aom/vp9/encoder/vp9_onyx_if.c

4130 строки
130 KiB
C
Исходник Обычный вид История

2010-05-18 19:58:33 +04:00
/*
* Copyright (c) 2010 The WebM project authors. All Rights Reserved.
2010-05-18 19:58:33 +04:00
*
* Use of this source code is governed by a BSD-style license
* that can be found in the LICENSE file in the root of the source
* tree. An additional intellectual property rights grant can be found
* in the file PATENTS. All contributing project authors may
* be found in the AUTHORS file in the root of the source tree.
2010-05-18 19:58:33 +04:00
*/
#include "vpx_config.h"
#include "vp9/common/vp9_onyxc_int.h"
#include "vp9/common/vp9_reconinter.h"
#include "vp9/encoder/vp9_onyx_int.h"
#include "vp9/common/vp9_systemdependent.h"
#include "vp9/encoder/vp9_quantize.h"
#include "vp9/common/vp9_alloccommon.h"
#include "vp9/encoder/vp9_mcomp.h"
#include "vp9/encoder/vp9_firstpass.h"
#include "vp9/encoder/vp9_psnr.h"
#include "vpx_scale/vpx_scale.h"
#include "vp9/common/vp9_extend.h"
#include "vp9/encoder/vp9_ratectrl.h"
#include "vp9/common/vp9_quant_common.h"
#include "vp9/encoder/vp9_segmentation.h"
#include "./vp9_rtcd.h"
#include "./vpx_scale_rtcd.h"
#if CONFIG_POSTPROC
#include "vp9/common/vp9_postproc.h"
#endif
2010-05-18 19:58:33 +04:00
#include "vpx_mem/vpx_mem.h"
#include "vp9/common/vp9_swapyv12buffer.h"
2010-05-18 19:58:33 +04:00
#include "vpx_ports/vpx_timer.h"
2011-10-05 14:26:00 +04:00
#include "vp9/common/vp9_seg_common.h"
#include "vp9/encoder/vp9_mbgraph.h"
#include "vp9/common/vp9_pred_common.h"
#include "vp9/encoder/vp9_rdopt.h"
#include "vp9/encoder/vp9_bitstream.h"
#include "vp9/encoder/vp9_picklpf.h"
#include "vp9/common/vp9_mvref_common.h"
#include "vp9/encoder/vp9_temporal_filter.h"
2010-05-18 19:58:33 +04:00
#include <math.h>
#include <stdio.h>
#include <limits.h>
extern void print_tree_update_probs();
static void set_default_lf_deltas(VP9_COMP *cpi);
2010-05-18 19:58:33 +04:00
#define DEFAULT_INTERP_FILTER SWITCHABLE
#define SEARCH_BEST_FILTER 0 /* to search exhaustively for
best filter */
#define RESET_FOREACH_FILTER 0 /* whether to reset the encoder state
before trying each new filter */
#define SHARP_FILTER_QTHRESH 0 /* Q threshold for 8-tap sharp filter */
#define ALTREF_HIGH_PRECISION_MV 1 /* whether to use high precision mv
for altref computation */
#define HIGH_PRECISION_MV_QTHRESH 200 /* Q threshold for use of high precision
mv. Choose a very high value for
now so that HIGH_PRECISION is always
chosen */
#if CONFIG_INTERNAL_STATS
2010-05-18 19:58:33 +04:00
#include "math.h"
extern double vp9_calc_ssim(YV12_BUFFER_CONFIG *source,
YV12_BUFFER_CONFIG *dest, int lumamask,
double *weight);
2010-05-18 19:58:33 +04:00
extern double vp9_calc_ssimg(YV12_BUFFER_CONFIG *source,
YV12_BUFFER_CONFIG *dest, double *ssim_y,
double *ssim_u, double *ssim_v);
2010-05-18 19:58:33 +04:00
#endif
// #define OUTPUT_YUV_REC
2010-05-18 19:58:33 +04:00
#ifdef OUTPUT_YUV_SRC
FILE *yuv_file;
#endif
#ifdef OUTPUT_YUV_REC
FILE *yuv_rec_file;
#endif
2010-05-18 19:58:33 +04:00
#if 0
FILE *framepsnr;
FILE *kf_list;
2010-05-18 19:58:33 +04:00
FILE *keyfile;
#endif
#if 0
extern int skip_true_count;
extern int skip_false_count;
#endif
#ifdef ENTROPY_STATS
extern int intra_mode_stats[VP9_KF_BINTRAMODES]
[VP9_KF_BINTRAMODES]
[VP9_KF_BINTRAMODES];
2010-05-18 19:58:33 +04:00
#endif
#ifdef NMV_STATS
extern void init_nmvstats();
extern void print_nmvstats();
#endif
2010-05-18 19:58:33 +04:00
#ifdef SPEEDSTATS
unsigned int frames_at_speed[16] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0};
#endif
#if defined(SECTIONBITS_OUTPUT)
extern unsigned __int64 Sectionbits[500];
#endif
2010-05-18 19:58:33 +04:00
#ifdef MODE_STATS
extern int64_t Sectionbits[500];
extern unsigned int y_modes[VP9_YMODES];
extern unsigned int i8x8_modes[VP9_I8X8_MODES];
extern unsigned int uv_modes[VP9_UV_MODES];
extern unsigned int uv_modes_y[VP9_YMODES][VP9_UV_MODES];
extern unsigned int b_modes[B_MODE_COUNT];
extern unsigned int inter_y_modes[MB_MODE_COUNT];
extern unsigned int inter_uv_modes[VP9_UV_MODES];
extern unsigned int inter_b_modes[B_MODE_COUNT];
2010-05-18 19:58:33 +04:00
#endif
extern void vp9_init_quantizer(VP9_COMP *cpi);
static int base_skip_false_prob[QINDEX_RANGE][3];
// Tables relating active max Q to active min Q
static int kf_low_motion_minq[QINDEX_RANGE];
static int kf_high_motion_minq[QINDEX_RANGE];
static int gf_low_motion_minq[QINDEX_RANGE];
static int gf_high_motion_minq[QINDEX_RANGE];
static int inter_minq[QINDEX_RANGE];
// Functions to compute the active minq lookup table entries based on a
// formulaic approach to facilitate easier adjustment of the Q tables.
// The formulae were derived from computing a 3rd order polynomial best
// fit to the original data (after plotting real maxq vs minq (not q index))
static int calculate_minq_index(double maxq,
double x3, double x2, double x, double c) {
int i;
double minqtarget;
minqtarget = ((x3 * maxq * maxq * maxq) +
(x2 * maxq * maxq) +
(x * maxq) +
c);
if (minqtarget > maxq)
minqtarget = maxq;
for (i = 0; i < QINDEX_RANGE; i++) {
if (minqtarget <= vp9_convert_qindex_to_q(i))
return i;
}
return QINDEX_RANGE - 1;
}
static void init_minq_luts(void) {
int i;
double maxq;
for (i = 0; i < QINDEX_RANGE; i++) {
maxq = vp9_convert_qindex_to_q(i);
kf_low_motion_minq[i] = calculate_minq_index(maxq,
0.0000003,
-0.000015,
0.074,
0.0);
kf_high_motion_minq[i] = calculate_minq_index(maxq,
0.0000004,
-0.000125,
0.14,
0.0);
gf_low_motion_minq[i] = calculate_minq_index(maxq,
0.0000015,
-0.0009,
0.33,
0.0);
gf_high_motion_minq[i] = calculate_minq_index(maxq,
0.0000021,
-0.00125,
0.45,
0.0);
inter_minq[i] = calculate_minq_index(maxq,
0.00000271,
-0.00113,
0.697,
0.0);
}
}
static void set_mvcost(MACROBLOCK *mb) {
if (mb->e_mbd.allow_high_precision_mv) {
mb->mvcost = mb->nmvcost_hp;
mb->mvsadcost = mb->nmvsadcost_hp;
} else {
mb->mvcost = mb->nmvcost;
mb->mvsadcost = mb->nmvsadcost;
}
}
static void init_base_skip_probs(void) {
int i;
double q;
int t;
for (i = 0; i < QINDEX_RANGE; i++) {
q = vp9_convert_qindex_to_q(i);
// Exponential decay caluclation of baseline skip prob with clamping
// Based on crude best fit of old table.
t = (int)(564.25 * pow(2.71828, (-0.012 * q)));
base_skip_false_prob[i][1] = clip_prob(t);
base_skip_false_prob[i][2] = clip_prob(t * 3 / 4);
base_skip_false_prob[i][0] = clip_prob(t * 5 / 4);
}
}
static void update_base_skip_probs(VP9_COMP *cpi) {
VP9_COMMON *cm = &cpi->common;
if (cm->frame_type != KEY_FRAME) {
vp9_update_skip_probs(cpi);
if (cpi->refresh_alt_ref_frame) {
int k;
for (k = 0; k < MBSKIP_CONTEXTS; ++k)
cpi->last_skip_false_probs[2][k] = cm->mbskip_pred_probs[k];
cpi->last_skip_probs_q[2] = cm->base_qindex;
} else if (cpi->refresh_golden_frame) {
int k;
for (k = 0; k < MBSKIP_CONTEXTS; ++k)
cpi->last_skip_false_probs[1][k] = cm->mbskip_pred_probs[k];
cpi->last_skip_probs_q[1] = cm->base_qindex;
} else {
int k;
for (k = 0; k < MBSKIP_CONTEXTS; ++k)
cpi->last_skip_false_probs[0][k] = cm->mbskip_pred_probs[k];
cpi->last_skip_probs_q[0] = cm->base_qindex;
// update the baseline table for the current q
for (k = 0; k < MBSKIP_CONTEXTS; ++k)
cpi->base_skip_false_prob[cm->base_qindex][k] =
cm->mbskip_pred_probs[k];
}
}
}
void vp9_initialize_enc() {
static int init_done = 0;
if (!init_done) {
vp9_initialize_common();
vp9_tokenize_initialize();
vp9_init_quant_tables();
vp9_init_me_luts();
init_minq_luts();
init_base_skip_probs();
init_done = 1;
}
2010-05-18 19:58:33 +04:00
}
#ifdef PACKET_TESTING
extern FILE *vpxlogc;
#endif
static void setup_features(VP9_COMP *cpi) {
MACROBLOCKD *xd = &cpi->mb.e_mbd;
// Set up default state for MB feature flags
xd->segmentation_enabled = 0; // Default segmentation disabled
xd->update_mb_segmentation_map = 0;
xd->update_mb_segmentation_data = 0;
vpx_memset(xd->mb_segment_tree_probs, 255, sizeof(xd->mb_segment_tree_probs));
vp9_clearall_segfeatures(xd);
xd->mode_ref_lf_delta_enabled = 0;
xd->mode_ref_lf_delta_update = 0;
vpx_memset(xd->ref_lf_deltas, 0, sizeof(xd->ref_lf_deltas));
vpx_memset(xd->mode_lf_deltas, 0, sizeof(xd->mode_lf_deltas));
vpx_memset(xd->last_ref_lf_deltas, 0, sizeof(xd->ref_lf_deltas));
vpx_memset(xd->last_mode_lf_deltas, 0, sizeof(xd->mode_lf_deltas));
2010-05-18 19:58:33 +04:00
set_default_lf_deltas(cpi);
2010-05-18 19:58:33 +04:00
}
static void dealloc_compressor_data(VP9_COMP *cpi) {
vpx_free(cpi->tplist);
cpi->tplist = NULL;
// Delete last frame MV storage buffers
vpx_free(cpi->lfmv);
cpi->lfmv = 0;
vpx_free(cpi->lf_ref_frame_sign_bias);
cpi->lf_ref_frame_sign_bias = 0;
vpx_free(cpi->lf_ref_frame);
cpi->lf_ref_frame = 0;
2010-05-18 19:58:33 +04:00
// Delete sementation map
vpx_free(cpi->segmentation_map);
cpi->segmentation_map = 0;
vpx_free(cpi->common.last_frame_seg_map);
cpi->common.last_frame_seg_map = 0;
vpx_free(cpi->coding_context.last_frame_seg_map_copy);
cpi->coding_context.last_frame_seg_map_copy = 0;
2010-05-18 19:58:33 +04:00
vpx_free(cpi->active_map);
cpi->active_map = 0;
2010-05-18 19:58:33 +04:00
vp9_de_alloc_frame_buffers(&cpi->common);
2010-05-18 19:58:33 +04:00
vp8_yv12_de_alloc_frame_buffer(&cpi->last_frame_uf);
vp8_yv12_de_alloc_frame_buffer(&cpi->scaled_source);
#if VP9_TEMPORAL_ALT_REF
vp8_yv12_de_alloc_frame_buffer(&cpi->alt_ref_buffer);
2010-05-18 19:58:33 +04:00
#endif
vp9_lookahead_destroy(cpi->lookahead);
2010-05-18 19:58:33 +04:00
vpx_free(cpi->tok);
cpi->tok = 0;
2010-05-18 19:58:33 +04:00
// Structure used to monitor GF usage
vpx_free(cpi->gf_active_flags);
cpi->gf_active_flags = 0;
// Activity mask based per mb zbin adjustments
vpx_free(cpi->mb_activity_map);
cpi->mb_activity_map = 0;
vpx_free(cpi->mb_norm_activity_map);
cpi->mb_norm_activity_map = 0;
vpx_free(cpi->mb.pip);
cpi->mb.pip = 0;
vpx_free(cpi->twopass.total_stats);
cpi->twopass.total_stats = 0;
vpx_free(cpi->twopass.total_left_stats);
cpi->twopass.total_left_stats = 0;
vpx_free(cpi->twopass.this_frame_stats);
cpi->twopass.this_frame_stats = 0;
2010-05-18 19:58:33 +04:00
}
// Computes a q delta (in "q index" terms) to get from a starting q value
// to a target value
// target q value
static int compute_qdelta(VP9_COMP *cpi, double qstart, double qtarget) {
int i;
int start_index = cpi->worst_quality;
int target_index = cpi->worst_quality;
// Convert the average q value to an index.
for (i = cpi->best_quality; i < cpi->worst_quality; i++) {
start_index = i;
if (vp9_convert_qindex_to_q(i) >= qstart)
break;
}
// Convert the q target to an index
for (i = cpi->best_quality; i < cpi->worst_quality; i++) {
target_index = i;
if (vp9_convert_qindex_to_q(i) >= qtarget)
break;
}
return target_index - start_index;
}
static void init_seg_features(VP9_COMP *cpi) {
VP9_COMMON *cm = &cpi->common;
MACROBLOCKD *xd = &cpi->mb.e_mbd;
int high_q = (int)(cpi->avg_q > 48.0);
int qi_delta;
// Disable and clear down for KF
if (cm->frame_type == KEY_FRAME) {
// Clear down the global segmentation map
vpx_memset(cpi->segmentation_map, 0, (cm->mb_rows * cm->mb_cols));
xd->update_mb_segmentation_map = 0;
xd->update_mb_segmentation_data = 0;
cpi->static_mb_pct = 0;
// Disable segmentation
vp9_disable_segmentation((VP9_PTR)cpi);
// Clear down the segment features.
vp9_clearall_segfeatures(xd);
} else if (cpi->refresh_alt_ref_frame) {
// If this is an alt ref frame
// Clear down the global segmentation map
vpx_memset(cpi->segmentation_map, 0, (cm->mb_rows * cm->mb_cols));
xd->update_mb_segmentation_map = 0;
xd->update_mb_segmentation_data = 0;
cpi->static_mb_pct = 0;
// Disable segmentation and individual segment features by default
vp9_disable_segmentation((VP9_PTR)cpi);
vp9_clearall_segfeatures(xd);
// Scan frames from current to arf frame.
