segment 1)
// The segment id may get reset to 0 later if the MB gets coded anything other than last frame 0,0
// as only (last frame 0,0) MBs are eligable for refresh : that is to say Mbs likely to be background blocks.
if (cpi->cyclic_refresh_map[i] == 0)
{
seg_map[i] = 1;
}
else
{
seg_map[i] = 0;
// Skip blocks that have been refreshed recently anyway.
if (cpi->cyclic_refresh_map[i] < 0)
//cpi->cyclic_refresh_map[i] = cpi->cyclic_refresh_map[i] / 16;
cpi->cyclic_refresh_map[i]++;
}
if (block_count > 0)
block_count--;
else
break;
}
// If we have gone through the frame reset to the start
cpi->cyclic_refresh_mode_index = i;
if (cpi->cyclic_refresh_mode_index >= mbs_in_frame)
cpi->cyclic_refresh_mode_index = 0;
}
// Set the segmentation Map
set_segmentation_map((VP8_PTR)cpi, seg_map);
// Activate segmentation.
enable_segmentation((VP8_PTR)cpi);
// Set up the quant segment data
feature_data[MB_LVL_ALT_Q][0] = 0;
feature_data[MB_LVL_ALT_Q][1] = (cpi->cyclic_refresh_q - Q);
feature_data[MB_LVL_ALT_Q][2] = 0;
feature_data[MB_LVL_ALT_Q][3] = 0;
// Set up the loop segment data
feature_data[MB_LVL_ALT_LF][0] = 0;
feature_data[MB_LVL_ALT_LF][1] = lf_adjustment;
feature_data[MB_LVL_ALT_LF][2] = 0;
feature_data[MB_LVL_ALT_LF][3] = 0;
// Initialise the feature data structure
// SEGMENT_DELTADATA 0, SEGMENT_ABSDATA 1
set_segment_data((VP8_PTR)cpi, &feature_data[0][0], SEGMENT_DELTADATA);
// Delete sementation map
if (seg_map != 0)
vpx_free(seg_map);
seg_map = 0;
}
Rename mode_ref_lf_test_function This function graduated from being a test func to something that's on by default. Rename it and remove some spurious comments that confuse its status. Change-Id: I689695a3ad29c35e9a72a43ec93766733ac6c20b
2010-09-29 19:53:08 +02:00
static void set_default_lf_deltas(VP8_COMP *cpi)
Initial WebM release
2010-05-18 17:58:33 +02:00
{
cpi->mb.e_mbd.mode_ref_lf_delta_enabled = 1;
cpi->mb.e_mbd.mode_ref_lf_delta_update = 1;
vpx_memset(cpi->mb.e_mbd.ref_lf_deltas, 0, sizeof(cpi->mb.e_mbd.ref_lf_deltas));
vpx_memset(cpi->mb.e_mbd.mode_lf_deltas, 0, sizeof(cpi->mb.e_mbd.mode_lf_deltas));
// Test of ref frame deltas
cpi->mb.e_mbd.ref_lf_deltas[INTRA_FRAME] = 2;
cpi->mb.e_mbd.ref_lf_deltas[LAST_FRAME] = 0;
cpi->mb.e_mbd.ref_lf_deltas[GOLDEN_FRAME] = -2;
cpi->mb.e_mbd.ref_lf_deltas[ALTREF_FRAME] = -2;
cpi->mb.e_mbd.mode_lf_deltas[0] = 4; // BPRED
cpi->mb.e_mbd.mode_lf_deltas[1] = -2; // Zero
cpi->mb.e_mbd.mode_lf_deltas[2] = 2; // New mv
cpi->mb.e_mbd.mode_lf_deltas[3] = 4; // Split mv
}
void vp8_set_speed_features(VP8_COMP *cpi)
{
SPEED_FEATURES *sf = &cpi->sf;
int Mode = cpi->compressor_speed;
int Speed = cpi->Speed;
int i;
VP8_COMMON *cm = &cpi->common;
quantizer: fix assertion in fast quantizer path The fast quantizer assembly code has not been updated to match the new exact quantizer, which was made the default in commit 6adbe09. Specifically, they are not aware of the potential for the coefficient to be scaled, which results in the quantized result exceeding the range of the DCT. This patch restores the previous behavior of using the non-shifted coefficients when in the fast quantizer code path, but unfortunately requires rebuilding the tables when switching between the two. Change-Id: I0a33f5b3850335011a06906f49fafed54dda9546
2010-11-11 18:41:07 +01:00
int last_improved_quant = sf->improved_quant;
Initial WebM release
2010-05-18 17:58:33 +02:00
// Initialise default mode frequency sampling variables
for (i = 0; i < MAX_MODES; i ++)
{
cpi->mode_check_freq[i] = 0;
cpi->mode_test_hit_counts[i] = 0;
cpi->mode_chosen_counts[i] = 0;
}
cpi->mbs_tested_so_far = 0;
// best quality
sf->RD = 1;
sf->search_method = NSTEP;
sf->improved_quant = 1;
sf->improved_dct = 1;
sf->auto_filter = 1;
sf->recode_loop = 1;
sf->quarter_pixel_search = 1;
sf->half_pixel_search = 1;
sf->full_freq[0] = 7;
sf->full_freq[1] = 7;
sf->min_fs_radius = 8;
sf->max_fs_radius = 32;
sf->iterative_sub_pixel = 1;
sf->optimize_coefficients = 1;
sf->first_step = 0;
sf->max_step_search_steps = MAX_MVSEARCH_STEPS;
cpi->do_full[0] = 0;
cpi->do_full[1] = 0;
// default thresholds to 0
for (i = 0; i < MAX_MODES; i++)
sf->thresh_mult[i] = 0;
switch (Mode)
{
#if !(CONFIG_REALTIME_ONLY)
case 0: // best quality mode
sf->thresh_mult[THR_ZEROMV ] = 0;
sf->thresh_mult[THR_ZEROG ] = 0;
sf->thresh_mult[THR_ZEROA ] = 0;
sf->thresh_mult[THR_NEARESTMV] = 0;
sf->thresh_mult[THR_NEARESTG ] = 0;
sf->thresh_mult[THR_NEARESTA ] = 0;
sf->thresh_mult[THR_NEARMV ] = 0;
sf->thresh_mult[THR_NEARG ] = 0;
sf->thresh_mult[THR_NEARA ] = 0;
sf->thresh_mult[THR_DC ] = 0;
sf->thresh_mult[THR_V_PRED ] = 1000;
sf->thresh_mult[THR_H_PRED ] = 1000;
sf->thresh_mult[THR_B_PRED ] = 2000;
sf->thresh_mult[THR_TM ] = 1000;
sf->thresh_mult[THR_NEWMV ] = 1000;
sf->thresh_mult[THR_NEWG ] = 1000;
sf->thresh_mult[THR_NEWA ] = 1000;
sf->thresh_mult[THR_SPLITMV ] = 2500;
sf->thresh_mult[THR_SPLITG ] = 5000;
sf->thresh_mult[THR_SPLITA ] = 5000;
sf->full_freq[0] = 7;
sf->full_freq[1] = 15;
sf->first_step = 0;
sf->max_step_search_steps = MAX_MVSEARCH_STEPS;
if (!(cpi->ref_frame_flags & VP8_LAST_FLAG))
{
sf->thresh_mult[THR_NEWMV ] = INT_MAX;
sf->thresh_mult[THR_NEARESTMV] = INT_MAX;
sf->thresh_mult[THR_ZEROMV ] = INT_MAX;
sf->thresh_mult[THR_NEARMV ] = INT_MAX;
sf->thresh_mult[THR_SPLITMV ] = INT_MAX;
}
if (!(cpi->ref_frame_flags & VP8_GOLD_FLAG))
{
sf->thresh_mult[THR_NEARESTG ] = INT_MAX;
sf->thresh_mult[THR_ZEROG ] = INT_MAX;
sf->thresh_mult[THR_NEARG ] = INT_MAX;
sf->thresh_mult[THR_NEWG ] = INT_MAX;
sf->thresh_mult[THR_SPLITG ] = INT_MAX;
}
Fix breakout thresh computation for golden & AltRef frames 1. Unavailability of each reference frame type should be tested independently, 2. Also, only the VP8_GOLD_FLAG needs to be tested before setting golden frame specific thresholds, and only VP8_ALT_FLAG needs testing before setting thresholds relevant to the AltRef frame. (Raised by gbvalor, in response to Issue 47) Change-Id: I6a06fc2a6592841d85422bc1661e33349bb6c3b8
2010-06-21 14:44:42 +02:00
if (!(cpi->ref_frame_flags & VP8_ALT_FLAG))
Initial WebM release
2010-05-18 17:58:33 +02:00
{
sf->thresh_mult[THR_NEARESTA ] = INT_MAX;
sf->thresh_mult[THR_ZEROA ] = INT_MAX;
sf->thresh_mult[THR_NEARA ] = INT_MAX;
sf->thresh_mult[THR_NEWA ] = INT_MAX;
sf->thresh_mult[THR_SPLITA ] = INT_MAX;
}
break;
case 1:
case 3:
sf->thresh_mult[THR_NEARESTMV] = 0;
sf->thresh_mult[THR_ZEROMV ] = 0;
sf->thresh_mult[THR_DC ] = 0;
sf->thresh_mult[THR_NEARMV ] = 0;
sf->thresh_mult[THR_V_PRED ] = 1000;
sf->thresh_mult[THR_H_PRED ] = 1000;
sf->thresh_mult[THR_B_PRED ] = 2500;
sf->thresh_mult[THR_TM ] = 1000;
sf->thresh_mult[THR_NEARESTG ] = 1000;
sf->thresh_mult[THR_NEARESTA ] = 1000;
sf->thresh_mult[THR_ZEROG ] = 1000;
sf->thresh_mult[THR_ZEROA ] = 1000;
sf->thresh_mult[THR_NEARG ] = 1000;
sf->thresh_mult[THR_NEARA ] = 1000;
Changed segmentation check order In SPLITMV, the 8x8 segment will be checked first. If the 8x8 rd is better than the best, we check the other segments. Otherwise bail. Adjustments to the thresh_mult were necessary to make up for the initial quality loss. The performance improved by 20% (average) for good quality, speed 0 and speed 1, while the overall quality remained the same. Change-Id: I717aef401323c8a254fba3e9777d2a316c774cc3
2010-12-16 23:01:27 +01:00
#if 1
sf->thresh_mult[THR_ZEROMV ] = 0;
sf->thresh_mult[THR_ZEROG ] = 0;
sf->thresh_mult[THR_ZEROA ] = 0;
sf->thresh_mult[THR_NEARESTMV] = 0;
sf->thresh_mult[THR_NEARESTG ] = 0;
sf->thresh_mult[THR_NEARESTA ] = 0;
sf->thresh_mult[THR_NEARMV ] = 0;
sf->thresh_mult[THR_NEARG ] = 0;
sf->thresh_mult[THR_NEARA ] = 0;
// sf->thresh_mult[THR_DC ] = 0;
// sf->thresh_mult[THR_V_PRED ] = 1000;
// sf->thresh_mult[THR_H_PRED ] = 1000;
// sf->thresh_mult[THR_B_PRED ] = 2000;
// sf->thresh_mult[THR_TM ] = 1000;
sf->thresh_mult[THR_NEWMV ] = 1000;
sf->thresh_mult[THR_NEWG ] = 1000;
sf->thresh_mult[THR_NEWA ] = 1000;
sf->thresh_mult[THR_SPLITMV ] = 1700;
sf->thresh_mult[THR_SPLITG ] = 4500;
sf->thresh_mult[THR_SPLITA ] = 4500;
#else
Initial WebM release
2010-05-18 17:58:33 +02:00
sf->thresh_mult[THR_NEWMV ] = 1500;
sf->thresh_mult[THR_NEWG ] = 1500;
sf->thresh_mult[THR_NEWA ] = 1500;
sf->thresh_mult[THR_SPLITMV ] = 5000;
sf->thresh_mult[THR_SPLITG ] = 10000;
sf->thresh_mult[THR_SPLITA ] = 10000;
Changed segmentation check order In SPLITMV, the 8x8 segment will be checked first. If the 8x8 rd is better than the best, we check the other segments. Otherwise bail. Adjustments to the thresh_mult were necessary to make up for the initial quality loss. The performance improved by 20% (average) for good quality, speed 0 and speed 1, while the overall quality remained the same. Change-Id: I717aef401323c8a254fba3e9777d2a316c774cc3
2010-12-16 23:01:27 +01:00
#endif
Initial WebM release
2010-05-18 17:58:33 +02:00
sf->full_freq[0] = 15;
sf->full_freq[1] = 31;
sf->first_step = 0;
sf->max_step_search_steps = MAX_MVSEARCH_STEPS;
if (!(cpi->ref_frame_flags & VP8_LAST_FLAG))
{
sf->thresh_mult[THR_NEWMV ] = INT_MAX;
sf->thresh_mult[THR_NEARESTMV] = INT_MAX;
sf->thresh_mult[THR_ZEROMV ] = INT_MAX;
sf->thresh_mult[THR_NEARMV ] = INT_MAX;
sf->thresh_mult[THR_SPLITMV ] = INT_MAX;
}
Fix breakout thresh computation for golden & AltRef frames 1. Unavailability of each reference frame type should be tested independently, 2. Also, only the VP8_GOLD_FLAG needs to be tested before setting golden frame specific thresholds, and only VP8_ALT_FLAG needs testing before setting thresholds relevant to the AltRef frame. (Raised by gbvalor, in response to Issue 47) Change-Id: I6a06fc2a6592841d85422bc1661e33349bb6c3b8
2010-06-21 14:44:42 +02:00
if (!(cpi->ref_frame_flags & VP8_GOLD_FLAG))
Initial WebM release
2010-05-18 17:58:33 +02:00
{
sf->thresh_mult[THR_NEARESTG ] = INT_MAX;
sf->thresh_mult[THR_ZEROG ] = INT_MAX;
sf->thresh_mult[THR_NEARG ] = INT_MAX;
sf->thresh_mult[THR_NEWG ] = INT_MAX;
sf->thresh_mult[THR_SPLITG ] = INT_MAX;
}
Fix breakout thresh computation for golden & AltRef frames 1. Unavailability of each reference frame type should be tested independently, 2. Also, only the VP8_GOLD_FLAG needs to be tested before setting golden frame specific thresholds, and only VP8_ALT_FLAG needs testing before setting thresholds relevant to the AltRef frame. (Raised by gbvalor, in response to Issue 47) Change-Id: I6a06fc2a6592841d85422bc1661e33349bb6c3b8
2010-06-21 14:44:42 +02:00
if (!(cpi->ref_frame_flags & VP8_ALT_FLAG))
Initial WebM release
2010-05-18 17:58:33 +02:00
{
sf->thresh_mult[THR_NEARESTA ] = INT_MAX;
sf->thresh_mult[THR_ZEROA ] = INT_MAX;
sf->thresh_mult[THR_NEARA ] = INT_MAX;
sf->thresh_mult[THR_NEWA ] = INT_MAX;
sf->thresh_mult[THR_SPLITA ] = INT_MAX;
}
if (Speed > 0)
{
Change to coefficient optimization rules. Allow coefficient optimization for good quality speed 0. Change-Id: Id0cb363df6823c6798671584fbba097916a7df2c
2010-09-29 14:22:05 +02:00
// Disable coefficient optimization above speed 0
sf->optimize_coefficients = 0;
Initial WebM release
2010-05-18 17:58:33 +02:00
cpi->mode_check_freq[THR_SPLITG] = 4;
cpi->mode_check_freq[THR_SPLITA] = 4;
Changed segmentation check order In SPLITMV, the 8x8 segment will be checked first. If the 8x8 rd is better than the best, we check the other segments. Otherwise bail. Adjustments to the thresh_mult were necessary to make up for the initial quality loss. The performance improved by 20% (average) for good quality, speed 0 and speed 1, while the overall quality remained the same. Change-Id: I717aef401323c8a254fba3e9777d2a316c774cc3
2010-12-16 23:01:27 +01:00
cpi->mode_check_freq[THR_SPLITMV] = 0;
Initial WebM release
2010-05-18 17:58:33 +02:00
sf->thresh_mult[THR_TM ] = 1500;
sf->thresh_mult[THR_V_PRED ] = 1500;
sf->thresh_mult[THR_H_PRED ] = 1500;
sf->thresh_mult[THR_B_PRED ] = 5000;
if (cpi->ref_frame_flags & VP8_LAST_FLAG)
{
sf->thresh_mult[THR_NEWMV ] = 2000;
sf->thresh_mult[THR_SPLITMV ] = 10000;
}
if (cpi->ref_frame_flags & VP8_GOLD_FLAG)
{
sf->thresh_mult[THR_NEARESTG ] = 1500;
sf->thresh_mult[THR_ZEROG ] = 1500;
sf->thresh_mult[THR_NEARG ] = 1500;
sf->thresh_mult[THR_NEWG ] = 2000;
sf->thresh_mult[THR_SPLITG ] = 20000;
}
if (cpi->ref_frame_flags & VP8_ALT_FLAG)
{
sf->thresh_mult[THR_NEARESTA ] = 1500;
sf->thresh_mult[THR_ZEROA ] = 1500;
sf->thresh_mult[THR_NEARA ] = 1500;
sf->thresh_mult[THR_NEWA ] = 2000;
sf->thresh_mult[THR_SPLITA ] = 20000;
}
sf->improved_quant = 0;
sf->improved_dct = 0;
sf->first_step = 1;
sf->max_step_search_steps = MAX_MVSEARCH_STEPS;
}
if (Speed > 1)
{
cpi->mode_check_freq[THR_SPLITG] = 15;
cpi->mode_check_freq[THR_SPLITA] = 15;
cpi->mode_check_freq[THR_SPLITMV] = 7;
sf->thresh_mult[THR_TM ] = 2000;
sf->thresh_mult[THR_V_PRED ] = 2000;
sf->thresh_mult[THR_H_PRED ] = 2000;
sf->thresh_mult[THR_B_PRED ] = 7500;
if (cpi->ref_frame_flags & VP8_LAST_FLAG)
{
sf->thresh_mult[THR_NEWMV ] = 2000;
sf->thresh_mult[THR_SPLITMV ] = 25000;
}
if (cpi->ref_frame_flags & VP8_GOLD_FLAG)
{
sf->thresh_mult[THR_NEARESTG ] = 2000;
sf->thresh_mult[THR_ZEROG ] = 2000;
sf->thresh_mult[THR_NEARG ] = 2000;
sf->thresh_mult[THR_NEWG ] = 2500;
sf->thresh_mult[THR_SPLITG ] = 50000;
}
if (cpi->ref_frame_flags & VP8_ALT_FLAG)
{
sf->thresh_mult[THR_NEARESTA ] = 2000;
sf->thresh_mult[THR_ZEROA ] = 2000;
sf->thresh_mult[THR_NEARA ] = 2000;
sf->thresh_mult[THR_NEWA ] = 2500;
sf->thresh_mult[THR_SPLITA ] = 50000;
}
Replaced recode loop test with a function call Replaced existing code to decide if a frame recode is required with a function call. This is to simplify addition of extra clauses that may be needed for the planned constrained quality mode. Also fixed a bug where by alt ref not considered in the test. Change-Id: I3d40bb21abe3e19f8456761e6849deb171738b60
2010-11-17 16:12:04 +01:00
// Only do recode loop on key frames, golden frames and
// alt ref frames
Initial WebM release
2010-05-18 17:58:33 +02:00
sf->recode_loop = 2;
sf->full_freq[0] = 31;
sf->full_freq[1] = 63;
}
if (Speed > 2)
{
sf->auto_filter = 0; // Faster selection of loop filter
cpi->mode_check_freq[THR_V_PRED] = 2;
cpi->mode_check_freq[THR_H_PRED] = 2;
cpi->mode_check_freq[THR_B_PRED] = 2;
if (cpi->ref_frame_flags & VP8_GOLD_FLAG)
{
cpi->mode_check_freq[THR_NEARG] = 2;
cpi->mode_check_freq[THR_NEWG] = 4;
}
if (cpi->ref_frame_flags & VP8_ALT_FLAG)
{
cpi->mode_check_freq[THR_NEARA] = 2;
cpi->mode_check_freq[THR_NEWA] = 4;
}
sf->thresh_mult[THR_SPLITA ] = INT_MAX;
sf->thresh_mult[THR_SPLITG ] = INT_MAX;
sf->thresh_mult[THR_SPLITMV ] = INT_MAX;
sf->full_freq[0] = 63;
sf->full_freq[1] = 127;
}
if (Speed > 3)
{
cpi->mode_check_freq[THR_V_PRED] = 0;
cpi->mode_check_freq[THR_H_PRED] = 0;
cpi->mode_check_freq[THR_B_PRED] = 0;
cpi->mode_check_freq[THR_NEARG] = 0;
cpi->mode_check_freq[THR_NEWG] = 0;
cpi->mode_check_freq[THR_NEARA] = 0;
cpi->mode_check_freq[THR_NEWA] = 0;
sf->auto_filter = 1;
sf->recode_loop = 0; // recode loop off
sf->RD = 0; // Turn rd off
sf->full_freq[0] = INT_MAX;
sf->full_freq[1] = INT_MAX;
}
if (Speed > 4)
{
sf->auto_filter = 0; // Faster selection of loop filter
cpi->mode_check_freq[THR_V_PRED] = 2;
cpi->mode_check_freq[THR_H_PRED] = 2;
cpi->mode_check_freq[THR_B_PRED] = 2;
if (cpi->ref_frame_flags & VP8_GOLD_FLAG)
{
cpi->mode_check_freq[THR_NEARG] = 2;
cpi->mode_check_freq[THR_NEWG] = 4;
}
if (cpi->ref_frame_flags & VP8_ALT_FLAG)
{
cpi->mode_check_freq[THR_NEARA] = 2;
cpi->mode_check_freq[THR_NEWA] = 4;
}
Fix breakout thresh computation for golden & AltRef frames 1. Unavailability of each reference frame type should be tested independently, 2. Also, only the VP8_GOLD_FLAG needs to be tested before setting golden frame specific thresholds, and only VP8_ALT_FLAG needs testing before setting thresholds relevant to the AltRef frame. (Raised by gbvalor, in response to Issue 47) Change-Id: I6a06fc2a6592841d85422bc1661e33349bb6c3b8
2010-06-21 14:44:42 +02:00
if (cpi->ref_frame_flags & VP8_GOLD_FLAG)
Initial WebM release
2010-05-18 17:58:33 +02:00
{
sf->thresh_mult[THR_NEARESTG ] = 2000;
sf->thresh_mult[THR_ZEROG ] = 2000;
sf->thresh_mult[THR_NEARG ] = 2000;
sf->thresh_mult[THR_NEWG ] = 4000;
}
Fix breakout thresh computation for golden & AltRef frames 1. Unavailability of each reference frame type should be tested independently, 2. Also, only the VP8_GOLD_FLAG needs to be tested before setting golden frame specific thresholds, and only VP8_ALT_FLAG needs testing before setting thresholds relevant to the AltRef frame. (Raised by gbvalor, in response to Issue 47) Change-Id: I6a06fc2a6592841d85422bc1661e33349bb6c3b8
2010-06-21 14:44:42 +02:00
if (cpi->ref_frame_flags & VP8_ALT_FLAG)
Initial WebM release
2010-05-18 17:58:33 +02:00
{
sf->thresh_mult[THR_NEARESTA ] = 2000;
sf->thresh_mult[THR_ZEROA ] = 2000;
sf->thresh_mult[THR_NEARA ] = 2000;
sf->thresh_mult[THR_NEWA ] = 4000;
}
}
break;
#endif
case 2:
sf->optimize_coefficients = 0;
sf->recode_loop = 0;
sf->auto_filter = 1;
sf->iterative_sub_pixel = 1;
sf->thresh_mult[THR_NEARESTMV] = 0;
sf->thresh_mult[THR_ZEROMV ] = 0;
sf->thresh_mult[THR_DC ] = 0;
sf->thresh_mult[THR_TM ] = 0;
sf->thresh_mult[THR_NEARMV ] = 0;
sf->thresh_mult[THR_V_PRED ] = 1000;
sf->thresh_mult[THR_H_PRED ] = 1000;
sf->thresh_mult[THR_B_PRED ] = 2500;
sf->thresh_mult[THR_NEARESTG ] = 1000;
sf->thresh_mult[THR_ZEROG ] = 1000;
sf->thresh_mult[THR_NEARG ] = 1000;
sf->thresh_mult[THR_NEARESTA ] = 1000;
sf->thresh_mult[THR_ZEROA ] = 1000;
sf->thresh_mult[THR_NEARA ] = 1000;
sf->thresh_mult[THR_NEWMV ] = 2000;
sf->thresh_mult[THR_NEWG ] = 2000;
sf->thresh_mult[THR_NEWA ] = 2000;
sf->thresh_mult[THR_SPLITMV ] = 5000;
sf->thresh_mult[THR_SPLITG ] = 10000;
sf->thresh_mult[THR_SPLITA ] = 10000;
sf->full_freq[0] = 15;
sf->full_freq[1] = 31;
sf->search_method = NSTEP;
Fix compiler warnings Change-Id: I2a97f08cc3c7808ce5be39e910cc5147ecf03a1d
2010-06-11 20:33:49 +02:00
if (!(cpi->ref_frame_flags & VP8_LAST_FLAG))
Initial WebM release
2010-05-18 17:58:33 +02:00
{
sf->thresh_mult[THR_NEWMV ] = INT_MAX;
sf->thresh_mult[THR_NEARESTMV] = INT_MAX;
sf->thresh_mult[THR_ZEROMV ] = INT_MAX;
sf->thresh_mult[THR_NEARMV ] = INT_MAX;
sf->thresh_mult[THR_SPLITMV ] = INT_MAX;
}
Fix compiler warnings Change-Id: I2a97f08cc3c7808ce5be39e910cc5147ecf03a1d
2010-06-11 20:33:49 +02:00
if (!(cpi->ref_frame_flags & VP8_GOLD_FLAG))
Initial WebM release
2010-05-18 17:58:33 +02:00
{
sf->thresh_mult[THR_NEARESTG ] = INT_MAX;
sf->thresh_mult[THR_ZEROG ] = INT_MAX;
sf->thresh_mult[THR_NEARG ] = INT_MAX;
sf->thresh_mult[THR_NEWG ] = INT_MAX;
sf->thresh_mult[THR_SPLITG ] = INT_MAX;
}
Fix compiler warnings Change-Id: I2a97f08cc3c7808ce5be39e910cc5147ecf03a1d
2010-06-11 20:33:49 +02:00
if (!(cpi->ref_frame_flags & VP8_ALT_FLAG))
Initial WebM release
2010-05-18 17:58:33 +02:00
{
sf->thresh_mult[THR_NEARESTA ] = INT_MAX;
sf->thresh_mult[THR_ZEROA ] = INT_MAX;
sf->thresh_mult[THR_NEARA ] = INT_MAX;
sf->thresh_mult[THR_NEWA ] = INT_MAX;
sf->thresh_mult[THR_SPLITA ] = INT_MAX;
}
if (Speed > 0)
{
cpi->mode_check_freq[THR_SPLITG] = 4;
cpi->mode_check_freq[THR_SPLITA] = 4;
cpi->mode_check_freq[THR_SPLITMV] = 2;
sf->thresh_mult[THR_DC ] = 0;
sf->thresh_mult[THR_TM ] = 1000;
sf->thresh_mult[THR_V_PRED ] = 2000;
sf->thresh_mult[THR_H_PRED ] = 2000;
sf->thresh_mult[THR_B_PRED ] = 5000;
if (cpi->ref_frame_flags & VP8_LAST_FLAG)
{
sf->thresh_mult[THR_NEARESTMV] = 0;
sf->thresh_mult[THR_ZEROMV ] = 0;
sf->thresh_mult[THR_NEARMV ] = 0;
sf->thresh_mult[THR_NEWMV ] = 2000;
sf->thresh_mult[THR_SPLITMV ] = 10000;
}
if (cpi->ref_frame_flags & VP8_GOLD_FLAG)
{
sf->thresh_mult[THR_NEARESTG ] = 1000;
sf->thresh_mult[THR_ZEROG ] = 1000;
sf->thresh_mult[THR_NEARG ] = 1000;
sf->thresh_mult[THR_NEWG ] = 2000;
sf->thresh_mult[THR_SPLITG ] = 20000;
}
if (cpi->ref_frame_flags & VP8_ALT_FLAG)
{
sf->thresh_mult[THR_NEARESTA ] = 1000;
sf->thresh_mult[THR_ZEROA ] = 1000;
sf->thresh_mult[THR_NEARA ] = 1000;
sf->thresh_mult[THR_NEWA ] = 2000;
sf->thresh_mult[THR_SPLITA ] = 20000;
}
sf->improved_quant = 0;
sf->improved_dct = 0;
}
if (Speed > 1)
{
cpi->mode_check_freq[THR_SPLITMV] = 7;
cpi->mode_check_freq[THR_SPLITG] = 15;
cpi->mode_check_freq[THR_SPLITA] = 15;
sf->thresh_mult[THR_TM ] = 2000;
sf->thresh_mult[THR_V_PRED ] = 2000;
sf->thresh_mult[THR_H_PRED ] = 2000;
sf->thresh_mult[THR_B_PRED ] = 5000;
if (cpi->ref_frame_flags & VP8_LAST_FLAG)
{
sf->thresh_mult[THR_NEWMV ] = 2000;
sf->thresh_mult[THR_SPLITMV ] = 25000;
}
if (cpi->ref_frame_flags & VP8_GOLD_FLAG)
{
sf->thresh_mult[THR_NEARESTG ] = 2000;
sf->thresh_mult[THR_ZEROG ] = 2000;
sf->thresh_mult[THR_NEARG ] = 2000;
sf->thresh_mult[THR_NEWG ] = 2500;
sf->thresh_mult[THR_SPLITG ] = 50000;
}
if (cpi->ref_frame_flags & VP8_ALT_FLAG)
{
sf->thresh_mult[THR_NEARESTA ] = 2000;
sf->thresh_mult[THR_ZEROA ] = 2000;
sf->thresh_mult[THR_NEARA ] = 2000;
sf->thresh_mult[THR_NEWA ] = 2500;
sf->thresh_mult[THR_SPLITA ] = 50000;
}
sf->full_freq[0] = 31;
sf->full_freq[1] = 63;
}
if (Speed > 2)
{
sf->auto_filter = 0; // Faster selection of loop filter
cpi->mode_check_freq[THR_V_PRED] = 2;
cpi->mode_check_freq[THR_H_PRED] = 2;
cpi->mode_check_freq[THR_B_PRED] = 2;
if (cpi->ref_frame_flags & VP8_GOLD_FLAG)
{
cpi->mode_check_freq[THR_NEARG] = 2;
cpi->mode_check_freq[THR_NEWG] = 4;
}
if (cpi->ref_frame_flags & VP8_ALT_FLAG)
{
cpi->mode_check_freq[THR_NEARA] = 2;
cpi->mode_check_freq[THR_NEWA] = 4;
}
sf->thresh_mult[THR_SPLITMV ] = INT_MAX;
sf->thresh_mult[THR_SPLITG ] = INT_MAX;
sf->thresh_mult[THR_SPLITA ] = INT_MAX;
sf->full_freq[0] = 63;
sf->full_freq[1] = 127;
}
if (Speed > 3)
{
sf->RD = 0;
sf->full_freq[0] = INT_MAX;
sf->full_freq[1] = INT_MAX;
sf->auto_filter = 1;
}
if (Speed > 4)
{
sf->auto_filter = 0; // Faster selection of loop filter
#if CONFIG_REALTIME_ONLY
sf->search_method = HEX;
#else
sf->search_method = DIAMOND;
#endif
cpi->mode_check_freq[THR_V_PRED] = 4;
cpi->mode_check_freq[THR_H_PRED] = 4;
cpi->mode_check_freq[THR_B_PRED] = 4;
if (cpi->ref_frame_flags & VP8_GOLD_FLAG)
{
cpi->mode_check_freq[THR_NEARG] = 2;
cpi->mode_check_freq[THR_NEWG] = 4;
}
if (cpi->ref_frame_flags & VP8_ALT_FLAG)
{
cpi->mode_check_freq[THR_NEARA] = 2;
cpi->mode_check_freq[THR_NEWA] = 4;
}
sf->thresh_mult[THR_TM ] = 2000;
sf->thresh_mult[THR_B_PRED ] = 5000;
if (cpi->ref_frame_flags & VP8_GOLD_FLAG)
{
sf->thresh_mult[THR_NEARESTG ] = 2000;
sf->thresh_mult[THR_ZEROG ] = 2000;
sf->thresh_mult[THR_NEARG ] = 2000;
sf->thresh_mult[THR_NEWG ] = 4000;
}
if (cpi->ref_frame_flags & VP8_ALT_FLAG)
{
sf->thresh_mult[THR_NEARESTA ] = 2000;
sf->thresh_mult[THR_ZEROA ] = 2000;
sf->thresh_mult[THR_NEARA ] = 2000;
sf->thresh_mult[THR_NEWA ] = 4000;
}
}
if (Speed > 5)
{
// Disable split MB intra prediction mode
sf->thresh_mult[THR_B_PRED] = INT_MAX;
}
if (Speed > 6)
{
unsigned int i, sum = 0;
unsigned int total_mbs = cm->MBs;
int thresh;
int total_skip;
int min = 2000;
sf->iterative_sub_pixel = 0;
if (cpi->oxcf.encode_breakout > 2000)
min = cpi->oxcf.encode_breakout;
min >>= 7;
for (i = 0; i < min; i++)
{
sum += cpi->error_bins[i];
}
total_skip = sum;
sum = 0;
// i starts from 2 to make sure thresh started from 2048
for (; i < 1024; i++)
{
sum += cpi->error_bins[i];
if (10 * sum >= (unsigned int)(cpi->Speed - 6)*(total_mbs - total_skip))
break;
}
i--;
thresh = (i << 7);
if (thresh < 2000)
thresh = 2000;
if (cpi->ref_frame_flags & VP8_LAST_FLAG)
{
sf->thresh_mult[THR_NEWMV] = thresh;
sf->thresh_mult[THR_NEARESTMV ] = thresh >> 1;
sf->thresh_mult[THR_NEARMV ] = thresh >> 1;
}
if (cpi->ref_frame_flags & VP8_GOLD_FLAG)
{
sf->thresh_mult[THR_NEWG] = thresh << 1;
sf->thresh_mult[THR_NEARESTG ] = thresh;
sf->thresh_mult[THR_NEARG ] = thresh;
}
if (cpi->ref_frame_flags & VP8_ALT_FLAG)
{
sf->thresh_mult[THR_NEWA] = thresh << 1;
sf->thresh_mult[THR_NEARESTA ] = thresh;
sf->thresh_mult[THR_NEARA ] = thresh;
}
// Disable other intra prediction modes
sf->thresh_mult[THR_TM] = INT_MAX;
sf->thresh_mult[THR_V_PRED] = INT_MAX;
sf->thresh_mult[THR_H_PRED] = INT_MAX;
}
if (Speed > 8)
{
sf->quarter_pixel_search = 0;
}
if (Speed > 9)
{
int Tmp = cpi->Speed - 8;
if (Tmp > 4)
Tmp = 4;
if (cpi->ref_frame_flags & VP8_GOLD_FLAG)
{
cpi->mode_check_freq[THR_ZEROG] = 1 << (Tmp - 1);
cpi->mode_check_freq[THR_NEARESTG] = 1 << (Tmp - 1);
cpi->mode_check_freq[THR_NEARG] = 1 << Tmp;
cpi->mode_check_freq[THR_NEWG] = 1 << (Tmp + 1);
}
if (cpi->ref_frame_flags & VP8_ALT_FLAG)
{
cpi->mode_check_freq[THR_ZEROA] = 1 << (Tmp - 1);
cpi->mode_check_freq[THR_NEARESTA] = 1 << (Tmp - 1);
cpi->mode_check_freq[THR_NEARA] = 1 << Tmp;
cpi->mode_check_freq[THR_NEWA] = 1 << (Tmp + 1);
}
cpi->mode_check_freq[THR_NEWMV] = 1 << (Tmp - 1);
}
cm->filter_type = NORMAL_LOOPFILTER;
if (Speed >= 14)
cm->filter_type = SIMPLE_LOOPFILTER;
if (Speed >= 15)
{
sf->half_pixel_search = 0; // This has a big hit on quality. Last resort
}
vpx_memset(cpi->error_bins, 0, sizeof(cpi->error_bins));
};
if (cpi->sf.search_method == NSTEP)
{
Swap alt/gold/new/last frame buffer ptrs instead of copying. At the end of the decode, frame buffers were being copied. The frames are not updated after the copy, they are just for reference on later frames. This change allows multiple references to the same frame buffer instead of copying it. Changes needed to be made to the encoder to handle this. The encoder is still doing frame buffer copies in similar places where pointer reference could be done. Change-Id: I7c38be4d23979cc49b5f17241ca3a78703803e66
2010-07-22 14:07:32 +02:00
vp8_init3smotion_compensation(&cpi->mb, cm->yv12_fb[cm->lst_fb_idx].y_stride);
Initial WebM release
2010-05-18 17:58:33 +02:00
}
else if (cpi->sf.search_method == DIAMOND)
{
Swap alt/gold/new/last frame buffer ptrs instead of copying. At the end of the decode, frame buffers were being copied. The frames are not updated after the copy, they are just for reference on later frames. This change allows multiple references to the same frame buffer instead of copying it. Changes needed to be made to the encoder to handle this. The encoder is still doing frame buffer copies in similar places where pointer reference could be done. Change-Id: I7c38be4d23979cc49b5f17241ca3a78703803e66