// This function re-enables segmentation if appropriate.
vp9_update_mbgraph_stats(cpi);
// If segmentation was enabled set those features needed for the
// arf itself.
if (xd->segmentation_enabled) {
xd->update_mb_segmentation_map = 1;
xd->update_mb_segmentation_data = 1;
qi_delta = compute_qdelta(cpi, cpi->avg_q, (cpi->avg_q * 0.875));
vp9_set_segdata(xd, 1, SEG_LVL_ALT_Q, (qi_delta - 2));
vp9_set_segdata(xd, 1, SEG_LVL_ALT_LF, -2);
vp9_enable_segfeature(xd, 1, SEG_LVL_ALT_Q);
vp9_enable_segfeature(xd, 1, SEG_LVL_ALT_LF);
2011-10-05 14:26:00 +04:00
// Where relevant assume segment data is delta data
xd->mb_segment_abs_delta = SEGMENT_DELTADATA;
}
}
// All other frames if segmentation has been enabled
else if (xd->segmentation_enabled) {
// First normal frame in a valid gf or alt ref group
if (cpi->common.frames_since_golden == 0) {
// Set up segment features for normal frames in an arf group
if (cpi->source_alt_ref_active) {
xd->update_mb_segmentation_map = 0;
xd->update_mb_segmentation_data = 1;
xd->mb_segment_abs_delta = SEGMENT_DELTADATA;
qi_delta = compute_qdelta(cpi, cpi->avg_q,
(cpi->avg_q * 1.125));
vp9_set_segdata(xd, 1, SEG_LVL_ALT_Q, (qi_delta + 2));
vp9_set_segdata(xd, 1, SEG_LVL_ALT_Q, 0);
vp9_enable_segfeature(xd, 1, SEG_LVL_ALT_Q);
vp9_set_segdata(xd, 1, SEG_LVL_ALT_LF, -2);
vp9_enable_segfeature(xd, 1, SEG_LVL_ALT_LF);
// Segment coding disabled for compred testing
if (high_q || (cpi->static_mb_pct == 100)) {
vp9_set_segref(xd, 1, ALTREF_FRAME);
vp9_enable_segfeature(xd, 1, SEG_LVL_REF_FRAME);
vp9_enable_segfeature(xd, 1, SEG_LVL_SKIP);
}
}
// Disable segmentation and clear down features if alt ref
// is not active for this group
else {
vp9_disable_segmentation((VP9_PTR)cpi);
vpx_memset(cpi->segmentation_map, 0,
(cm->mb_rows * cm->mb_cols));
xd->update_mb_segmentation_map = 0;
xd->update_mb_segmentation_data = 0;
vp9_clearall_segfeatures(xd);
}
}
// Special case where we are coding over the top of a previous
// alt ref frame.
// Segment coding disabled for compred testing
else if (cpi->is_src_frame_alt_ref) {
// Enable ref frame features for segment 0 as well
vp9_enable_segfeature(xd, 0, SEG_LVL_REF_FRAME);
vp9_enable_segfeature(xd, 1, SEG_LVL_REF_FRAME);
// All mbs should use ALTREF_FRAME
vp9_clear_segref(xd, 0);
vp9_set_segref(xd, 0, ALTREF_FRAME);
vp9_clear_segref(xd, 1);
vp9_set_segref(xd, 1, ALTREF_FRAME);
// Skip all MBs if high Q (0,0 mv and skip coeffs)
if (high_q) {
vp9_enable_segfeature(xd, 0, SEG_LVL_SKIP);
vp9_enable_segfeature(xd, 1, SEG_LVL_SKIP);
}
// Enable data udpate
xd->update_mb_segmentation_data = 1;
}
// All other frames.
else {
// No updates.. leave things as they are.
xd->update_mb_segmentation_map = 0;
xd->update_mb_segmentation_data = 0;
}
}
}
// DEBUG: Print out the segment id of each MB in the current frame.
static void print_seg_map(VP9_COMP *cpi) {
VP9_COMMON *cm = &cpi->common;
int row, col;
int map_index = 0;
FILE *statsfile;
statsfile = fopen("segmap.stt", "a");
fprintf(statsfile, "%10d\n",
cm->current_video_frame);
for (row = 0; row < cpi->common.mb_rows; row++) {
for (col = 0; col < cpi->common.mb_cols; col++) {
fprintf(statsfile, "%10d",
cpi->segmentation_map[map_index]);
map_index++;
}
fprintf(statsfile, "\n");
}
fprintf(statsfile, "\n");
fclose(statsfile);
}
static void update_reference_segmentation_map(VP9_COMP *cpi) {
VP9_COMMON *const cm = &cpi->common;
int row, col;
MODE_INFO *mi, *mi_ptr = cm->mi;
uint8_t *cache_ptr = cm->last_frame_seg_map, *cache;
for (row = 0; row < cm->mb_rows; row++) {
mi = mi_ptr;
cache = cache_ptr;
for (col = 0; col < cm->mb_cols; col++, mi++, cache++) {
cache[0] = mi->mbmi.segment_id;
}
mi_ptr += cm->mode_info_stride;
cache_ptr += cm->mb_cols;
}
}
static void set_default_lf_deltas(VP9_COMP *cpi) {
cpi->mb.e_mbd.mode_ref_lf_delta_enabled = 1;
cpi->mb.e_mbd.mode_ref_lf_delta_update = 1;
2010-05-18 19:58:33 +04:00
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));
2010-05-18 19:58:33 +04:00
// 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;
2010-05-18 19:58:33 +04:00
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
2010-05-18 19:58:33 +04:00
}
static void set_rd_speed_thresholds(VP9_COMP *cpi, int mode, int speed) {
SPEED_FEATURES *sf = &cpi->sf;
int speed_multiplier = speed + 1;
int i;
// Set baseline threshold values
for (i = 0; i < MAX_MODES; ++i) {
sf->thresh_mult[i] = (mode == 0) ? -500 : 0;
}
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 ] += speed_multiplier * 1000;
sf->thresh_mult[THR_NEARG ] += speed_multiplier * 1000;
sf->thresh_mult[THR_NEARA ] += speed_multiplier * 1000;
sf->thresh_mult[THR_DC ] = 0;
sf->thresh_mult[THR_TM ] += speed_multiplier * 1000;
sf->thresh_mult[THR_V_PRED ] += speed_multiplier * 1000;
sf->thresh_mult[THR_H_PRED ] += speed_multiplier * 1000;
sf->thresh_mult[THR_D45_PRED ] += speed_multiplier * 1000;
sf->thresh_mult[THR_D135_PRED] += speed_multiplier * 1000;
sf->thresh_mult[THR_D117_PRED] += speed_multiplier * 1000;
sf->thresh_mult[THR_D153_PRED] += speed_multiplier * 1000;
sf->thresh_mult[THR_D27_PRED ] += speed_multiplier * 1000;
sf->thresh_mult[THR_D63_PRED ] += speed_multiplier * 1000;
sf->thresh_mult[THR_B_PRED ] += speed_multiplier * 2500;
sf->thresh_mult[THR_I8X8_PRED] += speed_multiplier * 2500;
sf->thresh_mult[THR_NEWMV ] += speed_multiplier * 1000;
sf->thresh_mult[THR_NEWG ] += speed_multiplier * 1000;
sf->thresh_mult[THR_NEWA ] += speed_multiplier * 1000;
sf->thresh_mult[THR_SPLITMV ] += speed_multiplier * 2500;
sf->thresh_mult[THR_SPLITG ] += speed_multiplier * 2500;
sf->thresh_mult[THR_SPLITA ] += speed_multiplier * 2500;
sf->thresh_mult[THR_COMP_ZEROLG ] += speed_multiplier * 1000;
sf->thresh_mult[THR_COMP_ZEROLA ] += speed_multiplier * 1000;
sf->thresh_mult[THR_COMP_ZEROGA ] += speed_multiplier * 1000;
sf->thresh_mult[THR_COMP_NEARESTLG] += speed_multiplier * 1000;
sf->thresh_mult[THR_COMP_NEARESTLA] += speed_multiplier * 1000;
sf->thresh_mult[THR_COMP_NEARESTGA] += speed_multiplier * 1000;
sf->thresh_mult[THR_COMP_NEARLG ] += speed_multiplier * 1000;
sf->thresh_mult[THR_COMP_NEARLA ] += speed_multiplier * 1000;
sf->thresh_mult[THR_COMP_NEARGA ] += speed_multiplier * 1000;
sf->thresh_mult[THR_COMP_NEWLG ] += speed_multiplier * 2000;
sf->thresh_mult[THR_COMP_NEWLA ] += speed_multiplier * 2000;
sf->thresh_mult[THR_COMP_NEWGA ] += speed_multiplier * 2000;
sf->thresh_mult[THR_COMP_SPLITLA ] += speed_multiplier * 4500;
sf->thresh_mult[THR_COMP_SPLITGA ] += speed_multiplier * 4500;
sf->thresh_mult[THR_COMP_SPLITLG ] += speed_multiplier * 4500;
#if CONFIG_COMP_INTERINTRA_PRED
sf->thresh_mult[THR_COMP_INTERINTRA_ZEROL ] += speed_multiplier * 1000;
sf->thresh_mult[THR_COMP_INTERINTRA_ZEROG ] += speed_multiplier * 1000;
sf->thresh_mult[THR_COMP_INTERINTRA_ZEROA ] += speed_multiplier * 1000;
sf->thresh_mult[THR_COMP_INTERINTRA_NEARESTL] += speed_multiplier * 1000;
sf->thresh_mult[THR_COMP_INTERINTRA_NEARESTG] += speed_multiplier * 1000;
sf->thresh_mult[THR_COMP_INTERINTRA_NEARESTA] += speed_multiplier * 1000;
sf->thresh_mult[THR_COMP_INTERINTRA_NEARL ] += speed_multiplier * 1000;
sf->thresh_mult[THR_COMP_INTERINTRA_NEARG ] += speed_multiplier * 1000;
sf->thresh_mult[THR_COMP_INTERINTRA_NEARA ] += speed_multiplier * 1000;
sf->thresh_mult[THR_COMP_INTERINTRA_NEWL ] += speed_multiplier * 2000;
sf->thresh_mult[THR_COMP_INTERINTRA_NEWG ] += speed_multiplier * 2000;
sf->thresh_mult[THR_COMP_INTERINTRA_NEWA ] += speed_multiplier * 2000;
#endif
}
void vp9_set_speed_features(VP9_COMP *cpi) {
SPEED_FEATURES *sf = &cpi->sf;
int mode = cpi->compressor_speed;
int speed = cpi->Speed;
int i;
VP9_COMMON *cm = &cpi->common;
// Only modes 0 and 1 supported for now in experimental code basae
if (mode > 1)
mode = 1;
// 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;
}
// best quality defaults
sf->RD = 1;
sf->search_method = NSTEP;
sf->improved_dct = 1;
sf->auto_filter = 1;
sf->recode_loop = 1;
sf->quarter_pixel_search = 1;
sf->half_pixel_search = 1;
sf->iterative_sub_pixel = 1;
Add lossless compression mode. This commit adds lossless compression capability to the experimental branch. The lossless experiment can be enabled using --enable-lossless in configure. When the experiment is enabled, the encoder will use lossless compression mode by command line option --lossless, and the decoder automatically recognizes a losslessly encoded clip and decodes accordingly. To achieve the lossless coding, this commit has changed the following: 1. To encode at lossless mode, encoder forces the use of unit quantizer, i.e, Q 0, where effective quantization is 1. Encoder also disables the usage of 8x8 transform and allows only 4x4 transform; 2. At Q 0, the first order 4x4 DCT/IDCT have been switched over to a pair of forward and inverse Walsh-Hadamard Transform (http://goo.gl/EIsfy), with proper scaling applied to match the range of the original 4x4 DCT/IDCT pair; 3. At Q 0, the second order remains to use the previous walsh-hadamard transform pair. However, to maintain the reversibility in second order transform at Q 0, scaling down is applied to first order DC coefficients prior to forward transform, and scaling up is applied to the second order output prior to quantization. Symmetric upscaling and downscaling are added around inverse second order transform; 4. At lossless mode, encoder also disables a number of minor features to ensure no loss is introduced, these features includes: a. Trellis quantization optimization b. Loop filtering c. Aggressive zero-binning, rounding and zero-bin boosting d. Mode based zero-bin boosting Lossless coding test was performed on all clips within the derf set, to verify that the commit has achieved lossless compression for all clips. The average compression ratio is around 2.57 to 1. (http://goo.gl/dEShs) Change-Id: Ia3aba7dd09df40dd590f93b9aba134defbc64e34
2012-06-14 06:03:31 +04:00
#if CONFIG_LOSSLESS
sf->optimize_coefficients = 0;
Add lossless compression mode. This commit adds lossless compression capability to the experimental branch. The lossless experiment can be enabled using --enable-lossless in configure. When the experiment is enabled, the encoder will use lossless compression mode by command line option --lossless, and the decoder automatically recognizes a losslessly encoded clip and decodes accordingly. To achieve the lossless coding, this commit has changed the following: 1. To encode at lossless mode, encoder forces the use of unit quantizer, i.e, Q 0, where effective quantization is 1. Encoder also disables the usage of 8x8 transform and allows only 4x4 transform; 2. At Q 0, the first order 4x4 DCT/IDCT have been switched over to a pair of forward and inverse Walsh-Hadamard Transform (http://goo.gl/EIsfy), with proper scaling applied to match the range of the original 4x4 DCT/IDCT pair; 3. At Q 0, the second order remains to use the previous walsh-hadamard transform pair. However, to maintain the reversibility in second order transform at Q 0, scaling down is applied to first order DC coefficients prior to forward transform, and scaling up is applied to the second order output prior to quantization. Symmetric upscaling and downscaling are added around inverse second order transform; 4. At lossless mode, encoder also disables a number of minor features to ensure no loss is introduced, these features includes: a. Trellis quantization optimization b. Loop filtering c. Aggressive zero-binning, rounding and zero-bin boosting d. Mode based zero-bin boosting Lossless coding test was performed on all clips within the derf set, to verify that the commit has achieved lossless compression for all clips. The average compression ratio is around 2.57 to 1. (http://goo.gl/dEShs) Change-Id: Ia3aba7dd09df40dd590f93b9aba134defbc64e34
2012-06-14 06:03:31 +04:00
#else
sf->optimize_coefficients = 1;
Add lossless compression mode. This commit adds lossless compression capability to the experimental branch. The lossless experiment can be enabled using --enable-lossless in configure. When the experiment is enabled, the encoder will use lossless compression mode by command line option --lossless, and the decoder automatically recognizes a losslessly encoded clip and decodes accordingly. To achieve the lossless coding, this commit has changed the following: 1. To encode at lossless mode, encoder forces the use of unit quantizer, i.e, Q 0, where effective quantization is 1. Encoder also disables the usage of 8x8 transform and allows only 4x4 transform; 2. At Q 0, the first order 4x4 DCT/IDCT have been switched over to a pair of forward and inverse Walsh-Hadamard Transform (http://goo.gl/EIsfy), with proper scaling applied to match the range of the original 4x4 DCT/IDCT pair; 3. At Q 0, the second order remains to use the previous walsh-hadamard transform pair. However, to maintain the reversibility in second order transform at Q 0, scaling down is applied to first order DC coefficients prior to forward transform, and scaling up is applied to the second order output prior to quantization. Symmetric upscaling and downscaling are added around inverse second order transform; 4. At lossless mode, encoder also disables a number of minor features to ensure no loss is introduced, these features includes: a. Trellis quantization optimization b. Loop filtering c. Aggressive zero-binning, rounding and zero-bin boosting d. Mode based zero-bin boosting Lossless coding test was performed on all clips within the derf set, to verify that the commit has achieved lossless compression for all clips. The average compression ratio is around 2.57 to 1. (http://goo.gl/dEShs) Change-Id: Ia3aba7dd09df40dd590f93b9aba134defbc64e34
2012-06-14 06:03:31 +04:00
#endif
sf->no_skip_block4x4_search = 1;
sf->first_step = 0;
sf->max_step_search_steps = MAX_MVSEARCH_STEPS;
sf->splitmode_breakout = 0;
sf->mb16_breakout = 0;
2010-05-18 19:58:33 +04:00
// Set rd thresholds based on mode and speed setting
set_rd_speed_thresholds(cpi, mode, speed);
2010-05-18 19:58:33 +04:00
switch (mode) {
2010-05-18 19:58:33 +04:00
case 0: // best quality mode
sf->search_best_filter = SEARCH_BEST_FILTER;
break;
case 1:
sf->splitmode_breakout = 1;
sf->mb16_breakout = 0;
if (speed > 0) {
/* Disable coefficient optimization above speed 0 */
sf->optimize_coefficients = 0;
sf->no_skip_block4x4_search = 0;
sf->first_step = 1;
cpi->mode_check_freq[THR_SPLITG] = 2;
cpi->mode_check_freq[THR_SPLITA] = 2;
cpi->mode_check_freq[THR_SPLITMV] = 0;
cpi->mode_check_freq[THR_COMP_SPLITGA] = 2;
cpi->mode_check_freq[THR_COMP_SPLITLG] = 2;
cpi->mode_check_freq[THR_COMP_SPLITLA] = 0;
}
2010-05-18 19:58:33 +04:00
if (speed > 1) {
cpi->mode_check_freq[THR_SPLITG] = 4;
cpi->mode_check_freq[THR_SPLITA] = 4;
cpi->mode_check_freq[THR_SPLITMV] = 2;
2010-05-18 19:58:33 +04:00
cpi->mode_check_freq[THR_COMP_SPLITGA] = 4;
cpi->mode_check_freq[THR_COMP_SPLITLG] = 4;
cpi->mode_check_freq[THR_COMP_SPLITLA] = 2;
}
if (speed > 2) {
cpi->mode_check_freq[THR_SPLITG] = 15;
cpi->mode_check_freq[THR_SPLITA] = 15;
cpi->mode_check_freq[THR_SPLITMV] = 7;
cpi->mode_check_freq[THR_COMP_SPLITGA] = 15;
cpi->mode_check_freq[THR_COMP_SPLITLG] = 15;
cpi->mode_check_freq[THR_COMP_SPLITLA] = 7;
sf->improved_dct = 0;
// Only do recode loop on key frames, golden frames and
// alt ref frames
sf->recode_loop = 2;
}
2010-05-18 19:58:33 +04:00
break;
2010-05-18 19:58:33 +04:00
}; /* switch */
/* disable frame modes if flags not set */
if (!(cpi->ref_frame_flags & VP9_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 & VP9_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;
#if CONFIG_COMP_INTERINTRA_PRED
sf->thresh_mult[THR_COMP_INTERINTRA_ZEROG ] = INT_MAX;
sf->thresh_mult[THR_COMP_INTERINTRA_NEARESTG] = INT_MAX;
sf->thresh_mult[THR_COMP_INTERINTRA_NEARG ] = INT_MAX;
sf->thresh_mult[THR_COMP_INTERINTRA_NEWG ] = INT_MAX;
#endif
sf->thresh_mult[THR_SPLITG ] = INT_MAX;
}
if (!(cpi->ref_frame_flags & VP9_ALT_FLAG)) {
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;
#if CONFIG_COMP_INTERINTRA_PRED
sf->thresh_mult[THR_COMP_INTERINTRA_ZEROA ] = INT_MAX;
sf->thresh_mult[THR_COMP_INTERINTRA_NEARESTA] = INT_MAX;
sf->thresh_mult[THR_COMP_INTERINTRA_NEARA ] = INT_MAX;
sf->thresh_mult[THR_COMP_INTERINTRA_NEWA ] = INT_MAX;
#endif
sf->thresh_mult[THR_SPLITA ] = INT_MAX;
}
if ((cpi->ref_frame_flags & (VP9_LAST_FLAG | VP9_GOLD_FLAG)) != (VP9_LAST_FLAG | VP9_GOLD_FLAG)) {
sf->thresh_mult[THR_COMP_ZEROLG ] = INT_MAX;
sf->thresh_mult[THR_COMP_NEARESTLG] = INT_MAX;
sf->thresh_mult[THR_COMP_NEARLG ] = INT_MAX;
sf->thresh_mult[THR_COMP_NEWLG ] = INT_MAX;
sf->thresh_mult[THR_COMP_SPLITLG ] = INT_MAX;
}
if ((cpi->ref_frame_flags & (VP9_LAST_FLAG | VP9_ALT_FLAG)) != (VP9_LAST_FLAG | VP9_ALT_FLAG)) {
sf->thresh_mult[THR_COMP_ZEROLA ] = INT_MAX;
sf->thresh_mult[THR_COMP_NEARESTLA] = INT_MAX;
sf->thresh_mult[THR_COMP_NEARLA ] = INT_MAX;
sf->thresh_mult[THR_COMP_NEWLA ] = INT_MAX;
sf->thresh_mult[THR_COMP_SPLITLA ] = INT_MAX;
}
if ((cpi->ref_frame_flags & (VP9_GOLD_FLAG | VP9_ALT_FLAG)) != (VP9_GOLD_FLAG | VP9_ALT_FLAG)) {
sf->thresh_mult[THR_COMP_ZEROGA ] = INT_MAX;
sf->thresh_mult[THR_COMP_NEARESTGA] = INT_MAX;
sf->thresh_mult[THR_COMP_NEARGA ] = INT_MAX;
sf->thresh_mult[THR_COMP_NEWGA ] = INT_MAX;
sf->thresh_mult[THR_COMP_SPLITGA ] = INT_MAX;
}
#if CONFIG_COMP_INTERINTRA_PRED
if ((cpi->ref_frame_flags & VP9_LAST_FLAG) != VP9_LAST_FLAG) {
sf->thresh_mult[THR_COMP_INTERINTRA_ZEROL ] = INT_MAX;
sf->thresh_mult[THR_COMP_INTERINTRA_NEARESTL] = INT_MAX;
sf->thresh_mult[THR_COMP_INTERINTRA_NEARL ] = INT_MAX;
sf->thresh_mult[THR_COMP_INTERINTRA_NEWL ] = INT_MAX;
}
#endif
// Slow quant, dct and trellis not worthwhile for first pass
// so make sure they are always turned off.
if (cpi->pass == 1) {
sf->optimize_coefficients = 0;
sf->improved_dct = 0;
}
{
int y_stride = cm->yv12_fb[cm->active_ref_idx[cpi->lst_fb_idx]].y_stride;
if (cpi->sf.search_method == NSTEP) {
vp9_init3smotion_compensation(&cpi->mb, y_stride);
} else if (cpi->sf.search_method == DIAMOND) {
vp9_init_dsmotion_compensation(&cpi->mb, y_stride);
}
}
2010-05-18 19:58:33 +04:00
cpi->mb.vp9_short_fdct16x16 = vp9_short_fdct16x16;
cpi->mb.vp9_short_fdct8x8 = vp9_short_fdct8x8;
cpi->mb.vp9_short_fdct8x4 = vp9_short_fdct8x4;
cpi->mb.vp9_short_fdct4x4 = vp9_short_fdct4x4;
cpi->mb.short_walsh4x4 = vp9_short_walsh4x4;
cpi->mb.short_fhaar2x2 = vp9_short_fhaar2x2;
#if CONFIG_LOSSLESS
if (cpi->oxcf.lossless) {
cpi->mb.vp9_short_fdct8x4 = vp9_short_walsh8x4_x8;
cpi->mb.vp9_short_fdct4x4 = vp9_short_walsh4x4_x8;
cpi->mb.short_walsh4x4 = vp9_short_walsh4x4;
cpi->mb.short_fhaar2x2 = vp9_short_fhaar2x2;
cpi->mb.short_walsh4x4 = vp9_short_walsh4x4_lossless;
}
#endif
2010-05-18 19:58:33 +04:00
cpi->mb.quantize_b_4x4 = vp9_regular_quantize_b_4x4;
cpi->mb.quantize_b_4x4_pair = vp9_regular_quantize_b_4x4_pair;
cpi->mb.quantize_b_8x8 = vp9_regular_quantize_b_8x8;
cpi->mb.quantize_b_16x16 = vp9_regular_quantize_b_16x16;
cpi->mb.quantize_b_2x2 = vp9_regular_quantize_b_2x2;
vp9_init_quantizer(cpi);
2010-05-18 19:58:33 +04:00
if (cpi->sf.iterative_sub_pixel == 1) {
cpi->find_fractional_mv_step = vp9_find_best_sub_pixel_step_iteratively;
} else if (cpi->sf.quarter_pixel_search) {
cpi->find_fractional_mv_step = vp9_find_best_sub_pixel_step;
} else if (cpi->sf.half_pixel_search) {
cpi->find_fractional_mv_step = vp9_find_best_half_pixel_step;
}
2010-05-18 19:58:33 +04:00
if (cpi->sf.optimize_coefficients == 1 && cpi->pass != 1)
cpi->mb.optimize = 1;
else
cpi->mb.optimize = 0;
2010-05-18 19:58:33 +04:00
#ifdef SPEEDSTATS
frames_at_speed[cpi->Speed]++;
2010-05-18 19:58:33 +04:00
#endif
}
static void alloc_raw_frame_buffers(VP9_COMP *cpi) {
int width = (cpi->oxcf.Width + 15) & ~15;
int height = (cpi->oxcf.Height + 15) & ~15;
2010-05-18 19:58:33 +04:00
cpi->lookahead = vp9_lookahead_init(cpi->oxcf.Width, cpi->oxcf.Height,
cpi->oxcf.lag_in_frames);
if (!cpi->lookahead)
vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
"Failed to allocate lag buffers");
2010-05-18 19:58:33 +04:00
#if VP9_TEMPORAL_ALT_REF
2010-05-18 19:58:33 +04:00
if (vp8_yv12_alloc_frame_buffer(&cpi->alt_ref_buffer,
width, height, VP9BORDERINPIXELS))
vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
"Failed to allocate altref buffer");
2010-05-18 19:58:33 +04:00
#endif
}
static int alloc_partition_data(VP9_COMP *cpi) {
vpx_free(cpi->mb.pip);
cpi->mb.pip = vpx_calloc((cpi->common.mb_cols + 1) *
(cpi->common.mb_rows + 1),
sizeof(PARTITION_INFO));
if (!cpi->mb.pip)
return 1;
cpi->mb.pi = cpi->mb.pip + cpi->common.mode_info_stride + 1;
return 0;
}
void vp9_alloc_compressor_data(VP9_COMP *cpi) {
VP9_COMMON *cm = &cpi->common;
2010-05-18 19:58:33 +04:00
int width = cm->Width;
int height = cm->Height;
2010-05-18 19:58:33 +04:00
if (vp9_alloc_frame_buffers(cm, width, height))
vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
"Failed to allocate frame buffers");
2010-05-18 19:58:33 +04:00
if (alloc_partition_data(cpi))
vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
"Failed to allocate partition data");
if ((width & 0xf) != 0)
width += 16 - (width & 0xf);
2010-05-18 19:58:33 +04:00
if ((height & 0xf) != 0)
height += 16 - (height & 0xf);
2010-05-18 19:58:33 +04:00
if (vp8_yv12_alloc_frame_buffer(&cpi->last_frame_uf,
width, height, VP9BORDERINPIXELS))
vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
"Failed to allocate last frame buffer");
2010-05-18 19:58:33 +04:00
if (vp8_yv12_alloc_frame_buffer(&cpi->scaled_source,
width, height, VP9BORDERINPIXELS))
vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
"Failed to allocate scaled source buffer");
2010-05-18 19:58:33 +04:00
vpx_free(cpi->tok);
2010-05-18 19:58:33 +04:00
{
unsigned int tokens = cm->mb_rows * cm->mb_cols * 24 * 16;
2010-05-18 19:58:33 +04:00
CHECK_MEM_ERROR(cpi->tok, vpx_calloc(tokens, sizeof(*cpi->tok)));
}
2010-05-18 19:58:33 +04:00
// 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;
// Structures used to minitor GF usage
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;
vpx_free(cpi->mb_activity_map);
CHECK_MEM_ERROR(cpi->mb_activity_map,
vpx_calloc(sizeof(unsigned int),
cm->mb_rows * cm->mb_cols));
vpx_free(cpi->mb_norm_activity_map);
CHECK_MEM_ERROR(cpi->mb_norm_activity_map,
vpx_calloc(sizeof(unsigned int),
cm->mb_rows * cm->mb_cols));
vpx_free(cpi->twopass.total_stats);
cpi->twopass.total_stats = vpx_calloc(1, sizeof(FIRSTPASS_STATS));
vpx_free(cpi->twopass.total_left_stats);
cpi->twopass.total_left_stats = vpx_calloc(1, sizeof(FIRSTPASS_STATS));
vpx_free(cpi->twopass.this_frame_stats);
cpi->twopass.this_frame_stats = vpx_calloc(1, sizeof(FIRSTPASS_STATS));
if (!cpi->twopass.total_stats ||
!cpi->twopass.total_left_stats ||
!cpi->twopass.this_frame_stats)
vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
"Failed to allocate firstpass stats");
vpx_free(cpi->tplist);
CHECK_MEM_ERROR(cpi->tplist,
vpx_malloc(sizeof(TOKENLIST) * (cpi->common.mb_rows)));
2010-05-18 19:58:33 +04:00
}
static void update_frame_size(VP9_COMP *cpi) {
VP9_COMMON *cm = &cpi->common;
/* our internal buffers are always multiples of 16 */
int width = (cm->Width + 15) & ~15;
int height = (cm->Height + 15) & ~15;
cm->mb_rows = height >> 4;
cm->mb_cols = width >> 4;
cm->MBs = cm->mb_rows * cm->mb_cols;
cm->mode_info_stride = cm->mb_cols + 1;
memset(cm->mip, 0,
(cm->mb_cols + 1) * (cm->mb_rows + 1) * sizeof(MODE_INFO));
vp9_update_mode_info_border(cm, cm->mip);
cm->mi = cm->mip + cm->mode_info_stride + 1;
cm->prev_mi = cm->prev_mip + cm->mode_info_stride + 1;
vp9_update_mode_info_in_image(cm, cm->mi);
/* Update size of buffers local to this frame */
if (vp8_yv12_realloc_frame_buffer(&cpi->last_frame_uf,
width, height, VP9BORDERINPIXELS))
vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
"Failed to reallocate last frame buffer");
if (vp8_yv12_realloc_frame_buffer(&cpi->scaled_source,
width, height, VP9BORDERINPIXELS))
vpx_internal_error(&cpi->common.error, VPX_CODEC_MEM_ERROR,
"Failed to reallocate scaled source buffer");
}
// TODO perhaps change number of steps expose to outside world when setting
// max and min limits. Also this will likely want refining for the extended Q
// range.