2010-07-22 14:07:32 +02:00
vp8_init_dsmotion_compensation(&cpi->mb, cm->yv12_fb[cm->lst_fb_idx].y_stride);
Initial WebM release
2010-05-18 17:58:33 +02:00
}
if (cpi->sf.improved_dct)
{
cpi->mb.vp8_short_fdct8x4 = FDCT_INVOKE(&cpi->rtcd.fdct, short8x4);
cpi->mb.vp8_short_fdct4x4 = FDCT_INVOKE(&cpi->rtcd.fdct, short4x4);
}
else
{
cpi->mb.vp8_short_fdct8x4 = FDCT_INVOKE(&cpi->rtcd.fdct, fast8x4);
cpi->mb.vp8_short_fdct4x4 = FDCT_INVOKE(&cpi->rtcd.fdct, fast4x4);
}
cpi->mb.short_walsh4x4 = FDCT_INVOKE(&cpi->rtcd.fdct, walsh_short4x4);
if (cpi->sf.improved_quant)
{
cpi->mb.quantize_b = QUANTIZE_INVOKE(&cpi->rtcd.quantize, quantb);
}
else
{
cpi->mb.quantize_b = QUANTIZE_INVOKE(&cpi->rtcd.quantize, fastquantb);
}
quantizer: fix assertion in fast quantizer path The fast quantizer assembly code has not been updated to match the new exact quantizer, which was made the default in commit 6adbe09. Specifically, they are not aware of the potential for the coefficient to be scaled, which results in the quantized result exceeding the range of the DCT. This patch restores the previous behavior of using the non-shifted coefficients when in the fast quantizer code path, but unfortunately requires rebuilding the tables when switching between the two. Change-Id: I0a33f5b3850335011a06906f49fafed54dda9546
2010-11-11 18:41:07 +01:00
if (cpi->sf.improved_quant != last_improved_quant)
vp8cx_init_quantizer(cpi);
Initial WebM release
2010-05-18 17:58:33 +02:00
#if CONFIG_RUNTIME_CPU_DETECT
cpi->mb.e_mbd.rtcd = &cpi->common.rtcd;
#endif
if (cpi->sf.iterative_sub_pixel == 1)
{
cpi->find_fractional_mv_step = vp8_find_best_sub_pixel_step_iteratively;
}
else if (cpi->sf.quarter_pixel_search)
{
cpi->find_fractional_mv_step = vp8_find_best_sub_pixel_step;
}
else if (cpi->sf.half_pixel_search)
{
cpi->find_fractional_mv_step = vp8_find_best_half_pixel_step;
}
else
{
cpi->find_fractional_mv_step = vp8_skip_fractional_mv_step;
}
if (cpi->sf.optimize_coefficients == 1)
change to make use of more trellis quantization when a subsequent frame is encoded as an alt reference frame, it is unlikely that any mb in current frame will be used as reference for future frames, so we can enable quantization optimization even when the RD constant is slightly rate-biased. The change has an overall benefit between 0.1% to 0.2% bit savings on the test sets based on vpxssim scores. Change-Id: I9aa7bc5cd573ea84e3ee655d2834c18c4460ceea
2010-10-15 03:58:34 +02:00
cpi->mb.optimize = 1 + cpi->is_next_src_alt_ref;
Initial WebM release
2010-05-18 17:58:33 +02:00
else
cpi->mb.optimize = 0;
if (cpi->common.full_pixel)
cpi->find_fractional_mv_step = vp8_skip_fractional_mv_step;
#ifdef SPEEDSTATS
frames_at_speed[cpi->Speed]++;
#endif
}
static void alloc_raw_frame_buffers(VP8_COMP *cpi)
{
int i, buffers;
buffers = cpi->oxcf.lag_in_frames;
if (buffers > MAX_LAG_BUFFERS)
buffers = MAX_LAG_BUFFERS;
if (buffers < 1)
buffers = 1;
for (i = 0; i < buffers; i++)
if (vp8_yv12_alloc_frame_buffer(&cpi->src_buffer[i].source_buffer,
cpi->oxcf.Width, cpi->oxcf.Height,
16))
vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
"Failed to allocate lag buffer");
#if VP8_TEMPORAL_ALT_REF
if (vp8_yv12_alloc_frame_buffer(&cpi->alt_ref_buffer.source_buffer,
cpi->oxcf.Width, cpi->oxcf.Height, 16))
vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
"Failed to allocate altref buffer");
#endif
cpi->source_buffer_count = 0;
}
Reduced the size of MB_MODE_INFO Moved partition_bmi and partition_count out of MB_MODE_INFO and placed into MACROBLOCK. Also reduced the size of other members of the MB_MODE_INFO struct. For 1080p, the memory was reduced by 1,209,516 bytes. The decoder performance appeared to improve by 3% for the clip used. Note: The main goal for this change is to improve the decoder performance. The encoder will be revisited at a later date for further structure cleanup. Change-Id: I4733621292ee9cc3fffa4046cb3fd4d99bd14613
2010-09-02 22:17:52 +02:00
static int vp8_alloc_partition_data(VP8_COMP *cpi)
{
Fix a memory leak problem in encoder Deallocating the buffers before re-allocating them. The fix passed James Berry's test program for memory leak check. Change-Id: I18c3cf665412c0e313a523e3d435106c03ca438d
2010-12-06 23:21:37 +01:00
if(cpi->mb.pip)
vpx_free(cpi->mb.pip);
Reduced the size of MB_MODE_INFO Moved partition_bmi and partition_count out of MB_MODE_INFO and placed into MACROBLOCK. Also reduced the size of other members of the MB_MODE_INFO struct. For 1080p, the memory was reduced by 1,209,516 bytes. The decoder performance appeared to improve by 3% for the clip used. Note: The main goal for this change is to improve the decoder performance. The encoder will be revisited at a later date for further structure cleanup. Change-Id: I4733621292ee9cc3fffa4046cb3fd4d99bd14613
2010-09-02 22:17:52 +02:00
cpi->mb.pip = vpx_calloc((cpi->common.mb_cols + 1) *
(cpi->common.mb_rows + 1),
sizeof(PARTITION_INFO));
if(!cpi->mb.pip)
return ALLOC_FAILURE;
cpi->mb.pi = cpi->mb.pip + cpi->common.mode_info_stride + 1;
return 0;
}
Initial WebM release
2010-05-18 17:58:33 +02:00
void vp8_alloc_compressor_data(VP8_COMP *cpi)
{
VP8_COMMON *cm = & cpi->common;
int width = cm->Width;
int height = cm->Height;
if (vp8_alloc_frame_buffers(cm, width, height))
vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
"Failed to allocate frame buffers");
Reduced the size of MB_MODE_INFO Moved partition_bmi and partition_count out of MB_MODE_INFO and placed into MACROBLOCK. Also reduced the size of other members of the MB_MODE_INFO struct. For 1080p, the memory was reduced by 1,209,516 bytes. The decoder performance appeared to improve by 3% for the clip used. Note: The main goal for this change is to improve the decoder performance. The encoder will be revisited at a later date for further structure cleanup. Change-Id: I4733621292ee9cc3fffa4046cb3fd4d99bd14613
2010-09-02 22:17:52 +02:00
if (vp8_alloc_partition_data(cpi))
vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
"Failed to allocate partition data");
Initial WebM release
2010-05-18 17:58:33 +02:00
if ((width & 0xf) != 0)
width += 16 - (width & 0xf);
if ((height & 0xf) != 0)
height += 16 - (height & 0xf);
if (vp8_yv12_alloc_frame_buffer(&cpi->last_frame_uf,
width, height, VP8BORDERINPIXELS))
vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
"Failed to allocate last frame buffer");
if (vp8_yv12_alloc_frame_buffer(&cpi->scaled_source, width, height, 16))
vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
"Failed to allocate scaled source buffer");
if (cpi->tok != 0)
vpx_free(cpi->tok);
{
unsigned int tokens = cm->mb_rows * cm->mb_cols * 24 * 16;
CHECK_MEM_ERROR(cpi->tok, vpx_calloc(tokens, sizeof(*cpi->tok)));
}
// Data used for real time vc mode to see if gf needs refreshing
cpi->inter_zz_count = 0;
cpi->gf_bad_count = 0;
cpi->gf_update_recommended = 0;
Moved gf_active code to encoder only The gf_active code is only used by the encoder, so it was moved from common and decoder. Change-Id: Iada15acd5b2b33ff70c34668ca87d4cfd0d05025
2010-08-11 17:02:31 +02:00
// Structures used to minitor GF usage
if (cpi->gf_active_flags != 0)
vpx_free(cpi->gf_active_flags);
CHECK_MEM_ERROR(cpi->gf_active_flags, vpx_calloc(1, cm->mb_rows * cm->mb_cols));
cpi->gf_active_count = cm->mb_rows * cm->mb_cols;
Move firstpass motion map to stats packet The first implementation of the firstpass motion map for motion compensated temporal filtering created a file, fpmotionmap.stt, in the current working directory. This was not safe for multiple encoder instances. This patch merges this data into the first pass stats packet interface, so that it is handled like the other (numerical) firstpass stats. The new stats packet is defined as follows: Numerical Stats (16 doubles) -- 128 bytes Motion Map -- 1 byte / Macroblock Padding -- to align packet to 8 bytes The fpmotionmap.stt file can still be generated for debugging purposes in the same way that the textual version of the stats are available (defining OUTPUT_FPF in firstpass.c) Change-Id: I083ffbfd95e7d6a42bb4039ba0e81f678c8183ca
2010-10-14 22:40:12 +02:00
Fix a memory leak problem in encoder Deallocating the buffers before re-allocating them. The fix passed James Berry's test program for memory leak check. Change-Id: I18c3cf665412c0e313a523e3d435106c03ca438d
2010-12-06 23:21:37 +01:00
if(cpi->total_stats)
vpx_free(cpi->total_stats);
Move firstpass motion map to stats packet The first implementation of the firstpass motion map for motion compensated temporal filtering created a file, fpmotionmap.stt, in the current working directory. This was not safe for multiple encoder instances. This patch merges this data into the first pass stats packet interface, so that it is handled like the other (numerical) firstpass stats. The new stats packet is defined as follows: Numerical Stats (16 doubles) -- 128 bytes Motion Map -- 1 byte / Macroblock Padding -- to align packet to 8 bytes The fpmotionmap.stt file can still be generated for debugging purposes in the same way that the textual version of the stats are available (defining OUTPUT_FPF in firstpass.c) Change-Id: I083ffbfd95e7d6a42bb4039ba0e81f678c8183ca
2010-10-14 22:40:12 +02:00
cpi->total_stats = vpx_calloc(1, vp8_firstpass_stats_sz(cpi->common.MBs));
Fix a memory leak problem in encoder Deallocating the buffers before re-allocating them. The fix passed James Berry's test program for memory leak check. Change-Id: I18c3cf665412c0e313a523e3d435106c03ca438d
2010-12-06 23:21:37 +01:00
if(cpi->this_frame_stats)
vpx_free(cpi->this_frame_stats);
Move firstpass motion map to stats packet The first implementation of the firstpass motion map for motion compensated temporal filtering created a file, fpmotionmap.stt, in the current working directory. This was not safe for multiple encoder instances. This patch merges this data into the first pass stats packet interface, so that it is handled like the other (numerical) firstpass stats. The new stats packet is defined as follows: Numerical Stats (16 doubles) -- 128 bytes Motion Map -- 1 byte / Macroblock Padding -- to align packet to 8 bytes The fpmotionmap.stt file can still be generated for debugging purposes in the same way that the textual version of the stats are available (defining OUTPUT_FPF in firstpass.c) Change-Id: I083ffbfd95e7d6a42bb4039ba0e81f678c8183ca
2010-10-14 22:40:12 +02:00
cpi->this_frame_stats = vpx_calloc(1, vp8_firstpass_stats_sz(cpi->common.MBs));
Fix a memory leak problem in encoder Deallocating the buffers before re-allocating them. The fix passed James Berry's test program for memory leak check. Change-Id: I18c3cf665412c0e313a523e3d435106c03ca438d
2010-12-06 23:21:37 +01:00
Move firstpass motion map to stats packet The first implementation of the firstpass motion map for motion compensated temporal filtering created a file, fpmotionmap.stt, in the current working directory. This was not safe for multiple encoder instances. This patch merges this data into the first pass stats packet interface, so that it is handled like the other (numerical) firstpass stats. The new stats packet is defined as follows: Numerical Stats (16 doubles) -- 128 bytes Motion Map -- 1 byte / Macroblock Padding -- to align packet to 8 bytes The fpmotionmap.stt file can still be generated for debugging purposes in the same way that the textual version of the stats are available (defining OUTPUT_FPF in firstpass.c) Change-Id: I083ffbfd95e7d6a42bb4039ba0e81f678c8183ca
2010-10-14 22:40:12 +02:00
if(!cpi->total_stats || !cpi->this_frame_stats)
vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
"Failed to allocate firstpass stats");
Initial WebM release
2010-05-18 17:58:33 +02:00
}
// Quant MOD
static const int q_trans[] =
{
0, 1, 2, 3, 4, 5, 7, 8,
9, 10, 12, 13, 15, 17, 18, 19,
20, 21, 23, 24, 25, 26, 27, 28,
29, 30, 31, 33, 35, 37, 39, 41,
43, 45, 47, 49, 51, 53, 55, 57,
59, 61, 64, 67, 70, 73, 76, 79,
82, 85, 88, 91, 94, 97, 100, 103,
106, 109, 112, 115, 118, 121, 124, 127,
};
int vp8_reverse_trans(int x)
{
int i;
for (i = 0; i < 64; i++)
if (q_trans[i] >= x)
return i;
return 63;
};
void vp8_new_frame_rate(VP8_COMP *cpi, double framerate)
{
safety check to avoid divide by 0s
2010-10-14 22:19:06 +02:00
if(framerate < .1)
framerate = 30;
Initial WebM release
2010-05-18 17:58:33 +02:00
cpi->oxcf.frame_rate = framerate;
cpi->output_frame_rate = cpi->oxcf.frame_rate;
cpi->per_frame_bandwidth = (int)(cpi->oxcf.target_bandwidth / cpi->output_frame_rate);
cpi->av_per_frame_bandwidth = (int)(cpi->oxcf.target_bandwidth / cpi->output_frame_rate);
cpi->min_frame_bandwidth = (int)(cpi->av_per_frame_bandwidth * cpi->oxcf.two_pass_vbrmin_section / 100);
cpi->max_gf_interval = (int)(cpi->output_frame_rate / 2) + 2;
//cpi->max_gf_interval = (int)(cpi->output_frame_rate * 2 / 3) + 1;
//cpi->max_gf_interval = 24;
if (cpi->max_gf_interval < 12)
cpi->max_gf_interval = 12;
Fix GF interval for non-lagged ARFs When ARFs are enabled in non-lagged compress modes, the GF interval was being reset to zero. Non-lagged ARF updates were enabled in commit 63ccfbd, but this incorrect GF interval caused a quality regression. Change-Id: I615c3b493f4ce2127044f4e68d0bcb07d6b730c3
2010-09-09 18:57:23 +02:00
// Special conditions when altr ref frame enabled in lagged compress mode
if (cpi->oxcf.play_alternate && cpi->oxcf.lag_in_frames)
Initial WebM release
2010-05-18 17:58:33 +02:00
{
if (cpi->max_gf_interval > cpi->oxcf.lag_in_frames - 1)
cpi->max_gf_interval = cpi->oxcf.lag_in_frames - 1;
}
}
increase rate control buffer level precision The external API exposes the RC initial/optimal/full buffer level in milliseconds, but this value was truncated internally to seconds. This patch allows the use of the full precision during the conversion from time to bits. Change-Id: If8dd2a87614c05747f81432cbe75dd9e6ed2f04e
2010-08-20 17:04:10 +02:00
static int
rescale(int val, int num, int denom)
{
int64_t llnum = num;
int64_t llden = denom;
int64_t llval = val;
return llval * llnum / llden;
}
Initial WebM release
2010-05-18 17:58:33 +02:00
void vp8_init_config(VP8_PTR ptr, VP8_CONFIG *oxcf)
{
VP8_COMP *cpi = (VP8_COMP *)(ptr);
VP8_COMMON *cm = &cpi->common;
if (!cpi)
return;
cpi->auto_gold = 1;
cpi->auto_adjust_gold_quantizer = 1;
cpi->goldquantizer = 1;
cpi->goldfreq = 7;
cpi->auto_adjust_key_quantizer = 1;
cpi->keyquantizer = 1;
cm->version = oxcf->Version;
vp8_setup_version(cm);
if (oxcf == 0)
{
cpi->pass = 0;
cpi->auto_worst_q = 0;
cpi->oxcf.best_allowed_q = MINQ;
cpi->oxcf.worst_allowed_q = MAXQ;
cpi->oxcf.end_usage = USAGE_STREAM_FROM_SERVER;
increase rate control buffer level precision The external API exposes the RC initial/optimal/full buffer level in milliseconds, but this value was truncated internally to seconds. This patch allows the use of the full precision during the conversion from time to bits. Change-Id: If8dd2a87614c05747f81432cbe75dd9e6ed2f04e
2010-08-20 17:04:10 +02:00
cpi->oxcf.starting_buffer_level = 4000;
cpi->oxcf.optimal_buffer_level = 5000;
cpi->oxcf.maximum_buffer_size = 6000;
Initial WebM release
2010-05-18 17:58:33 +02:00
cpi->oxcf.under_shoot_pct = 90;
cpi->oxcf.allow_df = 0;
cpi->oxcf.drop_frames_water_mark = 20;
cpi->oxcf.allow_spatial_resampling = 0;
cpi->oxcf.resample_down_water_mark = 40;
cpi->oxcf.resample_up_water_mark = 60;
cpi->oxcf.fixed_q = cpi->interquantizer;
cpi->filter_type = NORMAL_LOOPFILTER;
if (cm->simpler_lpf)
cpi->filter_type = SIMPLE_LOOPFILTER;
cpi->compressor_speed = 1;
cpi->horiz_scale = 0;
cpi->vert_scale = 0;
cpi->oxcf.two_pass_vbrbias = 50;
cpi->oxcf.two_pass_vbrmax_section = 400;
cpi->oxcf.two_pass_vbrmin_section = 0;
cpi->oxcf.Sharpness = 0;
cpi->oxcf.noise_sensitivity = 0;
}
else
cpi->oxcf = *oxcf;
switch (cpi->oxcf.Mode)
{
case MODE_REALTIME:
cpi->pass = 0;
cpi->compressor_speed = 2;
if (cpi->oxcf.cpu_used < -16)
{
cpi->oxcf.cpu_used = -16;
}
if (cpi->oxcf.cpu_used > 16)
cpi->oxcf.cpu_used = 16;
break;
#if !(CONFIG_REALTIME_ONLY)
case MODE_GOODQUALITY:
cpi->pass = 0;
cpi->compressor_speed = 1;
if (cpi->oxcf.cpu_used < -5)
{
cpi->oxcf.cpu_used = -5;
}
if (cpi->oxcf.cpu_used > 5)
cpi->oxcf.cpu_used = 5;
break;
case MODE_BESTQUALITY:
cpi->pass = 0;
cpi->compressor_speed = 0;
break;
case MODE_FIRSTPASS:
cpi->pass = 1;
cpi->compressor_speed = 1;
break;
case MODE_SECONDPASS:
cpi->pass = 2;
cpi->compressor_speed = 1;
if (cpi->oxcf.cpu_used < -5)
{
cpi->oxcf.cpu_used = -5;
}
if (cpi->oxcf.cpu_used > 5)
cpi->oxcf.cpu_used = 5;
break;
case MODE_SECONDPASS_BEST:
cpi->pass = 2;
cpi->compressor_speed = 0;
break;
#endif
}
if (cpi->pass == 0)
cpi->auto_worst_q = 1;
cpi->oxcf.worst_allowed_q = q_trans[oxcf->worst_allowed_q];
cpi->oxcf.best_allowed_q = q_trans[oxcf->best_allowed_q];
if (oxcf->fixed_q >= 0)
{
if (oxcf->worst_allowed_q < 0)
cpi->oxcf.fixed_q = q_trans[0];
else
cpi->oxcf.fixed_q = q_trans[oxcf->worst_allowed_q];
if (oxcf->alt_q < 0)
cpi->oxcf.alt_q = q_trans[0];
else
cpi->oxcf.alt_q = q_trans[oxcf->alt_q];
if (oxcf->key_q < 0)
cpi->oxcf.key_q = q_trans[0];
else
cpi->oxcf.key_q = q_trans[oxcf->key_q];
if (oxcf->gold_q < 0)
cpi->oxcf.gold_q = q_trans[0];
else
cpi->oxcf.gold_q = q_trans[oxcf->gold_q];
}
cpi->baseline_gf_interval = cpi->oxcf.alt_freq ? cpi->oxcf.alt_freq : DEFAULT_GF_INTERVAL;
cpi->ref_frame_flags = VP8_ALT_FLAG | VP8_GOLD_FLAG | VP8_LAST_FLAG;
//cpi->use_golden_frame_only = 0;
//cpi->use_last_frame_only = 0;
cm->refresh_golden_frame = 0;
cm->refresh_last_frame = 1;
cm->refresh_entropy_probs = 1;
if (cpi->oxcf.token_partitions >= 0 && cpi->oxcf.token_partitions <= 3)
cm->multi_token_partition = (TOKEN_PARTITION) cpi->oxcf.token_partitions;
setup_features(cpi);
{
int i;
for (i = 0; i < MAX_MB_SEGMENTS; i++)
cpi->segment_encode_breakout[i] = cpi->oxcf.encode_breakout;
}
// At the moment the first order values may not be > MAXQ
if (cpi->oxcf.fixed_q > MAXQ)
cpi->oxcf.fixed_q = MAXQ;
// local file playback mode == really big buffer
if (cpi->oxcf.end_usage == USAGE_LOCAL_FILE_PLAYBACK)
{
increase rate control buffer level precision The external API exposes the RC initial/optimal/full buffer level in milliseconds, but this value was truncated internally to seconds. This patch allows the use of the full precision during the conversion from time to bits. Change-Id: If8dd2a87614c05747f81432cbe75dd9e6ed2f04e
2010-08-20 17:04:10 +02:00
cpi->oxcf.starting_buffer_level = 60000;
cpi->oxcf.optimal_buffer_level = 60000;
cpi->oxcf.maximum_buffer_size = 240000;
Initial WebM release
2010-05-18 17:58:33 +02:00
}
// Convert target bandwidth from Kbit/s to Bit/s
cpi->oxcf.target_bandwidth *= 1000;
increase rate control buffer level precision The external API exposes the RC initial/optimal/full buffer level in milliseconds, but this value was truncated internally to seconds. This patch allows the use of the full precision during the conversion from time to bits. Change-Id: If8dd2a87614c05747f81432cbe75dd9e6ed2f04e
2010-08-20 17:04:10 +02:00
cpi->oxcf.starting_buffer_level =
rescale(cpi->oxcf.starting_buffer_level,
cpi->oxcf.target_bandwidth, 1000);
Initial WebM release
2010-05-18 17:58:33 +02:00
if (cpi->oxcf.optimal_buffer_level == 0)
cpi->oxcf.optimal_buffer_level = cpi->oxcf.target_bandwidth / 8;
else
increase rate control buffer level precision The external API exposes the RC initial/optimal/full buffer level in milliseconds, but this value was truncated internally to seconds. This patch allows the use of the full precision during the conversion from time to bits. Change-Id: If8dd2a87614c05747f81432cbe75dd9e6ed2f04e
2010-08-20 17:04:10 +02:00
cpi->oxcf.optimal_buffer_level =
rescale(cpi->oxcf.optimal_buffer_level,
cpi->oxcf.target_bandwidth, 1000);
Initial WebM release
2010-05-18 17:58:33 +02:00
if (cpi->oxcf.maximum_buffer_size == 0)
cpi->oxcf.maximum_buffer_size = cpi->oxcf.target_bandwidth / 8;
else
increase rate control buffer level precision The external API exposes the RC initial/optimal/full buffer level in milliseconds, but this value was truncated internally to seconds. This patch allows the use of the full precision during the conversion from time to bits. Change-Id: If8dd2a87614c05747f81432cbe75dd9e6ed2f04e
2010-08-20 17:04:10 +02:00
cpi->oxcf.maximum_buffer_size =
rescale(cpi->oxcf.maximum_buffer_size,
cpi->oxcf.target_bandwidth, 1000);
Initial WebM release
2010-05-18 17:58:33 +02:00
cpi->buffer_level = cpi->oxcf.starting_buffer_level;
cpi->bits_off_target = cpi->oxcf.starting_buffer_level;
vp8_new_frame_rate(cpi, cpi->oxcf.frame_rate);
cpi->worst_quality = cpi->oxcf.worst_allowed_q;
cpi->active_worst_quality = cpi->oxcf.worst_allowed_q;
cpi->avg_frame_qindex = cpi->oxcf.worst_allowed_q;
cpi->best_quality = cpi->oxcf.best_allowed_q;
cpi->active_best_quality = cpi->oxcf.best_allowed_q;
cpi->buffered_mode = (cpi->oxcf.optimal_buffer_level > 0) ? TRUE : FALSE;
Badly placed initialization of rolling rate monitors. This affects control of the active quantizer range. Change-Id: I30511fc81ac9f75ff20d9f1372382423d56739da
2010-09-24 18:52:55 +02:00
cpi->rolling_target_bits = cpi->av_per_frame_bandwidth;
cpi->rolling_actual_bits = cpi->av_per_frame_bandwidth;
cpi->long_rolling_target_bits = cpi->av_per_frame_bandwidth;
cpi->long_rolling_actual_bits = cpi->av_per_frame_bandwidth;
Initial WebM release
2010-05-18 17:58:33 +02:00
cpi->total_actual_bits = 0;
cpi->total_target_vs_actual = 0;
// Only allow dropped frames in buffered mode
cpi->drop_frames_allowed = cpi->oxcf.allow_df && cpi->buffered_mode;
cm->filter_type = (LOOPFILTERTYPE) cpi->filter_type;
if (!cm->use_bilinear_mc_filter)
cm->mcomp_filter_type = SIXTAP;
else
cm->mcomp_filter_type = BILINEAR;
cpi->target_bandwidth = cpi->oxcf.target_bandwidth;
cm->Width = cpi->oxcf.Width ;
cm->Height = cpi->oxcf.Height ;
cpi->intra_frame_target = (4 * (cm->Width + cm->Height) / 15) * 1000; // As per VP8
cm->horiz_scale = cpi->horiz_scale;
cm->vert_scale = cpi->vert_scale ;
// VP8 sharpness level mapping 0-7 (vs 0-10 in general VPx dialogs)
if (cpi->oxcf.Sharpness > 7)
cpi->oxcf.Sharpness = 7;
cm->sharpness_level = cpi->oxcf.Sharpness;
if (cm->horiz_scale != NORMAL || cm->vert_scale != NORMAL)
{
int UNINITIALIZED_IS_SAFE(hr), UNINITIALIZED_IS_SAFE(hs);
int UNINITIALIZED_IS_SAFE(vr), UNINITIALIZED_IS_SAFE(vs);
Scale2Ratio(cm->horiz_scale, &hr, &hs);
Scale2Ratio(cm->vert_scale, &vr, &vs);
// always go to the next whole number
cm->Width = (hs - 1 + cpi->oxcf.Width * hr) / hs;
cm->Height = (vs - 1 + cpi->oxcf.Height * vr) / vs;
}
Swap alt/gold/new/last frame buffer ptrs instead of copying. At the end of the decode, frame buffers were being copied. The frames are not updated after the copy, they are just for reference on later frames. This change allows multiple references to the same frame buffer instead of copying it. Changes needed to be made to the encoder to handle this. The encoder is still doing frame buffer copies in similar places where pointer reference could be done. Change-Id: I7c38be4d23979cc49b5f17241ca3a78703803e66
2010-07-22 14:07:32 +02:00
if (((cm->Width + 15) & 0xfffffff0) != cm->yv12_fb[cm->lst_fb_idx].y_width ||
((cm->Height + 15) & 0xfffffff0) != cm->yv12_fb[cm->lst_fb_idx].y_height ||
cm->yv12_fb[cm->lst_fb_idx].y_width == 0)
Initial WebM release
2010-05-18 17:58:33 +02:00
{
alloc_raw_frame_buffers(cpi);
vp8_alloc_compressor_data(cpi);
}
// Clamp KF frame size to quarter of data rate
if (cpi->intra_frame_target > cpi->target_bandwidth >> 2)
cpi->intra_frame_target = cpi->target_bandwidth >> 2;
if (cpi->oxcf.fixed_q >= 0)
{
cpi->last_q[0] = cpi->oxcf.fixed_q;
cpi->last_q[1] = cpi->oxcf.fixed_q;
}
cpi->Speed = cpi->oxcf.cpu_used;
// force to allowlag to 0 if lag_in_frames is 0;
if (cpi->oxcf.lag_in_frames == 0)
{
cpi->oxcf.allow_lag = 0;
}
// Limit on lag buffers as these are not currently dynamically allocated
else if (cpi->oxcf.lag_in_frames > MAX_LAG_BUFFERS)
cpi->oxcf.lag_in_frames = MAX_LAG_BUFFERS;
// YX Temp
cpi->last_alt_ref_sei = -1;
cpi->is_src_frame_alt_ref = 0;
change to make use of more trellis quantization when a subsequent frame is encoded as an alt reference frame, it is unlikely that any mb in current frame will be used as reference for future frames, so we can enable quantization optimization even when the RD constant is slightly rate-biased. The change has an overall benefit between 0.1% to 0.2% bit savings on the test sets based on vpxssim scores. Change-Id: I9aa7bc5cd573ea84e3ee655d2834c18c4460ceea
2010-10-15 03:58:34 +02:00
cpi->is_next_src_alt_ref = 0;
Initial WebM release
2010-05-18 17:58:33 +02:00
#if 0
// Experimental RD Code
cpi->frame_distortion = 0;
cpi->last_frame_distortion = 0;
#endif
#if VP8_TEMPORAL_ALT_REF
Made AltRef filter adaptive & added motion compensation Modified AltRef temporal filter to adapt filter length based on macroblock coding modes selected during first-pass encode. Also added sub-pixel motion compensation to the AltRef filter.
2010-09-28 16:23:41 +02:00
cpi->use_weighted_temporal_filter = 0;
Initial WebM release
2010-05-18 17:58:33 +02:00
{
int i;
cpi->fixed_divide[0] = 0;
Made AltRef filter adaptive & added motion compensation Modified AltRef temporal filter to adapt filter length based on macroblock coding modes selected during first-pass encode. Also added sub-pixel motion compensation to the AltRef filter.
2010-09-28 16:23:41 +02:00
for (i = 1; i < 512; i++)
cpi->fixed_divide[i] = 0x80000 / i;
Initial WebM release
2010-05-18 17:58:33 +02:00
}
#endif
}
/*
* This function needs more clean up, i.e. be more tuned torwards
* change_config rather than init_config !!!!!!!!!!!!!!!!