//
// Table that converts 0-63 Q range values passed in outside to the Qindex
// range used internally.
static const int q_trans[] = {
0, 4, 8, 12, 16, 20, 24, 28,
32, 36, 40, 44, 48, 52, 56, 60,
64, 68, 72, 76, 80, 84, 88, 92,
96, 100, 104, 108, 112, 116, 120, 124,
128, 132, 136, 140, 144, 148, 152, 156,
160, 164, 168, 172, 176, 180, 184, 188,
192, 196, 200, 204, 208, 212, 216, 220,
224, 228, 232, 236, 240, 244, 249, 255,
};
2010-05-18 19:58:33 +04:00
int vp9_reverse_trans(int x) {
int i;
2010-05-18 19:58:33 +04:00
for (i = 0; i < 64; i++)
if (q_trans[i] >= x)
return i;
2010-05-18 19:58:33 +04:00
return 63;
2010-05-18 19:58:33 +04:00
};
void vp9_new_frame_rate(VP9_COMP *cpi, double framerate) {
if (framerate < .1)
framerate = 30;
2010-10-15 00:19:06 +04:00
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);
2010-05-18 19:58:33 +04:00
if (cpi->min_frame_bandwidth < FRAME_OVERHEAD_BITS)
cpi->min_frame_bandwidth = FRAME_OVERHEAD_BITS;
// Set Maximum gf/arf interval
cpi->max_gf_interval = ((int)(cpi->output_frame_rate / 2.0) + 2);
2010-05-18 19:58:33 +04:00
if (cpi->max_gf_interval < 12)
cpi->max_gf_interval = 12;
2010-05-18 19:58:33 +04:00
// Extended interval for genuinely static scenes
cpi->twopass.static_scene_max_gf_interval = cpi->key_frame_frequency >> 1;
2010-05-18 19:58:33 +04:00
// Special conditions when altr ref frame enabled in lagged compress mode
if (cpi->oxcf.play_alternate && cpi->oxcf.lag_in_frames) {
if (cpi->max_gf_interval > cpi->oxcf.lag_in_frames - 1)
cpi->max_gf_interval = cpi->oxcf.lag_in_frames - 1;
if (cpi->twopass.static_scene_max_gf_interval > cpi->oxcf.lag_in_frames - 1)
cpi->twopass.static_scene_max_gf_interval = cpi->oxcf.lag_in_frames - 1;
}
if (cpi->max_gf_interval > cpi->twopass.static_scene_max_gf_interval)
cpi->max_gf_interval = cpi->twopass.static_scene_max_gf_interval;
2010-05-18 19:58:33 +04:00
}
static int
rescale(int val, int num, int denom) {
int64_t llnum = num;
int64_t llden = denom;
int64_t llval = val;
return (int)(llval * llnum / llden);
}
static void init_config(VP9_PTR ptr, VP9_CONFIG *oxcf) {
VP9_COMP *cpi = (VP9_COMP *)(ptr);
[WIP] Add column-based tiling. This patch adds column-based tiling. The idea is to make each tile independently decodable (after reading the common frame header) and also independendly encodable (minus within-frame cost adjustments in the RD loop) to speed-up hardware & software en/decoders if they used multi-threading. Column-based tiling has the added advantage (over other tiling methods) that it minimizes realtime use-case latency, since all threads can start encoding data as soon as the first SB-row worth of data is available to the encoder. There is some test code that does random tile ordering in the decoder, to confirm that each tile is indeed independently decodable from other tiles in the same frame. At tile edges, all contexts assume default values (i.e. 0, 0 motion vector, no coefficients, DC intra4x4 mode), and motion vector search and ordering do not cross tiles in the same frame. t log Tile independence is not maintained between frames ATM, i.e. tile 0 of frame 1 is free to use motion vectors that point into any tile of frame 0. We support 1 (i.e. no tiling), 2 or 4 column-tiles. The loopfilter crosses tile boundaries. I discussed this briefly with Aki and he says that's OK. An in-loop loopfilter would need to do some sync between tile threads, but that shouldn't be a big issue. Resuls: with tiling disabled, we go up slightly because of improved edge use in the intra4x4 prediction. With 2 tiles, we lose about ~1% on derf, ~0.35% on HD and ~0.55% on STD/HD. With 4 tiles, we lose another ~1.5% on derf ~0.77% on HD and ~0.85% on STD/HD. Most of this loss is concentrated in the low-bitrate end of clips, and most of it is because of the loss of edges at tile boundaries and the resulting loss of intra predictors. TODO: - more tiles (perhaps allow row-based tiling also, and max. 8 tiles)? - maybe optionally (for EC purposes), motion vectors themselves should not cross tile edges, or we should emulate such borders as if they were off-frame, to limit error propagation to within one tile only. This doesn't have to be the default behaviour but could be an optional bitstream flag. Change-Id: I5951c3a0742a767b20bc9fb5af685d9892c2c96f
2013-02-01 21:35:28 +04:00
VP9_COMMON *const cm = &cpi->common;
2010-05-18 19:58:33 +04:00
cpi->oxcf = *oxcf;
2010-05-18 19:58:33 +04:00
cpi->goldfreq = 7;
2010-05-18 19:58:33 +04:00
cm->version = oxcf->Version;
vp9_setup_version(cm);
2010-05-18 19:58:33 +04:00
// change includes all joint functionality
vp9_change_config(ptr, oxcf);
2010-05-18 19:58:33 +04:00
// Initialize active best and worst q and average q values.
cpi->active_worst_quality = cpi->oxcf.worst_allowed_q;
cpi->active_best_quality = cpi->oxcf.best_allowed_q;
cpi->avg_frame_qindex = cpi->oxcf.worst_allowed_q;
2010-05-18 19:58:33 +04:00
// Initialise the starting buffer levels
cpi->buffer_level = cpi->oxcf.starting_buffer_level;
cpi->bits_off_target = cpi->oxcf.starting_buffer_level;
2010-05-18 19:58:33 +04:00
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;
2010-05-18 19:58:33 +04:00
cpi->total_actual_bits = 0;
cpi->total_target_vs_actual = 0;
2010-05-18 19:58:33 +04:00
cpi->static_mb_pct = 0;
cpi->lst_fb_idx = 0;
cpi->gld_fb_idx = 1;
cpi->alt_fb_idx = 2;
[WIP] Add column-based tiling. This patch adds column-based tiling. The idea is to make each tile independently decodable (after reading the common frame header) and also independendly encodable (minus within-frame cost adjustments in the RD loop) to speed-up hardware & software en/decoders if they used multi-threading. Column-based tiling has the added advantage (over other tiling methods) that it minimizes realtime use-case latency, since all threads can start encoding data as soon as the first SB-row worth of data is available to the encoder. There is some test code that does random tile ordering in the decoder, to confirm that each tile is indeed independently decodable from other tiles in the same frame. At tile edges, all contexts assume default values (i.e. 0, 0 motion vector, no coefficients, DC intra4x4 mode), and motion vector search and ordering do not cross tiles in the same frame. t log Tile independence is not maintained between frames ATM, i.e. tile 0 of frame 1 is free to use motion vectors that point into any tile of frame 0. We support 1 (i.e. no tiling), 2 or 4 column-tiles. The loopfilter crosses tile boundaries. I discussed this briefly with Aki and he says that's OK. An in-loop loopfilter would need to do some sync between tile threads, but that shouldn't be a big issue. Resuls: with tiling disabled, we go up slightly because of improved edge use in the intra4x4 prediction. With 2 tiles, we lose about ~1% on derf, ~0.35% on HD and ~0.55% on STD/HD. With 4 tiles, we lose another ~1.5% on derf ~0.77% on HD and ~0.85% on STD/HD. Most of this loss is concentrated in the low-bitrate end of clips, and most of it is because of the loss of edges at tile boundaries and the resulting loss of intra predictors. TODO: - more tiles (perhaps allow row-based tiling also, and max. 8 tiles)? - maybe optionally (for EC purposes), motion vectors themselves should not cross tile edges, or we should emulate such borders as if they were off-frame, to limit error propagation to within one tile only. This doesn't have to be the default behaviour but could be an optional bitstream flag. Change-Id: I5951c3a0742a767b20bc9fb5af685d9892c2c96f
2013-02-01 21:35:28 +04:00
cm->tile_columns = 1 << cpi->oxcf.tile_columns;
#if VP9_TEMPORAL_ALT_REF
{
int i;
2010-05-18 19:58:33 +04:00
cpi->fixed_divide[0] = 0;
2010-05-18 19:58:33 +04:00
for (i = 1; i < 512; i++)
cpi->fixed_divide[i] = 0x80000 / i;
}
2010-05-18 19:58:33 +04:00
#endif
}
void vp9_change_config(VP9_PTR ptr, VP9_CONFIG *oxcf) {
VP9_COMP *cpi = (VP9_COMP *)(ptr);
[WIP] Add column-based tiling. This patch adds column-based tiling. The idea is to make each tile independently decodable (after reading the common frame header) and also independendly encodable (minus within-frame cost adjustments in the RD loop) to speed-up hardware & software en/decoders if they used multi-threading. Column-based tiling has the added advantage (over other tiling methods) that it minimizes realtime use-case latency, since all threads can start encoding data as soon as the first SB-row worth of data is available to the encoder. There is some test code that does random tile ordering in the decoder, to confirm that each tile is indeed independently decodable from other tiles in the same frame. At tile edges, all contexts assume default values (i.e. 0, 0 motion vector, no coefficients, DC intra4x4 mode), and motion vector search and ordering do not cross tiles in the same frame. t log Tile independence is not maintained between frames ATM, i.e. tile 0 of frame 1 is free to use motion vectors that point into any tile of frame 0. We support 1 (i.e. no tiling), 2 or 4 column-tiles. The loopfilter crosses tile boundaries. I discussed this briefly with Aki and he says that's OK. An in-loop loopfilter would need to do some sync between tile threads, but that shouldn't be a big issue. Resuls: with tiling disabled, we go up slightly because of improved edge use in the intra4x4 prediction. With 2 tiles, we lose about ~1% on derf, ~0.35% on HD and ~0.55% on STD/HD. With 4 tiles, we lose another ~1.5% on derf ~0.77% on HD and ~0.85% on STD/HD. Most of this loss is concentrated in the low-bitrate end of clips, and most of it is because of the loss of edges at tile boundaries and the resulting loss of intra predictors. TODO: - more tiles (perhaps allow row-based tiling also, and max. 8 tiles)? - maybe optionally (for EC purposes), motion vectors themselves should not cross tile edges, or we should emulate such borders as if they were off-frame, to limit error propagation to within one tile only. This doesn't have to be the default behaviour but could be an optional bitstream flag. Change-Id: I5951c3a0742a767b20bc9fb5af685d9892c2c96f
2013-02-01 21:35:28 +04:00
VP9_COMMON *const cm = &cpi->common;
2010-05-18 19:58:33 +04:00
if (!cpi)
return;
2010-05-18 19:58:33 +04:00
if (!oxcf)
return;
2010-05-18 19:58:33 +04:00
if (cm->version != oxcf->Version) {
cm->version = oxcf->Version;
vp9_setup_version(cm);
}
2010-05-18 19:58:33 +04:00
cpi->oxcf = *oxcf;
2010-05-18 19:58:33 +04:00
switch (cpi->oxcf.Mode) {
// Real time and one pass deprecated in test code base
2010-05-18 19:58:33 +04:00
case MODE_FIRSTPASS:
cpi->pass = 1;
cpi->compressor_speed = 1;
break;
2010-05-18 19:58:33 +04:00
case MODE_SECONDPASS:
cpi->pass = 2;
cpi->compressor_speed = 1;
2010-05-18 19:58:33 +04:00
if (cpi->oxcf.cpu_used < -5) {
cpi->oxcf.cpu_used = -5;
}
2010-05-18 19:58:33 +04:00
if (cpi->oxcf.cpu_used > 5)
cpi->oxcf.cpu_used = 5;
break;
2010-05-18 19:58:33 +04:00
case MODE_SECONDPASS_BEST:
cpi->pass = 2;
cpi->compressor_speed = 0;
break;
}
2010-05-18 19:58:33 +04:00
cpi->oxcf.worst_allowed_q = q_trans[oxcf->worst_allowed_q];
cpi->oxcf.best_allowed_q = q_trans[oxcf->best_allowed_q];
cpi->oxcf.cq_level = q_trans[cpi->oxcf.cq_level];
2010-05-18 19:58:33 +04:00
cpi->mb.e_mbd.inv_xform4x4_1_x8 = vp9_short_idct4x4llm_1;
cpi->mb.e_mbd.inv_xform4x4_x8 = vp9_short_idct4x4llm;
cpi->mb.e_mbd.inv_walsh4x4_1 = vp9_short_inv_walsh4x4_1;
cpi->mb.e_mbd.inv_walsh4x4_lossless = vp9_short_inv_walsh4x4;
Add lossless compression mode. This commit adds lossless compression capability to the experimental branch. The lossless experiment can be enabled using --enable-lossless in configure. When the experiment is enabled, the encoder will use lossless compression mode by command line option --lossless, and the decoder automatically recognizes a losslessly encoded clip and decodes accordingly. To achieve the lossless coding, this commit has changed the following: 1. To encode at lossless mode, encoder forces the use of unit quantizer, i.e, Q 0, where effective quantization is 1. Encoder also disables the usage of 8x8 transform and allows only 4x4 transform; 2. At Q 0, the first order 4x4 DCT/IDCT have been switched over to a pair of forward and inverse Walsh-Hadamard Transform (http://goo.gl/EIsfy), with proper scaling applied to match the range of the original 4x4 DCT/IDCT pair; 3. At Q 0, the second order remains to use the previous walsh-hadamard transform pair. However, to maintain the reversibility in second order transform at Q 0, scaling down is applied to first order DC coefficients prior to forward transform, and scaling up is applied to the second order output prior to quantization. Symmetric upscaling and downscaling are added around inverse second order transform; 4. At lossless mode, encoder also disables a number of minor features to ensure no loss is introduced, these features includes: a. Trellis quantization optimization b. Loop filtering c. Aggressive zero-binning, rounding and zero-bin boosting d. Mode based zero-bin boosting Lossless coding test was performed on all clips within the derf set, to verify that the commit has achieved lossless compression for all clips. The average compression ratio is around 2.57 to 1. (http://goo.gl/dEShs) Change-Id: Ia3aba7dd09df40dd590f93b9aba134defbc64e34
2012-06-14 06:03:31 +04:00
#if CONFIG_LOSSLESS
cpi->oxcf.lossless = oxcf->lossless;
if (cpi->oxcf.lossless) {
cpi->mb.e_mbd.inv_xform4x4_1_x8 = vp9_short_inv_walsh4x4_1_x8;
cpi->mb.e_mbd.inv_xform4x4_x8 = vp9_short_inv_walsh4x4_x8;
cpi->mb.e_mbd.inv_walsh4x4_1 = vp9_short_inv_walsh4x4_1_lossless;
cpi->mb.e_mbd.inv_walsh4x4_lossless = vp9_short_inv_walsh4x4_lossless;
}
Add lossless compression mode. This commit adds lossless compression capability to the experimental branch. The lossless experiment can be enabled using --enable-lossless in configure. When the experiment is enabled, the encoder will use lossless compression mode by command line option --lossless, and the decoder automatically recognizes a losslessly encoded clip and decodes accordingly. To achieve the lossless coding, this commit has changed the following: 1. To encode at lossless mode, encoder forces the use of unit quantizer, i.e, Q 0, where effective quantization is 1. Encoder also disables the usage of 8x8 transform and allows only 4x4 transform; 2. At Q 0, the first order 4x4 DCT/IDCT have been switched over to a pair of forward and inverse Walsh-Hadamard Transform (http://goo.gl/EIsfy), with proper scaling applied to match the range of the original 4x4 DCT/IDCT pair; 3. At Q 0, the second order remains to use the previous walsh-hadamard transform pair. However, to maintain the reversibility in second order transform at Q 0, scaling down is applied to first order DC coefficients prior to forward transform, and scaling up is applied to the second order output prior to quantization. Symmetric upscaling and downscaling are added around inverse second order transform; 4. At lossless mode, encoder also disables a number of minor features to ensure no loss is introduced, these features includes: a. Trellis quantization optimization b. Loop filtering c. Aggressive zero-binning, rounding and zero-bin boosting d. Mode based zero-bin boosting Lossless coding test was performed on all clips within the derf set, to verify that the commit has achieved lossless compression for all clips. The average compression ratio is around 2.57 to 1. (http://goo.gl/dEShs) Change-Id: Ia3aba7dd09df40dd590f93b9aba134defbc64e34
2012-06-14 06:03:31 +04:00
#endif
cpi->baseline_gf_interval = DEFAULT_GF_INTERVAL;
2010-05-18 19:58:33 +04:00
cpi->ref_frame_flags = VP9_ALT_FLAG | VP9_GOLD_FLAG | VP9_LAST_FLAG;
2010-05-18 19:58:33 +04:00
// cpi->use_golden_frame_only = 0;
// cpi->use_last_frame_only = 0;
cpi->refresh_golden_frame = 0;
cpi->refresh_last_frame = 1;
cm->refresh_entropy_probs = 1;
2010-05-18 19:58:33 +04:00
setup_features(cpi);
cpi->mb.e_mbd.allow_high_precision_mv = 0; // Default mv precision adaptation
set_mvcost(&cpi->mb);
2010-05-18 19:58:33 +04:00
{
int i;
2010-05-18 19:58:33 +04:00
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) {
cpi->oxcf.starting_buffer_level = 60000;
cpi->oxcf.optimal_buffer_level = 60000;
cpi->oxcf.maximum_buffer_size = 240000;
}
// Convert target bandwidth from Kbit/s to Bit/s
cpi->oxcf.target_bandwidth *= 1000;
cpi->oxcf.starting_buffer_level =
rescale(cpi->oxcf.starting_buffer_level,
cpi->oxcf.target_bandwidth, 1000);
// Set or reset optimal and maximum buffer levels.
if (cpi->oxcf.optimal_buffer_level == 0)
cpi->oxcf.optimal_buffer_level = cpi->oxcf.target_bandwidth / 8;
else
cpi->oxcf.optimal_buffer_level =
rescale(cpi->oxcf.optimal_buffer_level,
cpi->oxcf.target_bandwidth, 1000);
if (cpi->oxcf.maximum_buffer_size == 0)
cpi->oxcf.maximum_buffer_size = cpi->oxcf.target_bandwidth / 8;
else
cpi->oxcf.maximum_buffer_size =
rescale(cpi->oxcf.maximum_buffer_size,
cpi->oxcf.target_bandwidth, 1000);
// Set up frame rate and related parameters rate control values.
vp9_new_frame_rate(cpi, cpi->oxcf.frame_rate);
// Set absolute upper and lower quality limits
cpi->worst_quality = cpi->oxcf.worst_allowed_q;
cpi->best_quality = cpi->oxcf.best_allowed_q;
// active values should only be modified if out of new range
if (cpi->active_worst_quality > cpi->oxcf.worst_allowed_q) {
cpi->active_worst_quality = cpi->oxcf.worst_allowed_q;
}
// less likely
else if (cpi->active_worst_quality < cpi->oxcf.best_allowed_q) {
cpi->active_worst_quality = cpi->oxcf.best_allowed_q;
}
if (cpi->active_best_quality < cpi->oxcf.best_allowed_q) {
cpi->active_best_quality = cpi->oxcf.best_allowed_q;
}
// less likely
else if (cpi->active_best_quality > cpi->oxcf.worst_allowed_q) {
cpi->active_best_quality = cpi->oxcf.worst_allowed_q;
}
cpi->buffered_mode = (cpi->oxcf.optimal_buffer_level > 0) ? TRUE : FALSE;
cpi->cq_target_quality = cpi->oxcf.cq_level;
if (!cm->use_bilinear_mc_filter)
cm->mcomp_filter_type = DEFAULT_INTERP_FILTER;
else
cm->mcomp_filter_type = BILINEAR;
2010-05-18 19:58:33 +04:00
cpi->target_bandwidth = cpi->oxcf.target_bandwidth;
2010-05-18 19:58:33 +04:00
cm->Width = cpi->oxcf.Width;
cm->Height = cpi->oxcf.Height;
2010-05-18 19:58:33 +04:00
cm->horiz_scale = cpi->horiz_scale;
cm->vert_scale = cpi->vert_scale;
2010-05-18 19:58:33 +04:00
// VP8 sharpness level mapping 0-7 (vs 0-10 in general VPx dialogs)
if (cpi->oxcf.Sharpness > 7)
cpi->oxcf.Sharpness = 7;
2010-05-18 19:58:33 +04:00
cm->sharpness_level = cpi->oxcf.Sharpness;
2010-05-18 19:58:33 +04:00
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);
2010-05-18 19:58:33 +04:00
Scale2Ratio(cm->horiz_scale, &hr, &hs);
Scale2Ratio(cm->vert_scale, &vr, &vs);
2010-05-18 19:58:33 +04:00
// 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;
}
2010-05-18 19:58:33 +04:00
// Increasing the size of the frame beyond the first seen frame, or some
// otherwise signalled maximum size, is not supported.
// TODO(jkoleszar): exit gracefully.
if (!cpi->initial_width) {
alloc_raw_frame_buffers(cpi);
vp9_alloc_compressor_data(cpi);
cpi->initial_width = cm->Width;
cpi->initial_height = cm->Height;
}
assert(cm->Width <= cpi->initial_width);
assert(cm->Height <= cpi->initial_height);
update_frame_size(cpi);
2010-05-18 19:58:33 +04:00
if (cpi->oxcf.fixed_q >= 0) {
cpi->last_q[0] = cpi->oxcf.fixed_q;
cpi->last_q[1] = cpi->oxcf.fixed_q;
cpi->last_boosted_qindex = cpi->oxcf.fixed_q;
}
2010-05-18 19:58:33 +04:00
cpi->Speed = cpi->oxcf.cpu_used;
2010-05-18 19:58:33 +04:00
// 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;
2010-05-18 19:58:33 +04:00
// YX Temp
cpi->alt_ref_source = NULL;
cpi->is_src_frame_alt_ref = 0;
2010-05-18 19:58:33 +04:00
#if 0
// Experimental RD Code
cpi->frame_distortion = 0;
cpi->last_frame_distortion = 0;
2010-05-18 19:58:33 +04:00
#endif
[WIP] Add column-based tiling. This patch adds column-based tiling. The idea is to make each tile independently decodable (after reading the common frame header) and also independendly encodable (minus within-frame cost adjustments in the RD loop) to speed-up hardware & software en/decoders if they used multi-threading. Column-based tiling has the added advantage (over other tiling methods) that it minimizes realtime use-case latency, since all threads can start encoding data as soon as the first SB-row worth of data is available to the encoder. There is some test code that does random tile ordering in the decoder, to confirm that each tile is indeed independently decodable from other tiles in the same frame. At tile edges, all contexts assume default values (i.e. 0, 0 motion vector, no coefficients, DC intra4x4 mode), and motion vector search and ordering do not cross tiles in the same frame. t log Tile independence is not maintained between frames ATM, i.e. tile 0 of frame 1 is free to use motion vectors that point into any tile of frame 0. We support 1 (i.e. no tiling), 2 or 4 column-tiles. The loopfilter crosses tile boundaries. I discussed this briefly with Aki and he says that's OK. An in-loop loopfilter would need to do some sync between tile threads, but that shouldn't be a big issue. Resuls: with tiling disabled, we go up slightly because of improved edge use in the intra4x4 prediction. With 2 tiles, we lose about ~1% on derf, ~0.35% on HD and ~0.55% on STD/HD. With 4 tiles, we lose another ~1.5% on derf ~0.77% on HD and ~0.85% on STD/HD. Most of this loss is concentrated in the low-bitrate end of clips, and most of it is because of the loss of edges at tile boundaries and the resulting loss of intra predictors. TODO: - more tiles (perhaps allow row-based tiling also, and max. 8 tiles)? - maybe optionally (for EC purposes), motion vectors themselves should not cross tile edges, or we should emulate such borders as if they were off-frame, to limit error propagation to within one tile only. This doesn't have to be the default behaviour but could be an optional bitstream flag. Change-Id: I5951c3a0742a767b20bc9fb5af685d9892c2c96f
2013-02-01 21:35:28 +04:00
cm->tile_columns = 1 << cpi->oxcf.tile_columns;
2010-05-18 19:58:33 +04:00
}
#define M_LOG2_E 0.693147180559945309417
#define log2f(x) (log (x) / (float) M_LOG2_E)
static void cal_nmvjointsadcost(int *mvjointsadcost) {
mvjointsadcost[0] = 600;
mvjointsadcost[1] = 300;
mvjointsadcost[2] = 300;
mvjointsadcost[0] = 300;
}
static void cal_nmvsadcosts(int *mvsadcost[2]) {
int i = 1;
mvsadcost [0] [0] = 0;
mvsadcost [1] [0] = 0;
do {
double z = 256 * (2 * (log2f(8 * 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);
}
static void cal_nmvsadcosts_hp(int *mvsadcost[2]) {
int i = 1;
mvsadcost [0] [0] = 0;
mvsadcost [1] [0] = 0;
do {
double z = 256 * (2 * (log2f(8 * 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);
}
VP9_PTR vp9_create_compressor(VP9_CONFIG *oxcf) {
int i;
volatile union {
VP9_COMP *cpi;
VP9_PTR ptr;
} ctx;
2010-05-18 19:58:33 +04:00
VP9_COMP *cpi;
VP9_COMMON *cm;
2010-05-18 19:58:33 +04:00
cpi = ctx.cpi = vpx_memalign(32, sizeof(VP9_COMP));
// Check that the CPI instance is valid
if (!cpi)
return 0;
2010-05-18 19:58:33 +04:00
cm = &cpi->common;
2010-05-18 19:58:33 +04:00
vpx_memset(cpi, 0, sizeof(VP9_COMP));
2010-05-18 19:58:33 +04:00
if (setjmp(cm->error.jmp)) {
VP9_PTR ptr = ctx.ptr;
2010-05-18 19:58:33 +04:00
ctx.cpi->common.error.setjmp = 0;
vp9_remove_compressor(&ptr);
return 0;
}
cpi->common.error.setjmp = 1;
CHECK_MEM_ERROR(cpi->mb.ss, vpx_calloc(sizeof(search_site), (MAX_MVSEARCH_STEPS * 8) + 1));
vp9_create_common(&cpi->common);
init_config((VP9_PTR)cpi, oxcf);
memcpy(cpi->base_skip_false_prob, base_skip_false_prob, sizeof(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;
cm->prob_last_coded = 128;
cm->prob_gf_coded = 128;
cm->prob_intra_coded = 63;
cm->sb32_coded = 200;
cm->sb64_coded = 200;
for (i = 0; i < COMP_PRED_CONTEXTS; i++)
cm->prob_comppred[i] = 128;
32x32 transform for superblocks. This adds Debargha's DCT/DWT hybrid and a regular 32x32 DCT, and adds code all over the place to wrap that in the bitstream/encoder/decoder/RD. Some implementation notes (these probably need careful review): - token range is extended by 1 bit, since the value range out of this transform is [-16384,16383]. - the coefficients coming out of the FDCT are manually scaled back by 1 bit, or else they won't fit in int16_t (they are 17 bits). Because of this, the RD error scoring does not right-shift the MSE score by two (unlike for 4x4/8x8/16x16). - to compensate for this loss in precision, the quantizer is halved also. This is currently a little hacky. - FDCT and IDCT is double-only right now. Needs a fixed-point impl. - There are no default probabilities for the 32x32 transform yet; I'm simply using the 16x16 luma ones. A future commit will add newly generated probabilities for all transforms. - No ADST version. I don't think we'll add one for this level; if an ADST is desired, transform-size selection can scale back to 16x16 or lower, and use an ADST at that level. Additional notes specific to Debargha's DWT/DCT hybrid: - coefficient scale is different for the top/left 16x16 (DCT-over-DWT) block than for the rest (DWT pixel differences) of the block. Therefore, RD error scoring isn't easily scalable between coefficient and pixel domain. Thus, unfortunately, we need to compute the RD distortion in the pixel domain until we figure out how to scale these appropriately. Change-Id: I00386f20f35d7fabb19aba94c8162f8aee64ef2b
2012-12-08 02:45:05 +04:00
for (i = 0; i < TX_SIZE_MAX_SB - 1; i++)
cm->prob_tx[i] = 128;
// 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->baseline_gf_interval = DEFAULT_GF_INTERVAL;
cpi->gold_is_last = 0;
cpi->alt_is_last = 0;
cpi->gold_is_alt = 0;
// allocate memory for storing last frame's MVs for MV prediction.
CHECK_MEM_ERROR(cpi->lfmv, vpx_calloc((cpi->common.mb_rows + 2) * (cpi->common.mb_cols + 2), sizeof(int_mv)));
CHECK_MEM_ERROR(cpi->lf_ref_frame_sign_bias, vpx_calloc((cpi->common.mb_rows + 2) * (cpi->common.mb_cols + 2), sizeof(int)));
CHECK_MEM_ERROR(cpi->lf_ref_frame, vpx_calloc((cpi->common.mb_rows + 2) * (cpi->common.mb_cols + 2), sizeof(int)));
// 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));
// And a copy in common for temporal coding
CHECK_MEM_ERROR(cm->last_frame_seg_map,
vpx_calloc((cpi->common.mb_rows * cpi->common.mb_cols), 1));
// And a place holder structure is the coding context
// for use if we want to save and restore it
CHECK_MEM_ERROR(cpi->coding_context.last_frame_seg_map_copy,
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;
for (i = 0; i < (sizeof(cpi->mbgraph_stats) /
sizeof(cpi->mbgraph_stats[0])); i++) {
CHECK_MEM_ERROR(cpi->mbgraph_stats[i].mb_stats,
vpx_calloc(cpi->common.mb_rows * cpi->common.mb_cols *
sizeof(*cpi->mbgraph_stats[i].mb_stats),
1));
}
2011-10-05 14:26:00 +04:00
2010-05-18 19:58:33 +04:00
#ifdef ENTROPY_STATS
if (cpi->pass != 1)
init_context_counters();
2010-05-18 19:58:33 +04:00
#endif
#ifdef MODE_STATS
vp9_zero(y_modes);
vp9_zero(i8x8_modes);
vp9_zero(uv_modes);
vp9_zero(uv_modes_y);
vp9_zero(b_modes);
vp9_zero(inter_y_modes);
vp9_zero(inter_uv_modes);
vp9_zero(inter_b_modes);
#endif
#ifdef NMV_STATS
init_nmvstats();
#endif
2010-05-18 19:58:33 +04:00
/*Initialize the feed-forward activity masking.*/
cpi->activity_avg = 90 << 12;
2010-05-18 19:58:33 +04:00
cpi->frames_since_key = 8; // Give a sensible default for the first frame.