* YX - 5/28/2009
*
*/
void vp8_change_config(VP8_PTR ptr, VP8_CONFIG *oxcf)
{
VP8_COMP *cpi = (VP8_COMP *)(ptr);
VP8_COMMON *cm = &cpi->common;
if (!cpi)
return;
if (!oxcf)
return;
if (cm->version != oxcf->Version)
{
cm->version = oxcf->Version;
vp8_setup_version(cm);
}
cpi->oxcf = *oxcf;
switch (cpi->oxcf.Mode)
{
case MODE_REALTIME:
cpi->pass = 0;
cpi->compressor_speed = 2;
if (cpi->oxcf.cpu_used < -16)
{
cpi->oxcf.cpu_used = -16;
}
if (cpi->oxcf.cpu_used > 16)
cpi->oxcf.cpu_used = 16;
break;
#if !(CONFIG_REALTIME_ONLY)
case MODE_GOODQUALITY:
cpi->pass = 0;
cpi->compressor_speed = 1;
if (cpi->oxcf.cpu_used < -5)
{
cpi->oxcf.cpu_used = -5;
}
if (cpi->oxcf.cpu_used > 5)
cpi->oxcf.cpu_used = 5;
break;
case MODE_BESTQUALITY:
cpi->pass = 0;
cpi->compressor_speed = 0;
break;
case MODE_FIRSTPASS:
cpi->pass = 1;
cpi->compressor_speed = 1;
break;
case MODE_SECONDPASS:
cpi->pass = 2;
cpi->compressor_speed = 1;
if (cpi->oxcf.cpu_used < -5)
{
cpi->oxcf.cpu_used = -5;
}
if (cpi->oxcf.cpu_used > 5)
cpi->oxcf.cpu_used = 5;
break;
case MODE_SECONDPASS_BEST:
cpi->pass = 2;
cpi->compressor_speed = 0;
break;
#endif
}
if (cpi->pass == 0)
cpi->auto_worst_q = 1;
cpi->oxcf.worst_allowed_q = q_trans[oxcf->worst_allowed_q];
cpi->oxcf.best_allowed_q = q_trans[oxcf->best_allowed_q];
if (oxcf->fixed_q >= 0)
{
if (oxcf->worst_allowed_q < 0)
cpi->oxcf.fixed_q = q_trans[0];
else
cpi->oxcf.fixed_q = q_trans[oxcf->worst_allowed_q];
if (oxcf->alt_q < 0)
cpi->oxcf.alt_q = q_trans[0];
else
cpi->oxcf.alt_q = q_trans[oxcf->alt_q];
if (oxcf->key_q < 0)
cpi->oxcf.key_q = q_trans[0];
else
cpi->oxcf.key_q = q_trans[oxcf->key_q];
if (oxcf->gold_q < 0)
cpi->oxcf.gold_q = q_trans[0];
else
cpi->oxcf.gold_q = q_trans[oxcf->gold_q];
}
cpi->baseline_gf_interval = cpi->oxcf.alt_freq ? cpi->oxcf.alt_freq : DEFAULT_GF_INTERVAL;
cpi->ref_frame_flags = VP8_ALT_FLAG | VP8_GOLD_FLAG | VP8_LAST_FLAG;
//cpi->use_golden_frame_only = 0;
//cpi->use_last_frame_only = 0;
cm->refresh_golden_frame = 0;
cm->refresh_last_frame = 1;
cm->refresh_entropy_probs = 1;
if (cpi->oxcf.token_partitions >= 0 && cpi->oxcf.token_partitions <= 3)
cm->multi_token_partition = (TOKEN_PARTITION) cpi->oxcf.token_partitions;
setup_features(cpi);
{
int i;
for (i = 0; i < MAX_MB_SEGMENTS; i++)
cpi->segment_encode_breakout[i] = cpi->oxcf.encode_breakout;
}
// At the moment the first order values may not be > MAXQ
if (cpi->oxcf.fixed_q > MAXQ)
cpi->oxcf.fixed_q = MAXQ;
// local file playback mode == really big buffer
if (cpi->oxcf.end_usage == USAGE_LOCAL_FILE_PLAYBACK)
{
increase rate control buffer level precision The external API exposes the RC initial/optimal/full buffer level in milliseconds, but this value was truncated internally to seconds. This patch allows the use of the full precision during the conversion from time to bits. Change-Id: If8dd2a87614c05747f81432cbe75dd9e6ed2f04e
2010-08-20 17:04:10 +02:00
cpi->oxcf.starting_buffer_level = 60000;
cpi->oxcf.optimal_buffer_level = 60000;
cpi->oxcf.maximum_buffer_size = 240000;
Initial WebM release
2010-05-18 17:58:33 +02:00
}
// Convert target bandwidth from Kbit/s to Bit/s
cpi->oxcf.target_bandwidth *= 1000;
increase rate control buffer level precision The external API exposes the RC initial/optimal/full buffer level in milliseconds, but this value was truncated internally to seconds. This patch allows the use of the full precision during the conversion from time to bits. Change-Id: If8dd2a87614c05747f81432cbe75dd9e6ed2f04e
2010-08-20 17:04:10 +02:00
cpi->oxcf.starting_buffer_level =
rescale(cpi->oxcf.starting_buffer_level,
cpi->oxcf.target_bandwidth, 1000);
Initial WebM release
2010-05-18 17:58:33 +02:00
if (cpi->oxcf.optimal_buffer_level == 0)
cpi->oxcf.optimal_buffer_level = cpi->oxcf.target_bandwidth / 8;
else
increase rate control buffer level precision The external API exposes the RC initial/optimal/full buffer level in milliseconds, but this value was truncated internally to seconds. This patch allows the use of the full precision during the conversion from time to bits. Change-Id: If8dd2a87614c05747f81432cbe75dd9e6ed2f04e
2010-08-20 17:04:10 +02:00
cpi->oxcf.optimal_buffer_level =
rescale(cpi->oxcf.optimal_buffer_level,
cpi->oxcf.target_bandwidth, 1000);
Initial WebM release
2010-05-18 17:58:33 +02:00
if (cpi->oxcf.maximum_buffer_size == 0)
cpi->oxcf.maximum_buffer_size = cpi->oxcf.target_bandwidth / 8;
else
increase rate control buffer level precision The external API exposes the RC initial/optimal/full buffer level in milliseconds, but this value was truncated internally to seconds. This patch allows the use of the full precision during the conversion from time to bits. Change-Id: If8dd2a87614c05747f81432cbe75dd9e6ed2f04e
2010-08-20 17:04:10 +02:00
cpi->oxcf.maximum_buffer_size =
rescale(cpi->oxcf.maximum_buffer_size,
cpi->oxcf.target_bandwidth, 1000);
Initial WebM release
2010-05-18 17:58:33 +02:00
cpi->buffer_level = cpi->oxcf.starting_buffer_level;
cpi->bits_off_target = cpi->oxcf.starting_buffer_level;
vp8_new_frame_rate(cpi, cpi->oxcf.frame_rate);
cpi->worst_quality = cpi->oxcf.worst_allowed_q;
cpi->active_worst_quality = cpi->oxcf.worst_allowed_q;
cpi->avg_frame_qindex = cpi->oxcf.worst_allowed_q;
cpi->best_quality = cpi->oxcf.best_allowed_q;
cpi->active_best_quality = cpi->oxcf.best_allowed_q;
cpi->buffered_mode = (cpi->oxcf.optimal_buffer_level > 0) ? TRUE : FALSE;
Badly placed initialization of rolling rate monitors. This affects control of the active quantizer range. Change-Id: I30511fc81ac9f75ff20d9f1372382423d56739da
2010-09-24 18:52:55 +02:00
cpi->rolling_target_bits = cpi->av_per_frame_bandwidth;
cpi->rolling_actual_bits = cpi->av_per_frame_bandwidth;
cpi->long_rolling_target_bits = cpi->av_per_frame_bandwidth;
cpi->long_rolling_actual_bits = cpi->av_per_frame_bandwidth;
Initial WebM release
2010-05-18 17:58:33 +02:00
cpi->total_actual_bits = 0;
cpi->total_target_vs_actual = 0;
// Only allow dropped frames in buffered mode
cpi->drop_frames_allowed = cpi->oxcf.allow_df && cpi->buffered_mode;
cm->filter_type = (LOOPFILTERTYPE) cpi->filter_type;
if (!cm->use_bilinear_mc_filter)
cm->mcomp_filter_type = SIXTAP;
else
cm->mcomp_filter_type = BILINEAR;
cpi->target_bandwidth = cpi->oxcf.target_bandwidth;
cm->Width = cpi->oxcf.Width ;
cm->Height = cpi->oxcf.Height ;
cm->horiz_scale = cpi->horiz_scale;
cm->vert_scale = cpi->vert_scale ;
cpi->intra_frame_target = (4 * (cm->Width + cm->Height) / 15) * 1000; // As per VP8
// VP8 sharpness level mapping 0-7 (vs 0-10 in general VPx dialogs)
if (cpi->oxcf.Sharpness > 7)
cpi->oxcf.Sharpness = 7;
cm->sharpness_level = cpi->oxcf.Sharpness;
if (cm->horiz_scale != NORMAL || cm->vert_scale != NORMAL)
{
int UNINITIALIZED_IS_SAFE(hr), UNINITIALIZED_IS_SAFE(hs);
int UNINITIALIZED_IS_SAFE(vr), UNINITIALIZED_IS_SAFE(vs);
Scale2Ratio(cm->horiz_scale, &hr, &hs);
Scale2Ratio(cm->vert_scale, &vr, &vs);
// always go to the next whole number
cm->Width = (hs - 1 + cpi->oxcf.Width * hr) / hs;
cm->Height = (vs - 1 + cpi->oxcf.Height * vr) / vs;
}
Swap alt/gold/new/last frame buffer ptrs instead of copying. At the end of the decode, frame buffers were being copied. The frames are not updated after the copy, they are just for reference on later frames. This change allows multiple references to the same frame buffer instead of copying it. Changes needed to be made to the encoder to handle this. The encoder is still doing frame buffer copies in similar places where pointer reference could be done. Change-Id: I7c38be4d23979cc49b5f17241ca3a78703803e66
2010-07-22 14:07:32 +02:00
if (((cm->Width + 15) & 0xfffffff0) != cm->yv12_fb[cm->lst_fb_idx].y_width ||
((cm->Height + 15) & 0xfffffff0) != cm->yv12_fb[cm->lst_fb_idx].y_height ||
cm->yv12_fb[cm->lst_fb_idx].y_width == 0)
Initial WebM release
2010-05-18 17:58:33 +02:00
{
alloc_raw_frame_buffers(cpi);
vp8_alloc_compressor_data(cpi);
}
// Clamp KF frame size to quarter of data rate
if (cpi->intra_frame_target > cpi->target_bandwidth >> 2)
cpi->intra_frame_target = cpi->target_bandwidth >> 2;
if (cpi->oxcf.fixed_q >= 0)
{
cpi->last_q[0] = cpi->oxcf.fixed_q;
cpi->last_q[1] = cpi->oxcf.fixed_q;
}
cpi->Speed = cpi->oxcf.cpu_used;
// force to allowlag to 0 if lag_in_frames is 0;
if (cpi->oxcf.lag_in_frames == 0)
{
cpi->oxcf.allow_lag = 0;
}
// Limit on lag buffers as these are not currently dynamically allocated
else if (cpi->oxcf.lag_in_frames > MAX_LAG_BUFFERS)
cpi->oxcf.lag_in_frames = MAX_LAG_BUFFERS;
// YX Temp
cpi->last_alt_ref_sei = -1;
cpi->is_src_frame_alt_ref = 0;
change to make use of more trellis quantization when a subsequent frame is encoded as an alt reference frame, it is unlikely that any mb in current frame will be used as reference for future frames, so we can enable quantization optimization even when the RD constant is slightly rate-biased. The change has an overall benefit between 0.1% to 0.2% bit savings on the test sets based on vpxssim scores. Change-Id: I9aa7bc5cd573ea84e3ee655d2834c18c4460ceea
2010-10-15 03:58:34 +02:00
cpi->is_next_src_alt_ref = 0;
Initial WebM release
2010-05-18 17:58:33 +02:00
#if 0
// Experimental RD Code
cpi->frame_distortion = 0;
cpi->last_frame_distortion = 0;
#endif
}
#define M_LOG2_E 0.693147180559945309417
#define log2f(x) (log (x) / (float) M_LOG2_E)
static void cal_mvsadcosts(int *mvsadcost[2])
{
int i = 1;
mvsadcost [0] [0] = 300;
mvsadcost [1] [0] = 300;
do
{
double z = 256 * (2 * (log2f(2 * i) + .6));
mvsadcost [0][i] = (int) z;
mvsadcost [1][i] = (int) z;
mvsadcost [0][-i] = (int) z;
mvsadcost [1][-i] = (int) z;
}
while (++i <= mv_max);
}
VP8_PTR vp8_create_compressor(VP8_CONFIG *oxcf)
{
int i;
volatile union
{
VP8_COMP *cpi;
VP8_PTR ptr;
} ctx;
cosmetics: trim trailing whitespace When the license headers were updated, they accidentally contained trailing whitespace, so unfortunately we have to touch all the files again. Change-Id: I236c05fade06589e417179c0444cb39b09e4200d
2010-06-18 18:39:21 +02:00
Initial WebM release
2010-05-18 17:58:33 +02:00
VP8_COMP *cpi;
VP8_COMMON *cm;
cpi = ctx.cpi = vpx_memalign(32, sizeof(VP8_COMP));
// Check that the CPI instance is valid
if (!cpi)
return 0;
cm = &cpi->common;
vpx_memset(cpi, 0, sizeof(VP8_COMP));
if (setjmp(cm->error.jmp))
{
VP8_PTR ptr = ctx.ptr;
ctx.cpi->common.error.setjmp = 0;
vp8_remove_compressor(&ptr);
return 0;
}
cpi->common.error.setjmp = 1;
CHECK_MEM_ERROR(cpi->rdtok, vpx_calloc(256 * 3 / 2, sizeof(TOKENEXTRA)));
CHECK_MEM_ERROR(cpi->mb.ss, vpx_calloc(sizeof(search_site), (MAX_MVSEARCH_STEPS * 8) + 1));
vp8_create_common(&cpi->common);
Add runtime CPU detection support for ARM. The primary goal is to allow a binary to be built which supports NEON, but can fall back to non-NEON routines, since some Android devices do not have NEON, even if they are otherwise ARMv7 (e.g., Tegra). The configure-generated flags HAVE_ARMV7, etc., are used to decide which versions of each function to build, and when CONFIG_RUNTIME_CPU_DETECT is enabled, the correct version is chosen at run time. In order for this to work, the CFLAGS must be set to something appropriate (e.g., without -mfpu=neon for ARMv7, and with appropriate -march and -mcpu for even earlier configurations), or the native C code will not be able to run. The ASFLAGS must remain set for the most advanced instruction set required at build time, since the ARM assembler will refuse to emit them otherwise. I have not attempted to make any changes to configure to do this automatically. Doing so will probably require the addition of new configure options. Many of the hooks for RTCD on ARM were already there, but a lot of the code had bit-rotted, and a good deal of the ARM-specific code is not integrated into the RTCD structs at all. I did not try to resolve the latter, merely to add the minimal amount of protection around them to allow RTCD to work. Those functions that were called based on an ifdef at the calling site were expanded to check the RTCD flags at that site, but they should be added to an RTCD struct somewhere in the future. The functions invoked with global function pointers still are, but these should be moved into an RTCD struct for thread safety (I believe every platform currently supported has atomic pointer stores, but this is not guaranteed). The encoder's boolhuff functions did not even have _c and armv7 suffixes, and the correct version was resolved at link time. The token packing functions did have appropriate suffixes, but the version was selected with a define, with no associated RTCD struct. However, for both of these, the only armv7 instruction they actually used was rbit, and this was completely superfluous, so I reworked them to avoid it. The only non-ARMv4 instruction remaining in them is clz, which is ARMv5 (not even ARMv5TE is required). Considering that there are no ARM-specific configs which are not at least ARMv5TE, I did not try to detect these at runtime, and simply enable them for ARMv5 and above. Finally, the NEON register saving code was completely non-reentrant, since it saved the registers to a global, static variable. I moved the storage for this onto the stack. A single binary built with this code was tested on an ARM11 (ARMv6) and a Cortex A8 (ARMv7 w/NEON), for both the encoder and decoder, and produced identical output, while using the correct accelerated functions on each. I did not test on any earlier processors. Change-Id: I45cbd63a614f4554c3b325c45d46c0806f009eaa
2010-10-21 00:39:11 +02:00
vp8_cmachine_specific_config(cpi);
Initial WebM release
2010-05-18 17:58:33 +02:00
vp8_init_config((VP8_PTR)cpi, oxcf);
memcpy(cpi->base_skip_false_prob, vp8cx_base_skip_false_prob, sizeof(vp8cx_base_skip_false_prob));
cpi->common.current_video_frame = 0;
cpi->kf_overspend_bits = 0;
cpi->kf_bitrate_adjustment = 0;
cpi->frames_till_gf_update_due = 0;
cpi->gf_overspend_bits = 0;
cpi->non_gf_bitrate_adjustment = 0;
cpi->prob_last_coded = 128;
cpi->prob_gf_coded = 128;
cpi->prob_intra_coded = 63;
// Prime the recent reference frame useage counters.
// Hereafter they will be maintained as a sort of moving average
cpi->recent_ref_frame_usage[INTRA_FRAME] = 1;
cpi->recent_ref_frame_usage[LAST_FRAME] = 1;
cpi->recent_ref_frame_usage[GOLDEN_FRAME] = 1;
cpi->recent_ref_frame_usage[ALTREF_FRAME] = 1;
// Set reference frame sign bias for ALTREF frame to 1 (for now)
cpi->common.ref_frame_sign_bias[ALTREF_FRAME] = 1;
cpi->gf_decay_rate = 0;
cpi->baseline_gf_interval = DEFAULT_GF_INTERVAL;
cpi->gold_is_last = 0 ;
cpi->alt_is_last = 0 ;
cpi->gold_is_alt = 0 ;
Improve MV prediction accuracy to achieve performance gain Add vp8_mv_pred() to better predict starting MV for NEWMV mode in vp8_rd_pick_inter_mode(). Set different search ranges according to MV prediction accuracy, which improves encoder performance without hurting the quality. Also, as Yaowu suggested, using diamond search result as full search starting point and therefore adjusting(reducing) full search range helps the performance. Change-Id: Ie4a3c8df87e697c1f4f6e2ddb693766bba1b77b6
2010-12-03 17:26:21 +01:00
// allocate memory for storing last frame's MVs for MV prediction.
CHECK_MEM_ERROR(cpi->lfmv, vpx_calloc((cpi->common.mb_rows+1) * (cpi->common.mb_cols+1), sizeof(int_mv)));
CHECK_MEM_ERROR(cpi->lf_ref_frame_sign_bias, vpx_calloc((cpi->common.mb_rows+1) * (cpi->common.mb_cols+1), sizeof(int)));
CHECK_MEM_ERROR(cpi->lf_ref_frame, vpx_calloc((cpi->common.mb_rows+1) * (cpi->common.mb_cols+1), sizeof(int)));
Initial WebM release
2010-05-18 17:58:33 +02:00
// Create the encoder segmentation map and set all entries to 0
CHECK_MEM_ERROR(cpi->segmentation_map, vpx_calloc(cpi->common.mb_rows * cpi->common.mb_cols, 1));
CHECK_MEM_ERROR(cpi->active_map, vpx_calloc(cpi->common.mb_rows * cpi->common.mb_cols, 1));
vpx_memset(cpi->active_map , 1, (cpi->common.mb_rows * cpi->common.mb_cols));
cpi->active_map_enabled = 0;
// Create the first pass motion map structure and set to 0
Made AltRef filter adaptive & added motion compensation Modified AltRef temporal filter to adapt filter length based on macroblock coding modes selected during first-pass encode. Also added sub-pixel motion compensation to the AltRef filter.
2010-09-28 16:23:41 +02:00
// Allocate space for maximum of 15 buffers
CHECK_MEM_ERROR(cpi->fp_motion_map, vpx_calloc(15*cpi->common.MBs, 1));
Initial WebM release
2010-05-18 17:58:33 +02:00
#if 0
// Experimental code for lagged and one pass
// Initialise one_pass GF frames stats
// Update stats used for GF selection
if (cpi->pass == 0)
{
cpi->one_pass_frame_index = 0;
for (i = 0; i < MAX_LAG_BUFFERS; i++)
{
cpi->one_pass_frame_stats[i].frames_so_far = 0;
cpi->one_pass_frame_stats[i].frame_intra_error = 0.0;
cpi->one_pass_frame_stats[i].frame_coded_error = 0.0;
cpi->one_pass_frame_stats[i].frame_pcnt_inter = 0.0;
cpi->one_pass_frame_stats[i].frame_pcnt_motion = 0.0;
cpi->one_pass_frame_stats[i].frame_mvr = 0.0;
cpi->one_pass_frame_stats[i].frame_mvr_abs = 0.0;
cpi->one_pass_frame_stats[i].frame_mvc = 0.0;
cpi->one_pass_frame_stats[i].frame_mvc_abs = 0.0;
}
}
#endif
// Should we use the cyclic refresh method.
// Currently this is tied to error resilliant mode
cpi->cyclic_refresh_mode_enabled = cpi->oxcf.error_resilient_mode;
cpi->cyclic_refresh_mode_max_mbs_perframe = (cpi->common.mb_rows * cpi->common.mb_cols) / 40;
cpi->cyclic_refresh_mode_index = 0;
cpi->cyclic_refresh_q = 32;
if (cpi->cyclic_refresh_mode_enabled)
{
CHECK_MEM_ERROR(cpi->cyclic_refresh_map, vpx_calloc((cpi->common.mb_rows * cpi->common.mb_cols), 1));
}
else
cpi->cyclic_refresh_map = (signed char *) NULL;
// Test function for segmentation
//segmentation_test_function((VP8_PTR) cpi);
#ifdef ENTROPY_STATS
init_context_counters();
#endif
Add simple version of activity masking. This uses MB variance to change the RDO weight for mode decision and quantization. Activity is normalized against the average for the frame, which is currently tracked using feed-forward statistics. This could also be used to adjust the quantizer for the entire frame, but that requires more extensive rate control changes. This does not yet attempt to adapt the quantizer within the frame, but the signaling cost means that will likely only be useful at very high rates. Change-Id: I26cd7c755cac3ff33cfe0688b1da50b2b87b9c93
2010-10-11 23:37:27 +02:00
/*Initialize the feed-forward activity masking.*/
cpi->activity_avg = 90<<12;
Initial WebM release
2010-05-18 17:58:33 +02:00
cpi->frames_since_key = 8; // Give a sensible default for the first frame.
cpi->key_frame_frequency = cpi->oxcf.key_freq;
cpi->source_alt_ref_pending = FALSE;
cpi->source_alt_ref_active = FALSE;
cpi->common.refresh_alt_ref_frame = 0;
cpi->b_calculate_psnr = CONFIG_PSNR;
#if CONFIG_PSNR
cpi->b_calculate_ssimg = 0;
cpi->count = 0;
cpi->bytes = 0;
if (cpi->b_calculate_psnr)
{
cpi->total_sq_error = 0.0;
cpi->total_sq_error2 = 0.0;
cpi->total_y = 0.0;
cpi->total_u = 0.0;
cpi->total_v = 0.0;
cpi->total = 0.0;
cpi->totalp_y = 0.0;
cpi->totalp_u = 0.0;
cpi->totalp_v = 0.0;
cpi->totalp = 0.0;
cpi->tot_recode_hits = 0;
cpi->summed_quality = 0;
cpi->summed_weights = 0;
}
if (cpi->b_calculate_ssimg)
{
cpi->total_ssimg_y = 0;
cpi->total_ssimg_u = 0;
cpi->total_ssimg_v = 0;
cpi->total_ssimg_all = 0;
}
#ifndef LLONG_MAX
#define LLONG_MAX 9223372036854775807LL
#endif
cpi->first_time_stamp_ever = LLONG_MAX;
#endif
cpi->frames_till_gf_update_due = 0;
cpi->key_frame_count = 1;
cpi->tot_key_frame_bits = 0;
cpi->ni_av_qi = cpi->oxcf.worst_allowed_q;
cpi->ni_tot_qi = 0;
cpi->ni_frames = 0;
cpi->total_byte_count = 0;
cpi->drop_frame = 0;
cpi->drop_count = 0;
cpi->max_drop_count = 0;
cpi->max_consec_dropped_frames = 4;
cpi->rate_correction_factor = 1.0;
cpi->key_frame_rate_correction_factor = 1.0;
cpi->gf_rate_correction_factor = 1.0;
cpi->est_max_qcorrection_factor = 1.0;
cpi->mb.mvcost[0] = &cpi->mb.mvcosts[0][mv_max+1];
cpi->mb.mvcost[1] = &cpi->mb.mvcosts[1][mv_max+1];
cpi->mb.mvsadcost[0] = &cpi->mb.mvsadcosts[0][mv_max+1];
cpi->mb.mvsadcost[1] = &cpi->mb.mvsadcosts[1][mv_max+1];
cal_mvsadcosts(cpi->mb.mvsadcost);
for (i = 0; i < KEY_FRAME_CONTEXT; i++)
{
cpi->prior_key_frame_size[i] = cpi->intra_frame_target;
cpi->prior_key_frame_distance[i] = (int)cpi->output_frame_rate;
}
cpi->check_freq[0] = 15;
cpi->check_freq[1] = 15;
#ifdef OUTPUT_YUV_SRC
yuv_file = fopen("bd.yuv", "ab");
#endif
#if 0
framepsnr = fopen("framepsnr.stt", "a");
kf_list = fopen("kf_list.stt", "w");
#endif
cpi->output_pkt_list = oxcf->output_pkt_list;
#if !(CONFIG_REALTIME_ONLY)
if (cpi->pass == 1)
{
vp8_init_first_pass(cpi);
}
else if (cpi->pass == 2)
{
Move firstpass motion map to stats packet The first implementation of the firstpass motion map for motion compensated temporal filtering created a file, fpmotionmap.stt, in the current working directory. This was not safe for multiple encoder instances. This patch merges this data into the first pass stats packet interface, so that it is handled like the other (numerical) firstpass stats. The new stats packet is defined as follows: Numerical Stats (16 doubles) -- 128 bytes Motion Map -- 1 byte / Macroblock Padding -- to align packet to 8 bytes The fpmotionmap.stt file can still be generated for debugging purposes in the same way that the textual version of the stats are available (defining OUTPUT_FPF in firstpass.c) Change-Id: I083ffbfd95e7d6a42bb4039ba0e81f678c8183ca
2010-10-14 22:40:12 +02:00
size_t packet_sz = vp8_firstpass_stats_sz(cpi->common.MBs);
int packets = oxcf->two_pass_stats_in.sz / packet_sz;
Initial WebM release
2010-05-18 17:58:33 +02:00
cpi->stats_in = oxcf->two_pass_stats_in.buf;
Move firstpass motion map to stats packet The first implementation of the firstpass motion map for motion compensated temporal filtering created a file, fpmotionmap.stt, in the current working directory. This was not safe for multiple encoder instances. This patch merges this data into the first pass stats packet interface, so that it is handled like the other (numerical) firstpass stats. The new stats packet is defined as follows: Numerical Stats (16 doubles) -- 128 bytes Motion Map -- 1 byte / Macroblock Padding -- to align packet to 8 bytes The fpmotionmap.stt file can still be generated for debugging purposes in the same way that the textual version of the stats are available (defining OUTPUT_FPF in firstpass.c) Change-Id: I083ffbfd95e7d6a42bb4039ba0e81f678c8183ca
2010-10-14 22:40:12 +02:00
cpi->stats_in_end = (void*)((char *)cpi->stats_in
+ (packets - 1) * packet_sz);
Initial WebM release
2010-05-18 17:58:33 +02:00
vp8_init_second_pass(cpi);
}
#endif
if (cpi->compressor_speed == 2)
{
cpi->cpu_freq = 0; //vp8_get_processor_freq();
cpi->avg_encode_time = 0;
cpi->avg_pick_mode_time = 0;
}
vp8_set_speed_features(cpi);
// Set starting values of RD threshold multipliers (128 = *1)
for (i = 0; i < MAX_MODES; i++)
{
cpi->rd_thresh_mult[i] = 128;
}
#ifdef ENTROPY_STATS
init_mv_ref_counts();
#endif
vp8cx_create_encoder_threads(cpi);
Add half-pixel variance RTCD functions NEON has optimized 16x16 half-pixel variance functions, but they were not part of the RTCD framework. Add these functions to RTCD, so that other platforms can make use of this optimization in the future and special-case ARM code can be removed. A number of functions were taking two variance functions as parameters. These functions were changed to take a single parameter, a pointer to a struct containing all the variance functions for that block size. This provides additional flexibility for calling additional variance functions (the half-pixel special case, for example) and by initializing the table for all block sizes, we don't have to construct this function pointer table for each macroblock. Change-Id: I78289ff36b2715f9a7aa04d5f6fbe3d23acdc29c
2010-10-26 21:34:16 +02:00
cpi->fn_ptr[BLOCK_16X16].sdf = VARIANCE_INVOKE(&cpi->rtcd.variance, sad16x16);
cpi->fn_ptr[BLOCK_16X16].vf = VARIANCE_INVOKE(&cpi->rtcd.variance, var16x16);
cpi->fn_ptr[BLOCK_16X16].svf = VARIANCE_INVOKE(&cpi->rtcd.variance, subpixvar16x16);
cpi->fn_ptr[BLOCK_16X16].svf_halfpix_h = VARIANCE_INVOKE(&cpi->rtcd.variance, halfpixvar16x16_h);
cpi->fn_ptr[BLOCK_16X16].svf_halfpix_v = VARIANCE_INVOKE(&cpi->rtcd.variance, halfpixvar16x16_v);
cpi->fn_ptr[BLOCK_16X16].svf_halfpix_hv = VARIANCE_INVOKE(&cpi->rtcd.variance, halfpixvar16x16_hv);
cpi->fn_ptr[BLOCK_16X16].sdx3f = VARIANCE_INVOKE(&cpi->rtcd.variance, sad16x16x3);
Full search SAD function optimization in SSE4.1 Use mpsadbw, and calculate 8 sad at once. Function list: vp8_sad16x16x8_sse4 vp8_sad16x8x8_sse4 vp8_sad8x16x8_sse4 vp8_sad8x8x8_sse4 vp8_sad4x4x8_sse4 (test clip: tulip) For best quality mode, this gave encoder a 5% performance boost. For good quality mode with speed=1, this gave encoder a 3% performance boost. Change-Id: I083b5a39d39144f88dcbccbef95da6498e490134
2010-10-27 14:45:24 +02:00
cpi->fn_ptr[BLOCK_16X16].sdx8f = VARIANCE_INVOKE(&cpi->rtcd.variance, sad16x16x8);
Add half-pixel variance RTCD functions NEON has optimized 16x16 half-pixel variance functions, but they were not part of the RTCD framework. Add these functions to RTCD, so that other platforms can make use of this optimization in the future and special-case ARM code can be removed. A number of functions were taking two variance functions as parameters. These functions were changed to take a single parameter, a pointer to a struct containing all the variance functions for that block size. This provides additional flexibility for calling additional variance functions (the half-pixel special case, for example) and by initializing the table for all block sizes, we don't have to construct this function pointer table for each macroblock. Change-Id: I78289ff36b2715f9a7aa04d5f6fbe3d23acdc29c
2010-10-26 21:34:16 +02:00
cpi->fn_ptr[BLOCK_16X16].sdx4df = VARIANCE_INVOKE(&cpi->rtcd.variance, sad16x16x4d);
cpi->fn_ptr[BLOCK_16X8].sdf = VARIANCE_INVOKE(&cpi->rtcd.variance, sad16x8);
cpi->fn_ptr[BLOCK_16X8].vf = VARIANCE_INVOKE(&cpi->rtcd.variance, var16x8);
cpi->fn_ptr[BLOCK_16X8].svf = VARIANCE_INVOKE(&cpi->rtcd.variance, subpixvar16x8);
cpi->fn_ptr[BLOCK_16X8].svf_halfpix_h = NULL;
cpi->fn_ptr[BLOCK_16X8].svf_halfpix_v = NULL;
cpi->fn_ptr[BLOCK_16X8].svf_halfpix_hv = NULL;
cpi->fn_ptr[BLOCK_16X8].sdx3f = VARIANCE_INVOKE(&cpi->rtcd.variance, sad16x8x3);
Full search SAD function optimization in SSE4.1 Use mpsadbw, and calculate 8 sad at once. Function list: vp8_sad16x16x8_sse4 vp8_sad16x8x8_sse4 vp8_sad8x16x8_sse4 vp8_sad8x8x8_sse4 vp8_sad4x4x8_sse4 (test clip: tulip) For best quality mode, this gave encoder a 5% performance boost. For good quality mode with speed=1, this gave encoder a 3% performance boost. Change-Id: I083b5a39d39144f88dcbccbef95da6498e490134
2010-10-27 14:45:24 +02:00
cpi->fn_ptr[BLOCK_16X8].sdx8f = VARIANCE_INVOKE(&cpi->rtcd.variance, sad16x8x8);
Add half-pixel variance RTCD functions NEON has optimized 16x16 half-pixel variance functions, but they were not part of the RTCD framework. Add these functions to RTCD, so that other platforms can make use of this optimization in the future and special-case ARM code can be removed. A number of functions were taking two variance functions as parameters. These functions were changed to take a single parameter, a pointer to a struct containing all the variance functions for that block size. This provides additional flexibility for calling additional variance functions (the half-pixel special case, for example) and by initializing the table for all block sizes, we don't have to construct this function pointer table for each macroblock. Change-Id: I78289ff36b2715f9a7aa04d5f6fbe3d23acdc29c
2010-10-26 21:34:16 +02:00
cpi->fn_ptr[BLOCK_16X8].sdx4df = VARIANCE_INVOKE(&cpi->rtcd.variance, sad16x8x4d);
cpi->fn_ptr[BLOCK_8X16].sdf = VARIANCE_INVOKE(&cpi->rtcd.variance, sad8x16);
cpi->fn_ptr[BLOCK_8X16].vf = VARIANCE_INVOKE(&cpi->rtcd.variance, var8x16);
cpi->fn_ptr[BLOCK_8X16].svf = VARIANCE_INVOKE(&cpi->rtcd.variance, subpixvar8x16);
cpi->fn_ptr[BLOCK_8X16].svf_halfpix_h = NULL;
cpi->fn_ptr[BLOCK_8X16].svf_halfpix_v = NULL;
cpi->fn_ptr[BLOCK_8X16].svf_halfpix_hv = NULL;
cpi->fn_ptr[BLOCK_8X16].sdx3f = VARIANCE_INVOKE(&cpi->rtcd.variance, sad8x16x3);
Full search SAD function optimization in SSE4.1 Use mpsadbw, and calculate 8 sad at once. Function list: vp8_sad16x16x8_sse4 vp8_sad16x8x8_sse4 vp8_sad8x16x8_sse4 vp8_sad8x8x8_sse4 vp8_sad4x4x8_sse4 (test clip: tulip) For best quality mode, this gave encoder a 5% performance boost. For good quality mode with speed=1, this gave encoder a 3% performance boost. Change-Id: I083b5a39d39144f88dcbccbef95da6498e490134
2010-10-27 14:45:24 +02:00
cpi->fn_ptr[BLOCK_8X16].sdx8f = VARIANCE_INVOKE(&cpi->rtcd.variance, sad8x16x8);
Add half-pixel variance RTCD functions NEON has optimized 16x16 half-pixel variance functions, but they were not part of the RTCD framework. Add these functions to RTCD, so that other platforms can make use of this optimization in the future and special-case ARM code can be removed. A number of functions were taking two variance functions as parameters. These functions were changed to take a single parameter, a pointer to a struct containing all the variance functions for that block size. This provides additional flexibility for calling additional variance functions (the half-pixel special case, for example) and by initializing the table for all block sizes, we don't have to construct this function pointer table for each macroblock. Change-Id: I78289ff36b2715f9a7aa04d5f6fbe3d23acdc29c
2010-10-26 21:34:16 +02:00
cpi->fn_ptr[BLOCK_8X16].sdx4df = VARIANCE_INVOKE(&cpi->rtcd.variance, sad8x16x4d);
cpi->fn_ptr[BLOCK_8X8].sdf = VARIANCE_INVOKE(&cpi->rtcd.variance, sad8x8);
cpi->fn_ptr[BLOCK_8X8].vf = VARIANCE_INVOKE(&cpi->rtcd.variance, var8x8);
cpi->fn_ptr[BLOCK_8X8].svf = VARIANCE_INVOKE(&cpi->rtcd.variance, subpixvar8x8);
cpi->fn_ptr[BLOCK_8X8].svf_halfpix_h = NULL;
cpi->fn_ptr[BLOCK_8X8].svf_halfpix_v = NULL;
cpi->fn_ptr[BLOCK_8X8].svf_halfpix_hv = NULL;
cpi->fn_ptr[BLOCK_8X8].sdx3f = VARIANCE_INVOKE(&cpi->rtcd.variance, sad8x8x3);
Full search SAD function optimization in SSE4.1 Use mpsadbw, and calculate 8 sad at once. Function list: vp8_sad16x16x8_sse4 vp8_sad16x8x8_sse4 vp8_sad8x16x8_sse4 vp8_sad8x8x8_sse4 vp8_sad4x4x8_sse4 (test clip: tulip) For best quality mode, this gave encoder a 5% performance boost. For good quality mode with speed=1, this gave encoder a 3% performance boost. Change-Id: I083b5a39d39144f88dcbccbef95da6498e490134
2010-10-27 14:45:24 +02:00
cpi->fn_ptr[BLOCK_8X8].sdx8f = VARIANCE_INVOKE(&cpi->rtcd.variance, sad8x8x8);
Add half-pixel variance RTCD functions NEON has optimized 16x16 half-pixel variance functions, but they were not part of the RTCD framework. Add these functions to RTCD, so that other platforms can make use of this optimization in the future and special-case ARM code can be removed. A number of functions were taking two variance functions as parameters. These functions were changed to take a single parameter, a pointer to a struct containing all the variance functions for that block size. This provides additional flexibility for calling additional variance functions (the half-pixel special case, for example) and by initializing the table for all block sizes, we don't have to construct this function pointer table for each macroblock. Change-Id: I78289ff36b2715f9a7aa04d5f6fbe3d23acdc29c
2010-10-26 21:34:16 +02:00
cpi->fn_ptr[BLOCK_8X8].sdx4df = VARIANCE_INVOKE(&cpi->rtcd.variance, sad8x8x4d);
cpi->fn_ptr[BLOCK_4X4].sdf = VARIANCE_INVOKE(&cpi->rtcd.variance, sad4x4);
cpi->fn_ptr[BLOCK_4X4].vf = VARIANCE_INVOKE(&cpi->rtcd.variance, var4x4);
cpi->fn_ptr[BLOCK_4X4].svf = VARIANCE_INVOKE(&cpi->rtcd.variance, subpixvar4x4);
cpi->fn_ptr[BLOCK_4X4].svf_halfpix_h = NULL;
cpi->fn_ptr[BLOCK_4X4].svf_halfpix_v = NULL;
cpi->fn_ptr[BLOCK_4X4].svf_halfpix_hv = NULL;
cpi->fn_ptr[BLOCK_4X4].sdx3f = VARIANCE_INVOKE(&cpi->rtcd.variance, sad4x4x3);
Full search SAD function optimization in SSE4.1 Use mpsadbw, and calculate 8 sad at once. Function list: vp8_sad16x16x8_sse4 vp8_sad16x8x8_sse4 vp8_sad8x16x8_sse4 vp8_sad8x8x8_sse4 vp8_sad4x4x8_sse4 (test clip: tulip) For best quality mode, this gave encoder a 5% performance boost. For good quality mode with speed=1, this gave encoder a 3% performance boost. Change-Id: I083b5a39d39144f88dcbccbef95da6498e490134
2010-10-27 14:45:24 +02:00
cpi->fn_ptr[BLOCK_4X4].sdx8f = VARIANCE_INVOKE(&cpi->rtcd.variance, sad4x4x8);
Add half-pixel variance RTCD functions NEON has optimized 16x16 half-pixel variance functions, but they were not part of the RTCD framework. Add these functions to RTCD, so that other platforms can make use of this optimization in the future and special-case ARM code can be removed. A number of functions were taking two variance functions as parameters. These functions were changed to take a single parameter, a pointer to a struct containing all the variance functions for that block size. This provides additional flexibility for calling additional variance functions (the half-pixel special case, for example) and by initializing the table for all block sizes, we don't have to construct this function pointer table for each macroblock. Change-Id: I78289ff36b2715f9a7aa04d5f6fbe3d23acdc29c
2010-10-26 21:34:16 +02:00
cpi->fn_ptr[BLOCK_4X4].sdx4df = VARIANCE_INVOKE(&cpi->rtcd.variance, sad4x4x4d);
Initial WebM release
2010-05-18 17:58:33 +02:00
#if !(CONFIG_REALTIME_ONLY)
cpi->full_search_sad = SEARCH_INVOKE(&cpi->rtcd.search, full_search);
#endif
cpi->diamond_search_sad = SEARCH_INVOKE(&cpi->rtcd.search, diamond_search);
cpi->ready_for_new_frame = 1;
cpi->source_encode_index = 0;
// make sure frame 1 is okay
cpi->error_bins[0] = cpi->common.MBs;
//vp8cx_init_quantizer() is first called here. Add check in vp8cx_frame_init_quantizer() so that vp8cx_init_quantizer is only called later
//when needed. This will avoid unnecessary calls of vp8cx_init_quantizer() for every frame.