cpi->key_frame_frequency = cpi->oxcf.key_freq;
cpi->this_key_frame_forced = FALSE;
cpi->next_key_frame_forced = FALSE;
2010-05-18 19:58:33 +04:00
cpi->source_alt_ref_pending = FALSE;
cpi->source_alt_ref_active = FALSE;
cpi->refresh_alt_ref_frame = 0;
2010-05-18 19:58:33 +04:00
cpi->b_calculate_psnr = CONFIG_INTERNAL_STATS;
#if CONFIG_INTERNAL_STATS
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;
}
2010-05-18 19:58:33 +04:00
#endif
cpi->first_time_stamp_ever = INT64_MAX;
2010-05-18 19:58:33 +04:00
cpi->frames_till_gf_update_due = 0;
cpi->key_frame_count = 1;
2010-05-18 19:58:33 +04:00
cpi->ni_av_qi = cpi->oxcf.worst_allowed_q;
cpi->ni_tot_qi = 0;
cpi->ni_frames = 0;
cpi->tot_q = 0.0;
cpi->avg_q = vp9_convert_qindex_to_q(cpi->oxcf.worst_allowed_q);
cpi->total_byte_count = 0;
2010-05-18 19:58:33 +04:00
cpi->rate_correction_factor = 1.0;
cpi->key_frame_rate_correction_factor = 1.0;
cpi->gf_rate_correction_factor = 1.0;
cpi->twopass.est_max_qcorrection_factor = 1.0;
2010-05-18 19:58:33 +04:00
cal_nmvjointsadcost(cpi->mb.nmvjointsadcost);
cpi->mb.nmvcost[0] = &cpi->mb.nmvcosts[0][MV_MAX];
cpi->mb.nmvcost[1] = &cpi->mb.nmvcosts[1][MV_MAX];
cpi->mb.nmvsadcost[0] = &cpi->mb.nmvsadcosts[0][MV_MAX];
cpi->mb.nmvsadcost[1] = &cpi->mb.nmvsadcosts[1][MV_MAX];
cal_nmvsadcosts(cpi->mb.nmvsadcost);
cpi->mb.nmvcost_hp[0] = &cpi->mb.nmvcosts_hp[0][MV_MAX];
cpi->mb.nmvcost_hp[1] = &cpi->mb.nmvcosts_hp[1][MV_MAX];
cpi->mb.nmvsadcost_hp[0] = &cpi->mb.nmvsadcosts_hp[0][MV_MAX];
cpi->mb.nmvsadcost_hp[1] = &cpi->mb.nmvsadcosts_hp[1][MV_MAX];
cal_nmvsadcosts_hp(cpi->mb.nmvsadcost_hp);
for (i = 0; i < KEY_FRAME_CONTEXT; i++) {
cpi->prior_key_frame_distance[i] = (int)cpi->output_frame_rate;
}
2010-05-18 19:58:33 +04:00
#ifdef OUTPUT_YUV_SRC
yuv_file = fopen("bd.yuv", "ab");
2010-05-18 19:58:33 +04:00
#endif
#ifdef OUTPUT_YUV_REC
yuv_rec_file = fopen("rec.yuv", "wb");
#endif
2010-05-18 19:58:33 +04:00
#if 0
framepsnr = fopen("framepsnr.stt", "a");
kf_list = fopen("kf_list.stt", "w");
2010-05-18 19:58:33 +04:00
#endif
cpi->output_pkt_list = oxcf->output_pkt_list;
2010-05-18 19:58:33 +04:00
if (cpi->pass == 1) {
vp9_init_first_pass(cpi);
} else if (cpi->pass == 2) {
size_t packet_sz = sizeof(FIRSTPASS_STATS);
int packets = (int)(oxcf->two_pass_stats_in.sz / packet_sz);
2010-05-18 19:58:33 +04:00
cpi->twopass.stats_in_start = oxcf->two_pass_stats_in.buf;
cpi->twopass.stats_in = cpi->twopass.stats_in_start;
cpi->twopass.stats_in_end = (void *)((char *)cpi->twopass.stats_in
+ (packets - 1) * packet_sz);
vp9_init_second_pass(cpi);
}
2010-05-18 19:58:33 +04:00
vp9_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;
}
2010-05-18 19:58:33 +04:00
#ifdef ENTROPY_STATS
init_mv_ref_counts();
2010-05-18 19:58:33 +04:00
#endif
#define BFP(BT, SDF, VF, SVF, SVFHH, SVFHV, SVFHHV, SDX3F, SDX8F, SDX4DF) \
cpi->fn_ptr[BT].sdf = SDF; \
cpi->fn_ptr[BT].vf = VF; \
cpi->fn_ptr[BT].svf = SVF; \
cpi->fn_ptr[BT].svf_halfpix_h = SVFHH; \
cpi->fn_ptr[BT].svf_halfpix_v = SVFHV; \
cpi->fn_ptr[BT].svf_halfpix_hv = SVFHHV; \
cpi->fn_ptr[BT].sdx3f = SDX3F; \
cpi->fn_ptr[BT].sdx8f = SDX8F; \
cpi->fn_ptr[BT].sdx4df = SDX4DF;
BFP(BLOCK_32X32, vp9_sad32x32, vp9_variance32x32, vp9_sub_pixel_variance32x32,
vp9_variance_halfpixvar32x32_h, vp9_variance_halfpixvar32x32_v,
vp9_variance_halfpixvar32x32_hv, vp9_sad32x32x3, vp9_sad32x32x8,
vp9_sad32x32x4d)
BFP(BLOCK_64X64, vp9_sad64x64, vp9_variance64x64, vp9_sub_pixel_variance64x64,
vp9_variance_halfpixvar64x64_h, vp9_variance_halfpixvar64x64_v,
vp9_variance_halfpixvar64x64_hv, vp9_sad64x64x3, vp9_sad64x64x8,
vp9_sad64x64x4d)
BFP(BLOCK_16X16, vp9_sad16x16, vp9_variance16x16, vp9_sub_pixel_variance16x16,
vp9_variance_halfpixvar16x16_h, vp9_variance_halfpixvar16x16_v,
vp9_variance_halfpixvar16x16_hv, vp9_sad16x16x3, vp9_sad16x16x8,
vp9_sad16x16x4d)
BFP(BLOCK_16X8, vp9_sad16x8, vp9_variance16x8, vp9_sub_pixel_variance16x8,
NULL, NULL, NULL, vp9_sad16x8x3, vp9_sad16x8x8, vp9_sad16x8x4d)
BFP(BLOCK_8X16, vp9_sad8x16, vp9_variance8x16, vp9_sub_pixel_variance8x16,
NULL, NULL, NULL, vp9_sad8x16x3, vp9_sad8x16x8, vp9_sad8x16x4d)
BFP(BLOCK_8X8, vp9_sad8x8, vp9_variance8x8, vp9_sub_pixel_variance8x8,
NULL, NULL, NULL, vp9_sad8x8x3, vp9_sad8x8x8, vp9_sad8x8x4d)
BFP(BLOCK_4X4, vp9_sad4x4, vp9_variance4x4, vp9_sub_pixel_variance4x4,
NULL, NULL, NULL, vp9_sad4x4x3, vp9_sad4x4x8, vp9_sad4x4x4d)
2010-05-18 19:58:33 +04:00
#if ARCH_X86 || ARCH_X86_64
cpi->fn_ptr[BLOCK_16X16].copymem = vp9_copy32xn;
cpi->fn_ptr[BLOCK_16X8].copymem = vp9_copy32xn;
cpi->fn_ptr[BLOCK_8X16].copymem = vp9_copy32xn;
cpi->fn_ptr[BLOCK_8X8].copymem = vp9_copy32xn;
cpi->fn_ptr[BLOCK_4X4].copymem = vp9_copy32xn;
#endif
cpi->full_search_sad = vp9_full_search_sad;
cpi->diamond_search_sad = vp9_diamond_search_sad;
cpi->refining_search_sad = vp9_refining_search_sad;
2010-05-18 19:58:33 +04:00
// make sure frame 1 is okay
cpi->error_bins[0] = cpi->common.MBs;
2010-05-18 19:58:33 +04:00
/* vp9_init_quantizer() is first called here. Add check in
* vp9_frame_init_quantizer() so that vp9_init_quantizer is only
* called later when needed. This will avoid unnecessary calls of
* vp9_init_quantizer() for every frame.
*/
vp9_init_quantizer(cpi);
vp9_loop_filter_init(cm);
cpi->common.error.setjmp = 0;
vp9_zero(cpi->y_uv_mode_count)
return (VP9_PTR) cpi;
2010-05-18 19:58:33 +04:00
}
void vp9_remove_compressor(VP9_PTR *ptr) {
VP9_COMP *cpi = (VP9_COMP *)(*ptr);
int i;
2010-05-18 19:58:33 +04:00
if (!cpi)
return;
2010-05-18 19:58:33 +04:00
if (cpi && (cpi->common.current_video_frame > 0)) {
if (cpi->pass == 2) {
vp9_end_second_pass(cpi);
}
2010-05-18 19:58:33 +04:00
#ifdef ENTROPY_STATS
if (cpi->pass != 1) {
print_context_counters();
print_tree_update_probs();
print_mode_context();
}
2010-05-18 19:58:33 +04:00
#endif
#ifdef NMV_STATS
if (cpi->pass != 1)
print_nmvstats();
#endif
2010-05-18 19:58:33 +04:00
#if CONFIG_INTERNAL_STATS
2010-05-18 19:58:33 +04:00
vp9_clear_system_state();
// printf("\n8x8-4x4:%d-%d\n", cpi->t8x8_count, cpi->t4x4_count);
if (cpi->pass != 1) {
FILE *f = fopen("opsnr.stt", "a");
double time_encoded = (cpi->last_end_time_stamp_seen
- 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 defined(MODE_STATS)
print_mode_contexts(&cpi->common);
#endif
if (cpi->b_calculate_psnr) {
YV12_BUFFER_CONFIG *lst_yv12 =
&cpi->common.yv12_fb[cpi->common.active_ref_idx[cpi->lst_fb_idx]];
double samples = 3.0 / 2 * cpi->count * lst_yv12->y_width * lst_yv12->y_height;
double total_psnr = vp9_mse2psnr(samples, 255.0, cpi->total_sq_error);
double total_psnr2 = vp9_mse2psnr(samples, 255.0, cpi->total_sq_error2);
double total_ssim = 100 * pow(cpi->summed_quality / cpi->summed_weights, 8.0);
fprintf(f, "Bitrate\tAVGPsnr\tGLBPsnr\tAVPsnrP\tGLPsnrP\tVPXSSIM\t Time(ms)\n");
fprintf(f, "%7.2f\t%7.3f\t%7.3f\t%7.3f\t%7.3f\t%7.3f\t%8.0f\n",
dr, cpi->total / cpi->count, total_psnr, cpi->totalp / cpi->count, total_psnr2, total_ssim,
total_encode_time);
// fprintf(f, "%7.3f\t%7.3f\t%7.3f\t%7.3f\t%7.3f\t%7.3f\t%8.0f %10ld\n",
// dr, cpi->total / cpi->count, total_psnr, cpi->totalp / cpi->count, total_psnr2, total_ssim,
// total_encode_time, cpi->tot_recode_hits);
}
2010-05-18 19:58:33 +04:00
if (cpi->b_calculate_ssimg) {
fprintf(f, "BitRate\tSSIM_Y\tSSIM_U\tSSIM_V\tSSIM_A\t Time(ms)\n");
fprintf(f, "%7.2f\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);
// fprintf(f, "%7.3f\t%6.4f\t%6.4f\t%6.4f\t%6.4f\t%8.0f %10ld\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, cpi->tot_recode_hits);
}
2010-05-18 19:58:33 +04:00
fclose(f);
}
2010-05-18 19:58:33 +04:00
#endif
#ifdef MODE_STATS
{
extern int count_mb_seg[4];
char modes_stats_file[250];
FILE *f;
double dr = (double)cpi->oxcf.frame_rate * (double)cpi->bytes * (double)8 / (double)cpi->count / (double)1000;
sprintf(modes_stats_file, "modes_q%03d.stt", cpi->common.base_qindex);
f = fopen(modes_stats_file, "w");
fprintf(f, "intra_mode in Intra Frames:\n");
{
int i;
fprintf(f, "Y: ");
for (i = 0; i < VP9_YMODES; i++) fprintf(f, " %8d,", y_modes[i]);
fprintf(f, "\n");
}
{
int i;
fprintf(f, "I8: ");
for (i = 0; i < VP9_I8X8_MODES; i++) fprintf(f, " %8d,", i8x8_modes[i]);
fprintf(f, "\n");
}
{
int i;
fprintf(f, "UV: ");
for (i = 0; i < VP9_UV_MODES; i++) fprintf(f, " %8d,", uv_modes[i]);
fprintf(f, "\n");
}
{
int i, j;
fprintf(f, "KeyFrame Y-UV:\n");
for (i = 0; i < VP9_YMODES; i++) {
fprintf(f, "%2d:", i);
for (j = 0; j < VP9_UV_MODES; j++) fprintf(f, "%8d, ", uv_modes_y[i][j]);
fprintf(f, "\n");
}
}
{
int i, j;
fprintf(f, "Inter Y-UV:\n");
for (i = 0; i < VP9_YMODES; i++) {
fprintf(f, "%2d:", i);
for (j = 0; j < VP9_UV_MODES; j++) fprintf(f, "%8d, ", cpi->y_uv_mode_count[i][j]);
fprintf(f, "\n");
}
}
{
int i;
2010-05-18 19:58:33 +04:00
fprintf(f, "B: ");
for (i = 0; i < VP9_NKF_BINTRAMODES; i++)
fprintf(f, "%8d, ", b_modes[i]);
2010-05-18 19:58:33 +04:00
fprintf(f, "\n");
2010-05-18 19:58:33 +04:00
}
2010-05-18 19:58:33 +04:00
fprintf(f, "Modes in Inter Frames:\n");
{
int i;
fprintf(f, "Y: ");
for (i = 0; i < MB_MODE_COUNT; i++) fprintf(f, " %8d,", inter_y_modes[i]);
fprintf(f, "\n");
}
{
int i;
fprintf(f, "UV: ");
for (i = 0; i < VP9_UV_MODES; i++) fprintf(f, " %8d,", inter_uv_modes[i]);
fprintf(f, "\n");
}
{
int i;
fprintf(f, "B: ");
for (i = 0; i < B_MODE_COUNT; 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);
}
2010-05-18 19:58:33 +04:00
#endif
#ifdef ENTROPY_STATS
{
int i, j, k;
FILE *fmode = fopen("vp9_modecontext.c", "w");
2010-05-18 19:58:33 +04:00
fprintf(fmode, "\n#include \"vp9_entropymode.h\"\n\n");
fprintf(fmode, "const unsigned int vp9_kf_default_bmode_counts ");
fprintf(fmode, "[VP9_KF_BINTRAMODES][VP9_KF_BINTRAMODES]"
"[VP9_KF_BINTRAMODES] =\n{\n");
2010-05-18 19:58:33 +04:00
for (i = 0; i < VP9_KF_BINTRAMODES; i++) {
2010-05-18 19:58:33 +04:00
fprintf(fmode, " { // Above Mode : %d\n", i);
2010-05-18 19:58:33 +04:00
for (j = 0; j < VP9_KF_BINTRAMODES; j++) {
2010-05-18 19:58:33 +04:00
fprintf(fmode, " {");
2010-05-18 19:58:33 +04:00
for (k = 0; k < VP9_KF_BINTRAMODES; k++) {
if (!intra_mode_stats[i][j][k])
fprintf(fmode, " %5d, ", 1);
else
fprintf(fmode, " %5d, ", intra_mode_stats[i][j][k]);
}
2010-05-18 19:58:33 +04:00
fprintf(fmode, "}, // left_mode %d\n", j);
2010-05-18 19:58:33 +04:00
}
2010-05-18 19:58:33 +04:00
fprintf(fmode, " },\n");
2010-05-18 19:58:33 +04:00
}
2010-05-18 19:58:33 +04:00
fprintf(fmode, "};\n");
fclose(fmode);
}
2010-05-18 19:58:33 +04:00
#endif
#if defined(SECTIONBITS_OUTPUT)
if (0) {
int i;
FILE *f = fopen("tokenbits.stt", "a");
2010-05-18 19:58:33 +04:00
for (i = 0; i < 28; i++)
fprintf(f, "%8d", (int)(Sectionbits[i] / 256));
2010-05-18 19:58:33 +04:00
fprintf(f, "\n");
fclose(f);
}
2010-05-18 19:58:33 +04:00
#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);
}
2010-05-18 19:58:33 +04:00
#endif
}
2010-05-18 19:58:33 +04:00
dealloc_compressor_data(cpi);
vpx_free(cpi->mb.ss);
vpx_free(cpi->tok);
2010-05-18 19:58:33 +04:00
for (i = 0; i < sizeof(cpi->mbgraph_stats) / sizeof(cpi->mbgraph_stats[0]); i++) {
vpx_free(cpi->mbgraph_stats[i].mb_stats);
}
2011-10-05 14:26:00 +04:00
vp9_remove_common(&cpi->common);
vpx_free(cpi);
*ptr = 0;
2010-05-18 19:58:33 +04:00
#ifdef OUTPUT_YUV_SRC
fclose(yuv_file);
2010-05-18 19:58:33 +04:00
#endif
#ifdef OUTPUT_YUV_REC
fclose(yuv_rec_file);
#endif
2010-05-18 19:58:33 +04:00
#if 0
if (keyfile)
fclose(keyfile);
2010-05-18 19:58:33 +04:00
if (framepsnr)
fclose(framepsnr);
2010-05-18 19:58:33 +04:00
if (kf_list)
fclose(kf_list);
2010-05-18 19:58:33 +04:00
#endif
}
static uint64_t calc_plane_error(uint8_t *orig, int orig_stride,
uint8_t *recon, int recon_stride,
unsigned int cols, unsigned int rows) {
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;
vp9_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;
uint8_t *border_orig = orig;
uint8_t *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;
2010-05-18 19:58:33 +04:00
}
border_orig += orig_stride;
border_recon += recon_stride;
}
2010-05-18 19:58:33 +04:00
}
orig += orig_stride * 16;
recon += recon_stride * 16;
}
2010-05-18 19:58:33 +04:00
/* Handle odd-sized height */
for (; row < rows; row++) {
for (col = 0; col < cols; col++) {