vp8cx_init_quantizer(cpi);
{
vp8_init_loop_filter(cm);
cm->last_frame_type = KEY_FRAME;
cm->last_filter_type = cm->filter_type;
cm->last_sharpness_level = cm->sharpness_level;
}
cpi->common.error.setjmp = 0;
return (VP8_PTR) cpi;
}
void vp8_remove_compressor(VP8_PTR *ptr)
{
VP8_COMP *cpi = (VP8_COMP *)(*ptr);
if (!cpi)
return;
if (cpi && (cpi->common.current_video_frame > 0))
{
#if !(CONFIG_REALTIME_ONLY)
if (cpi->pass == 2)
{
vp8_end_second_pass(cpi);
}
#endif
#ifdef ENTROPY_STATS
print_context_counters();
print_tree_update_probs();
print_mode_context();
#endif
#if CONFIG_PSNR
if (cpi->pass != 1)
{
FILE *f = fopen("opsnr.stt", "a");
double time_encoded = (cpi->source_end_time_stamp - cpi->first_time_stamp_ever) / 10000000.000;
double total_encode_time = (cpi->time_receive_data + cpi->time_compress_data) / 1000.000;
double dr = (double)cpi->bytes * (double) 8 / (double)1000 / time_encoded;
if (cpi->b_calculate_psnr)
{
Swap alt/gold/new/last frame buffer ptrs instead of copying. At the end of the decode, frame buffers were being copied. The frames are not updated after the copy, they are just for reference on later frames. This change allows multiple references to the same frame buffer instead of copying it. Changes needed to be made to the encoder to handle this. The encoder is still doing frame buffer copies in similar places where pointer reference could be done. Change-Id: I7c38be4d23979cc49b5f17241ca3a78703803e66
2010-07-22 14:07:32 +02:00
YV12_BUFFER_CONFIG *lst_yv12 = &cpi->common.yv12_fb[cpi->common.lst_fb_idx];
double samples = 3.0 / 2 * cpi->count * lst_yv12->y_width * lst_yv12->y_height;
Initial WebM release
2010-05-18 17:58:33 +02:00
double total_psnr = vp8_mse2psnr(samples, 255.0, cpi->total_sq_error);
double total_psnr2 = vp8_mse2psnr(samples, 255.0, cpi->total_sq_error2);
double total_ssim = 100 * pow(cpi->summed_quality / cpi->summed_weights, 8.0);
Fix stats format and correct data size and bit rate output Change-ID: I093abe6094589a0d73f6ca85b825678a19e68285
2010-05-27 19:32:40 +02:00
fprintf(f, "Bitrate\tAVGPsnr\tGLBPsnr\tAVPsnrP\tGLPsnrP\tVPXSSIM\t Time(us)\n");
Initial WebM release
2010-05-18 17:58:33 +02:00
fprintf(f, "%7.3f\t%7.3f\t%7.3f\t%7.3f\t%7.3f\t%7.3f %8.0f\n",
dr, cpi->total / cpi->count, total_psnr, cpi->totalp / cpi->count, total_psnr2, total_ssim,
total_encode_time);
}
if (cpi->b_calculate_ssimg)
{
fprintf(f, "BitRate\tSSIM_Y\tSSIM_U\tSSIM_V\tSSIM_A\t Time(us)\n");
fprintf(f, "%7.3f\t%6.4f\t%6.4f\t%6.4f\t%6.4f\t%8.0f\n", dr,
cpi->total_ssimg_y / cpi->count, cpi->total_ssimg_u / cpi->count,
cpi->total_ssimg_v / cpi->count, cpi->total_ssimg_all / cpi->count, total_encode_time);
}
fclose(f);
#if 0
f = fopen("qskip.stt", "a");
fprintf(f, "minq:%d -maxq:%d skipture:skipfalse = %d:%d\n", cpi->oxcf.best_allowed_q, cpi->oxcf.worst_allowed_q, skiptruecount, skipfalsecount);
fclose(f);
#endif
}
#endif
#ifdef SPEEDSTATS
if (cpi->compressor_speed == 2)
{
int i;
FILE *f = fopen("cxspeed.stt", "a");
cnt_pm /= cpi->common.MBs;
for (i = 0; i < 16; i++)
fprintf(f, "%5d", frames_at_speed[i]);
fprintf(f, "\n");
//fprintf(f, "%10d PM %10d %10d %10d EF %10d %10d %10d\n", cpi->Speed, cpi->avg_pick_mode_time, (tot_pm/cnt_pm), cnt_pm, cpi->avg_encode_time, 0, 0);
fclose(f);
}
#endif
#ifdef MODE_STATS
{
extern int count_mb_seg[4];
FILE *f = fopen("modes.stt", "a");
double dr = (double)cpi->oxcf.frame_rate * (double)bytes * (double)8 / (double)count / (double)1000 ;
fprintf(f, "intra_mode in Intra Frames:\n");
fprintf(f, "Y: %8d, %8d, %8d, %8d, %8d\n", y_modes[0], y_modes[1], y_modes[2], y_modes[3], y_modes[4]);
fprintf(f, "UV:%8d, %8d, %8d, %8d\n", uv_modes[0], uv_modes[1], uv_modes[2], uv_modes[3]);
fprintf(f, "B: ");
{
int i;
for (i = 0; i < 10; i++)
fprintf(f, "%8d, ", b_modes[i]);
fprintf(f, "\n");
}
fprintf(f, "Modes in Inter Frames:\n");
fprintf(f, "Y: %8d, %8d, %8d, %8d, %8d, %8d, %8d, %8d, %8d, %8d\n",
inter_y_modes[0], inter_y_modes[1], inter_y_modes[2], inter_y_modes[3], inter_y_modes[4],
inter_y_modes[5], inter_y_modes[6], inter_y_modes[7], inter_y_modes[8], inter_y_modes[9]);
fprintf(f, "UV:%8d, %8d, %8d, %8d\n", inter_uv_modes[0], inter_uv_modes[1], inter_uv_modes[2], inter_uv_modes[3]);
fprintf(f, "B: ");
{
int i;
for (i = 0; i < 15; i++)
fprintf(f, "%8d, ", inter_b_modes[i]);
fprintf(f, "\n");
}
fprintf(f, "P:%8d, %8d, %8d, %8d\n", count_mb_seg[0], count_mb_seg[1], count_mb_seg[2], count_mb_seg[3]);
fprintf(f, "PB:%8d, %8d, %8d, %8d\n", inter_b_modes[LEFT4X4], inter_b_modes[ABOVE4X4], inter_b_modes[ZERO4X4], inter_b_modes[NEW4X4]);
fclose(f);
}
#endif
#ifdef ENTROPY_STATS
{
int i, j, k;
FILE *fmode = fopen("modecontext.c", "w");
fprintf(fmode, "\n#include \"entropymode.h\"\n\n");
fprintf(fmode, "const unsigned int vp8_kf_default_bmode_counts ");
fprintf(fmode, "[VP8_BINTRAMODES] [VP8_BINTRAMODES] [VP8_BINTRAMODES] =\n{\n");
for (i = 0; i < 10; i++)
{
fprintf(fmode, " { //Above Mode : %d\n", i);
for (j = 0; j < 10; j++)
{
fprintf(fmode, " {");
for (k = 0; k < 10; k++)
{
if (!intra_mode_stats[i][j][k])
fprintf(fmode, " %5d, ", 1);
else
fprintf(fmode, " %5d, ", intra_mode_stats[i][j][k]);
}
fprintf(fmode, "}, // left_mode %d\n", j);
}
fprintf(fmode, " },\n");
}
fprintf(fmode, "};\n");
Fix leaked file descriptor with ENTROPY_STATS cppcheck found a leaked file descriptor in the debugging code enabled by defining ENTROPY_STATS. Fixes issue #60. Change-Id: I0c1d0669cb94d44fed77860f97b82763be06b7cb
2010-10-25 19:14:11 +02:00
fclose(fmode);
Initial WebM release
2010-05-18 17:58:33 +02:00
}
#endif
#if defined(SECTIONBITS_OUTPUT)
if (0)
{
int i;
FILE *f = fopen("tokenbits.stt", "a");
for (i = 0; i < 28; i++)
fprintf(f, "%8d", (int)(Sectionbits[i] / 256));
fprintf(f, "\n");
fclose(f);
}
#endif
#if 0
{
printf("\n_pick_loop_filter_level:%d\n", cpi->time_pick_lpf / 1000);
printf("\n_frames recive_data encod_mb_row compress_frame Total\n");
printf("%6d %10ld %10ld %10ld %10ld\n", cpi->common.current_video_frame, cpi->time_receive_data / 1000, cpi->time_encode_mb_row / 1000, cpi->time_compress_data / 1000, (cpi->time_receive_data + cpi->time_compress_data) / 1000);
}
#endif
}
vp8cx_remove_encoder_threads(cpi);
vp8_dealloc_compressor_data(cpi);
vpx_free(cpi->mb.ss);
vpx_free(cpi->tok);
vpx_free(cpi->rdtok);
vpx_free(cpi->cyclic_refresh_map);
vp8_remove_common(&cpi->common);
vpx_free(cpi);
*ptr = 0;
#ifdef OUTPUT_YUV_SRC
fclose(yuv_file);
#endif
#if 0
if (keyfile)
fclose(keyfile);
if (framepsnr)
fclose(framepsnr);
if (kf_list)
fclose(kf_list);
#endif
}
static uint64_t calc_plane_error(unsigned char *orig, int orig_stride,
unsigned char *recon, int recon_stride,
unsigned int cols, unsigned int rows,
vp8_variance_rtcd_vtable_t *rtcd)
{
unsigned int row, col;
uint64_t total_sse = 0;
int diff;
for (row = 0; row + 16 <= rows; row += 16)
{
for (col = 0; col + 16 <= cols; col += 16)
{
unsigned int sse;
VARIANCE_INVOKE(rtcd, mse16x16)(orig + col, orig_stride,
recon + col, recon_stride,
&sse);
total_sse += sse;
}
/* Handle odd-sized width */
if (col < cols)
{
unsigned int border_row, border_col;
unsigned char *border_orig = orig;
unsigned char *border_recon = recon;
for (border_row = 0; border_row < 16; border_row++)
{
for (border_col = col; border_col < cols; border_col++)
{
diff = border_orig[border_col] - border_recon[border_col];
total_sse += diff * diff;
}
border_orig += orig_stride;
border_recon += recon_stride;
}
}
orig += orig_stride * 16;
recon += recon_stride * 16;
}
/* Handle odd-sized height */
for (; row < rows; row++)
{
for (col = 0; col < cols; col++)
{
diff = orig[col] - recon[col];
total_sse += diff * diff;
}
orig += orig_stride;
recon += recon_stride;
}
return total_sse;
}
static void generate_psnr_packet(VP8_COMP *cpi)
{
YV12_BUFFER_CONFIG *orig = cpi->Source;
YV12_BUFFER_CONFIG *recon = cpi->common.frame_to_show;
struct vpx_codec_cx_pkt pkt;
uint64_t sse;
int i;
unsigned int width = cpi->common.Width;
unsigned int height = cpi->common.Height;
pkt.kind = VPX_CODEC_PSNR_PKT;
sse = calc_plane_error(orig->y_buffer, orig->y_stride,
recon->y_buffer, recon->y_stride,
width, height,
IF_RTCD(&cpi->rtcd.variance));
pkt.data.psnr.sse[0] = sse;
pkt.data.psnr.sse[1] = sse;
pkt.data.psnr.samples[0] = width * height;
pkt.data.psnr.samples[1] = width * height;
width = (width + 1) / 2;
height = (height + 1) / 2;
sse = calc_plane_error(orig->u_buffer, orig->uv_stride,
recon->u_buffer, recon->uv_stride,
width, height,
IF_RTCD(&cpi->rtcd.variance));
pkt.data.psnr.sse[0] += sse;
pkt.data.psnr.sse[2] = sse;
pkt.data.psnr.samples[0] += width * height;
pkt.data.psnr.samples[2] = width * height;
sse = calc_plane_error(orig->v_buffer, orig->uv_stride,
recon->v_buffer, recon->uv_stride,
width, height,
IF_RTCD(&cpi->rtcd.variance));
pkt.data.psnr.sse[0] += sse;
pkt.data.psnr.sse[3] = sse;
pkt.data.psnr.samples[0] += width * height;
pkt.data.psnr.samples[3] = width * height;
for (i = 0; i < 4; i++)
pkt.data.psnr.psnr[i] = vp8_mse2psnr(pkt.data.psnr.samples[i], 255.0,
pkt.data.psnr.sse[i]);
vpx_codec_pkt_list_add(cpi->output_pkt_list, &pkt);
}
int vp8_use_as_reference(VP8_PTR ptr, int ref_frame_flags)
{
VP8_COMP *cpi = (VP8_COMP *)(ptr);
if (ref_frame_flags > 7)
return -1 ;
cpi->ref_frame_flags = ref_frame_flags;
return 0;
}
int vp8_update_reference(VP8_PTR ptr, int ref_frame_flags)
{
VP8_COMP *cpi = (VP8_COMP *)(ptr);
if (ref_frame_flags > 7)
return -1 ;
cpi->common.refresh_golden_frame = 0;
cpi->common.refresh_alt_ref_frame = 0;
cpi->common.refresh_last_frame = 0;
if (ref_frame_flags & VP8_LAST_FLAG)
cpi->common.refresh_last_frame = 1;
if (ref_frame_flags & VP8_GOLD_FLAG)
cpi->common.refresh_golden_frame = 1;
if (ref_frame_flags & VP8_ALT_FLAG)
cpi->common.refresh_alt_ref_frame = 1;
return 0;
}
int vp8_get_reference(VP8_PTR ptr, VP8_REFFRAME ref_frame_flag, YV12_BUFFER_CONFIG *sd)
{
VP8_COMP *cpi = (VP8_COMP *)(ptr);
VP8_COMMON *cm = &cpi->common;
Swap alt/gold/new/last frame buffer ptrs instead of copying. At the end of the decode, frame buffers were being copied. The frames are not updated after the copy, they are just for reference on later frames. This change allows multiple references to the same frame buffer instead of copying it. Changes needed to be made to the encoder to handle this. The encoder is still doing frame buffer copies in similar places where pointer reference could be done. Change-Id: I7c38be4d23979cc49b5f17241ca3a78703803e66
2010-07-22 14:07:32 +02:00
int ref_fb_idx;
Initial WebM release
2010-05-18 17:58:33 +02:00
if (ref_frame_flag == VP8_LAST_FLAG)
Swap alt/gold/new/last frame buffer ptrs instead of copying. At the end of the decode, frame buffers were being copied. The frames are not updated after the copy, they are just for reference on later frames. This change allows multiple references to the same frame buffer instead of copying it. Changes needed to be made to the encoder to handle this. The encoder is still doing frame buffer copies in similar places where pointer reference could be done. Change-Id: I7c38be4d23979cc49b5f17241ca3a78703803e66
2010-07-22 14:07:32 +02:00
ref_fb_idx = cm->lst_fb_idx;
Initial WebM release
2010-05-18 17:58:33 +02:00
else if (ref_frame_flag == VP8_GOLD_FLAG)
Swap alt/gold/new/last frame buffer ptrs instead of copying. At the end of the decode, frame buffers were being copied. The frames are not updated after the copy, they are just for reference on later frames. This change allows multiple references to the same frame buffer instead of copying it. Changes needed to be made to the encoder to handle this. The encoder is still doing frame buffer copies in similar places where pointer reference could be done. Change-Id: I7c38be4d23979cc49b5f17241ca3a78703803e66
2010-07-22 14:07:32 +02:00
ref_fb_idx = cm->gld_fb_idx;
Initial WebM release
2010-05-18 17:58:33 +02:00
else if (ref_frame_flag == VP8_ALT_FLAG)
Swap alt/gold/new/last frame buffer ptrs instead of copying. At the end of the decode, frame buffers were being copied. The frames are not updated after the copy, they are just for reference on later frames. This change allows multiple references to the same frame buffer instead of copying it. Changes needed to be made to the encoder to handle this. The encoder is still doing frame buffer copies in similar places where pointer reference could be done. Change-Id: I7c38be4d23979cc49b5f17241ca3a78703803e66
2010-07-22 14:07:32 +02:00
ref_fb_idx = cm->alt_fb_idx;
Initial WebM release
2010-05-18 17:58:33 +02:00
else
return -1;
Swap alt/gold/new/last frame buffer ptrs instead of copying. At the end of the decode, frame buffers were being copied. The frames are not updated after the copy, they are just for reference on later frames. This change allows multiple references to the same frame buffer instead of copying it. Changes needed to be made to the encoder to handle this. The encoder is still doing frame buffer copies in similar places where pointer reference could be done. Change-Id: I7c38be4d23979cc49b5f17241ca3a78703803e66
2010-07-22 14:07:32 +02:00
vp8_yv12_copy_frame_ptr(&cm->yv12_fb[ref_fb_idx], sd);
Initial WebM release
2010-05-18 17:58:33 +02:00
return 0;
}
int vp8_set_reference(VP8_PTR ptr, VP8_REFFRAME ref_frame_flag, YV12_BUFFER_CONFIG *sd)
{
VP8_COMP *cpi = (VP8_COMP *)(ptr);
VP8_COMMON *cm = &cpi->common;
Swap alt/gold/new/last frame buffer ptrs instead of copying. At the end of the decode, frame buffers were being copied. The frames are not updated after the copy, they are just for reference on later frames. This change allows multiple references to the same frame buffer instead of copying it. Changes needed to be made to the encoder to handle this. The encoder is still doing frame buffer copies in similar places where pointer reference could be done. Change-Id: I7c38be4d23979cc49b5f17241ca3a78703803e66
2010-07-22 14:07:32 +02:00
int ref_fb_idx;
Initial WebM release
2010-05-18 17:58:33 +02:00
Swap alt/gold/new/last frame buffer ptrs instead of copying. At the end of the decode, frame buffers were being copied. The frames are not updated after the copy, they are just for reference on later frames. This change allows multiple references to the same frame buffer instead of copying it. Changes needed to be made to the encoder to handle this. The encoder is still doing frame buffer copies in similar places where pointer reference could be done. Change-Id: I7c38be4d23979cc49b5f17241ca3a78703803e66
2010-07-22 14:07:32 +02:00
if (ref_frame_flag == VP8_LAST_FLAG)
ref_fb_idx = cm->lst_fb_idx;
Initial WebM release
2010-05-18 17:58:33 +02:00
else if (ref_frame_flag == VP8_GOLD_FLAG)
Swap alt/gold/new/last frame buffer ptrs instead of copying. At the end of the decode, frame buffers were being copied. The frames are not updated after the copy, they are just for reference on later frames. This change allows multiple references to the same frame buffer instead of copying it. Changes needed to be made to the encoder to handle this. The encoder is still doing frame buffer copies in similar places where pointer reference could be done. Change-Id: I7c38be4d23979cc49b5f17241ca3a78703803e66
2010-07-22 14:07:32 +02:00
ref_fb_idx = cm->gld_fb_idx;
Initial WebM release
2010-05-18 17:58:33 +02:00
else if (ref_frame_flag == VP8_ALT_FLAG)
Swap alt/gold/new/last frame buffer ptrs instead of copying. At the end of the decode, frame buffers were being copied. The frames are not updated after the copy, they are just for reference on later frames. This change allows multiple references to the same frame buffer instead of copying it. Changes needed to be made to the encoder to handle this. The encoder is still doing frame buffer copies in similar places where pointer reference could be done. Change-Id: I7c38be4d23979cc49b5f17241ca3a78703803e66
2010-07-22 14:07:32 +02:00
ref_fb_idx = cm->alt_fb_idx;
Initial WebM release
2010-05-18 17:58:33 +02:00
else
return -1;
Swap alt/gold/new/last frame buffer ptrs instead of copying. At the end of the decode, frame buffers were being copied. The frames are not updated after the copy, they are just for reference on later frames. This change allows multiple references to the same frame buffer instead of copying it. Changes needed to be made to the encoder to handle this. The encoder is still doing frame buffer copies in similar places where pointer reference could be done. Change-Id: I7c38be4d23979cc49b5f17241ca3a78703803e66
2010-07-22 14:07:32 +02:00
vp8_yv12_copy_frame_ptr(sd, &cm->yv12_fb[ref_fb_idx]);
Initial WebM release
2010-05-18 17:58:33 +02:00
return 0;
}
int vp8_update_entropy(VP8_PTR comp, int update)
{
VP8_COMP *cpi = (VP8_COMP *) comp;
VP8_COMMON *cm = &cpi->common;
cm->refresh_entropy_probs = update;
return 0;
}
disable compilation of debugging code This patch avoids compiling some debugging code in onyx_if.c. The most significant fix is to avoid generating code for vp8_write_yuv_frame, which is never called. Some other code was removed by the dead code elimination performed by the compiler, and this patch does it with the preprocessor instead. There are advantages both ways. Change-Id: I044fd43179d2e947553f0d6f2cad5b40907ac458
2010-09-24 17:10:25 +02:00
#if OUTPUT_YUV_SRC
Initial WebM release
2010-05-18 17:58:33 +02:00
void vp8_write_yuv_frame(const char *name, YV12_BUFFER_CONFIG *s)
{
FILE *yuv_file = fopen(name, "ab");
unsigned char *src = s->y_buffer;
int h = s->y_height;
do
{
fwrite(src, s->y_width, 1, yuv_file);
src += s->y_stride;
}
while (--h);
src = s->u_buffer;
h = s->uv_height;
do
{
fwrite(src, s->uv_width, 1, yuv_file);
src += s->uv_stride;
}
while (--h);
src = s->v_buffer;
h = s->uv_height;
do
{
fwrite(src, s->uv_width, 1, yuv_file);
src += s->uv_stride;
}
while (--h);
fclose(yuv_file);
}
disable compilation of debugging code This patch avoids compiling some debugging code in onyx_if.c. The most significant fix is to avoid generating code for vp8_write_yuv_frame, which is never called. Some other code was removed by the dead code elimination performed by the compiler, and this patch does it with the preprocessor instead. There are advantages both ways. Change-Id: I044fd43179d2e947553f0d6f2cad5b40907ac458
2010-09-24 17:10:25 +02:00
#endif
Initial WebM release
2010-05-18 17:58:33 +02:00
static void scale_and_extend_source(YV12_BUFFER_CONFIG *sd, VP8_COMP *cpi)
{
VP8_COMMON *cm = &cpi->common;
// are we resizing the image
if (cm->horiz_scale != 0 || cm->vert_scale != 0)
{
#if CONFIG_SPATIAL_RESAMPLING
int UNINITIALIZED_IS_SAFE(hr), UNINITIALIZED_IS_SAFE(hs);
int UNINITIALIZED_IS_SAFE(vr), UNINITIALIZED_IS_SAFE(vs);
int tmp_height;
if (cm->vert_scale == 3)
tmp_height = 9;
else
tmp_height = 11;
Scale2Ratio(cm->horiz_scale, &hr, &hs);
Scale2Ratio(cm->vert_scale, &vr, &vs);
vp8_scale_frame(sd, &cpi->scaled_source, cm->temp_scale_frame.y_buffer,
tmp_height, hs, hr, vs, vr, 0);
cpi->Source = &cpi->scaled_source;
#endif
}
// we may need to copy to a buffer so we can extend the image...
Swap alt/gold/new/last frame buffer ptrs instead of copying. At the end of the decode, frame buffers were being copied. The frames are not updated after the copy, they are just for reference on later frames. This change allows multiple references to the same frame buffer instead of copying it. Changes needed to be made to the encoder to handle this. The encoder is still doing frame buffer copies in similar places where pointer reference could be done. Change-Id: I7c38be4d23979cc49b5f17241ca3a78703803e66
2010-07-22 14:07:32 +02:00
else if (cm->Width != cm->yv12_fb[cm->lst_fb_idx].y_width ||
cm->Height != cm->yv12_fb[cm->lst_fb_idx].y_height)
Initial WebM release
2010-05-18 17:58:33 +02:00
{
//vp8_yv12_copy_frame_ptr(sd, &cpi->scaled_source);
#if HAVE_ARMV7
Add runtime CPU detection support for ARM. The primary goal is to allow a binary to be built which supports NEON, but can fall back to non-NEON routines, since some Android devices do not have NEON, even if they are otherwise ARMv7 (e.g., Tegra). The configure-generated flags HAVE_ARMV7, etc., are used to decide which versions of each function to build, and when CONFIG_RUNTIME_CPU_DETECT is enabled, the correct version is chosen at run time. In order for this to work, the CFLAGS must be set to something appropriate (e.g., without -mfpu=neon for ARMv7, and with appropriate -march and -mcpu for even earlier configurations), or the native C code will not be able to run. The ASFLAGS must remain set for the most advanced instruction set required at build time, since the ARM assembler will refuse to emit them otherwise. I have not attempted to make any changes to configure to do this automatically. Doing so will probably require the addition of new configure options. Many of the hooks for RTCD on ARM were already there, but a lot of the code had bit-rotted, and a good deal of the ARM-specific code is not integrated into the RTCD structs at all. I did not try to resolve the latter, merely to add the minimal amount of protection around them to allow RTCD to work. Those functions that were called based on an ifdef at the calling site were expanded to check the RTCD flags at that site, but they should be added to an RTCD struct somewhere in the future. The functions invoked with global function pointers still are, but these should be moved into an RTCD struct for thread safety (I believe every platform currently supported has atomic pointer stores, but this is not guaranteed). The encoder's boolhuff functions did not even have _c and armv7 suffixes, and the correct version was resolved at link time. The token packing functions did have appropriate suffixes, but the version was selected with a define, with no associated RTCD struct. However, for both of these, the only armv7 instruction they actually used was rbit, and this was completely superfluous, so I reworked them to avoid it. The only non-ARMv4 instruction remaining in them is clz, which is ARMv5 (not even ARMv5TE is required). Considering that there are no ARM-specific configs which are not at least ARMv5TE, I did not try to detect these at runtime, and simply enable them for ARMv5 and above. Finally, the NEON register saving code was completely non-reentrant, since it saved the registers to a global, static variable. I moved the storage for this onto the stack. A single binary built with this code was tested on an ARM11 (ARMv6) and a Cortex A8 (ARMv7 w/NEON), for both the encoder and decoder, and produced identical output, while using the correct accelerated functions on each. I did not test on any earlier processors. Change-Id: I45cbd63a614f4554c3b325c45d46c0806f009eaa
2010-10-21 00:39:11 +02:00
#if CONFIG_RUNTIME_CPU_DETECT
if (cm->rtcd.flags & HAS_NEON)
#endif
{
vp8_yv12_copy_src_frame_func_neon(sd, &cpi->scaled_source);
}
#if CONFIG_RUNTIME_CPU_DETECT
else
#endif
#endif
#if !HAVE_ARMV7 || CONFIG_RUNTIME_CPU_DETECT
{
vp8_yv12_copy_frame_ptr(sd, &cpi->scaled_source);
}
Initial WebM release
2010-05-18 17:58:33 +02:00
#endif
cpi->Source = &cpi->scaled_source;
}
vp8_extend_to_multiple_of16(cpi->Source, cm->Width, cm->Height);
}
static void resize_key_frame(VP8_COMP *cpi)
{
#if CONFIG_SPATIAL_RESAMPLING
VP8_COMMON *cm = &cpi->common;
// Do we need to apply resampling for one pass cbr.
// In one pass this is more limited than in two pass cbr
// The test and any change is only made one per key frame sequence
if (cpi->oxcf.allow_spatial_resampling && (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER))
{
int UNINITIALIZED_IS_SAFE(hr), UNINITIALIZED_IS_SAFE(hs);
int UNINITIALIZED_IS_SAFE(vr), UNINITIALIZED_IS_SAFE(vs);
int new_width, new_height;
// If we are below the resample DOWN watermark then scale down a notch.
if (cpi->buffer_level < (cpi->oxcf.resample_down_water_mark * cpi->oxcf.optimal_buffer_level / 100))
{
cm->horiz_scale = (cm->horiz_scale < ONETWO) ? cm->horiz_scale + 1 : ONETWO;
cm->vert_scale = (cm->vert_scale < ONETWO) ? cm->vert_scale + 1 : ONETWO;
}
// Should we now start scaling back up
else if (cpi->buffer_level > (cpi->oxcf.resample_up_water_mark * cpi->oxcf.optimal_buffer_level / 100))
{
cm->horiz_scale = (cm->horiz_scale > NORMAL) ? cm->horiz_scale - 1 : NORMAL;
cm->vert_scale = (cm->vert_scale > NORMAL) ? cm->vert_scale - 1 : NORMAL;
}
// Get the new hieght and width
Scale2Ratio(cm->horiz_scale, &hr, &hs);
Scale2Ratio(cm->vert_scale, &vr, &vs);
new_width = ((hs - 1) + (cpi->oxcf.Width * hr)) / hs;
new_height = ((vs - 1) + (cpi->oxcf.Height * vr)) / vs;
// If the image size has changed we need to reallocate the buffers
// and resample the source image
if ((cm->Width != new_width) || (cm->Height != new_height))
{
cm->Width = new_width;
cm->Height = new_height;
vp8_alloc_compressor_data(cpi);
scale_and_extend_source(cpi->un_scaled_source, cpi);
}
}
#endif
}
// return of 0 means drop frame
static int pick_frame_size(VP8_COMP *cpi)
{
VP8_COMMON *cm = &cpi->common;
// First Frame is a special case
if (cm->current_video_frame == 0)
{
#if !(CONFIG_REALTIME_ONLY)
if (cpi->pass == 2)
vp8_calc_auto_iframe_target_size(cpi);
// 1 Pass there is no information on which to base size so use bandwidth per second * fixed fraction
else
#endif
cpi->this_frame_target = cpi->oxcf.target_bandwidth / 2;
// in error resilient mode the first frame is bigger since it likely contains
// all the static background
if (cpi->oxcf.error_resilient_mode == 1 || (cpi->compressor_speed == 2))
{
cpi->this_frame_target *= 3; // 5;
}
// Key frame from VFW/auto-keyframe/first frame
cm->frame_type = KEY_FRAME;
}
Improved Force Key Frame Behaviour These changes improve the behaviour of the code with forced key frames sent in by a calling application. The sizing of the frames is still suboptimal for two pass in particular but the behaviour is much better than it was. Change-Id: I35fae610c67688ccc69d11f385e87dfc884e65a1
2010-08-20 13:27:26 +02:00
// Special case for forced key frames
// The frame sizing here is still far from ideal for 2 pass.
else if (cm->frame_flags & FRAMEFLAGS_KEY)
Initial WebM release
2010-05-18 17:58:33 +02:00
{
cm->frame_type = KEY_FRAME;
resize_key_frame(cpi);
Improved Force Key Frame Behaviour These changes improve the behaviour of the code with forced key frames sent in by a calling application. The sizing of the frames is still suboptimal for two pass in particular but the behaviour is much better than it was. Change-Id: I35fae610c67688ccc69d11f385e87dfc884e65a1
2010-08-20 13:27:26 +02:00
vp8_calc_iframe_target_size(cpi);
}
else if (cm->frame_type == KEY_FRAME)
{
vp8_calc_auto_iframe_target_size(cpi);
Initial WebM release
2010-05-18 17:58:33 +02:00
}
else
{
// INTER frame: compute target frame size
cm->frame_type = INTER_FRAME;
vp8_calc_pframe_target_size(cpi);
// Check if we're dropping the frame:
if (cpi->drop_frame)
{
cpi->drop_frame = FALSE;
cpi->drop_count++;
return 0;
}
}
// Note target_size in bits * 256 per MB
cpi->target_bits_per_mb = (cpi->this_frame_target * 256) / cpi->common.MBs;
return 1;
}
static void set_quantizer(VP8_COMP *cpi, int Q)
{
VP8_COMMON *cm = &cpi->common;
MACROBLOCKD *mbd = &cpi->mb.e_mbd;
cm->base_qindex = Q;
cm->y1dc_delta_q = 0;
cm->y2dc_delta_q = 0;
cm->y2ac_delta_q = 0;
cm->uvdc_delta_q = 0;
cm->uvac_delta_q = 0;
// Set Segment specific quatizers
mbd->segment_feature_data[MB_LVL_ALT_Q][0] = cpi->segment_feature_data[MB_LVL_ALT_Q][0];
mbd->segment_feature_data[MB_LVL_ALT_Q][1] = cpi->segment_feature_data[MB_LVL_ALT_Q][1];
mbd->segment_feature_data[MB_LVL_ALT_Q][2] = cpi->segment_feature_data[MB_LVL_ALT_Q][2];
mbd->segment_feature_data[MB_LVL_ALT_Q][3] = cpi->segment_feature_data[MB_LVL_ALT_Q][3];
}
static void update_alt_ref_frame_and_stats(VP8_COMP *cpi)
{
VP8_COMMON *cm = &cpi->common;
// Update the golden frame buffer
Swap alt/gold/new/last frame buffer ptrs instead of copying. At the end of the decode, frame buffers were being copied. The frames are not updated after the copy, they are just for reference on later frames. This change allows multiple references to the same frame buffer instead of copying it. Changes needed to be made to the encoder to handle this. The encoder is still doing frame buffer copies in similar places where pointer reference could be done. Change-Id: I7c38be4d23979cc49b5f17241ca3a78703803e66
2010-07-22 14:07:32 +02:00
vp8_yv12_copy_frame_ptr(cm->frame_to_show, &cm->yv12_fb[cm->alt_fb_idx]);
Initial WebM release
2010-05-18 17:58:33 +02:00
// Select an interval before next GF or altref
if (!cpi->auto_gold)
cpi->frames_till_gf_update_due = cpi->goldfreq;
if ((cpi->pass != 2) && cpi->frames_till_gf_update_due)
{
cpi->current_gf_interval = cpi->frames_till_gf_update_due;
// Set the bits per frame that we should try and recover in subsequent inter frames
// to account for the extra GF spend... note that his does not apply for GF updates
// that occur coincident with a key frame as the extra cost of key frames is dealt
// with elsewhere.
cpi->gf_overspend_bits += cpi->projected_frame_size;
cpi->non_gf_bitrate_adjustment = cpi->gf_overspend_bits / cpi->frames_till_gf_update_due;
}
// Update data structure that monitors level of reference to last GF
Moved gf_active code to encoder only The gf_active code is only used by the encoder, so it was moved from common and decoder. Change-Id: Iada15acd5b2b33ff70c34668ca87d4cfd0d05025
2010-08-11 17:02:31 +02:00
vpx_memset(cpi->gf_active_flags, 1, (cm->mb_rows * cm->mb_cols));
cpi->gf_active_count = cm->mb_rows * cm->mb_cols;
Initial WebM release
2010-05-18 17:58:33 +02:00
// this frame refreshes means next frames don't unless specified by user
cpi->common.frames_since_golden = 0;
// Clear the alternate reference update pending flag.
cpi->source_alt_ref_pending = FALSE;
// Set the alternate refernce frame active flag
cpi->source_alt_ref_active = TRUE;
}
static void update_golden_frame_and_stats(VP8_COMP *cpi)
{
VP8_COMMON *cm = &cpi->common;
// Update the Golden frame reconstruction buffer if signalled and the GF usage counts.
if (cm->refresh_golden_frame)
{
// Update the golden frame buffer
Swap alt/gold/new/last frame buffer ptrs instead of copying. At the end of the decode, frame buffers were being copied. The frames are not updated after the copy, they are just for reference on later frames. This change allows multiple references to the same frame buffer instead of copying it. Changes needed to be made to the encoder to handle this. The encoder is still doing frame buffer copies in similar places where pointer reference could be done. Change-Id: I7c38be4d23979cc49b5f17241ca3a78703803e66
2010-07-22 14:07:32 +02:00
vp8_yv12_copy_frame_ptr(cm->frame_to_show, &cm->yv12_fb[cm->gld_fb_idx]);
Initial WebM release
2010-05-18 17:58:33 +02:00
// Select an interval before next GF
if (!cpi->auto_gold)
cpi->frames_till_gf_update_due = cpi->goldfreq;
if ((cpi->pass != 2) && (cpi->frames_till_gf_update_due > 0))
{
cpi->current_gf_interval = cpi->frames_till_gf_update_due;
// Set the bits per frame that we should try and recover in subsequent inter frames
// to account for the extra GF spend... note that his does not apply for GF updates
// that occur coincident with a key frame as the extra cost of key frames is dealt
// with elsewhere.
if ((cm->frame_type != KEY_FRAME) && !cpi->source_alt_ref_active)
{
// Calcluate GF bits to be recovered
// Projected size - av frame bits available for inter frames for clip as a whole
cpi->gf_overspend_bits += (cpi->projected_frame_size - cpi->inter_frame_target);
}
cpi->non_gf_bitrate_adjustment = cpi->gf_overspend_bits / cpi->frames_till_gf_update_due;
}
// Update data structure that monitors level of reference to last GF
Moved gf_active code to encoder only The gf_active code is only used by the encoder, so it was moved from common and decoder. Change-Id: Iada15acd5b2b33ff70c34668ca87d4cfd0d05025
2010-08-11 17:02:31 +02:00
vpx_memset(cpi->gf_active_flags, 1, (cm->mb_rows * cm->mb_cols));
cpi->gf_active_count = cm->mb_rows * cm->mb_cols;
Initial WebM release
2010-05-18 17:58:33 +02:00
// this frame refreshes means next frames don't unless specified by user
cm->refresh_golden_frame = 0;
cpi->common.frames_since_golden = 0;
//if ( cm->frame_type == KEY_FRAME )
//{
cpi->recent_ref_frame_usage[INTRA_FRAME] = 1;
cpi->recent_ref_frame_usage[LAST_FRAME] = 1;
cpi->recent_ref_frame_usage[GOLDEN_FRAME] = 1;
cpi->recent_ref_frame_usage[ALTREF_FRAME] = 1;
//}
//else
//{
// // Carry a potrtion of count over to begining of next gf sequence
// cpi->recent_ref_frame_usage[INTRA_FRAME] >>= 5;
// cpi->recent_ref_frame_usage[LAST_FRAME] >>= 5;
// cpi->recent_ref_frame_usage[GOLDEN_FRAME] >>= 5;
// cpi->recent_ref_frame_usage[ALTREF_FRAME] >>= 5;
//}
// ******** Fixed Q test code only ************
// If we are going to use the ALT reference for the next group of frames set a flag to say so.
if (cpi->oxcf.fixed_q >= 0 &&
cpi->oxcf.play_alternate && !cpi->common.refresh_alt_ref_frame)
{
cpi->source_alt_ref_pending = TRUE;
cpi->frames_till_gf_update_due = cpi->baseline_gf_interval;
}
if (!cpi->source_alt_ref_pending)
cpi->source_alt_ref_active = FALSE;
// Decrement count down till next gf
if (cpi->frames_till_gf_update_due > 0)
cpi->frames_till_gf_update_due--;
}
else if (!cpi->common.refresh_alt_ref_frame)
{
// Decrement count down till next gf
if (cpi->frames_till_gf_update_due > 0)
cpi->frames_till_gf_update_due--;
if (cpi->common.frames_till_alt_ref_frame)
cpi->common.frames_till_alt_ref_frame --;
cpi->common.frames_since_golden ++;
if (cpi->common.frames_since_golden > 1)
{
cpi->recent_ref_frame_usage[INTRA_FRAME] += cpi->count_mb_ref_frame_usage[INTRA_FRAME];
cpi->recent_ref_frame_usage[LAST_FRAME] += cpi->count_mb_ref_frame_usage[LAST_FRAME];
cpi->recent_ref_frame_usage[GOLDEN_FRAME] += cpi->count_mb_ref_frame_usage[GOLDEN_FRAME];
cpi->recent_ref_frame_usage[ALTREF_FRAME] += cpi->count_mb_ref_frame_usage[ALTREF_FRAME];
}
}
}
// This function updates the reference frame probability estimates that
// will be used during mode selection
static void update_rd_ref_frame_probs(VP8_COMP *cpi)
{
VP8_COMMON *cm = &cpi->common;
#if 0
const int *const rfct = cpi->recent_ref_frame_usage;
const int rf_intra = rfct[INTRA_FRAME];
const int rf_inter = rfct[LAST_FRAME] + rfct[GOLDEN_FRAME] + rfct[ALTREF_FRAME];
if (cm->frame_type == KEY_FRAME)
{
cpi->prob_intra_coded = 255;
cpi->prob_last_coded = 128;
cpi->prob_gf_coded = 128;
}
else if (!(rf_intra + rf_inter))
{
// This is a trap in case this function is called with cpi->recent_ref_frame_usage[] blank.
cpi->prob_intra_coded = 63;
cpi->prob_last_coded = 128;
cpi->prob_gf_coded = 128;
}
else
{
cpi->prob_intra_coded = (rf_intra * 255) / (rf_intra + rf_inter);
if (cpi->prob_intra_coded < 1)
cpi->prob_intra_coded = 1;
if ((cm->frames_since_golden > 0) || cpi->source_alt_ref_active)
{
cpi->prob_last_coded = rf_inter ? (rfct[LAST_FRAME] * 255) / rf_inter : 128;
if (cpi->prob_last_coded < 1)
cpi->prob_last_coded = 1;
cpi->prob_gf_coded = (rfct[GOLDEN_FRAME] + rfct[ALTREF_FRAME])
? (rfct[GOLDEN_FRAME] * 255) / (rfct[GOLDEN_FRAME] + rfct[ALTREF_FRAME]) : 128;
if (cpi->prob_gf_coded < 1)
cpi->prob_gf_coded = 1;
}
}
#else
const int *const rfct = cpi->count_mb_ref_frame_usage;
const int rf_intra = rfct[INTRA_FRAME];
const int rf_inter = rfct[LAST_FRAME] + rfct[GOLDEN_FRAME] + rfct[ALTREF_FRAME];
if (cm->frame_type == KEY_FRAME)
{
cpi->prob_intra_coded = 255;
cpi->prob_last_coded = 128;
cpi->prob_gf_coded = 128;
}
else if (!(rf_intra + rf_inter))
{
// This is a trap in case this function is called with cpi->recent_ref_frame_usage[] blank.
cpi->prob_intra_coded = 63;
cpi->prob_last_coded = 128;
cpi->prob_gf_coded = 128;
}
else
{
cpi->prob_intra_coded = (rf_intra * 255) / (rf_intra + rf_inter);
if (cpi->prob_intra_coded < 1)
cpi->prob_intra_coded = 1;
cpi->prob_last_coded = rf_inter ? (rfct[LAST_FRAME] * 255) / rf_inter : 128;
if (cpi->prob_last_coded < 1)
cpi->prob_last_coded = 1;
cpi->prob_gf_coded = (rfct[GOLDEN_FRAME] + rfct[ALTREF_FRAME])
? (rfct[GOLDEN_FRAME] * 255) / (rfct[GOLDEN_FRAME] + rfct[ALTREF_FRAME]) : 128;
if (cpi->prob_gf_coded < 1)
cpi->prob_gf_coded = 1;
}
// update reference frame costs since we can do better than what we got last frame.
if (cpi->common.refresh_alt_ref_frame)
{
cpi->prob_intra_coded += 40;
cpi->prob_last_coded = 200;
cpi->prob_gf_coded = 1;
}
else if (cpi->common.frames_since_golden == 0)
{
cpi->prob_last_coded = 214;
cpi->prob_gf_coded = 1;
}
else if (cpi->common.frames_since_golden == 1)
{
cpi->prob_last_coded = 192;
cpi->prob_gf_coded = 220;
}
else if (cpi->source_alt_ref_active)
{
//int dist = cpi->common.frames_till_alt_ref_frame + cpi->common.frames_since_golden;
cpi->prob_gf_coded -= 20;
if (cpi->prob_gf_coded < 10)
cpi->prob_gf_coded = 10;
}
#endif
}
// 1 = key, 0 = inter
static int decide_key_frame(VP8_COMP *cpi)
{
VP8_COMMON *cm = &cpi->common;
int code_key_frame = FALSE;
cpi->kf_boost = 0;
if (cpi->Speed > 11)
return FALSE;
// Clear down mmx registers
vp8_clear_system_state(); //__asm emms;
if ((cpi->compressor_speed == 2) && (cpi->Speed >= 5) && (cpi->sf.RD == 0))
{
double change = 1.0 * abs((int)(cpi->intra_error - cpi->last_intra_error)) / (1 + cpi->last_intra_error);
double change2 = 1.0 * abs((int)(cpi->prediction_error - cpi->last_prediction_error)) / (1 + cpi->last_prediction_error);
double minerror = cm->MBs * 256;
#if 0
if (10 * cpi->intra_error / (1 + cpi->prediction_error) < 15
&& cpi->prediction_error > minerror
&& (change > .25 || change2 > .25))
{
FILE *f = fopen("intra_inter.stt", "a");
if (cpi->prediction_error <= 0)
cpi->prediction_error = 1;
fprintf(f, "%d %d %d %d %14.4f\n",
cm->current_video_frame,
(int) cpi->prediction_error,
(int) cpi->intra_error,
(int)((10 * cpi->intra_error) / cpi->prediction_error),
change);
fclose(f);
}
#endif
cpi->last_intra_error = cpi->intra_error;
cpi->last_prediction_error = cpi->prediction_error;
if (10 * cpi->intra_error / (1 + cpi->prediction_error) < 15
&& cpi->prediction_error > minerror
&& (change > .25 || change2 > .25))
{
/*(change > 1.4 || change < .75)&& cpi->this_frame_percent_intra > cpi->last_frame_percent_intra + 3*/
return TRUE;
}
return FALSE;
}
// If the following are true we might as well code a key frame
if (((cpi->this_frame_percent_intra == 100) &&
(cpi->this_frame_percent_intra > (cpi->last_frame_percent_intra + 2))) ||
((cpi->this_frame_percent_intra > 95) &&
(cpi->this_frame_percent_intra >= (cpi->last_frame_percent_intra + 5))))
{
code_key_frame = TRUE;
}
// in addition if the following are true and this is not a golden frame then code a key frame
// Note that on golden frames there often seems to be a pop in intra useage anyway hence this
// restriction is designed to prevent spurious key frames. The Intra pop needs to be investigated.
else if (((cpi->this_frame_percent_intra > 60) &&
(cpi->this_frame_percent_intra > (cpi->last_frame_percent_intra * 2))) ||
((cpi->this_frame_percent_intra > 75) &&
(cpi->this_frame_percent_intra > (cpi->last_frame_percent_intra * 3 / 2))) ||
((cpi->this_frame_percent_intra > 90) &&
(cpi->this_frame_percent_intra > (cpi->last_frame_percent_intra + 10))))
{
if (!cm->refresh_golden_frame)
code_key_frame = TRUE;
}
return code_key_frame;
}
#if !(CONFIG_REALTIME_ONLY)
static void Pass1Encode(VP8_COMP *cpi, unsigned long *size, unsigned char *dest, unsigned int *frame_flags)
{
(void) size;
(void) dest;
(void) frame_flags;
set_quantizer(cpi, 26);
scale_and_extend_source(cpi->un_scaled_source, cpi);
vp8_first_pass(cpi);
}
#endif
#if 0
void write_cx_frame_to_file(YV12_BUFFER_CONFIG *frame, int this_frame)
{
// write the frame
FILE *yframe;
int i;
char filename[255];
sprintf(filename, "cx\\y%04d.raw", this_frame);
yframe = fopen(filename, "wb");
for (i = 0; i < frame->y_height; i++)
fwrite(frame->y_buffer + i * frame->y_stride, frame->y_width, 1, yframe);
fclose(yframe);
sprintf(filename, "cx\\u%04d.raw", this_frame);
yframe = fopen(filename, "wb");
for (i = 0; i < frame->uv_height; i++)
fwrite(frame->u_buffer + i * frame->uv_stride, frame->uv_width, 1, yframe);
fclose(yframe);
sprintf(filename, "cx\\v%04d.raw", this_frame);
yframe = fopen(filename, "wb");
for (i = 0; i < frame->uv_height; i++)
fwrite(frame->v_buffer + i * frame->uv_stride, frame->uv_width, 1, yframe);
fclose(yframe);
}
#endif
// return of 0 means drop frame
Replaced recode loop test with a function call Replaced existing code to decide if a frame recode is required with a function call. This is to simplify addition of extra clauses that may be needed for the planned constrained quality mode. Also fixed a bug where by alt ref not considered in the test. Change-Id: I3d40bb21abe3e19f8456761e6849deb171738b60
2010-11-17 16:12:04 +01:00
// Function to test for conditions that indeicate we should loop
// back and recode a frame.
static BOOL recode_loop_test( VP8_COMP *cpi,
int high_limit, int low_limit,
int q, int maxq, int minq )
{
BOOL force_recode = FALSE;
VP8_COMMON *cm = &cpi->common;
// Is frame recode allowed at all
// Yes if either recode mode 1 is selected or mode two is selcted
// and the frame is a key frame. golden frame or alt_ref_frame
if ( (cpi->sf.recode_loop == 1) ||
( (cpi->sf.recode_loop == 2) &&
( (cm->frame_type == KEY_FRAME) ||
cm->refresh_golden_frame ||
cm->refresh_alt_ref_frame ) ) )
{
// General over and under shoot tests
if ( ((cpi->projected_frame_size > high_limit) && (q < maxq)) ||
((cpi->projected_frame_size < low_limit) && (q > minq)) )
{
force_recode = TRUE;
}
// Specific rate control mode related tests
// TBD
}
return force_recode;
}
Made AltRef filter adaptive & added motion compensation Modified AltRef temporal filter to adapt filter length based on macroblock coding modes selected during first-pass encode. Also added sub-pixel motion compensation to the AltRef filter.
2010-09-28 16:23:41 +02:00
static void encode_frame_to_data_rate
(
VP8_COMP *cpi,
unsigned long *size,
unsigned char *dest,
unsigned int *frame_flags
)
Initial WebM release
2010-05-18 17:58:33 +02:00
{
int Q;
int frame_over_shoot_limit;
int frame_under_shoot_limit;
int Loop = FALSE;
int loop_count;
int this_q;
int last_zbin_oq;
int q_low;
int q_high;
int zbin_oq_high;
int zbin_oq_low = 0;
int top_index;
int bottom_index;
VP8_COMMON *cm = &cpi->common;
int active_worst_qchanged = FALSE;
int overshoot_seen = FALSE;
int undershoot_seen = FALSE;
int drop_mark = cpi->oxcf.drop_frames_water_mark * cpi->oxcf.optimal_buffer_level / 100;
int drop_mark75 = drop_mark * 2 / 3;
int drop_mark50 = drop_mark / 4;
int drop_mark25 = drop_mark / 8;
// Clear down mmx registers to allow floating point in what follows
vp8_clear_system_state();
// Test code for segmentation of gf/arf (0,0)
//segmentation_test_function((VP8_PTR) cpi);
// For an alt ref frame in 2 pass we skip the call to the second pass function that sets the target bandwidth
#if !(CONFIG_REALTIME_ONLY)
if (cpi->pass == 2)
{
if (cpi->common.refresh_alt_ref_frame)
{
cpi->per_frame_bandwidth = cpi->gf_bits; // Per frame bit target for the alt ref frame
cpi->target_bandwidth = cpi->gf_bits * cpi->output_frame_rate; // per second target bitrate
}
}
else
#endif
cpi->per_frame_bandwidth = (int)(cpi->target_bandwidth / cpi->output_frame_rate);
// Default turn off buffer to buffer copying
cm->copy_buffer_to_gf = 0;
cm->copy_buffer_to_arf = 0;
// Clear zbin over-quant value and mode boost values.
cpi->zbin_over_quant = 0;
cpi->zbin_mode_boost = 0;
Tuning for the more exact quantizer. Small changes to the default zero bin and rounding tables. Though the tables are currently the same for the Y1 and Y2 cases I have left them as separate tables in case we want to tune this later. There is now some adjustment of the zbin based on the prediction mode. Previously this was restricted to an adjustment for gf/arf 0,0 MV. The exact quantizer now marginal outperforms and is the default. The overall average gain is about 0.5% Change-Id: I5e4353f3d5326dde4e86823684b236a1e9ea7f47
2010-11-08 16:28:54 +01:00
// Enable or disable mode based tweaking of the zbin
// For 2 Pass Only used where GF/ARF prediction quality
// is above a threshold
cpi->zbin_mode_boost = 0;
Initial WebM release
2010-05-18 17:58:33 +02:00
cpi->zbin_mode_boost_enabled = TRUE;
Tuning for the more exact quantizer. Small changes to the default zero bin and rounding tables. Though the tables are currently the same for the Y1 and Y2 cases I have left them as separate tables in case we want to tune this later. There is now some adjustment of the zbin based on the prediction mode. Previously this was restricted to an adjustment for gf/arf 0,0 MV. The exact quantizer now marginal outperforms and is the default. The overall average gain is about 0.5% Change-Id: I5e4353f3d5326dde4e86823684b236a1e9ea7f47
2010-11-08 16:28:54 +01:00
if (cpi->pass == 2)
{
if ( cpi->gfu_boost <= 400 )
{
cpi->zbin_mode_boost_enabled = FALSE;
}
}
Initial WebM release
2010-05-18 17:58:33 +02:00
// Current default encoder behaviour for the altref sign bias
if (cpi->source_alt_ref_active)
cpi->common.ref_frame_sign_bias[ALTREF_FRAME] = 1;
else
cpi->common.ref_frame_sign_bias[ALTREF_FRAME] = 0;
// Check to see if a key frame is signalled
// For two pass with auto key frame enabled cm->frame_type may already be set, but not for one pass.
if ((cm->current_video_frame == 0) ||
(cm->frame_flags & FRAMEFLAGS_KEY) ||
(cpi->oxcf.auto_key && (cpi->frames_since_key % cpi->key_frame_frequency == 0)))
{
// Key frame from VFW/auto-keyframe/first frame
cm->frame_type = KEY_FRAME;
}
// Set default state for segment and mode based loop filter update flags
cpi->mb.e_mbd.update_mb_segmentation_map = 0;
cpi->mb.e_mbd.update_mb_segmentation_data = 0;
cpi->mb.e_mbd.mode_ref_lf_delta_update = 0;
// Set various flags etc to special state if it is a key frame
if (cm->frame_type == KEY_FRAME)
{
int i;
Fix loopfilter delta zero transitions Loopfilter deltas are initialized to zero on keyframes in the decoder. The values then persist from the previous frame unless an update bit is set in the bitstream. This data is not included in the entropy data saved by the 'refresh entropy' bit in the bitstream, so it is effectively an additional contextual element beyond the 3 ref-frames and the entropy data. The encoder was treating this delta update bit as update-if-nonzero, meaning that the value would be refreshed even if it hadn't changed, and more significantly, if the correct value for the delta changed to zero, the update wouldn't be sent, and the decoder would preserve the last (presumably non-zero) value. This patch updates the encoder to send an update only if the value has changed from the previously transmitted value. It also forces the value to be transmitted in error resilient mode, to account for lost context in the event of lost frames. Change-Id: I56671d5b42965d0166ac226765dbfce3e5301868
2010-09-29 19:04:04 +02:00
// Reset the loop filter deltas and segmentation map
setup_features(cpi);
Initial WebM release
2010-05-18 17:58:33 +02:00
// If segmentation is enabled force a map update for key frames
if (cpi->mb.e_mbd.segmentation_enabled)
{
cpi->mb.e_mbd.update_mb_segmentation_map = 1;
cpi->mb.e_mbd.update_mb_segmentation_data = 1;
}
// The alternate reference frame cannot be active for a key frame
cpi->source_alt_ref_active = FALSE;
// Reset the RD threshold multipliers to default of * 1 (128)
for (i = 0; i < MAX_MODES; i++)
{
cpi->rd_thresh_mult[i] = 128;
}
}
// Test code for segmentation
//if ( (cm->frame_type == KEY_FRAME) || ((cm->current_video_frame % 2) == 0))
//if ( (cm->current_video_frame % 2) == 0 )
// enable_segmentation((VP8_PTR)cpi);
//else
// disable_segmentation((VP8_PTR)cpi);
#if 0
// Experimental code for lagged compress and one pass
// Initialise one_pass GF frames stats
// Update stats used for GF selection
//if ( cpi->pass == 0 )
{
cpi->one_pass_frame_index = cm->current_video_frame % MAX_LAG_BUFFERS;
cpi->one_pass_frame_stats[cpi->one_pass_frame_index ].frames_so_far = 0;
cpi->one_pass_frame_stats[cpi->one_pass_frame_index ].frame_intra_error = 0.0;
cpi->one_pass_frame_stats[cpi->one_pass_frame_index ].frame_coded_error = 0.0;
cpi->one_pass_frame_stats[cpi->one_pass_frame_index ].frame_pcnt_inter = 0.0;
cpi->one_pass_frame_stats[cpi->one_pass_frame_index ].frame_pcnt_motion = 0.0;
cpi->one_pass_frame_stats[cpi->one_pass_frame_index ].frame_mvr = 0.0;
cpi->one_pass_frame_stats[cpi->one_pass_frame_index ].frame_mvr_abs = 0.0;
cpi->one_pass_frame_stats[cpi->one_pass_frame_index ].frame_mvc = 0.0;
cpi->one_pass_frame_stats[cpi->one_pass_frame_index ].frame_mvc_abs = 0.0;
}
#endif
update_rd_ref_frame_probs(cpi);
if (cpi->drop_frames_allowed)
{
// The reset to decimation 0 is only done here for one pass.
// Once it is set two pass leaves decimation on till the next kf.
if ((cpi->buffer_level > drop_mark) && (cpi->decimation_factor > 0))
cpi->decimation_factor --;
if (cpi->buffer_level > drop_mark75 && cpi->decimation_factor > 0)
cpi->decimation_factor = 1;
else if (cpi->buffer_level < drop_mark25 && (cpi->decimation_factor == 2 || cpi->decimation_factor == 3))
{
cpi->decimation_factor = 3;
}
else if (cpi->buffer_level < drop_mark50 && (cpi->decimation_factor == 1 || cpi->decimation_factor == 2))
{
cpi->decimation_factor = 2;
}
else if (cpi->buffer_level < drop_mark75 && (cpi->decimation_factor == 0 || cpi->decimation_factor == 1))
{
cpi->decimation_factor = 1;
}
//vpx_log("Encoder: Decimation Factor: %d \n",cpi->decimation_factor);
}
// The following decimates the frame rate according to a regular pattern (i.e. to 1/2 or 2/3 frame rate)
// This can be used to help prevent buffer under-run in CBR mode. Alternatively it might be desirable in
// some situations to drop frame rate but throw more bits at each frame.
//
// Note that dropping a key frame can be problematic if spatial resampling is also active
if (cpi->decimation_factor > 0)
{
switch (cpi->decimation_factor)
{
case 1:
cpi->per_frame_bandwidth = cpi->per_frame_bandwidth * 3 / 2;
break;
case 2:
cpi->per_frame_bandwidth = cpi->per_frame_bandwidth * 5 / 4;
break;
case 3:
cpi->per_frame_bandwidth = cpi->per_frame_bandwidth * 5 / 4;
break;
}
// Note that we should not throw out a key frame (especially when spatial resampling is enabled).
if ((cm->frame_type == KEY_FRAME)) // && cpi->oxcf.allow_spatial_resampling )
{
cpi->decimation_count = cpi->decimation_factor;
}
else if (cpi->decimation_count > 0)
{
cpi->decimation_count --;
cpi->bits_off_target += cpi->av_per_frame_bandwidth;
cm->current_video_frame++;
cpi->frames_since_key++;
#if CONFIG_PSNR
cpi->count ++;
#endif
cpi->buffer_level = cpi->bits_off_target;
return;
}
else
cpi->decimation_count = cpi->decimation_factor;
}
// Decide how big to make the frame
if (!pick_frame_size(cpi))
{
cm->current_video_frame++;
cpi->frames_since_key++;
return;
}
// Reduce active_worst_allowed_q for CBR if our buffer is getting too full.
// This has a knock on effect on active best quality as well.
// For CBR if the buffer reaches its maximum level then we can no longer
// save up bits for later frames so we might as well use them up
// on the current frame.
if ((cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER) &&
(cpi->buffer_level >= cpi->oxcf.optimal_buffer_level) && cpi->buffered_mode)
{
int Adjustment = cpi->active_worst_quality / 4; // Max adjustment is 1/4
if (Adjustment)
{
int buff_lvl_step;
int tmp_lvl = cpi->buffer_level;
if (cpi->buffer_level < cpi->oxcf.maximum_buffer_size)
{
buff_lvl_step = (cpi->oxcf.maximum_buffer_size - cpi->oxcf.optimal_buffer_level) / Adjustment;
if (buff_lvl_step)
{
Adjustment = (cpi->buffer_level - cpi->oxcf.optimal_buffer_level) / buff_lvl_step;
cpi->active_worst_quality -= Adjustment;
}
}
else
{
cpi->active_worst_quality -= Adjustment;
}
}
}
// Set an active best quality and if necessary active worst quality
if (cpi->pass == 2 || (cm->current_video_frame > 150))
{
int Q;
int i;
int bpm_target;
Control of active min quantizer for two pass. Create look up tables for controlling the active quantizer range. Some initial tuning to improve quality circa 0.5% on test set. Clean up of some stats output code Change-Id: Ia698a8525f8b8129a503cadace3ee73fe888f543
2010-09-29 13:03:19 +02:00
//int tmp;
vp8_clear_system_state();
Initial WebM release
2010-05-18 17:58:33 +02:00
Q = cpi->active_worst_quality;
if ((cm->frame_type == KEY_FRAME) || cm->refresh_golden_frame || cpi->common.refresh_alt_ref_frame)
{
if (cm->frame_type != KEY_FRAME)
{
if (cpi->avg_frame_qindex < cpi->active_worst_quality)
Q = cpi->avg_frame_qindex;
Tune effect of motion on KF/GF boost in two pass; This code adjust the impact of the amount and speed of motion on GF and KF boost. Sections with lots of slow motion will tend to have a somewhat bigger boost and sections with fast motion may have less. There is a knock on effect to the selection of the active quantizer range. This will likely require further tuning but helps with a couple of particularly bad edge cases. Change-Id: Ic2449cda7305672b69acf42fc0a845b77ac98d40
2010-10-02 18:31:46 +02:00
if ( cpi->gfu_boost > 1000 )
Control of active min quantizer for two pass. Create look up tables for controlling the active quantizer range. Some initial tuning to improve quality circa 0.5% on test set. Clean up of some stats output code Change-Id: Ia698a8525f8b8129a503cadace3ee73fe888f543
2010-09-29 13:03:19 +02:00
cpi->active_best_quality = gf_low_motion_minq[Q];
else if ( cpi->gfu_boost < 400 )
cpi->active_best_quality = gf_high_motion_minq[Q];
Initial WebM release
2010-05-18 17:58:33 +02:00
else
Control of active min quantizer for two pass. Create look up tables for controlling the active quantizer range. Some initial tuning to improve quality circa 0.5% on test set. Clean up of some stats output code Change-Id: Ia698a8525f8b8129a503cadace3ee73fe888f543
2010-09-29 13:03:19 +02:00
cpi->active_best_quality = gf_mid_motion_minq[Q];
/*cpi->active_best_quality = gf_arf_minq[Q];
tmp = (cpi->gfu_boost > 1000) ? 600 : cpi->gfu_boost - 400;
//tmp = (cpi->gfu_boost > 1000) ? 600 :
//(cpi->gfu_boost < 400) ? 0 : cpi->gfu_boost - 400;
tmp = 128 - (tmp >> 4);
cpi->active_best_quality = (cpi->active_best_quality * tmp)>>7;*/
}
// KEY FRAMES
else
{
Tune effect of motion on KF/GF boost in two pass; This code adjust the impact of the amount and speed of motion on GF and KF boost. Sections with lots of slow motion will tend to have a somewhat bigger boost and sections with fast motion may have less. There is a knock on effect to the selection of the active quantizer range. This will likely require further tuning but helps with a couple of particularly bad edge cases. Change-Id: Ic2449cda7305672b69acf42fc0a845b77ac98d40
2010-10-02 18:31:46 +02:00
if (cpi->gfu_boost > 600)
Control of active min quantizer for two pass. Create look up tables for controlling the active quantizer range. Some initial tuning to improve quality circa 0.5% on test set. Clean up of some stats output code Change-Id: Ia698a8525f8b8129a503cadace3ee73fe888f543
2010-09-29 13:03:19 +02:00
cpi->active_best_quality = kf_low_motion_minq[Q];
else
cpi->active_best_quality = kf_high_motion_minq[Q];
}
Initial WebM release
2010-05-18 17:58:33 +02:00
}
else
{
Control of active min quantizer for two pass. Create look up tables for controlling the active quantizer range. Some initial tuning to improve quality circa 0.5% on test set. Clean up of some stats output code Change-Id: Ia698a8525f8b8129a503cadace3ee73fe888f543
2010-09-29 13:03:19 +02:00
cpi->active_best_quality = inter_minq[Q];
Initial WebM release
2010-05-18 17:58:33 +02:00
}
// If CBR and the buffer is as full then it is reasonable to allow higher quality on the frames
// to prevent bits just going to waste.
if (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER)
{
// Note that the use of >= here elliminates the risk of a devide by 0 error in the else if clause
if (cpi->buffer_level >= cpi->oxcf.maximum_buffer_size)
cpi->active_best_quality = cpi->best_quality;
else if (cpi->buffer_level > cpi->oxcf.optimal_buffer_level)
{
int Fraction = ((cpi->buffer_level - cpi->oxcf.optimal_buffer_level) * 128) / (cpi->oxcf.maximum_buffer_size - cpi->oxcf.optimal_buffer_level);
int min_qadjustment = ((cpi->active_best_quality - cpi->best_quality) * Fraction) / 128;
cpi->active_best_quality -= min_qadjustment;
}
}
}
// Clip the active best and worst quality values to limits
if (cpi->active_worst_quality > cpi->worst_quality)
cpi->active_worst_quality = cpi->worst_quality;
if (cpi->active_best_quality < cpi->best_quality)
cpi->active_best_quality = cpi->best_quality;
else if (cpi->active_best_quality > cpi->active_worst_quality)
cpi->active_best_quality = cpi->active_worst_quality;
// Determine initial Q to try
Q = vp8_regulate_q(cpi, cpi->this_frame_target);
last_zbin_oq = cpi->zbin_over_quant;
// Set highest allowed value for Zbin over quant
if (cm->frame_type == KEY_FRAME)
zbin_oq_high = 0; //ZBIN_OQ_MAX/16
else if (cm->refresh_alt_ref_frame || (cm->refresh_golden_frame && !cpi->source_alt_ref_active))
zbin_oq_high = 16;
else
zbin_oq_high = ZBIN_OQ_MAX;
// Setup background Q adjustment for error resilliant mode
if (cpi->cyclic_refresh_mode_enabled)
cyclic_background_refresh(cpi, Q, 0);
vp8_compute_frame_size_bounds(cpi, &frame_under_shoot_limit, &frame_over_shoot_limit);
Correct q_low and q_high limits for the recode loop Corrected the initial Q range limits for the recode loop to reflect the current allowed range for the frame. In experimental work on constrained quality this bug was causing unnecessary recodes. Change-Id: I7e256fbfa681293b0223fe21ec329933d76c229f
2010-12-09 16:02:04 +01:00
// Limit Q range for the adaptive loop.