diff = orig[col] - recon[col];
total_sse += diff * diff;
2010-05-18 19:58:33 +04:00
}
orig += orig_stride;
recon += recon_stride;
}
return total_sse;
2010-05-18 19:58:33 +04:00
}
static void generate_psnr_packet(VP9_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);
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);
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);
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] = vp9_mse2psnr(pkt.data.psnr.samples[i], 255.0,
(double)pkt.data.psnr.sse[i]);
vpx_codec_pkt_list_add(cpi->output_pkt_list, &pkt);
2010-05-18 19:58:33 +04:00
}
int vp9_use_as_reference(VP9_PTR ptr, int ref_frame_flags) {
VP9_COMP *cpi = (VP9_COMP *)(ptr);
2010-05-18 19:58:33 +04:00
if (ref_frame_flags > 7)
return -1;
2010-05-18 19:58:33 +04:00
cpi->ref_frame_flags = ref_frame_flags;
return 0;
2010-05-18 19:58:33 +04:00
}
int vp9_update_reference(VP9_PTR ptr, int ref_frame_flags) {
VP9_COMP *cpi = (VP9_COMP *)(ptr);
2010-05-18 19:58:33 +04:00
if (ref_frame_flags > 7)
return -1;
2010-05-18 19:58:33 +04:00
cpi->refresh_golden_frame = 0;
cpi->refresh_alt_ref_frame = 0;
cpi->refresh_last_frame = 0;
2010-05-18 19:58:33 +04:00
if (ref_frame_flags & VP9_LAST_FLAG)
cpi->refresh_last_frame = 1;
2010-05-18 19:58:33 +04:00
if (ref_frame_flags & VP9_GOLD_FLAG)
cpi->refresh_golden_frame = 1;
2010-05-18 19:58:33 +04:00
if (ref_frame_flags & VP9_ALT_FLAG)
cpi->refresh_alt_ref_frame = 1;
2010-05-18 19:58:33 +04:00
return 0;
2010-05-18 19:58:33 +04:00
}
int vp9_get_reference_enc(VP9_PTR ptr, VP9_REFFRAME ref_frame_flag,
YV12_BUFFER_CONFIG *sd) {
VP9_COMP *cpi = (VP9_COMP *)(ptr);
VP9_COMMON *cm = &cpi->common;
int ref_fb_idx;
2010-05-18 19:58:33 +04:00
if (ref_frame_flag == VP9_LAST_FLAG)
ref_fb_idx = cm->active_ref_idx[cpi->lst_fb_idx];
else if (ref_frame_flag == VP9_GOLD_FLAG)
ref_fb_idx = cm->active_ref_idx[cpi->gld_fb_idx];
else if (ref_frame_flag == VP9_ALT_FLAG)
ref_fb_idx = cm->active_ref_idx[cpi->alt_fb_idx];
else
return -1;
vp8_yv12_copy_frame(&cm->yv12_fb[ref_fb_idx], sd);
2010-05-18 19:58:33 +04:00
return 0;
2010-05-18 19:58:33 +04:00
}
int vp9_set_reference_enc(VP9_PTR ptr, VP9_REFFRAME ref_frame_flag,
YV12_BUFFER_CONFIG *sd) {
VP9_COMP *cpi = (VP9_COMP *)(ptr);
VP9_COMMON *cm = &cpi->common;
2010-05-18 19:58:33 +04:00
int ref_fb_idx;
if (ref_frame_flag == VP9_LAST_FLAG)
ref_fb_idx = cm->active_ref_idx[cpi->lst_fb_idx];
else if (ref_frame_flag == VP9_GOLD_FLAG)
ref_fb_idx = cm->active_ref_idx[cpi->gld_fb_idx];
else if (ref_frame_flag == VP9_ALT_FLAG)
ref_fb_idx = cm->active_ref_idx[cpi->alt_fb_idx];
else
return -1;
2010-05-18 19:58:33 +04:00
vp8_yv12_copy_frame(sd, &cm->yv12_fb[ref_fb_idx]);
2010-05-18 19:58:33 +04:00
return 0;
}
int vp9_update_entropy(VP9_PTR comp, int update) {
VP9_COMP *cpi = (VP9_COMP *) comp;
VP9_COMMON *cm = &cpi->common;
cm->refresh_entropy_probs = update;
2010-05-18 19:58:33 +04:00
return 0;
}
2010-05-18 19:58:33 +04:00
#ifdef OUTPUT_YUV_SRC
void vp9_write_yuv_frame(YV12_BUFFER_CONFIG *s) {
uint8_t *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);
}
#endif
2010-05-18 19:58:33 +04:00
#ifdef OUTPUT_YUV_REC
void vp9_write_yuv_rec_frame(VP9_COMMON *cm) {
YV12_BUFFER_CONFIG *s = cm->frame_to_show;
uint8_t *src = s->y_buffer;
int h = cm->Height;
do {
fwrite(src, s->y_width, 1, yuv_rec_file);
src += s->y_stride;
} while (--h);
src = s->u_buffer;
h = (cm->Height + 1) / 2;
do {
fwrite(src, s->uv_width, 1, yuv_rec_file);
src += s->uv_stride;
} while (--h);
src = s->v_buffer;
h = (cm->Height + 1) / 2;
do {
fwrite(src, s->uv_width, 1, yuv_rec_file);
src += s->uv_stride;
} while (--h);
[WIP] Add column-based tiling. This patch adds column-based tiling. The idea is to make each tile independently decodable (after reading the common frame header) and also independendly encodable (minus within-frame cost adjustments in the RD loop) to speed-up hardware & software en/decoders if they used multi-threading. Column-based tiling has the added advantage (over other tiling methods) that it minimizes realtime use-case latency, since all threads can start encoding data as soon as the first SB-row worth of data is available to the encoder. There is some test code that does random tile ordering in the decoder, to confirm that each tile is indeed independently decodable from other tiles in the same frame. At tile edges, all contexts assume default values (i.e. 0, 0 motion vector, no coefficients, DC intra4x4 mode), and motion vector search and ordering do not cross tiles in the same frame. t log Tile independence is not maintained between frames ATM, i.e. tile 0 of frame 1 is free to use motion vectors that point into any tile of frame 0. We support 1 (i.e. no tiling), 2 or 4 column-tiles. The loopfilter crosses tile boundaries. I discussed this briefly with Aki and he says that's OK. An in-loop loopfilter would need to do some sync between tile threads, but that shouldn't be a big issue. Resuls: with tiling disabled, we go up slightly because of improved edge use in the intra4x4 prediction. With 2 tiles, we lose about ~1% on derf, ~0.35% on HD and ~0.55% on STD/HD. With 4 tiles, we lose another ~1.5% on derf ~0.77% on HD and ~0.85% on STD/HD. Most of this loss is concentrated in the low-bitrate end of clips, and most of it is because of the loss of edges at tile boundaries and the resulting loss of intra predictors. TODO: - more tiles (perhaps allow row-based tiling also, and max. 8 tiles)? - maybe optionally (for EC purposes), motion vectors themselves should not cross tile edges, or we should emulate such borders as if they were off-frame, to limit error propagation to within one tile only. This doesn't have to be the default behaviour but could be an optional bitstream flag. Change-Id: I5951c3a0742a767b20bc9fb5af685d9892c2c96f
2013-02-01 21:35:28 +04:00
fflush(yuv_rec_file);
}
#endif
static void update_alt_ref_frame_stats(VP9_COMP *cpi) {
VP9_COMMON *cm = &cpi->common;
// Update data structure that monitors level of reference to last GF
vpx_memset(cpi->gf_active_flags, 1, (cm->mb_rows * cm->mb_cols));
cpi->gf_active_count = cm->mb_rows * cm->mb_cols;
// 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;
2010-05-18 19:58:33 +04:00
}
static void update_golden_frame_stats(VP9_COMP *cpi) {
VP9_COMMON *cm = &cpi->common;
// Update the Golden frame usage counts.
if (cpi->refresh_golden_frame) {
2010-05-18 19:58:33 +04:00
// Update data structure that monitors level of reference to last GF
vpx_memset(cpi->gf_active_flags, 1, (cm->mb_rows * cm->mb_cols));
cpi->gf_active_count = cm->mb_rows * cm->mb_cols;
2010-05-18 19:58:33 +04:00
// this frame refreshes means next frames don't unless specified by user
cpi->refresh_golden_frame = 0;
2010-05-18 19:58:33 +04:00
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->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->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];
}
}
2010-05-18 19:58:33 +04:00
}
static int find_fp_qindex() {
int i;
for (i = 0; i < QINDEX_RANGE; i++) {
if (vp9_convert_qindex_to_q(i) >= 30.0) {
break;
}
}
experiment extending the quantizer range Prior to this change, VP8 min quantizer is 4, which caps the highest quality around 51DB. This experimental change extends the min quantizer to 1, removes the cap and allows the highest quality to be around ~73DB, consistent with the fdct/idct round trip error. To test this change, at configure time use options: --enable-experimental --enable-extend_qrange The following is a brief log of changes in each of the patch sets patch set 1: In this commit, the quantization/dequantization constants are kept unchanged, instead scaling factor 4 is rolled into fdct/idct. Fixed Q0 encoding tests on mobile: Before: 9560.567kbps Overall PSNR:50.255DB VPXSSIM:98.288 Now: 18035.774kbps Overall PSNR:73.022DB VPXSSIM:99.991 patch set 2: regenerated dc/ac quantizer lookup tables based on the scaling factor rolled in the fdct/idct. Also slightly extended the range towards the high quantizer end. patch set 3: slightly tweaked the quantizer tables and generated bits_per_mb table based on Paul's suggestions. patch set 4: fix a typo in idct, re-calculated tables relating active max Q to active min Q patch set 5: added rdmult lookup table based on Q patch set 6: fix rdmult scale: dct coefficient has scaled up by 4 patch set 7: make transform coefficients to be within 16bits patch set 8: normalize 2nd order quantizers patch set 9: fix mis-spellings patch set 10: change the configure script and macros to allow experimental code to be enabled at configure time with --enable-extend_qrange patch set 11: rebase for merge Change-Id: Ib50641ddd44aba2a52ed890222c309faa31cc59c
2010-12-02 02:50:14 +03:00
if (i == QINDEX_RANGE)
i--;
return i;
}
2010-05-18 19:58:33 +04:00
static void Pass1Encode(VP9_COMP *cpi, unsigned long *size, unsigned char *dest, unsigned int *frame_flags) {
(void) size;
(void) dest;
(void) frame_flags;
vp9_set_quantizer(cpi, find_fp_qindex());
vp9_first_pass(cpi);
2010-05-18 19:58:33 +04:00
}
#define WRITE_RECON_BUFFER 0
#if WRITE_RECON_BUFFER
void write_cx_frame_to_file(YV12_BUFFER_CONFIG *frame, int this_frame) {
2010-05-18 19:58:33 +04:00
// write the frame
FILE *yframe;
int i;
char filename[255];
2010-05-18 19:58:33 +04:00
sprintf(filename, "cx\\y%04d.raw", this_frame);
yframe = fopen(filename, "wb");
2010-05-18 19:58:33 +04:00
for (i = 0; i < frame->y_height; i++)
fwrite(frame->y_buffer + i * frame->y_stride,
frame->y_width, 1, yframe);
2010-05-18 19:58:33 +04:00
fclose(yframe);
sprintf(filename, "cx\\u%04d.raw", this_frame);
yframe = fopen(filename, "wb");
2010-05-18 19:58:33 +04:00
for (i = 0; i < frame->uv_height; i++)
fwrite(frame->u_buffer + i * frame->uv_stride,
frame->uv_width, 1, yframe);
2010-05-18 19:58:33 +04:00
fclose(yframe);
sprintf(filename, "cx\\v%04d.raw", this_frame);
yframe = fopen(filename, "wb");
2010-05-18 19:58:33 +04:00
for (i = 0; i < frame->uv_height; i++)
fwrite(frame->v_buffer + i * frame->uv_stride,
frame->uv_width, 1, yframe);
2010-05-18 19:58:33 +04:00
fclose(yframe);
2010-05-18 19:58:33 +04:00
}
#endif
static double compute_edge_pixel_proportion(YV12_BUFFER_CONFIG *frame) {
#define EDGE_THRESH 128
int i, j;
int num_edge_pels = 0;
int num_pels = (frame->y_height - 2) * (frame->y_width - 2);
uint8_t *prev = frame->y_buffer + 1;
uint8_t *curr = frame->y_buffer + 1 + frame->y_stride;
uint8_t *next = frame->y_buffer + 1 + 2 * frame->y_stride;
for (i = 1; i < frame->y_height - 1; i++) {
for (j = 1; j < frame->y_width - 1; j++) {
/* Sobel hor and ver gradients */
int v = 2 * (curr[1] - curr[-1]) + (prev[1] - prev[-1]) + (next[1] - next[-1]);
int h = 2 * (prev[0] - next[0]) + (prev[1] - next[1]) + (prev[-1] - next[-1]);
h = (h < 0 ? -h : h);
v = (v < 0 ? -v : v);
if (h > EDGE_THRESH || v > EDGE_THRESH) num_edge_pels++;
curr++;
prev++;
next++;
}
curr += frame->y_stride - frame->y_width + 2;
prev += frame->y_stride - frame->y_width + 2;
next += frame->y_stride - frame->y_width + 2;
}
return (double)num_edge_pels / (double)num_pels;
}
// Function to test for conditions that indicate we should loop
// back and recode a frame.
static int recode_loop_test(VP9_COMP *cpi,
int high_limit, int low_limit,
int q, int maxq, int minq) {
int force_recode = FALSE;
VP9_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) ||
cpi->refresh_golden_frame ||
cpi->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;
}
// Special Constrained quality tests
else if (cpi->oxcf.end_usage == USAGE_CONSTRAINED_QUALITY) {
// Undershoot and below auto cq level
if ((q > cpi->cq_target_quality) &&
(cpi->projected_frame_size <
((cpi->this_frame_target * 7) >> 3))) {
force_recode = TRUE;
}
// Severe undershoot and between auto and user cq level
else if ((q > cpi->oxcf.cq_level) &&
(cpi->projected_frame_size < cpi->min_frame_bandwidth) &&
(cpi->active_best_quality > cpi->oxcf.cq_level)) {
force_recode = TRUE;
cpi->active_best_quality = cpi->oxcf.cq_level;
}
}
}
return force_recode;
}
static void update_reference_frames(VP9_COMP * const cpi) {
VP9_COMMON * const cm = &cpi->common;
// At this point the new frame has been encoded.