Initial WebM release
2010-05-18 17:58:33 +02:00
bottom_index = cpi->active_best_quality;
top_index = cpi->active_worst_quality;
Correct q_low and q_high limits for the recode loop Corrected the initial Q range limits for the recode loop to reflect the current allowed range for the frame. In experimental work on constrained quality this bug was causing unnecessary recodes. Change-Id: I7e256fbfa681293b0223fe21ec329933d76c229f
2010-12-09 16:02:04 +01:00
q_low = cpi->active_best_quality;
q_high = cpi->active_worst_quality;
Initial WebM release
2010-05-18 17:58:33 +02:00
vp8_save_coding_context(cpi);
loop_count = 0;
scale_and_extend_source(cpi->un_scaled_source, cpi);
#if !(CONFIG_REALTIME_ONLY) && CONFIG_POSTPROC
if (cpi->oxcf.noise_sensitivity > 0)
{
unsigned char *src;
int l = 0;
switch (cpi->oxcf.noise_sensitivity)
{
case 1:
l = 20;
break;
case 2:
l = 40;
break;
case 3:
l = 60;
break;
case 4:
l = 80;
break;
case 5:
l = 100;
break;
case 6:
l = 150;
break;
}
if (cm->frame_type == KEY_FRAME)
{
vp8_de_noise(cpi->Source, cpi->Source, l , 1, 0, RTCD(postproc));
}
else
{
vp8_de_noise(cpi->Source, cpi->Source, l , 1, 0, RTCD(postproc));
src = cpi->Source->y_buffer;
if (cpi->Source->y_stride < 0)
{
src += cpi->Source->y_stride * (cpi->Source->y_height - 1);
}
}
}
#endif
#ifdef OUTPUT_YUV_SRC
vp8_write_yuv_frame(cpi->Source);
#endif
do
{
vp8_clear_system_state(); //__asm emms;
/*
if(cpi->is_src_frame_alt_ref)
Q = 127;
*/
set_quantizer(cpi, Q);
this_q = Q;
// setup skip prob for costing in mode/mv decision
if (cpi->common.mb_no_coeff_skip)
{
cpi->prob_skip_false = cpi->base_skip_false_prob[Q];
if (cm->frame_type != KEY_FRAME)
{
if (cpi->common.refresh_alt_ref_frame)
{
if (cpi->last_skip_false_probs[2] != 0)
cpi->prob_skip_false = cpi->last_skip_false_probs[2];
/*
if(cpi->last_skip_false_probs[2]!=0 && abs(Q- cpi->last_skip_probs_q[2])<=16 )
cpi->prob_skip_false = cpi->last_skip_false_probs[2];
else if (cpi->last_skip_false_probs[2]!=0)
cpi->prob_skip_false = (cpi->last_skip_false_probs[2] + cpi->prob_skip_false ) / 2;
*/
}
else if (cpi->common.refresh_golden_frame)
{
if (cpi->last_skip_false_probs[1] != 0)
cpi->prob_skip_false = cpi->last_skip_false_probs[1];
/*
if(cpi->last_skip_false_probs[1]!=0 && abs(Q- cpi->last_skip_probs_q[1])<=16 )
cpi->prob_skip_false = cpi->last_skip_false_probs[1];
else if (cpi->last_skip_false_probs[1]!=0)
cpi->prob_skip_false = (cpi->last_skip_false_probs[1] + cpi->prob_skip_false ) / 2;
*/
}
else
{
if (cpi->last_skip_false_probs[0] != 0)
cpi->prob_skip_false = cpi->last_skip_false_probs[0];
/*
if(cpi->last_skip_false_probs[0]!=0 && abs(Q- cpi->last_skip_probs_q[0])<=16 )
cpi->prob_skip_false = cpi->last_skip_false_probs[0];
else if(cpi->last_skip_false_probs[0]!=0)
cpi->prob_skip_false = (cpi->last_skip_false_probs[0] + cpi->prob_skip_false ) / 2;
*/
}
//as this is for cost estimate, let's make sure it does not go extreme eitehr way
if (cpi->prob_skip_false < 5)
cpi->prob_skip_false = 5;
if (cpi->prob_skip_false > 250)
cpi->prob_skip_false = 250;
if (cpi->is_src_frame_alt_ref)
cpi->prob_skip_false = 1;
}
#if 0
if (cpi->pass != 1)
{
FILE *f = fopen("skip.stt", "a");
fprintf(f, "%d, %d, %4d ", cpi->common.refresh_golden_frame, cpi->common.refresh_alt_ref_frame, cpi->prob_skip_false);
fclose(f);
}
#endif
}
if (cm->frame_type == KEY_FRAME)
vp8_setup_key_frame(cpi);
// transform / motion compensation build reconstruction frame
vp8_encode_frame(cpi);
cpi->projected_frame_size -= vp8_estimate_entropy_savings(cpi);
cpi->projected_frame_size = (cpi->projected_frame_size > 0) ? cpi->projected_frame_size : 0;
vp8_clear_system_state(); //__asm emms;
// Test to see if the stats generated for this frame indicate that we should have coded a key frame
// (assuming that we didn't)!
if (cpi->pass != 2 && cpi->oxcf.auto_key && cm->frame_type != KEY_FRAME)
{
if (decide_key_frame(cpi))
{
vp8_calc_auto_iframe_target_size(cpi);
// Reset all our sizing numbers and recode
cm->frame_type = KEY_FRAME;
// Clear the Alt reference frame active flag when we have a key frame
cpi->source_alt_ref_active = FALSE;
Fix loopfilter delta zero transitions Loopfilter deltas are initialized to zero on keyframes in the decoder. The values then persist from the previous frame unless an update bit is set in the bitstream. This data is not included in the entropy data saved by the 'refresh entropy' bit in the bitstream, so it is effectively an additional contextual element beyond the 3 ref-frames and the entropy data. The encoder was treating this delta update bit as update-if-nonzero, meaning that the value would be refreshed even if it hadn't changed, and more significantly, if the correct value for the delta changed to zero, the update wouldn't be sent, and the decoder would preserve the last (presumably non-zero) value. This patch updates the encoder to send an update only if the value has changed from the previously transmitted value. It also forces the value to be transmitted in error resilient mode, to account for lost context in the event of lost frames. Change-Id: I56671d5b42965d0166ac226765dbfce3e5301868
2010-09-29 19:04:04 +02:00
// Reset the loop filter deltas and segmentation map
setup_features(cpi);
Initial WebM release
2010-05-18 17:58:33 +02:00
// If segmentation is enabled force a map update for key frames
if (cpi->mb.e_mbd.segmentation_enabled)
{
cpi->mb.e_mbd.update_mb_segmentation_map = 1;
cpi->mb.e_mbd.update_mb_segmentation_data = 1;
}
vp8_restore_coding_context(cpi);
Q = vp8_regulate_q(cpi, cpi->this_frame_target);
vp8_compute_frame_size_bounds(cpi, &frame_under_shoot_limit, &frame_over_shoot_limit);
Correct q_low and q_high limits for the recode loop Corrected the initial Q range limits for the recode loop to reflect the current allowed range for the frame. In experimental work on constrained quality this bug was causing unnecessary recodes. Change-Id: I7e256fbfa681293b0223fe21ec329933d76c229f
2010-12-09 16:02:04 +01:00
// Limit Q range for the adaptive loop.
Initial WebM release
2010-05-18 17:58:33 +02:00
bottom_index = cpi->active_best_quality;
top_index = cpi->active_worst_quality;
Correct q_low and q_high limits for the recode loop Corrected the initial Q range limits for the recode loop to reflect the current allowed range for the frame. In experimental work on constrained quality this bug was causing unnecessary recodes. Change-Id: I7e256fbfa681293b0223fe21ec329933d76c229f
2010-12-09 16:02:04 +01:00
q_low = cpi->active_best_quality;
q_high = cpi->active_worst_quality;
Initial WebM release
2010-05-18 17:58:33 +02:00
loop_count++;
Loop = TRUE;
resize_key_frame(cpi);
continue;
}
}
vp8_clear_system_state();
if (frame_over_shoot_limit == 0)
frame_over_shoot_limit = 1;
// Are we are overshooting and up against the limit of active max Q.
if (((cpi->pass != 2) || (cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER)) &&
(Q == cpi->active_worst_quality) &&
(cpi->active_worst_quality < cpi->worst_quality) &&
(cpi->projected_frame_size > frame_over_shoot_limit))
{
int over_size_percent = ((cpi->projected_frame_size - frame_over_shoot_limit) * 100) / frame_over_shoot_limit;
// If so is there any scope for relaxing it
while ((cpi->active_worst_quality < cpi->worst_quality) && (over_size_percent > 0))
{
cpi->active_worst_quality++;
top_index = cpi->active_worst_quality;
over_size_percent = (int)(over_size_percent * 0.96); // Assume 1 qstep = about 4% on frame size.
}
// If we have updated the active max Q do not call vp8_update_rate_correction_factors() this loop.
active_worst_qchanged = TRUE;
}
else
active_worst_qchanged = FALSE;
#if !(CONFIG_REALTIME_ONLY)
// Is the projected frame size out of range and are we allowed to attempt to recode.
Replaced recode loop test with a function call Replaced existing code to decide if a frame recode is required with a function call. This is to simplify addition of extra clauses that may be needed for the planned constrained quality mode. Also fixed a bug where by alt ref not considered in the test. Change-Id: I3d40bb21abe3e19f8456761e6849deb171738b60
2010-11-17 16:12:04 +01:00
if ( recode_loop_test( cpi,
frame_over_shoot_limit, frame_under_shoot_limit,
Q, top_index, bottom_index ) )
Initial WebM release
2010-05-18 17:58:33 +02:00
{
int last_q = Q;
int Retries = 0;
// Frame size out of permitted range:
// Update correction factor & compute new Q to try...
Replaced recode loop test with a function call Replaced existing code to decide if a frame recode is required with a function call. This is to simplify addition of extra clauses that may be needed for the planned constrained quality mode. Also fixed a bug where by alt ref not considered in the test. Change-Id: I3d40bb21abe3e19f8456761e6849deb171738b60
2010-11-17 16:12:04 +01:00
// Frame is too large
if (cpi->projected_frame_size > cpi->this_frame_target)
Initial WebM release
2010-05-18 17:58:33 +02:00
{
//if ( cpi->zbin_over_quant == 0 )
q_low = (Q < q_high) ? (Q + 1) : q_high; // Raise Qlow as to at least the current value
if (cpi->zbin_over_quant > 0) // If we are using over quant do the same for zbin_oq_low
zbin_oq_low = (cpi->zbin_over_quant < zbin_oq_high) ? (cpi->zbin_over_quant + 1) : zbin_oq_high;
//if ( undershoot_seen || (Q == MAXQ) )
if (undershoot_seen)
{
// Update rate_correction_factor unless cpi->active_worst_quality has changed.
if (!active_worst_qchanged)
vp8_update_rate_correction_factors(cpi, 1);
Q = (q_high + q_low + 1) / 2;
// Adjust cpi->zbin_over_quant (only allowed when Q is max)
if (Q < MAXQ)
cpi->zbin_over_quant = 0;
else
{
zbin_oq_low = (cpi->zbin_over_quant < zbin_oq_high) ? (cpi->zbin_over_quant + 1) : zbin_oq_high;
cpi->zbin_over_quant = (zbin_oq_high + zbin_oq_low) / 2;
}
}
else
{
// Update rate_correction_factor unless cpi->active_worst_quality has changed.
if (!active_worst_qchanged)
vp8_update_rate_correction_factors(cpi, 0);
Q = vp8_regulate_q(cpi, cpi->this_frame_target);
while (((Q < q_low) || (cpi->zbin_over_quant < zbin_oq_low)) && (Retries < 10))
{
vp8_update_rate_correction_factors(cpi, 0);
Q = vp8_regulate_q(cpi, cpi->this_frame_target);
Retries ++;
}
}
overshoot_seen = TRUE;
}
Replaced recode loop test with a function call Replaced existing code to decide if a frame recode is required with a function call. This is to simplify addition of extra clauses that may be needed for the planned constrained quality mode. Also fixed a bug where by alt ref not considered in the test. Change-Id: I3d40bb21abe3e19f8456761e6849deb171738b60
2010-11-17 16:12:04 +01:00
// Frame is too small
Initial WebM release
2010-05-18 17:58:33 +02:00
else
{
if (cpi->zbin_over_quant == 0)
q_high = (Q > q_low) ? (Q - 1) : q_low; // Lower q_high if not using over quant
else // else lower zbin_oq_high
zbin_oq_high = (cpi->zbin_over_quant > zbin_oq_low) ? (cpi->zbin_over_quant - 1) : zbin_oq_low;
if (overshoot_seen)
{
// Update rate_correction_factor unless cpi->active_worst_quality has changed.
if (!active_worst_qchanged)
vp8_update_rate_correction_factors(cpi, 1);
Q = (q_high + q_low) / 2;
// Adjust cpi->zbin_over_quant (only allowed when Q is max)
if (Q < MAXQ)
cpi->zbin_over_quant = 0;
else
cpi->zbin_over_quant = (zbin_oq_high + zbin_oq_low) / 2;
}
else
{
// Update rate_correction_factor unless cpi->active_worst_quality has changed.
if (!active_worst_qchanged)
vp8_update_rate_correction_factors(cpi, 0);
Q = vp8_regulate_q(cpi, cpi->this_frame_target);
while (((Q > q_high) || (cpi->zbin_over_quant > zbin_oq_high)) && (Retries < 10))
{
vp8_update_rate_correction_factors(cpi, 0);
Q = vp8_regulate_q(cpi, cpi->this_frame_target);
Retries ++;
}
}
undershoot_seen = TRUE;
}
// Clamp Q to upper and lower limits:
if (Q > q_high)
Q = q_high;
else if (Q < q_low)
Q = q_low;
// Clamp cpi->zbin_over_quant
cpi->zbin_over_quant = (cpi->zbin_over_quant < zbin_oq_low) ? zbin_oq_low : (cpi->zbin_over_quant > zbin_oq_high) ? zbin_oq_high : cpi->zbin_over_quant;
//Loop = ((Q != last_q) || (last_zbin_oq != cpi->zbin_over_quant)) ? TRUE : FALSE;
Loop = ((Q != last_q)) ? TRUE : FALSE;
last_zbin_oq = cpi->zbin_over_quant;
}
else
#endif
Loop = FALSE;
if (cpi->is_src_frame_alt_ref)
Loop = FALSE;
if (Loop == TRUE)
{
vp8_restore_coding_context(cpi);
loop_count++;
#if CONFIG_PSNR
cpi->tot_recode_hits++;
#endif
}
}
while (Loop == TRUE);
#if 0
// Experimental code for lagged and one pass
// Update stats used for one pass GF selection
{
/*
int frames_so_far;
double frame_intra_error;
double frame_coded_error;
double frame_pcnt_inter;
double frame_pcnt_motion;
double frame_mvr;
double frame_mvr_abs;
double frame_mvc;
double frame_mvc_abs;
*/
cpi->one_pass_frame_stats[cpi->one_pass_frame_index].frame_coded_error = (double)cpi->prediction_error;
cpi->one_pass_frame_stats[cpi->one_pass_frame_index].frame_intra_error = (double)cpi->intra_error;
cpi->one_pass_frame_stats[cpi->one_pass_frame_index].frame_pcnt_inter = (double)(100 - cpi->this_frame_percent_intra) / 100.0;
}
#endif
Merge "Improve MV prediction accuracy to achieve performance gain"
2010-12-06 18:41:09 +01:00
// Update the GF useage maps.
// This is done after completing the compression of a frame when all modes etc. are finalized but before loop filter
vp8_update_gf_useage_maps(cpi, cm, &cpi->mb);
Fix a bug in motion search code The MV's range is 256. Since the new motion search uses a different starting MV than the center ref MV, a MV range checking needs to be done to avoid corruption. Change-Id: I8ae0721d1bd203639e13891e2e54a2e87276f306
2010-12-14 17:00:25 +01:00
Improve MV prediction accuracy to achieve performance gain Add vp8_mv_pred() to better predict starting MV for NEWMV mode in vp8_rd_pick_inter_mode(). Set different search ranges according to MV prediction accuracy, which improves encoder performance without hurting the quality. Also, as Yaowu suggested, using diamond search result as full search starting point and therefore adjusting(reducing) full search range helps the performance. Change-Id: Ie4a3c8df87e697c1f4f6e2ddb693766bba1b77b6
2010-12-03 17:26:21 +01:00
// This frame's MVs are saved and will be used in next frame's MV prediction.
if(cm->show_frame) //do not save for altref frame
{
int mb_row;
int mb_col;
MODE_INFO *tmp = cm->mip; //point to beginning of allocated MODE_INFO arrays.
//static int last_video_frame = 0;
if(cm->frame_type != KEY_FRAME)
{
for (mb_row = 0; mb_row < cm->mb_rows+1; mb_row ++)
{
for (mb_col = 0; mb_col < cm->mb_cols+1; mb_col ++)
{
if(tmp->mbmi.ref_frame != INTRA_FRAME)
cpi->lfmv[mb_col + mb_row*(cm->mode_info_stride)].as_int = tmp->mbmi.mv.as_int;
cpi->lf_ref_frame_sign_bias[mb_col + mb_row*(cm->mode_info_stride)] = cm->ref_frame_sign_bias[tmp->mbmi.ref_frame];
cpi->lf_ref_frame[mb_col + mb_row*(cm->mode_info_stride)] = tmp->mbmi.ref_frame;
tmp++;
}
}
}
}
Initial WebM release
2010-05-18 17:58:33 +02:00
// Update the GF useage maps.
// This is done after completing the compression of a frame when all modes etc. are finalized but before loop filter
Moved gf_active code to encoder only The gf_active code is only used by the encoder, so it was moved from common and decoder. Change-Id: Iada15acd5b2b33ff70c34668ca87d4cfd0d05025
2010-08-11 17:02:31 +02:00
vp8_update_gf_useage_maps(cpi, cm, &cpi->mb);
Initial WebM release
2010-05-18 17:58:33 +02:00
if (cm->frame_type == KEY_FRAME)
cm->refresh_last_frame = 1;
disable compilation of debugging code This patch avoids compiling some debugging code in onyx_if.c. The most significant fix is to avoid generating code for vp8_write_yuv_frame, which is never called. Some other code was removed by the dead code elimination performed by the compiler, and this patch does it with the preprocessor instead. There are advantages both ways. Change-Id: I044fd43179d2e947553f0d6f2cad5b40907ac458
2010-09-24 17:10:25 +02:00
#if 0
Initial WebM release
2010-05-18 17:58:33 +02:00
{
FILE *f = fopen("gfactive.stt", "a");
Moved gf_active code to encoder only The gf_active code is only used by the encoder, so it was moved from common and decoder. Change-Id: Iada15acd5b2b33ff70c34668ca87d4cfd0d05025
2010-08-11 17:02:31 +02:00
fprintf(f, "%8d %8d %8d %8d %8d\n", cm->current_video_frame, (100 * cpi->gf_active_count) / (cpi->common.mb_rows * cpi->common.mb_cols), cpi->this_iiratio, cpi->next_iiratio, cm->refresh_golden_frame);
Initial WebM release
2010-05-18 17:58:33 +02:00
fclose(f);
}
disable compilation of debugging code This patch avoids compiling some debugging code in onyx_if.c. The most significant fix is to avoid generating code for vp8_write_yuv_frame, which is never called. Some other code was removed by the dead code elimination performed by the compiler, and this patch does it with the preprocessor instead. There are advantages both ways. Change-Id: I044fd43179d2e947553f0d6f2cad5b40907ac458
2010-09-24 17:10:25 +02:00
#endif
Initial WebM release
2010-05-18 17:58:33 +02:00
// For inter frames the current default behaviour is that when cm->refresh_golden_frame is set we copy the old GF over to the ARF buffer
// This is purely an encoder descision at present.
if (!cpi->oxcf.error_resilient_mode && cm->refresh_golden_frame)
cm->copy_buffer_to_arf = 2;
else
cm->copy_buffer_to_arf = 0;
if (cm->refresh_last_frame)
{
Swap alt/gold/new/last frame buffer ptrs instead of copying. At the end of the decode, frame buffers were being copied. The frames are not updated after the copy, they are just for reference on later frames. This change allows multiple references to the same frame buffer instead of copying it. Changes needed to be made to the encoder to handle this. The encoder is still doing frame buffer copies in similar places where pointer reference could be done. Change-Id: I7c38be4d23979cc49b5f17241ca3a78703803e66
2010-07-22 14:07:32 +02:00
vp8_swap_yv12_buffer(&cm->yv12_fb[cm->lst_fb_idx], &cm->yv12_fb[cm->new_fb_idx]);
cm->frame_to_show = &cm->yv12_fb[cm->lst_fb_idx];
Initial WebM release
2010-05-18 17:58:33 +02:00
}
else
Swap alt/gold/new/last frame buffer ptrs instead of copying. At the end of the decode, frame buffers were being copied. The frames are not updated after the copy, they are just for reference on later frames. This change allows multiple references to the same frame buffer instead of copying it. Changes needed to be made to the encoder to handle this. The encoder is still doing frame buffer copies in similar places where pointer reference could be done. Change-Id: I7c38be4d23979cc49b5f17241ca3a78703803e66
2010-07-22 14:07:32 +02:00
cm->frame_to_show = &cm->yv12_fb[cm->new_fb_idx];
Initial WebM release
2010-05-18 17:58:33 +02:00
//#pragma omp parallel sections
{
//#pragma omp section
{
struct vpx_usec_timer timer;
vpx_usec_timer_start(&timer);
if (cpi->sf.auto_filter == 0)
vp8cx_pick_filter_level_fast(cpi->Source, cpi);
else
vp8cx_pick_filter_level(cpi->Source, cpi);
vpx_usec_timer_mark(&timer);
cpi->time_pick_lpf += vpx_usec_timer_elapsed(&timer);
if (cm->no_lpf)
cm->filter_level = 0;
if (cm->filter_level > 0)
{
vp8cx_set_alt_lf_level(cpi, cm->filter_level);
vp8_loop_filter_frame(cm, &cpi->mb.e_mbd, cm->filter_level);
cm->last_frame_type = cm->frame_type;
cm->last_filter_type = cm->filter_type;
cm->last_sharpness_level = cm->sharpness_level;
}
vp8_yv12_extend_frame_borders_ptr(cm->frame_to_show);
if (cpi->oxcf.error_resilient_mode == 1)
{
cm->refresh_entropy_probs = 0;
}
}
//#pragma omp section
{
// build the bitstream
vp8_pack_bitstream(cpi, dest, size);
}
}
{
Swap alt/gold/new/last frame buffer ptrs instead of copying. At the end of the decode, frame buffers were being copied. The frames are not updated after the copy, they are just for reference on later frames. This change allows multiple references to the same frame buffer instead of copying it. Changes needed to be made to the encoder to handle this. The encoder is still doing frame buffer copies in similar places where pointer reference could be done. Change-Id: I7c38be4d23979cc49b5f17241ca3a78703803e66
2010-07-22 14:07:32 +02:00
YV12_BUFFER_CONFIG *lst_yv12 = &cm->yv12_fb[cm->lst_fb_idx];
YV12_BUFFER_CONFIG *new_yv12 = &cm->yv12_fb[cm->new_fb_idx];
YV12_BUFFER_CONFIG *gld_yv12 = &cm->yv12_fb[cm->gld_fb_idx];
YV12_BUFFER_CONFIG *alt_yv12 = &cm->yv12_fb[cm->alt_fb_idx];
// At this point the new frame has been encoded coded.
// If any buffer copy / swaping is signalled it should be done here.
if (cm->frame_type == KEY_FRAME)
{
vp8_yv12_copy_frame_ptr(cm->frame_to_show, gld_yv12);
vp8_yv12_copy_frame_ptr(cm->frame_to_show, alt_yv12);
}
else // For non key frames
Initial WebM release
2010-05-18 17:58:33 +02:00
{
Swap alt/gold/new/last frame buffer ptrs instead of copying. At the end of the decode, frame buffers were being copied. The frames are not updated after the copy, they are just for reference on later frames. This change allows multiple references to the same frame buffer instead of copying it. Changes needed to be made to the encoder to handle this. The encoder is still doing frame buffer copies in similar places where pointer reference could be done. Change-Id: I7c38be4d23979cc49b5f17241ca3a78703803e66
2010-07-22 14:07:32 +02:00
// Code to copy between reference buffers
if (cm->copy_buffer_to_arf)
Initial WebM release
2010-05-18 17:58:33 +02:00
{
Swap alt/gold/new/last frame buffer ptrs instead of copying. At the end of the decode, frame buffers were being copied. The frames are not updated after the copy, they are just for reference on later frames. This change allows multiple references to the same frame buffer instead of copying it. Changes needed to be made to the encoder to handle this. The encoder is still doing frame buffer copies in similar places where pointer reference could be done. Change-Id: I7c38be4d23979cc49b5f17241ca3a78703803e66
2010-07-22 14:07:32 +02:00
if (cm->copy_buffer_to_arf == 1)
{
if (cm->refresh_last_frame)
// We copy new_frame here because last and new buffers will already have been swapped if cm->refresh_last_frame is set.
vp8_yv12_copy_frame_ptr(new_yv12, alt_yv12);
else
vp8_yv12_copy_frame_ptr(lst_yv12, alt_yv12);
}
else if (cm->copy_buffer_to_arf == 2)
vp8_yv12_copy_frame_ptr(gld_yv12, alt_yv12);
Initial WebM release
2010-05-18 17:58:33 +02:00
}
Swap alt/gold/new/last frame buffer ptrs instead of copying. At the end of the decode, frame buffers were being copied. The frames are not updated after the copy, they are just for reference on later frames. This change allows multiple references to the same frame buffer instead of copying it. Changes needed to be made to the encoder to handle this. The encoder is still doing frame buffer copies in similar places where pointer reference could be done. Change-Id: I7c38be4d23979cc49b5f17241ca3a78703803e66
2010-07-22 14:07:32 +02:00
if (cm->copy_buffer_to_gf)
Initial WebM release
2010-05-18 17:58:33 +02:00
{
Swap alt/gold/new/last frame buffer ptrs instead of copying. At the end of the decode, frame buffers were being copied. The frames are not updated after the copy, they are just for reference on later frames. This change allows multiple references to the same frame buffer instead of copying it. Changes needed to be made to the encoder to handle this. The encoder is still doing frame buffer copies in similar places where pointer reference could be done. Change-Id: I7c38be4d23979cc49b5f17241ca3a78703803e66
2010-07-22 14:07:32 +02:00
if (cm->copy_buffer_to_gf == 1)
{
if (cm->refresh_last_frame)
// We copy new_frame here because last and new buffers will already have been swapped if cm->refresh_last_frame is set.
vp8_yv12_copy_frame_ptr(new_yv12, gld_yv12);
else
vp8_yv12_copy_frame_ptr(lst_yv12, gld_yv12);
}
else if (cm->copy_buffer_to_gf == 2)
vp8_yv12_copy_frame_ptr(alt_yv12, gld_yv12);
Initial WebM release
2010-05-18 17:58:33 +02:00
}
}
}
// Update rate control heuristics
cpi->total_byte_count += (*size);
cpi->projected_frame_size = (*size) << 3;
if (!active_worst_qchanged)
vp8_update_rate_correction_factors(cpi, 2);
cpi->last_q[cm->frame_type] = cm->base_qindex;
if (cm->frame_type == KEY_FRAME)
{
vp8_adjust_key_frame_context(cpi);
}
// Keep a record of ambient average Q.
if (cm->frame_type == KEY_FRAME)
cpi->avg_frame_qindex = cm->base_qindex;
else
cpi->avg_frame_qindex = (2 + 3 * cpi->avg_frame_qindex + cm->base_qindex) >> 2;
// Keep a record from which we can calculate the average Q excluding GF updates and key frames
if ((cm->frame_type != KEY_FRAME) && !cm->refresh_golden_frame && !cm->refresh_alt_ref_frame)
{
cpi->ni_frames++;
// Calculate the average Q for normal inter frames (not key or GFU frames)
// This is used as a basis for setting active worst quality.
if (cpi->ni_frames > 150)
{
cpi->ni_tot_qi += Q;
cpi->ni_av_qi = (cpi->ni_tot_qi / cpi->ni_frames);
}
// Early in the clip ... average the current frame Q value with the default
// entered by the user as a dampening measure
else
{
cpi->ni_tot_qi += Q;
cpi->ni_av_qi = ((cpi->ni_tot_qi / cpi->ni_frames) + cpi->worst_quality + 1) / 2;
}
// If the average Q is higher than what was used in the last frame
// (after going through the recode loop to keep the frame size within range)
// then use the last frame value - 1.
// The -1 is designed to stop Q and hence the data rate, from progressively
// falling away during difficult sections, but at the same time reduce the number of
// itterations around the recode loop.
if (Q > cpi->ni_av_qi)
cpi->ni_av_qi = Q - 1;
}
#if 0
// If the frame was massively oversize and we are below optimal buffer level drop next frame
if ((cpi->drop_frames_allowed) &&
(cpi->oxcf.end_usage == USAGE_STREAM_FROM_SERVER) &&
(cpi->buffer_level < cpi->oxcf.drop_frames_water_mark * cpi->oxcf.optimal_buffer_level / 100) &&
(cpi->projected_frame_size > (4 * cpi->this_frame_target)))
{
cpi->drop_frame = TRUE;
}
#endif
// Set the count for maximum consequative dropped frames based upon the ratio of
// this frame size to the target average per frame bandwidth.
// (cpi->av_per_frame_bandwidth > 0) is just a sanity check to prevent / 0.
if (cpi->drop_frames_allowed && (cpi->av_per_frame_bandwidth > 0))
{
cpi->max_drop_count = cpi->projected_frame_size / cpi->av_per_frame_bandwidth;
if (cpi->max_drop_count > cpi->max_consec_dropped_frames)
cpi->max_drop_count = cpi->max_consec_dropped_frames;
}
// Update the buffer level variable.
if (cpi->common.refresh_alt_ref_frame)
cpi->bits_off_target -= cpi->projected_frame_size;
else
cpi->bits_off_target += cpi->av_per_frame_bandwidth - cpi->projected_frame_size;
// Rolling monitors of whether we are over or underspending used to help regulate min and Max Q in two pass.
cpi->rolling_target_bits = ((cpi->rolling_target_bits * 3) + cpi->this_frame_target + 2) / 4;
cpi->rolling_actual_bits = ((cpi->rolling_actual_bits * 3) + cpi->projected_frame_size + 2) / 4;
cpi->long_rolling_target_bits = ((cpi->long_rolling_target_bits * 31) + cpi->this_frame_target + 16) / 32;
cpi->long_rolling_actual_bits = ((cpi->long_rolling_actual_bits * 31) + cpi->projected_frame_size + 16) / 32;
// Actual bits spent
cpi->total_actual_bits += cpi->projected_frame_size;
// Debug stats
cpi->total_target_vs_actual += (cpi->this_frame_target - cpi->projected_frame_size);
cpi->buffer_level = cpi->bits_off_target;
// Update bits left to the kf and gf groups to account for overshoot or undershoot on these frames
if (cm->frame_type == KEY_FRAME)
{
cpi->kf_group_bits += cpi->this_frame_target - cpi->projected_frame_size;
if (cpi->kf_group_bits < 0)
cpi->kf_group_bits = 0 ;
}
else if (cm->refresh_golden_frame || cm->refresh_alt_ref_frame)
{
cpi->gf_group_bits += cpi->this_frame_target - cpi->projected_frame_size;
if (cpi->gf_group_bits < 0)
cpi->gf_group_bits = 0 ;
}
if (cm->frame_type != KEY_FRAME)
{
if (cpi->common.refresh_alt_ref_frame)
{
cpi->last_skip_false_probs[2] = cpi->prob_skip_false;
cpi->last_skip_probs_q[2] = cm->base_qindex;
}
else if (cpi->common.refresh_golden_frame)
{
cpi->last_skip_false_probs[1] = cpi->prob_skip_false;
cpi->last_skip_probs_q[1] = cm->base_qindex;
}
else
{
cpi->last_skip_false_probs[0] = cpi->prob_skip_false;
cpi->last_skip_probs_q[0] = cm->base_qindex;
//update the baseline
cpi->base_skip_false_prob[cm->base_qindex] = cpi->prob_skip_false;
}
}
disable compilation of debugging code This patch avoids compiling some debugging code in onyx_if.c. The most significant fix is to avoid generating code for vp8_write_yuv_frame, which is never called. Some other code was removed by the dead code elimination performed by the compiler, and this patch does it with the preprocessor instead. There are advantages both ways. Change-Id: I044fd43179d2e947553f0d6f2cad5b40907ac458
2010-09-24 17:10:25 +02:00
#if 0 && CONFIG_PSNR
Initial WebM release
2010-05-18 17:58:33 +02:00
{
FILE *f = fopen("tmp.stt", "a");
vp8_clear_system_state(); //__asm emms;
if (cpi->total_coded_error_left != 0.0)
Control of active min quantizer for two pass. Create look up tables for controlling the active quantizer range. Some initial tuning to improve quality circa 0.5% on test set. Clean up of some stats output code Change-Id: Ia698a8525f8b8129a503cadace3ee73fe888f543
2010-09-29 13:03:19 +02:00
fprintf(f, "%10d %10d %10d %10d %10d %10d %10d %10d %6ld %6ld"
"%6ld %6ld %5ld %5ld %5ld %8ld %8.2f %10d %10.3f"
"%10.3f %8ld\n",
cpi->common.current_video_frame, cpi->this_frame_target,
cpi->projected_frame_size,
(cpi->projected_frame_size - cpi->this_frame_target),
(int)cpi->total_target_vs_actual,
(cpi->oxcf.starting_buffer_level-cpi->bits_off_target),
(int)cpi->total_actual_bits, cm->base_qindex,
cpi->active_best_quality, cpi->active_worst_quality,
cpi->avg_frame_qindex, cpi->zbin_over_quant,
cm->refresh_golden_frame, cm->refresh_alt_ref_frame,
cm->frame_type, cpi->gfu_boost,
cpi->est_max_qcorrection_factor, (int)cpi->bits_left,
cpi->total_coded_error_left,
(double)cpi->bits_left / cpi->total_coded_error_left,
cpi->tot_recode_hits);
Initial WebM release
2010-05-18 17:58:33 +02:00
else
Control of active min quantizer for two pass. Create look up tables for controlling the active quantizer range. Some initial tuning to improve quality circa 0.5% on test set. Clean up of some stats output code Change-Id: Ia698a8525f8b8129a503cadace3ee73fe888f543
2010-09-29 13:03:19 +02:00
fprintf(f, "%10d %10d %10d %10d %10d %10d %10d %10d %6ld %6ld"
"%6ld %6ld %5ld %5ld %5ld %8ld %8.2f %10d %10.3f"
"%8ld\n",
cpi->common.current_video_frame,
cpi->this_frame_target, cpi->projected_frame_size,
(cpi->projected_frame_size - cpi->this_frame_target),
(int)cpi->total_target_vs_actual,
(cpi->oxcf.starting_buffer_level-cpi->bits_off_target),
(int)cpi->total_actual_bits, cm->base_qindex,
cpi->active_best_quality, cpi->active_worst_quality,
cpi->avg_frame_qindex, cpi->zbin_over_quant,
cm->refresh_golden_frame, cm->refresh_alt_ref_frame,
cm->frame_type, cpi->gfu_boost,
cpi->est_max_qcorrection_factor, (int)cpi->bits_left,
cpi->total_coded_error_left, cpi->tot_recode_hits);
Initial WebM release
2010-05-18 17:58:33 +02:00
fclose(f);
{
FILE *fmodes = fopen("Modes.stt", "a");
int i;
Control of active min quantizer for two pass. Create look up tables for controlling the active quantizer range. Some initial tuning to improve quality circa 0.5% on test set. Clean up of some stats output code Change-Id: Ia698a8525f8b8129a503cadace3ee73fe888f543
2010-09-29 13:03:19 +02:00
fprintf(fmodes, "%6d:%1d:%1d:%1d ",
cpi->common.current_video_frame,
cm->frame_type, cm->refresh_golden_frame,
cm->refresh_alt_ref_frame);
Initial WebM release
2010-05-18 17:58:33 +02:00
for (i = 0; i < MAX_MODES; i++)
fprintf(fmodes, "%5d ", cpi->mode_chosen_counts[i]);
fprintf(fmodes, "\n");
fclose(fmodes);
}
}
#endif
// If this was a kf or Gf note the Q
if ((cm->frame_type == KEY_FRAME) || cm->refresh_golden_frame || cm->refresh_alt_ref_frame)
cm->last_kf_gf_q = cm->base_qindex;
if (cm->refresh_golden_frame == 1)
cm->frame_flags = cm->frame_flags | FRAMEFLAGS_GOLDEN;
else
cm->frame_flags = cm->frame_flags&~FRAMEFLAGS_GOLDEN;
if (cm->refresh_alt_ref_frame == 1)
cm->frame_flags = cm->frame_flags | FRAMEFLAGS_ALTREF;
else
cm->frame_flags = cm->frame_flags&~FRAMEFLAGS_ALTREF;
if (cm->refresh_last_frame & cm->refresh_golden_frame) // both refreshed
cpi->gold_is_last = 1;
else if (cm->refresh_last_frame ^ cm->refresh_golden_frame) // 1 refreshed but not the other
cpi->gold_is_last = 0;
if (cm->refresh_last_frame & cm->refresh_alt_ref_frame) // both refreshed
cpi->alt_is_last = 1;
else if (cm->refresh_last_frame ^ cm->refresh_alt_ref_frame) // 1 refreshed but not the other
cpi->alt_is_last = 0;
if (cm->refresh_alt_ref_frame & cm->refresh_golden_frame) // both refreshed
cpi->gold_is_alt = 1;
else if (cm->refresh_alt_ref_frame ^ cm->refresh_golden_frame) // 1 refreshed but not the other
cpi->gold_is_alt = 0;
cpi->ref_frame_flags = VP8_ALT_FLAG | VP8_GOLD_FLAG | VP8_LAST_FLAG;
if (cpi->gold_is_last)
Changed unary operator from ! to ~ Since the intent is to reset the appropriate bit in ref_frame_flags not to test a logic condition. Prior result would always have been ref_frame_flags being set to 0. (Issue reported by dgohman, issue 47) Change-Id: I2c12502ed74c73cf38e98c9680e0249c29e16433
2010-06-18 16:58:18 +02:00
cpi->ref_frame_flags &= ~VP8_GOLD_FLAG;
Initial WebM release
2010-05-18 17:58:33 +02:00
if (cpi->alt_is_last)
Changed unary operator from ! to ~ Since the intent is to reset the appropriate bit in ref_frame_flags not to test a logic condition. Prior result would always have been ref_frame_flags being set to 0. (Issue reported by dgohman, issue 47) Change-Id: I2c12502ed74c73cf38e98c9680e0249c29e16433
2010-06-18 16:58:18 +02:00
cpi->ref_frame_flags &= ~VP8_ALT_FLAG;
Initial WebM release
2010-05-18 17:58:33 +02:00
if (cpi->gold_is_alt)
Changed unary operator from ! to ~ Since the intent is to reset the appropriate bit in ref_frame_flags not to test a logic condition. Prior result would always have been ref_frame_flags being set to 0. (Issue reported by dgohman, issue 47) Change-Id: I2c12502ed74c73cf38e98c9680e0249c29e16433
2010-06-18 16:58:18 +02:00
cpi->ref_frame_flags &= ~VP8_ALT_FLAG;
Initial WebM release
2010-05-18 17:58:33 +02:00
if (cpi->oxcf.error_resilient_mode)
{
// Is this an alternate reference update
if (cpi->common.refresh_alt_ref_frame)
Swap alt/gold/new/last frame buffer ptrs instead of copying. At the end of the decode, frame buffers were being copied. The frames are not updated after the copy, they are just for reference on later frames. This change allows multiple references to the same frame buffer instead of copying it. Changes needed to be made to the encoder to handle this. The encoder is still doing frame buffer copies in similar places where pointer reference could be done. Change-Id: I7c38be4d23979cc49b5f17241ca3a78703803e66
2010-07-22 14:07:32 +02:00
vp8_yv12_copy_frame_ptr(cm->frame_to_show, &cm->yv12_fb[cm->alt_fb_idx]);
Initial WebM release
2010-05-18 17:58:33 +02:00
if (cpi->common.refresh_golden_frame)
Swap alt/gold/new/last frame buffer ptrs instead of copying. At the end of the decode, frame buffers were being copied. The frames are not updated after the copy, they are just for reference on later frames. This change allows multiple references to the same frame buffer instead of copying it. Changes needed to be made to the encoder to handle this. The encoder is still doing frame buffer copies in similar places where pointer reference could be done. Change-Id: I7c38be4d23979cc49b5f17241ca3a78703803e66
2010-07-22 14:07:32 +02:00
vp8_yv12_copy_frame_ptr(cm->frame_to_show, &cm->yv12_fb[cm->gld_fb_idx]);
Initial WebM release
2010-05-18 17:58:33 +02:00
}
else
{
Set refresh_alt_ref_frame on keyframe encode. On a keyframe alt ref and golden are refreshed. The flag was not being set and so on the frame after a keyframe, motion search would occur on the alt ref frame. This is not necessary because the alt ref frame identical to the last frame in this scenario. Handle corner case where a forward alt-ref frame is put directly after a keyframe. Change-Id: I9be4cf290d694f8cf2f9a31852014b5ccf1504d3
2010-11-18 19:40:58 +01:00
if (cpi->oxcf.play_alternate && cpi->common.refresh_alt_ref_frame && (cpi->common.frame_type != KEY_FRAME))
Initial WebM release
2010-05-18 17:58:33 +02:00
// Update the alternate reference frame and stats as appropriate.
update_alt_ref_frame_and_stats(cpi);
else
// Update the Golden frame and golden frame and stats as appropriate.
update_golden_frame_and_stats(cpi);
}
if (cm->frame_type == KEY_FRAME)
{
// Tell the caller that the frame was coded as a key frame
*frame_flags = cm->frame_flags | FRAMEFLAGS_KEY;
// As this frame is a key frame the next defaults to an inter frame.
cm->frame_type = INTER_FRAME;
cpi->last_frame_percent_intra = 100;
}
else
{
*frame_flags = cm->frame_flags&~FRAMEFLAGS_KEY;
cpi->last_frame_percent_intra = cpi->this_frame_percent_intra;
}
// Clear the one shot update flags for segmentation map and mode/ref loop filter deltas.
cpi->mb.e_mbd.update_mb_segmentation_map = 0;
cpi->mb.e_mbd.update_mb_segmentation_data = 0;
cpi->mb.e_mbd.mode_ref_lf_delta_update = 0;
// Dont increment frame counters if this was an altref buffer update not a real frame
if (cm->show_frame)
{
cm->current_video_frame++;
cpi->frames_since_key++;
}
// reset to normal state now that we are done.
disable compilation of debugging code This patch avoids compiling some debugging code in onyx_if.c. The most significant fix is to avoid generating code for vp8_write_yuv_frame, which is never called. Some other code was removed by the dead code elimination performed by the compiler, and this patch does it with the preprocessor instead. There are advantages both ways. Change-Id: I044fd43179d2e947553f0d6f2cad5b40907ac458
2010-09-24 17:10:25 +02:00
#if 0
Initial WebM release
2010-05-18 17:58:33 +02:00
{
char filename[512];
FILE *recon_file;
sprintf(filename, "enc%04d.yuv", (int) cm->current_video_frame);
recon_file = fopen(filename, "wb");
Swap alt/gold/new/last frame buffer ptrs instead of copying. At the end of the decode, frame buffers were being copied. The frames are not updated after the copy, they are just for reference on later frames. This change allows multiple references to the same frame buffer instead of copying it. Changes needed to be made to the encoder to handle this. The encoder is still doing frame buffer copies in similar places where pointer reference could be done. Change-Id: I7c38be4d23979cc49b5f17241ca3a78703803e66
2010-07-22 14:07:32 +02:00
fwrite(cm->yv12_fb[cm->lst_fb_idx].buffer_alloc,
cm->yv12_fb[cm->lst_fb_idx].frame_size, 1, recon_file);
Initial WebM release
2010-05-18 17:58:33 +02:00
fclose(recon_file);
}
disable compilation of debugging code This patch avoids compiling some debugging code in onyx_if.c. The most significant fix is to avoid generating code for vp8_write_yuv_frame, which is never called. Some other code was removed by the dead code elimination performed by the compiler, and this patch does it with the preprocessor instead. There are advantages both ways. Change-Id: I044fd43179d2e947553f0d6f2cad5b40907ac458
2010-09-24 17:10:25 +02:00
#endif
Initial WebM release
2010-05-18 17:58:33 +02:00
// DEBUG
//vp8_write_yuv_frame("encoder_recon.yuv", cm->frame_to_show);
}
int vp8_is_gf_update_needed(VP8_PTR ptr)
{
VP8_COMP *cpi = (VP8_COMP *) ptr;
int ret_val;
ret_val = cpi->gf_update_recommended;
cpi->gf_update_recommended = 0;
return ret_val;
}
void vp8_check_gf_quality(VP8_COMP *cpi)
{
VP8_COMMON *cm = &cpi->common;
Moved gf_active code to encoder only The gf_active code is only used by the encoder, so it was moved from common and decoder. Change-Id: Iada15acd5b2b33ff70c34668ca87d4cfd0d05025
2010-08-11 17:02:31 +02:00
int gf_active_pct = (100 * cpi->gf_active_count) / (cm->mb_rows * cm->mb_cols);
Initial WebM release
2010-05-18 17:58:33 +02:00
int gf_ref_usage_pct = (cpi->count_mb_ref_frame_usage[GOLDEN_FRAME] * 100) / (cm->mb_rows * cm->mb_cols);
int last_ref_zz_useage = (cpi->inter_zz_count * 100) / (cm->mb_rows * cm->mb_cols);
// Gf refresh is not currently being signalled
if (cpi->gf_update_recommended == 0)
{
if (cpi->common.frames_since_golden > 7)
{
// Low use of gf
if ((gf_active_pct < 10) || ((gf_active_pct + gf_ref_usage_pct) < 15))
{
// ...but last frame zero zero usage is reasonbable so a new gf might be appropriate
if (last_ref_zz_useage >= 25)
{
cpi->gf_bad_count ++;
if (cpi->gf_bad_count >= 8) // Check that the condition is stable
{
cpi->gf_update_recommended = 1;
cpi->gf_bad_count = 0;
}
}
else
cpi->gf_bad_count = 0; // Restart count as the background is not stable enough
}
else
cpi->gf_bad_count = 0; // Gf useage has picked up so reset count
}
}
// If the signal is set but has not been read should we cancel it.
else if (last_ref_zz_useage < 15)
{
cpi->gf_update_recommended = 0;
cpi->gf_bad_count = 0;
}
#if 0
{
FILE *f = fopen("gfneeded.stt", "a");
fprintf(f, "%10d %10d %10d %10d %10ld \n",
cm->current_video_frame,
cpi->common.frames_since_golden,
gf_active_pct, gf_ref_usage_pct,
cpi->gf_update_recommended);
fclose(f);
}
#endif
}
#if !(CONFIG_REALTIME_ONLY)
static void Pass2Encode(VP8_COMP *cpi, unsigned long *size, unsigned char *dest, unsigned int *frame_flags)
{
if (!cpi->common.refresh_alt_ref_frame)
vp8_second_pass(cpi);
encode_frame_to_data_rate(cpi, size, dest, frame_flags);
cpi->bits_left -= 8 * *size;
if (!cpi->common.refresh_alt_ref_frame)
Fixed an encoder debug/relese mismatch in x86_64-win64-vs8 Visual c++ compiler uses xmm registers for floating point operations for 64 bit architecture, therefore its calling convention requires the preservation of xmm6-xmm15 in any function that have used these registers. However, the sse2 functions, that were originally written for 32 bit windows, may have used xmm6 and xmm7 without preserving the content. In this particular case, the compiler used xmm6 to save the variable "two_pass_min_rate", the value of the variable is mucked up by our sse2 optimized loop filter functions, hence the results of release/debug mismatching.
2010-05-20 00:48:00 +02:00
{
double two_pass_min_rate = (double)(cpi->oxcf.target_bandwidth
*cpi->oxcf.two_pass_vbrmin_section / 100);
Initial WebM release
2010-05-18 17:58:33 +02:00
cpi->bits_left += (long long)(two_pass_min_rate / cpi->oxcf.frame_rate);
Fixed an encoder debug/relese mismatch in x86_64-win64-vs8 Visual c++ compiler uses xmm registers for floating point operations for 64 bit architecture, therefore its calling convention requires the preservation of xmm6-xmm15 in any function that have used these registers. However, the sse2 functions, that were originally written for 32 bit windows, may have used xmm6 and xmm7 without preserving the content. In this particular case, the compiler used xmm6 to save the variable "two_pass_min_rate", the value of the variable is mucked up by our sse2 optimized loop filter functions, hence the results of release/debug mismatching.
2010-05-20 00:48:00 +02:00
}
Initial WebM release
2010-05-18 17:58:33 +02:00
}
#endif
//For ARM NEON, d8-d15 are callee-saved registers, and need to be saved by us.
#if HAVE_ARMV7
extern void vp8_push_neon(INT64 *store);
extern void vp8_pop_neon(INT64 *store);
#endif
int vp8_receive_raw_frame(VP8_PTR ptr, unsigned int frame_flags, YV12_BUFFER_CONFIG *sd, INT64 time_stamp, INT64 end_time)
{
Add runtime CPU detection support for ARM. The primary goal is to allow a binary to be built which supports NEON, but can fall back to non-NEON routines, since some Android devices do not have NEON, even if they are otherwise ARMv7 (e.g., Tegra). The configure-generated flags HAVE_ARMV7, etc., are used to decide which versions of each function to build, and when CONFIG_RUNTIME_CPU_DETECT is enabled, the correct version is chosen at run time. In order for this to work, the CFLAGS must be set to something appropriate (e.g., without -mfpu=neon for ARMv7, and with appropriate -march and -mcpu for even earlier configurations), or the native C code will not be able to run. The ASFLAGS must remain set for the most advanced instruction set required at build time, since the ARM assembler will refuse to emit them otherwise. I have not attempted to make any changes to configure to do this automatically. Doing so will probably require the addition of new configure options. Many of the hooks for RTCD on ARM were already there, but a lot of the code had bit-rotted, and a good deal of the ARM-specific code is not integrated into the RTCD structs at all. I did not try to resolve the latter, merely to add the minimal amount of protection around them to allow RTCD to work. Those functions that were called based on an ifdef at the calling site were expanded to check the RTCD flags at that site, but they should be added to an RTCD struct somewhere in the future. The functions invoked with global function pointers still are, but these should be moved into an RTCD struct for thread safety (I believe every platform currently supported has atomic pointer stores, but this is not guaranteed). The encoder's boolhuff functions did not even have _c and armv7 suffixes, and the correct version was resolved at link time. The token packing functions did have appropriate suffixes, but the version was selected with a define, with no associated RTCD struct. However, for both of these, the only armv7 instruction they actually used was rbit, and this was completely superfluous, so I reworked them to avoid it. The only non-ARMv4 instruction remaining in them is clz, which is ARMv5 (not even ARMv5TE is required). Considering that there are no ARM-specific configs which are not at least ARMv5TE, I did not try to detect these at runtime, and simply enable them for ARMv5 and above. Finally, the NEON register saving code was completely non-reentrant, since it saved the registers to a global, static variable. I moved the storage for this onto the stack. A single binary built with this code was tested on an ARM11 (ARMv6) and a Cortex A8 (ARMv7 w/NEON), for both the encoder and decoder, and produced identical output, while using the correct accelerated functions on each. I did not test on any earlier processors. Change-Id: I45cbd63a614f4554c3b325c45d46c0806f009eaa
2010-10-21 00:39:11 +02:00
INT64 store_reg[8];
Initial WebM release
2010-05-18 17:58:33 +02:00
VP8_COMP *cpi = (VP8_COMP *) ptr;
VP8_COMMON *cm = &cpi->common;
struct vpx_usec_timer timer;
if (!cpi)
return -1;
#if HAVE_ARMV7
Add runtime CPU detection support for ARM. The primary goal is to allow a binary to be built which supports NEON, but can fall back to non-NEON routines, since some Android devices do not have NEON, even if they are otherwise ARMv7 (e.g., Tegra). The configure-generated flags HAVE_ARMV7, etc., are used to decide which versions of each function to build, and when CONFIG_RUNTIME_CPU_DETECT is enabled, the correct version is chosen at run time. In order for this to work, the CFLAGS must be set to something appropriate (e.g., without -mfpu=neon for ARMv7, and with appropriate -march and -mcpu for even earlier configurations), or the native C code will not be able to run. The ASFLAGS must remain set for the most advanced instruction set required at build time, since the ARM assembler will refuse to emit them otherwise. I have not attempted to make any changes to configure to do this automatically. Doing so will probably require the addition of new configure options. Many of the hooks for RTCD on ARM were already there, but a lot of the code had bit-rotted, and a good deal of the ARM-specific code is not integrated into the RTCD structs at all. I did not try to resolve the latter, merely to add the minimal amount of protection around them to allow RTCD to work. Those functions that were called based on an ifdef at the calling site were expanded to check the RTCD flags at that site, but they should be added to an RTCD struct somewhere in the future. The functions invoked with global function pointers still are, but these should be moved into an RTCD struct for thread safety (I believe every platform currently supported has atomic pointer stores, but this is not guaranteed). The encoder's boolhuff functions did not even have _c and armv7 suffixes, and the correct version was resolved at link time. The token packing functions did have appropriate suffixes, but the version was selected with a define, with no associated RTCD struct. However, for both of these, the only armv7 instruction they actually used was rbit, and this was completely superfluous, so I reworked them to avoid it. The only non-ARMv4 instruction remaining in them is clz, which is ARMv5 (not even ARMv5TE is required). Considering that there are no ARM-specific configs which are not at least ARMv5TE, I did not try to detect these at runtime, and simply enable them for ARMv5 and above. Finally, the NEON register saving code was completely non-reentrant, since it saved the registers to a global, static variable. I moved the storage for this onto the stack. A single binary built with this code was tested on an ARM11 (ARMv6) and a Cortex A8 (ARMv7 w/NEON), for both the encoder and decoder, and produced identical output, while using the correct accelerated functions on each. I did not test on any earlier processors. Change-Id: I45cbd63a614f4554c3b325c45d46c0806f009eaa
2010-10-21 00:39:11 +02:00
#if CONFIG_RUNTIME_CPU_DETECT
if (cm->rtcd.flags & HAS_NEON)
#endif
{
vp8_push_neon(store_reg);
}
Initial WebM release
2010-05-18 17:58:33 +02:00
#endif
vpx_usec_timer_start(&timer);
// no more room for frames;
if (cpi->source_buffer_count != 0 && cpi->source_buffer_count >= cpi->oxcf.lag_in_frames)
{
#if HAVE_ARMV7
Add runtime CPU detection support for ARM. The primary goal is to allow a binary to be built which supports NEON, but can fall back to non-NEON routines, since some Android devices do not have NEON, even if they are otherwise ARMv7 (e.g., Tegra). The configure-generated flags HAVE_ARMV7, etc., are used to decide which versions of each function to build, and when CONFIG_RUNTIME_CPU_DETECT is enabled, the correct version is chosen at run time. In order for this to work, the CFLAGS must be set to something appropriate (e.g., without -mfpu=neon for ARMv7, and with appropriate -march and -mcpu for even earlier configurations), or the native C code will not be able to run. The ASFLAGS must remain set for the most advanced instruction set required at build time, since the ARM assembler will refuse to emit them otherwise. I have not attempted to make any changes to configure to do this automatically. Doing so will probably require the addition of new configure options. Many of the hooks for RTCD on ARM were already there, but a lot of the code had bit-rotted, and a good deal of the ARM-specific code is not integrated into the RTCD structs at all. I did not try to resolve the latter, merely to add the minimal amount of protection around them to allow RTCD to work. Those functions that were called based on an ifdef at the calling site were expanded to check the RTCD flags at that site, but they should be added to an RTCD struct somewhere in the future. The functions invoked with global function pointers still are, but these should be moved into an RTCD struct for thread safety (I believe every platform currently supported has atomic pointer stores, but this is not guaranteed). The encoder's boolhuff functions did not even have _c and armv7 suffixes, and the correct version was resolved at link time. The token packing functions did have appropriate suffixes, but the version was selected with a define, with no associated RTCD struct. However, for both of these, the only armv7 instruction they actually used was rbit, and this was completely superfluous, so I reworked them to avoid it. The only non-ARMv4 instruction remaining in them is clz, which is ARMv5 (not even ARMv5TE is required). Considering that there are no ARM-specific configs which are not at least ARMv5TE, I did not try to detect these at runtime, and simply enable them for ARMv5 and above. Finally, the NEON register saving code was completely non-reentrant, since it saved the registers to a global, static variable. I moved the storage for this onto the stack. A single binary built with this code was tested on an ARM11 (ARMv6) and a Cortex A8 (ARMv7 w/NEON), for both the encoder and decoder, and produced identical output, while using the correct accelerated functions on each. I did not test on any earlier processors. Change-Id: I45cbd63a614f4554c3b325c45d46c0806f009eaa
2010-10-21 00:39:11 +02:00
#if CONFIG_RUNTIME_CPU_DETECT
if (cm->rtcd.flags & HAS_NEON)
#endif
{
vp8_pop_neon(store_reg);
}
Initial WebM release
2010-05-18 17:58:33 +02:00
#endif
return -1;
}
//printf("in-cpi->source_buffer_count: %d\n", cpi->source_buffer_count);
cm->clr_type = sd->clrtype;
// make a copy of the frame for use later...
#if !(CONFIG_REALTIME_ONLY)
if (cpi->oxcf.allow_lag)
{
int which_buffer = cpi->source_encode_index - 1;
SOURCE_SAMPLE *s;
if (which_buffer == -1)
which_buffer = cpi->oxcf.lag_in_frames - 1;
if (cpi->source_buffer_count < cpi->oxcf.lag_in_frames - 1)
which_buffer = cpi->source_buffer_count;
s = &cpi->src_buffer[which_buffer];
s->source_time_stamp = time_stamp;
s->source_end_time_stamp = end_time;
s->source_frame_flags = frame_flags;
vp8_yv12_copy_frame_ptr(sd, &s->source_buffer);
cpi->source_buffer_count ++;
}
else
#endif
{
SOURCE_SAMPLE *s;
s = &cpi->src_buffer[0];
s->source_end_time_stamp = end_time;
s->source_time_stamp = time_stamp;
s->source_frame_flags = frame_flags;
#if HAVE_ARMV7
Add runtime CPU detection support for ARM. The primary goal is to allow a binary to be built which supports NEON, but can fall back to non-NEON routines, since some Android devices do not have NEON, even if they are otherwise ARMv7 (e.g., Tegra). The configure-generated flags HAVE_ARMV7, etc., are used to decide which versions of each function to build, and when CONFIG_RUNTIME_CPU_DETECT is enabled, the correct version is chosen at run time. In order for this to work, the CFLAGS must be set to something appropriate (e.g., without -mfpu=neon for ARMv7, and with appropriate -march and -mcpu for even earlier configurations), or the native C code will not be able to run. The ASFLAGS must remain set for the most advanced instruction set required at build time, since the ARM assembler will refuse to emit them otherwise. I have not attempted to make any changes to configure to do this automatically. Doing so will probably require the addition of new configure options. Many of the hooks for RTCD on ARM were already there, but a lot of the code had bit-rotted, and a good deal of the ARM-specific code is not integrated into the RTCD structs at all. I did not try to resolve the latter, merely to add the minimal amount of protection around them to allow RTCD to work. Those functions that were called based on an ifdef at the calling site were expanded to check the RTCD flags at that site, but they should be added to an RTCD struct somewhere in the future. The functions invoked with global function pointers still are, but these should be moved into an RTCD struct for thread safety (I believe every platform currently supported has atomic pointer stores, but this is not guaranteed). The encoder's boolhuff functions did not even have _c and armv7 suffixes, and the correct version was resolved at link time. The token packing functions did have appropriate suffixes, but the version was selected with a define, with no associated RTCD struct. However, for both of these, the only armv7 instruction they actually used was rbit, and this was completely superfluous, so I reworked them to avoid it. The only non-ARMv4 instruction remaining in them is clz, which is ARMv5 (not even ARMv5TE is required). Considering that there are no ARM-specific configs which are not at least ARMv5TE, I did not try to detect these at runtime, and simply enable them for ARMv5 and above. Finally, the NEON register saving code was completely non-reentrant, since it saved the registers to a global, static variable. I moved the storage for this onto the stack. A single binary built with this code was tested on an ARM11 (ARMv6) and a Cortex A8 (ARMv7 w/NEON), for both the encoder and decoder, and produced identical output, while using the correct accelerated functions on each. I did not test on any earlier processors. Change-Id: I45cbd63a614f4554c3b325c45d46c0806f009eaa
2010-10-21 00:39:11 +02:00
#if CONFIG_RUNTIME_CPU_DETECT
if (cm->rtcd.flags & HAS_NEON)
#endif
{
vp8_yv12_copy_src_frame_func_neon(sd, &s->source_buffer);
}
#if CONFIG_RUNTIME_CPU_DETECT
else
#endif
#endif
#if !HAVE_ARMV7 || CONFIG_RUNTIME_CPU_DETECT
{
vp8_yv12_copy_frame_ptr(sd, &s->source_buffer);
}
Initial WebM release
2010-05-18 17:58:33 +02:00
#endif
cpi->source_buffer_count = 1;
}
vpx_usec_timer_mark(&timer);
cpi->time_receive_data += vpx_usec_timer_elapsed(&timer);
#if HAVE_ARMV7
Add runtime CPU detection support for ARM. The primary goal is to allow a binary to be built which supports NEON, but can fall back to non-NEON routines, since some Android devices do not have NEON, even if they are otherwise ARMv7 (e.g., Tegra). The configure-generated flags HAVE_ARMV7, etc., are used to decide which versions of each function to build, and when CONFIG_RUNTIME_CPU_DETECT is enabled, the correct version is chosen at run time. In order for this to work, the CFLAGS must be set to something appropriate (e.g., without -mfpu=neon for ARMv7, and with appropriate -march and -mcpu for even earlier configurations), or the native C code will not be able to run. The ASFLAGS must remain set for the most advanced instruction set required at build time, since the ARM assembler will refuse to emit them otherwise. I have not attempted to make any changes to configure to do this automatically. Doing so will probably require the addition of new configure options. Many of the hooks for RTCD on ARM were already there, but a lot of the code had bit-rotted, and a good deal of the ARM-specific code is not integrated into the RTCD structs at all. I did not try to resolve the latter, merely to add the minimal amount of protection around them to allow RTCD to work. Those functions that were called based on an ifdef at the calling site were expanded to check the RTCD flags at that site, but they should be added to an RTCD struct somewhere in the future. The functions invoked with global function pointers still are, but these should be moved into an RTCD struct for thread safety (I believe every platform currently supported has atomic pointer stores, but this is not guaranteed). The encoder's boolhuff functions did not even have _c and armv7 suffixes, and the correct version was resolved at link time. The token packing functions did have appropriate suffixes, but the version was selected with a define, with no associated RTCD struct. However, for both of these, the only armv7 instruction they actually used was rbit, and this was completely superfluous, so I reworked them to avoid it. The only non-ARMv4 instruction remaining in them is clz, which is ARMv5 (not even ARMv5TE is required). Considering that there are no ARM-specific configs which are not at least ARMv5TE, I did not try to detect these at runtime, and simply enable them for ARMv5 and above. Finally, the NEON register saving code was completely non-reentrant, since it saved the registers to a global, static variable. I moved the storage for this onto the stack. A single binary built with this code was tested on an ARM11 (ARMv6) and a Cortex A8 (ARMv7 w/NEON), for both the encoder and decoder, and produced identical output, while using the correct accelerated functions on each. I did not test on any earlier processors. Change-Id: I45cbd63a614f4554c3b325c45d46c0806f009eaa
2010-10-21 00:39:11 +02:00
#if CONFIG_RUNTIME_CPU_DETECT
if (cm->rtcd.flags & HAS_NEON)
#endif
{
vp8_pop_neon(store_reg);
}
Initial WebM release
2010-05-18 17:58:33 +02:00
#endif
return 0;
}
int vp8_get_compressed_data(VP8_PTR ptr, unsigned int *frame_flags, unsigned long *size, unsigned char *dest, INT64 *time_stamp, INT64 *time_end, int flush)
{
Add runtime CPU detection support for ARM. The primary goal is to allow a binary to be built which supports NEON, but can fall back to non-NEON routines, since some Android devices do not have NEON, even if they are otherwise ARMv7 (e.g., Tegra). The configure-generated flags HAVE_ARMV7, etc., are used to decide which versions of each function to build, and when CONFIG_RUNTIME_CPU_DETECT is enabled, the correct version is chosen at run time. In order for this to work, the CFLAGS must be set to something appropriate (e.g., without -mfpu=neon for ARMv7, and with appropriate -march and -mcpu for even earlier configurations), or the native C code will not be able to run. The ASFLAGS must remain set for the most advanced instruction set required at build time, since the ARM assembler will refuse to emit them otherwise. I have not attempted to make any changes to configure to do this automatically. Doing so will probably require the addition of new configure options. Many of the hooks for RTCD on ARM were already there, but a lot of the code had bit-rotted, and a good deal of the ARM-specific code is not integrated into the RTCD structs at all. I did not try to resolve the latter, merely to add the minimal amount of protection around them to allow RTCD to work. Those functions that were called based on an ifdef at the calling site were expanded to check the RTCD flags at that site, but they should be added to an RTCD struct somewhere in the future. The functions invoked with global function pointers still are, but these should be moved into an RTCD struct for thread safety (I believe every platform currently supported has atomic pointer stores, but this is not guaranteed). The encoder's boolhuff functions did not even have _c and armv7 suffixes, and the correct version was resolved at link time. The token packing functions did have appropriate suffixes, but the version was selected with a define, with no associated RTCD struct. However, for both of these, the only armv7 instruction they actually used was rbit, and this was completely superfluous, so I reworked them to avoid it. The only non-ARMv4 instruction remaining in them is clz, which is ARMv5 (not even ARMv5TE is required). Considering that there are no ARM-specific configs which are not at least ARMv5TE, I did not try to detect these at runtime, and simply enable them for ARMv5 and above. Finally, the NEON register saving code was completely non-reentrant, since it saved the registers to a global, static variable. I moved the storage for this onto the stack. A single binary built with this code was tested on an ARM11 (ARMv6) and a Cortex A8 (ARMv7 w/NEON), for both the encoder and decoder, and produced identical output, while using the correct accelerated functions on each. I did not test on any earlier processors. Change-Id: I45cbd63a614f4554c3b325c45d46c0806f009eaa
2010-10-21 00:39:11 +02:00
INT64 store_reg[8];
Initial WebM release
2010-05-18 17:58:33 +02:00
VP8_COMP *cpi = (VP8_COMP *) ptr;
VP8_COMMON *cm = &cpi->common;
struct vpx_usec_timer tsctimer;
struct vpx_usec_timer ticktimer;
struct vpx_usec_timer cmptimer;
if (!cpi)
return -1;
#if HAVE_ARMV7
Add runtime CPU detection support for ARM. The primary goal is to allow a binary to be built which supports NEON, but can fall back to non-NEON routines, since some Android devices do not have NEON, even if they are otherwise ARMv7 (e.g., Tegra). The configure-generated flags HAVE_ARMV7, etc., are used to decide which versions of each function to build, and when CONFIG_RUNTIME_CPU_DETECT is enabled, the correct version is chosen at run time. In order for this to work, the CFLAGS must be set to something appropriate (e.g., without -mfpu=neon for ARMv7, and with appropriate -march and -mcpu for even earlier configurations), or the native C code will not be able to run. The ASFLAGS must remain set for the most advanced instruction set required at build time, since the ARM assembler will refuse to emit them otherwise. I have not attempted to make any changes to configure to do this automatically. Doing so will probably require the addition of new configure options. Many of the hooks for RTCD on ARM were already there, but a lot of the code had bit-rotted, and a good deal of the ARM-specific code is not integrated into the RTCD structs at all. I did not try to resolve the latter, merely to add the minimal amount of protection around them to allow RTCD to work. Those functions that were called based on an ifdef at the calling site were expanded to check the RTCD flags at that site, but they should be added to an RTCD struct somewhere in the future. The functions invoked with global function pointers still are, but these should be moved into an RTCD struct for thread safety (I believe every platform currently supported has atomic pointer stores, but this is not guaranteed). The encoder's boolhuff functions did not even have _c and armv7 suffixes, and the correct version was resolved at link time. The token packing functions did have appropriate suffixes, but the version was selected with a define, with no associated RTCD struct. However, for both of these, the only armv7 instruction they actually used was rbit, and this was completely superfluous, so I reworked them to avoid it. The only non-ARMv4 instruction remaining in them is clz, which is ARMv5 (not even ARMv5TE is required). Considering that there are no ARM-specific configs which are not at least ARMv5TE, I did not try to detect these at runtime, and simply enable them for ARMv5 and above. Finally, the NEON register saving code was completely non-reentrant, since it saved the registers to a global, static variable. I moved the storage for this onto the stack. A single binary built with this code was tested on an ARM11 (ARMv6) and a Cortex A8 (ARMv7 w/NEON), for both the encoder and decoder, and produced identical output, while using the correct accelerated functions on each. I did not test on any earlier processors. Change-Id: I45cbd63a614f4554c3b325c45d46c0806f009eaa
2010-10-21 00:39:11 +02:00
#if CONFIG_RUNTIME_CPU_DETECT
if (cm->rtcd.flags & HAS_NEON)
#endif
{
vp8_push_neon(store_reg);
}
Initial WebM release
2010-05-18 17:58:33 +02:00
#endif
vpx_usec_timer_start(&cmptimer);
// flush variable tells us that even though we have less than 10 frames
// in our buffer we need to start producing compressed frames.
// Probably because we are at the end of a file....
if ((cpi->source_buffer_count == cpi->oxcf.lag_in_frames && cpi->oxcf.lag_in_frames > 0)
|| (!cpi->oxcf.allow_lag && cpi->source_buffer_count > 0)
|| (flush && cpi->source_buffer_count > 0))
{
SOURCE_SAMPLE *s;
s = &cpi->src_buffer[cpi->source_encode_index];
cpi->source_time_stamp = s->source_time_stamp;
cpi->source_end_time_stamp = s->source_end_time_stamp;
#if !(CONFIG_REALTIME_ONLY)
// Should we code an alternate reference frame
if (cpi->oxcf.error_resilient_mode == 0 &&
cpi->oxcf.play_alternate &&
cpi->source_alt_ref_pending &&
(cpi->frames_till_gf_update_due < cpi->source_buffer_count) &&
cpi->oxcf.lag_in_frames != 0)
{
cpi->last_alt_ref_sei = (cpi->source_encode_index + cpi->frames_till_gf_update_due) % cpi->oxcf.lag_in_frames;
#if VP8_TEMPORAL_ALT_REF
if (cpi->oxcf.arnr_max_frames > 0)
{
#if 0
// my attempt at a loop that tests the results of strength filter.
int start_frame = cpi->last_alt_ref_sei - 3;
int i, besti = -1, pastin = cpi->oxcf.arnr_strength;
int besterr;
if (start_frame < 0)
start_frame += cpi->oxcf.lag_in_frames;
besterr = vp8_calc_low_ss_err(&cpi->src_buffer[cpi->last_alt_ref_sei].source_buffer,
&cpi->src_buffer[start_frame].source_buffer, IF_RTCD(&cpi->rtcd.variance));
for (i = 0; i < 7; i++)
{
int thiserr;
cpi->oxcf.arnr_strength = i;
vp8cx_temp_filter_c(cpi);
thiserr = vp8_calc_low_ss_err(&cpi->alt_ref_buffer.source_buffer,
&cpi->src_buffer[start_frame].source_buffer, IF_RTCD(&cpi->rtcd.variance));
if (10 * thiserr < besterr * 8)
{
besterr = thiserr;
besti = i;
}
}
if (besti != -1)
{
cpi->oxcf.arnr_strength = besti;
vp8cx_temp_filter_c(cpi);
s = &cpi->alt_ref_buffer;
// FWG not sure if I need to copy this data for the Alt Ref frame
s->source_time_stamp = cpi->src_buffer[cpi->last_alt_ref_sei].source_time_stamp;
s->source_end_time_stamp = cpi->src_buffer[cpi->last_alt_ref_sei].source_end_time_stamp;
s->source_frame_flags = cpi->src_buffer[cpi->last_alt_ref_sei].source_frame_flags;
}
else
s = &cpi->src_buffer[cpi->last_alt_ref_sei];
#else
vp8cx_temp_filter_c(cpi);
s = &cpi->alt_ref_buffer;
// FWG not sure if I need to copy this data for the Alt Ref frame
s->source_time_stamp = cpi->src_buffer[cpi->last_alt_ref_sei].source_time_stamp;
s->source_end_time_stamp = cpi->src_buffer[cpi->last_alt_ref_sei].source_end_time_stamp;
s->source_frame_flags = cpi->src_buffer[cpi->last_alt_ref_sei].source_frame_flags;
#endif
}
else
#endif
s = &cpi->src_buffer[cpi->last_alt_ref_sei];
cm->frames_till_alt_ref_frame = cpi->frames_till_gf_update_due;
cm->refresh_alt_ref_frame = 1;
cm->refresh_golden_frame = 0;
cm->refresh_last_frame = 0;
cm->show_frame = 0;
cpi->source_alt_ref_pending = FALSE; // Clear Pending altf Ref flag.
cpi->is_src_frame_alt_ref = 0;
change to make use of more trellis quantization when a subsequent frame is encoded as an alt reference frame, it is unlikely that any mb in current frame will be used as reference for future frames, so we can enable quantization optimization even when the RD constant is slightly rate-biased. The change has an overall benefit between 0.1% to 0.2% bit savings on the test sets based on vpxssim scores. Change-Id: I9aa7bc5cd573ea84e3ee655d2834c18c4460ceea
2010-10-15 03:58:34 +02:00
cpi->is_next_src_alt_ref = 0;
Initial WebM release
2010-05-18 17:58:33 +02:00
}
else
#endif
{
cm->show_frame = 1;
#if !(CONFIG_REALTIME_ONLY)
if (cpi->oxcf.allow_lag)
{
if (cpi->source_encode_index == cpi->last_alt_ref_sei)
{
cpi->is_src_frame_alt_ref = 1;
cpi->last_alt_ref_sei = -1;
}
else
cpi->is_src_frame_alt_ref = 0;
cpi->source_encode_index = (cpi->source_encode_index + 1) % cpi->oxcf.lag_in_frames;
change to make use of more trellis quantization when a subsequent frame is encoded as an alt reference frame, it is unlikely that any mb in current frame will be used as reference for future frames, so we can enable quantization optimization even when the RD constant is slightly rate-biased. The change has an overall benefit between 0.1% to 0.2% bit savings on the test sets based on vpxssim scores. Change-Id: I9aa7bc5cd573ea84e3ee655d2834c18c4460ceea
2010-10-15 03:58:34 +02:00
if(cpi->source_encode_index == cpi->last_alt_ref_sei)
cpi->is_next_src_alt_ref = 1;
else
cpi->is_next_src_alt_ref = 0;
Initial WebM release
2010-05-18 17:58:33 +02:00
}
#endif
cpi->source_buffer_count--;
}
cpi->un_scaled_source = &s->source_buffer;
cpi->Source = &s->source_buffer;
cpi->source_frame_flags = s->source_frame_flags;
*time_stamp = cpi->source_time_stamp;
*time_end = cpi->source_end_time_stamp;
}
else
{
*size = 0;
#if !(CONFIG_REALTIME_ONLY)
if (flush && cpi->pass == 1 && !cpi->first_pass_done)
{
vp8_end_first_pass(cpi); /* get last stats packet */
cpi->first_pass_done = 1;
}
#endif
#if HAVE_ARMV7
Add runtime CPU detection support for ARM. The primary goal is to allow a binary to be built which supports NEON, but can fall back to non-NEON routines, since some Android devices do not have NEON, even if they are otherwise ARMv7 (e.g., Tegra). The configure-generated flags HAVE_ARMV7, etc., are used to decide which versions of each function to build, and when CONFIG_RUNTIME_CPU_DETECT is enabled, the correct version is chosen at run time. In order for this to work, the CFLAGS must be set to something appropriate (e.g., without -mfpu=neon for ARMv7, and with appropriate -march and -mcpu for even earlier configurations), or the native C code will not be able to run. The ASFLAGS must remain set for the most advanced instruction set required at build time, since the ARM assembler will refuse to emit them otherwise. I have not attempted to make any changes to configure to do this automatically. Doing so will probably require the addition of new configure options. Many of the hooks for RTCD on ARM were already there, but a lot of the code had bit-rotted, and a good deal of the ARM-specific code is not integrated into the RTCD structs at all. I did not try to resolve the latter, merely to add the minimal amount of protection around them to allow RTCD to work. Those functions that were called based on an ifdef at the calling site were expanded to check the RTCD flags at that site, but they should be added to an RTCD struct somewhere in the future. The functions invoked with global function pointers still are, but these should be moved into an RTCD struct for thread safety (I believe every platform currently supported has atomic pointer stores, but this is not guaranteed). The encoder's boolhuff functions did not even have _c and armv7 suffixes, and the correct version was resolved at link time. The token packing functions did have appropriate suffixes, but the version was selected with a define, with no associated RTCD struct. However, for both of these, the only armv7 instruction they actually used was rbit, and this was completely superfluous, so I reworked them to avoid it. The only non-ARMv4 instruction remaining in them is clz, which is ARMv5 (not even ARMv5TE is required). Considering that there are no ARM-specific configs which are not at least ARMv5TE, I did not try to detect these at runtime, and simply enable them for ARMv5 and above. Finally, the NEON register saving code was completely non-reentrant, since it saved the registers to a global, static variable. I moved the storage for this onto the stack. A single binary built with this code was tested on an ARM11 (ARMv6) and a Cortex A8 (ARMv7 w/NEON), for both the encoder and decoder, and produced identical output, while using the correct accelerated functions on each. I did not test on any earlier processors. Change-Id: I45cbd63a614f4554c3b325c45d46c0806f009eaa
2010-10-21 00:39:11 +02:00
#if CONFIG_RUNTIME_CPU_DETECT
if (cm->rtcd.flags & HAS_NEON)
#endif
{
vp8_pop_neon(store_reg);
}
Initial WebM release
2010-05-18 17:58:33 +02:00
#endif
return -1;
}
*frame_flags = cpi->source_frame_flags;
if (cpi->source_time_stamp < cpi->first_time_stamp_ever)
improve average framerate calculation Change Ice204e86 identified a problem with bitrate undershoot due to low precision in the timestamps passed to the library. This patch takes a different approach by calculating the duration of this frame and passing it to the library, rather than using a fixed duration and letting the library average it out with higher precision timestamps. This part of the fix only applies to vpxenc. This patch also attempts to fix the problem for generic applications that may have made the same mistake vpxenc did. Instead of calculating this frame's duration by the difference of this frame's and the last frame's start time, we use the end times instead. This allows the framerate calculation to scavenge "unclaimed" time from the last frame. For instance: start | end | calculated duration ======+=======+==================== 0ms 33ms 33ms 33ms 66ms 33ms 66ms 99ms 33ms 100ms 133ms 34ms Change-Id: I92be4b3518e0bd530e97f90e69e75330a4c413fc
2010-11-04 20:05:45 +01:00
{
Initial WebM release
2010-05-18 17:58:33 +02:00
cpi->first_time_stamp_ever = cpi->source_time_stamp;
improve average framerate calculation Change Ice204e86 identified a problem with bitrate undershoot due to low precision in the timestamps passed to the library. This patch takes a different approach by calculating the duration of this frame and passing it to the library, rather than using a fixed duration and letting the library average it out with higher precision timestamps. This part of the fix only applies to vpxenc. This patch also attempts to fix the problem for generic applications that may have made the same mistake vpxenc did. Instead of calculating this frame's duration by the difference of this frame's and the last frame's start time, we use the end times instead. This allows the framerate calculation to scavenge "unclaimed" time from the last frame. For instance: start | end | calculated duration ======+=======+==================== 0ms 33ms 33ms 33ms 66ms 33ms 66ms 99ms 33ms 100ms 133ms 34ms Change-Id: I92be4b3518e0bd530e97f90e69e75330a4c413fc
2010-11-04 20:05:45 +01:00
cpi->last_end_time_stamp_seen = cpi->source_time_stamp;
}
Initial WebM release
2010-05-18 17:58:33 +02:00
// adjust frame rates based on timestamps given
if (!cm->refresh_alt_ref_frame)
{
improve average framerate calculation Change Ice204e86 identified a problem with bitrate undershoot due to low precision in the timestamps passed to the library. This patch takes a different approach by calculating the duration of this frame and passing it to the library, rather than using a fixed duration and letting the library average it out with higher precision timestamps. This part of the fix only applies to vpxenc. This patch also attempts to fix the problem for generic applications that may have made the same mistake vpxenc did. Instead of calculating this frame's duration by the difference of this frame's and the last frame's start time, we use the end times instead. This allows the framerate calculation to scavenge "unclaimed" time from the last frame. For instance: start | end | calculated duration ======+=======+==================== 0ms 33ms 33ms 33ms 66ms 33ms 66ms 99ms 33ms 100ms 133ms 34ms Change-Id: I92be4b3518e0bd530e97f90e69e75330a4c413fc
2010-11-04 20:05:45 +01:00
if (cpi->source_time_stamp == cpi->first_time_stamp_ever)
Initial WebM release
2010-05-18 17:58:33 +02:00
{
double this_fps = 10000000.000 / (cpi->source_end_time_stamp - cpi->source_time_stamp);
vp8_new_frame_rate(cpi, this_fps);
}
else
{
improve average framerate calculation Change Ice204e86 identified a problem with bitrate undershoot due to low precision in the timestamps passed to the library. This patch takes a different approach by calculating the duration of this frame and passing it to the library, rather than using a fixed duration and letting the library average it out with higher precision timestamps. This part of the fix only applies to vpxenc. This patch also attempts to fix the problem for generic applications that may have made the same mistake vpxenc did. Instead of calculating this frame's duration by the difference of this frame's and the last frame's start time, we use the end times instead. This allows the framerate calculation to scavenge "unclaimed" time from the last frame. For instance: start | end | calculated duration ======+=======+==================== 0ms 33ms 33ms 33ms 66ms 33ms 66ms 99ms 33ms 100ms 133ms 34ms Change-Id: I92be4b3518e0bd530e97f90e69e75330a4c413fc
2010-11-04 20:05:45 +01:00
long long nanosecs = cpi->source_end_time_stamp
- cpi->last_end_time_stamp_seen;
Initial WebM release
2010-05-18 17:58:33 +02:00
double this_fps = 10000000.000 / nanosecs;
vp8_new_frame_rate(cpi, (7 * cpi->oxcf.frame_rate + this_fps) / 8);
}
cpi->last_time_stamp_seen = cpi->source_time_stamp;
improve average framerate calculation Change Ice204e86 identified a problem with bitrate undershoot due to low precision in the timestamps passed to the library. This patch takes a different approach by calculating the duration of this frame and passing it to the library, rather than using a fixed duration and letting the library average it out with higher precision timestamps. This part of the fix only applies to vpxenc. This patch also attempts to fix the problem for generic applications that may have made the same mistake vpxenc did. Instead of calculating this frame's duration by the difference of this frame's and the last frame's start time, we use the end times instead. This allows the framerate calculation to scavenge "unclaimed" time from the last frame. For instance: start | end | calculated duration ======+=======+==================== 0ms 33ms 33ms 33ms 66ms 33ms 66ms 99ms 33ms 100ms 133ms 34ms Change-Id: I92be4b3518e0bd530e97f90e69e75330a4c413fc
2010-11-04 20:05:45 +01:00
cpi->last_end_time_stamp_seen = cpi->source_end_time_stamp;
Initial WebM release
2010-05-18 17:58:33 +02:00
}
if (cpi->compressor_speed == 2)
{
vp8_check_gf_quality(cpi);
}
if (!cpi)
{
#if HAVE_ARMV7
Add runtime CPU detection support for ARM. The primary goal is to allow a binary to be built which supports NEON, but can fall back to non-NEON routines, since some Android devices do not have NEON, even if they are otherwise ARMv7 (e.g., Tegra). The configure-generated flags HAVE_ARMV7, etc., are used to decide which versions of each function to build, and when CONFIG_RUNTIME_CPU_DETECT is enabled, the correct version is chosen at run time. In order for this to work, the CFLAGS must be set to something appropriate (e.g., without -mfpu=neon for ARMv7, and with appropriate -march and -mcpu for even earlier configurations), or the native C code will not be able to run. The ASFLAGS must remain set for the most advanced instruction set required at build time, since the ARM assembler will refuse to emit them otherwise. I have not attempted to make any changes to configure to do this automatically. Doing so will probably require the addition of new configure options. Many of the hooks for RTCD on ARM were already there, but a lot of the code had bit-rotted, and a good deal of the ARM-specific code is not integrated into the RTCD structs at all. I did not try to resolve the latter, merely to add the minimal amount of protection around them to allow RTCD to work. Those functions that were called based on an ifdef at the calling site were expanded to check the RTCD flags at that site, but they should be added to an RTCD struct somewhere in the future. The functions invoked with global function pointers still are, but these should be moved into an RTCD struct for thread safety (I believe every platform currently supported has atomic pointer stores, but this is not guaranteed). The encoder's boolhuff functions did not even have _c and armv7 suffixes, and the correct version was resolved at link time. The token packing functions did have appropriate suffixes, but the version was selected with a define, with no associated RTCD struct. However, for both of these, the only armv7 instruction they actually used was rbit, and this was completely superfluous, so I reworked them to avoid it. The only non-ARMv4 instruction remaining in them is clz, which is ARMv5 (not even ARMv5TE is required). Considering that there are no ARM-specific configs which are not at least ARMv5TE, I did not try to detect these at runtime, and simply enable them for ARMv5 and above. Finally, the NEON register saving code was completely non-reentrant, since it saved the registers to a global, static variable. I moved the storage for this onto the stack. A single binary built with this code was tested on an ARM11 (ARMv6) and a Cortex A8 (ARMv7 w/NEON), for both the encoder and decoder, and produced identical output, while using the correct accelerated functions on each. I did not test on any earlier processors. Change-Id: I45cbd63a614f4554c3b325c45d46c0806f009eaa
2010-10-21 00:39:11 +02:00
#if CONFIG_RUNTIME_CPU_DETECT
if (cm->rtcd.flags & HAS_NEON)
#endif
{
vp8_pop_neon(store_reg);
}
Initial WebM release
2010-05-18 17:58:33 +02:00
#endif
return 0;
}
if (cpi->compressor_speed == 2)
{
vpx_usec_timer_start(&tsctimer);
vpx_usec_timer_start(&ticktimer);
}
// start with a 0 size frame
*size = 0;
// Clear down mmx registers
vp8_clear_system_state(); //__asm emms;
cm->frame_type = INTER_FRAME;
cm->frame_flags = *frame_flags;
#if 0
if (cm->refresh_alt_ref_frame)
{
//cm->refresh_golden_frame = 1;
cm->refresh_golden_frame = 0;
cm->refresh_last_frame = 0;
}
else
{
cm->refresh_golden_frame = 0;
cm->refresh_last_frame = 1;
}
#endif
#if !(CONFIG_REALTIME_ONLY)
if (cpi->pass == 1)
{
Pass1Encode(cpi, size, dest, frame_flags);
}
else if (cpi->pass == 2)
{
Pass2Encode(cpi, size, dest, frame_flags);
}
else
#endif
encode_frame_to_data_rate(cpi, size, dest, frame_flags);
if (cpi->compressor_speed == 2)
{
unsigned int duration, duration2;
vpx_usec_timer_mark(&tsctimer);
vpx_usec_timer_mark(&ticktimer);
duration = vpx_usec_timer_elapsed(&ticktimer);
duration2 = (unsigned int)((double)duration / 2);
if (cm->frame_type != KEY_FRAME)
{
if (cpi->avg_encode_time == 0)
cpi->avg_encode_time = duration;
else
cpi->avg_encode_time = (7 * cpi->avg_encode_time + duration) >> 3;
}
if (duration2)
{
//if(*frame_flags!=1)
{
if (cpi->avg_pick_mode_time == 0)
cpi->avg_pick_mode_time = duration2;
else
cpi->avg_pick_mode_time = (7 * cpi->avg_pick_mode_time + duration2) >> 3;
}
}
}
if (cm->refresh_entropy_probs == 0)
{
vpx_memcpy(&cm->fc, &cm->lfc, sizeof(cm->fc));
}
// if its a dropped frame honor the requests on subsequent frames
if (*size > 0)
{
// return to normal state
cm->refresh_entropy_probs = 1;
cm->refresh_alt_ref_frame = 0;
cm->refresh_golden_frame = 0;
cm->refresh_last_frame = 1;
cm->frame_type = INTER_FRAME;
}
cpi->ready_for_new_frame = 1;
vpx_usec_timer_mark(&cmptimer);
cpi->time_compress_data += vpx_usec_timer_elapsed(&cmptimer);
if (cpi->b_calculate_psnr && cpi->pass != 1 && cm->show_frame)
generate_psnr_packet(cpi);
#if CONFIG_PSNR
if (cpi->pass != 1)
{
cpi->bytes += *size;
if (cm->show_frame)
{
cpi->count ++;
if (cpi->b_calculate_psnr)
{
double y, u, v;
double sq_error;
double frame_psnr = vp8_calc_psnr(cpi->Source, cm->frame_to_show, &y, &u, &v, &sq_error);
cpi->total_y += y;
cpi->total_u += u;
cpi->total_v += v;
cpi->total_sq_error += sq_error;
cpi->total += frame_psnr;
{
double y2, u2, v2, frame_psnr2, frame_ssim2 = 0;
double weight = 0;
vp8_deblock(cm->frame_to_show, &cm->post_proc_buffer, cm->filter_level * 10 / 6, 1, 0, IF_RTCD(&cm->rtcd.postproc));
vp8_clear_system_state();
frame_psnr2 = vp8_calc_psnr(cpi->Source, &cm->post_proc_buffer, &y2, &u2, &v2, &sq_error);
frame_ssim2 = vp8_calc_ssim(cpi->Source, &cm->post_proc_buffer, 1, &weight);
cpi->summed_quality += frame_ssim2 * weight;
cpi->summed_weights += weight;
cpi->totalp_y += y2;
cpi->totalp_u += u2;
cpi->totalp_v += v2;
cpi->totalp += frame_psnr2;
cpi->total_sq_error2 += sq_error;
}
}
if (cpi->b_calculate_ssimg)
{
double y, u, v, frame_all;
frame_all = vp8_calc_ssimg(cpi->Source, cm->frame_to_show, &y, &u, &v);
cpi->total_ssimg_y += y;
cpi->total_ssimg_u += u;
cpi->total_ssimg_v += v;
cpi->total_ssimg_all += frame_all;
}
}
}
#if 0
if (cpi->common.frame_type != 0 && cpi->common.base_qindex == cpi->oxcf.worst_allowed_q)
{
skiptruecount += cpi->skip_true_count;
skipfalsecount += cpi->skip_false_count;
}
#endif
#if 0
if (cpi->pass != 1)
{
FILE *f = fopen("skip.stt", "a");
fprintf(f, "frame:%4d flags:%4x Q:%4d P:%4d Size:%5d\n", cpi->common.current_video_frame, *frame_flags, cpi->common.base_qindex, cpi->prob_skip_false, *size);
if (cpi->is_src_frame_alt_ref == 1)
fprintf(f, "skipcount: %4d framesize: %d\n", cpi->skip_true_count , *size);
fclose(f);
}
#endif
#endif
#if HAVE_ARMV7
Add runtime CPU detection support for ARM. The primary goal is to allow a binary to be built which supports NEON, but can fall back to non-NEON routines, since some Android devices do not have NEON, even if they are otherwise ARMv7 (e.g., Tegra). The configure-generated flags HAVE_ARMV7, etc., are used to decide which versions of each function to build, and when CONFIG_RUNTIME_CPU_DETECT is enabled, the correct version is chosen at run time. In order for this to work, the CFLAGS must be set to something appropriate (e.g., without -mfpu=neon for ARMv7, and with appropriate -march and -mcpu for even earlier configurations), or the native C code will not be able to run. The ASFLAGS must remain set for the most advanced instruction set required at build time, since the ARM assembler will refuse to emit them otherwise. I have not attempted to make any changes to configure to do this automatically. Doing so will probably require the addition of new configure options. Many of the hooks for RTCD on ARM were already there, but a lot of the code had bit-rotted, and a good deal of the ARM-specific code is not integrated into the RTCD structs at all. I did not try to resolve the latter, merely to add the minimal amount of protection around them to allow RTCD to work. Those functions that were called based on an ifdef at the calling site were expanded to check the RTCD flags at that site, but they should be added to an RTCD struct somewhere in the future. The functions invoked with global function pointers still are, but these should be moved into an RTCD struct for thread safety (I believe every platform currently supported has atomic pointer stores, but this is not guaranteed). The encoder's boolhuff functions did not even have _c and armv7 suffixes, and the correct version was resolved at link time. The token packing functions did have appropriate suffixes, but the version was selected with a define, with no associated RTCD struct. However, for both of these, the only armv7 instruction they actually used was rbit, and this was completely superfluous, so I reworked them to avoid it. The only non-ARMv4 instruction remaining in them is clz, which is ARMv5 (not even ARMv5TE is required). Considering that there are no ARM-specific configs which are not at least ARMv5TE, I did not try to detect these at runtime, and simply enable them for ARMv5 and above. Finally, the NEON register saving code was completely non-reentrant, since it saved the registers to a global, static variable. I moved the storage for this onto the stack. A single binary built with this code was tested on an ARM11 (ARMv6) and a Cortex A8 (ARMv7 w/NEON), for both the encoder and decoder, and produced identical output, while using the correct accelerated functions on each. I did not test on any earlier processors. Change-Id: I45cbd63a614f4554c3b325c45d46c0806f009eaa
2010-10-21 00:39:11 +02:00
#if CONFIG_RUNTIME_CPU_DETECT
if (cm->rtcd.flags & HAS_NEON)
#endif
{
vp8_pop_neon(store_reg);
}
Initial WebM release
2010-05-18 17:58:33 +02:00
#endif
return 0;
}
postproc : Re-work posproc calling to allow more flags. Debugging in postproc needs more flags to allow for specific block types to be turned on or off in the visualizations. Must be enabled with --enable-postproc-visualizer during configuration time. Change-Id: Ia74f357ddc3ad4fb8082afd3a64f62384e4fcb2d
2010-11-05 00:03:36 +01:00
int vp8_get_preview_raw_frame(VP8_PTR comp, YV12_BUFFER_CONFIG *dest, vp8_ppflags_t *flags)
Initial WebM release
2010-05-18 17:58:33 +02:00
{
VP8_COMP *cpi = (VP8_COMP *) comp;
if (cpi->common.refresh_alt_ref_frame)
return -1;
else
{
int ret;
#if CONFIG_POSTPROC
postproc : Re-work posproc calling to allow more flags. Debugging in postproc needs more flags to allow for specific block types to be turned on or off in the visualizations. Must be enabled with --enable-postproc-visualizer during configuration time. Change-Id: Ia74f357ddc3ad4fb8082afd3a64f62384e4fcb2d
2010-11-05 00:03:36 +01:00
ret = vp8_post_proc_frame(&cpi->common, dest, flags);
Initial WebM release
2010-05-18 17:58:33 +02:00
#else
if (cpi->common.frame_to_show)
{
*dest = *cpi->common.frame_to_show;
dest->y_width = cpi->common.Width;
dest->y_height = cpi->common.Height;
dest->uv_height = cpi->common.Height / 2;
ret = 0;
}
else
{
ret = -1;
}
#endif //!CONFIG_POSTPROC
vp8_clear_system_state();
return ret;
}
}
int vp8_set_roimap(VP8_PTR comp, unsigned char *map, unsigned int rows, unsigned int cols, int delta_q[4], int delta_lf[4], unsigned int threshold[4])
{
VP8_COMP *cpi = (VP8_COMP *) comp;
signed char feature_data[MB_LVL_MAX][MAX_MB_SEGMENTS];
if (cpi->common.mb_rows != rows || cpi->common.mb_cols != cols)
return -1;
if (!map)
{
disable_segmentation((VP8_PTR)cpi);
return 0;
}
// Set the segmentation Map
set_segmentation_map((VP8_PTR)cpi, map);
// Activate segmentation.
enable_segmentation((VP8_PTR)cpi);
// Set up the quant segment data
feature_data[MB_LVL_ALT_Q][0] = delta_q[0];
feature_data[MB_LVL_ALT_Q][1] = delta_q[1];
feature_data[MB_LVL_ALT_Q][2] = delta_q[2];
feature_data[MB_LVL_ALT_Q][3] = delta_q[3];
// Set up the loop segment data s
feature_data[MB_LVL_ALT_LF][0] = delta_lf[0];
feature_data[MB_LVL_ALT_LF][1] = delta_lf[1];
feature_data[MB_LVL_ALT_LF][2] = delta_lf[2];
feature_data[MB_LVL_ALT_LF][3] = delta_lf[3];
cpi->segment_encode_breakout[0] = threshold[0];
cpi->segment_encode_breakout[1] = threshold[1];
cpi->segment_encode_breakout[2] = threshold[2];
cpi->segment_encode_breakout[3] = threshold[3];
// Initialise the feature data structure
// SEGMENT_DELTADATA 0, SEGMENT_ABSDATA 1
set_segment_data((VP8_PTR)cpi, &feature_data[0][0], SEGMENT_DELTADATA);
return 0;
}
int vp8_set_active_map(VP8_PTR comp, unsigned char *map, unsigned int rows, unsigned int cols)
{
VP8_COMP *cpi = (VP8_COMP *) comp;
if (rows == cpi->common.mb_rows && cols == cpi->common.mb_cols)
{
if (map)
{
vpx_memcpy(cpi->active_map, map, rows * cols);
cpi->active_map_enabled = 1;
}
else
cpi->active_map_enabled = 0;
return 0;
}
else
{
//cpi->active_map_enabled = 0;
return -1 ;
}
}
int vp8_set_internal_size(VP8_PTR comp, VPX_SCALING horiz_mode, VPX_SCALING vert_mode)
{
VP8_COMP *cpi = (VP8_COMP *) comp;
remove warning was having: "vp8/encoder/onyx_if.c:5365: warning: comparison of unsigned expression >= 0 is always true"
2010-11-18 01:50:02 +01:00
if (horiz_mode <= ONETWO)
Initial WebM release
2010-05-18 17:58:33 +02:00
cpi->common.horiz_scale = horiz_mode;
else
return -1;
remove warning was having: "vp8/encoder/onyx_if.c:5365: warning: comparison of unsigned expression >= 0 is always true"
2010-11-18 01:50:02 +01:00
if (vert_mode <= ONETWO)
Initial WebM release
2010-05-18 17:58:33 +02:00
cpi->common.vert_scale = vert_mode;
else
return -1;
return 0;
}
int vp8_calc_ss_err(YV12_BUFFER_CONFIG *source, YV12_BUFFER_CONFIG *dest, const vp8_variance_rtcd_vtable_t *rtcd)
{
int i, j;
int Total = 0;
unsigned char *src = source->y_buffer;
unsigned char *dst = dest->y_buffer;
(void)rtcd;
// Loop through the Y plane raw and reconstruction data summing (square differences)
for (i = 0; i < source->y_height; i += 16)
{
for (j = 0; j < source->y_width; j += 16)
{
unsigned int sse;
Total += VARIANCE_INVOKE(rtcd, mse16x16)(src + j, source->y_stride, dst + j, dest->y_stride, &sse);
}
src += 16 * source->y_stride;
dst += 16 * dest->y_stride;
}
return Total;
}
int vp8_calc_low_ss_err(YV12_BUFFER_CONFIG *source, YV12_BUFFER_CONFIG *dest, const vp8_variance_rtcd_vtable_t *rtcd)
{
int i, j;
int Total = 0;
unsigned char *src = source->y_buffer;
unsigned char *dst = dest->y_buffer;
(void)rtcd;
// Loop through the Y plane raw and reconstruction data summing (square differences)
for (i = 0; i < source->y_height; i += 16)
{
for (j = 0; j < source->y_width; j += 16)
{
remove unneeded variables
2010-05-19 21:15:32 +02:00
unsigned int sse;
Initial WebM release
2010-05-18 17:58:33 +02:00
VARIANCE_INVOKE(rtcd, mse16x16)(src + j, source->y_stride, dst + j, dest->y_stride, &sse);
if (sse < 8096)
Total += sse;
}
src += 16 * source->y_stride;
dst += 16 * dest->y_stride;
}
return Total;
}
int vp8_get_speed(VP8_PTR c)
{
VP8_COMP *cpi = (VP8_COMP *) c;
return cpi->Speed;
}
int vp8_get_quantizer(VP8_PTR c)
{
VP8_COMP *cpi = (VP8_COMP *) c;
return cpi->common.base_qindex;
}
Reference in New Issue Copy Permalink