// If any buffer copy / swapping is signaled it should be done here.
if (cm->frame_type == KEY_FRAME) {
ref_cnt_fb(cm->fb_idx_ref_cnt,
&cm->active_ref_idx[cpi->gld_fb_idx], cm->new_fb_idx);
ref_cnt_fb(cm->fb_idx_ref_cnt,
&cm->active_ref_idx[cpi->alt_fb_idx], cm->new_fb_idx);
} else if (cpi->refresh_golden_frame && !cpi->refresh_alt_ref_frame) {
/* Preserve the previously existing golden frame and update the frame in
* the alt ref slot instead. This is highly specific to the current use of
* alt-ref as a forward reference, and this needs to be generalized as
* other uses are implemented (like RTC/temporal scaling)
*
* The update to the buffer in the alt ref slot was signalled in
* vp9_pack_bitstream(), now swap the buffer pointers so that it's treated
* as the golden frame next time.
*/
int tmp;
ref_cnt_fb(cm->fb_idx_ref_cnt,
&cm->active_ref_idx[cpi->alt_fb_idx], cm->new_fb_idx);
tmp = cpi->alt_fb_idx;
cpi->alt_fb_idx = cpi->gld_fb_idx;
cpi->gld_fb_idx = tmp;
} else { /* For non key/golden frames */
if (cpi->refresh_alt_ref_frame) {
ref_cnt_fb(cm->fb_idx_ref_cnt,
&cm->active_ref_idx[cpi->alt_fb_idx], cm->new_fb_idx);
}
if (cpi->refresh_golden_frame) {
ref_cnt_fb(cm->fb_idx_ref_cnt,
&cm->active_ref_idx[cpi->gld_fb_idx], cm->new_fb_idx);
}
}
if (cpi->refresh_last_frame) {
ref_cnt_fb(cm->fb_idx_ref_cnt,
&cm->active_ref_idx[cpi->lst_fb_idx], cm->new_fb_idx);
}
}
static void loopfilter_frame(VP9_COMP *cpi, VP9_COMMON *cm) {
if (cm->no_lpf) {
cm->filter_level = 0;
}
Add lossless compression mode. This commit adds lossless compression capability to the experimental branch. The lossless experiment can be enabled using --enable-lossless in configure. When the experiment is enabled, the encoder will use lossless compression mode by command line option --lossless, and the decoder automatically recognizes a losslessly encoded clip and decodes accordingly. To achieve the lossless coding, this commit has changed the following: 1. To encode at lossless mode, encoder forces the use of unit quantizer, i.e, Q 0, where effective quantization is 1. Encoder also disables the usage of 8x8 transform and allows only 4x4 transform; 2. At Q 0, the first order 4x4 DCT/IDCT have been switched over to a pair of forward and inverse Walsh-Hadamard Transform (http://goo.gl/EIsfy), with proper scaling applied to match the range of the original 4x4 DCT/IDCT pair; 3. At Q 0, the second order remains to use the previous walsh-hadamard transform pair. However, to maintain the reversibility in second order transform at Q 0, scaling down is applied to first order DC coefficients prior to forward transform, and scaling up is applied to the second order output prior to quantization. Symmetric upscaling and downscaling are added around inverse second order transform; 4. At lossless mode, encoder also disables a number of minor features to ensure no loss is introduced, these features includes: a. Trellis quantization optimization b. Loop filtering c. Aggressive zero-binning, rounding and zero-bin boosting d. Mode based zero-bin boosting Lossless coding test was performed on all clips within the derf set, to verify that the commit has achieved lossless compression for all clips. The average compression ratio is around 2.57 to 1. (http://goo.gl/dEShs) Change-Id: Ia3aba7dd09df40dd590f93b9aba134defbc64e34
2012-06-14 06:03:31 +04:00
#if CONFIG_LOSSLESS
else if (cpi->oxcf.lossless) {
cm->filter_level = 0;
}
Add lossless compression mode. This commit adds lossless compression capability to the experimental branch. The lossless experiment can be enabled using --enable-lossless in configure. When the experiment is enabled, the encoder will use lossless compression mode by command line option --lossless, and the decoder automatically recognizes a losslessly encoded clip and decodes accordingly. To achieve the lossless coding, this commit has changed the following: 1. To encode at lossless mode, encoder forces the use of unit quantizer, i.e, Q 0, where effective quantization is 1. Encoder also disables the usage of 8x8 transform and allows only 4x4 transform; 2. At Q 0, the first order 4x4 DCT/IDCT have been switched over to a pair of forward and inverse Walsh-Hadamard Transform (http://goo.gl/EIsfy), with proper scaling applied to match the range of the original 4x4 DCT/IDCT pair; 3. At Q 0, the second order remains to use the previous walsh-hadamard transform pair. However, to maintain the reversibility in second order transform at Q 0, scaling down is applied to first order DC coefficients prior to forward transform, and scaling up is applied to the second order output prior to quantization. Symmetric upscaling and downscaling are added around inverse second order transform; 4. At lossless mode, encoder also disables a number of minor features to ensure no loss is introduced, these features includes: a. Trellis quantization optimization b. Loop filtering c. Aggressive zero-binning, rounding and zero-bin boosting d. Mode based zero-bin boosting Lossless coding test was performed on all clips within the derf set, to verify that the commit has achieved lossless compression for all clips. The average compression ratio is around 2.57 to 1. (http://goo.gl/dEShs) Change-Id: Ia3aba7dd09df40dd590f93b9aba134defbc64e34
2012-06-14 06:03:31 +04:00
#endif
else {
struct vpx_usec_timer timer;
vp9_clear_system_state();
vpx_usec_timer_start(&timer);
if (cpi->sf.auto_filter == 0)
vp9_pick_filter_level_fast(cpi->Source, cpi);
else
vp9_pick_filter_level(cpi->Source, cpi);
vpx_usec_timer_mark(&timer);
cpi->time_pick_lpf += vpx_usec_timer_elapsed(&timer);
}
if (cm->filter_level > 0) {
vp9_set_alt_lf_level(cpi, cm->filter_level);
vp9_loop_filter_frame(cm, &cpi->mb.e_mbd, cm->filter_level, 0);
}
vp8_yv12_extend_frame_borders(cm->frame_to_show);
}
void select_interp_filter_type(VP9_COMP *cpi) {
int i;
int high_filter_index = 0;
unsigned int thresh;
unsigned int high_count = 0;
unsigned int count_sum = 0;
unsigned int *hist = cpi->best_switchable_interp_count;
if (DEFAULT_INTERP_FILTER != SWITCHABLE) {
cpi->common.mcomp_filter_type = DEFAULT_INTERP_FILTER;
return;
}
// TODO(agrange): Look at using RD criteria to select the interpolation
// filter to use for the next frame rather than this simpler counting scheme.
// Select the interpolation filter mode for the next frame
// based on the selection frequency seen in the current frame.
for (i = 0; i < VP9_SWITCHABLE_FILTERS; ++i) {
unsigned int count = hist[i];
count_sum += count;
if (count > high_count) {
high_count = count;
high_filter_index = i;
}
}
thresh = (unsigned int)(0.80 * count_sum);
if (high_count > thresh) {
// One filter accounts for 80+% of cases so force the next
// frame to use this filter exclusively using frame-level flag.
cpi->common.mcomp_filter_type = vp9_switchable_interp[high_filter_index];
} else {
// Use a MB-level switchable filter selection strategy.
cpi->common.mcomp_filter_type = SWITCHABLE;
}
}
#if CONFIG_COMP_INTERINTRA_PRED
static void select_interintra_mode(VP9_COMP *cpi) {
static const double threshold = 0.01;
VP9_COMMON *cm = &cpi->common;
// FIXME(debargha): Make this RD based
int sum = cpi->interintra_select_count[1] + cpi->interintra_select_count[0];
if (sum) {
double fraction = (double) cpi->interintra_select_count[1] / sum;
// printf("fraction: %f\n", fraction);
cm->use_interintra = (fraction > threshold);
}
}
#endif
static void encode_frame_to_data_rate(VP9_COMP *cpi,
unsigned long *size,
unsigned char *dest,
unsigned int *frame_flags) {
VP9_COMMON *cm = &cpi->common;
MACROBLOCKD *xd = &cpi->mb.e_mbd;
int Q;
int frame_over_shoot_limit;
int frame_under_shoot_limit;
int Loop = FALSE;
int loop_count;
int q_low;
int q_high;
int zbin_oq_high;
int zbin_oq_low = 0;
int top_index;
int bottom_index;
int active_worst_qchanged = FALSE;
int overshoot_seen = FALSE;
int undershoot_seen = FALSE;
SPEED_FEATURES *sf = &cpi->sf;
#if RESET_FOREACH_FILTER
int q_low0;
int q_high0;
int zbin_oq_high0;
int zbin_oq_low0 = 0;
int Q0;
int last_zbin_oq;
int last_zbin_oq0;
int active_best_quality0;
int active_worst_quality0;
double rate_correction_factor0;
double gf_rate_correction_factor0;
#endif
/* list of filters to search over */
int mcomp_filters_to_search[] = {
#if CONFIG_ENABLE_6TAP
EIGHTTAP, EIGHTTAP_SHARP, SIXTAP, SWITCHABLE
#else
EIGHTTAP, EIGHTTAP_SHARP, EIGHTTAP_SMOOTH, SWITCHABLE
#endif
};
int mcomp_filters = sizeof(mcomp_filters_to_search) /
sizeof(*mcomp_filters_to_search);
int mcomp_filter_index = 0;
int64_t mcomp_filter_cost[4];
// Clear down mmx registers to allow floating point in what follows
vp9_clear_system_state();
2010-05-18 19:58:33 +04:00
// For an alt ref frame in 2 pass we skip the call to the second
// pass function that sets the target bandwidth so must set it here
if (cpi->refresh_alt_ref_frame) {
cpi->per_frame_bandwidth = cpi->twopass.gf_bits; // Per frame bit target for the alt ref frame
// per second target bitrate
cpi->target_bandwidth = (int)(cpi->twopass.gf_bits *
cpi->output_frame_rate);
}
2010-05-18 19:58:33 +04:00
// Clear zbin over-quant value and mode boost values.
cpi->zbin_over_quant = 0;
cpi->zbin_mode_boost = 0;
2010-05-18 19:58:33 +04:00
// 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;
Add lossless compression mode. This commit adds lossless compression capability to the experimental branch. The lossless experiment can be enabled using --enable-lossless in configure. When the experiment is enabled, the encoder will use lossless compression mode by command line option --lossless, and the decoder automatically recognizes a losslessly encoded clip and decodes accordingly. To achieve the lossless coding, this commit has changed the following: 1. To encode at lossless mode, encoder forces the use of unit quantizer, i.e, Q 0, where effective quantization is 1. Encoder also disables the usage of 8x8 transform and allows only 4x4 transform; 2. At Q 0, the first order 4x4 DCT/IDCT have been switched over to a pair of forward and inverse Walsh-Hadamard Transform (http://goo.gl/EIsfy), with proper scaling applied to match the range of the original 4x4 DCT/IDCT pair; 3. At Q 0, the second order remains to use the previous walsh-hadamard transform pair. However, to maintain the reversibility in second order transform at Q 0, scaling down is applied to first order DC coefficients prior to forward transform, and scaling up is applied to the second order output prior to quantization. Symmetric upscaling and downscaling are added around inverse second order transform; 4. At lossless mode, encoder also disables a number of minor features to ensure no loss is introduced, these features includes: a. Trellis quantization optimization b. Loop filtering c. Aggressive zero-binning, rounding and zero-bin boosting d. Mode based zero-bin boosting Lossless coding test was performed on all clips within the derf set, to verify that the commit has achieved lossless compression for all clips. The average compression ratio is around 2.57 to 1. (http://goo.gl/dEShs) Change-Id: Ia3aba7dd09df40dd590f93b9aba134defbc64e34
2012-06-14 06:03:31 +04:00
#if CONFIG_LOSSLESS
cpi->zbin_mode_boost_enabled = FALSE;
Add lossless compression mode. This commit adds lossless compression capability to the experimental branch. The lossless experiment can be enabled using --enable-lossless in configure. When the experiment is enabled, the encoder will use lossless compression mode by command line option --lossless, and the decoder automatically recognizes a losslessly encoded clip and decodes accordingly. To achieve the lossless coding, this commit has changed the following: 1. To encode at lossless mode, encoder forces the use of unit quantizer, i.e, Q 0, where effective quantization is 1. Encoder also disables the usage of 8x8 transform and allows only 4x4 transform; 2. At Q 0, the first order 4x4 DCT/IDCT have been switched over to a pair of forward and inverse Walsh-Hadamard Transform (http://goo.gl/EIsfy), with proper scaling applied to match the range of the original 4x4 DCT/IDCT pair; 3. At Q 0, the second order remains to use the previous walsh-hadamard transform pair. However, to maintain the reversibility in second order transform at Q 0, scaling down is applied to first order DC coefficients prior to forward transform, and scaling up is applied to the second order output prior to quantization. Symmetric upscaling and downscaling are added around inverse second order transform; 4. At lossless mode, encoder also disables a number of minor features to ensure no loss is introduced, these features includes: a. Trellis quantization optimization b. Loop filtering c. Aggressive zero-binning, rounding and zero-bin boosting d. Mode based zero-bin boosting Lossless coding test was performed on all clips within the derf set, to verify that the commit has achieved lossless compression for all clips. The average compression ratio is around 2.57 to 1. (http://goo.gl/dEShs) Change-Id: Ia3aba7dd09df40dd590f93b9aba134defbc64e34
2012-06-14 06:03:31 +04:00
#else
cpi->zbin_mode_boost_enabled = TRUE;
Add lossless compression mode. This commit adds lossless compression capability to the experimental branch. The lossless experiment can be enabled using --enable-lossless in configure. When the experiment is enabled, the encoder will use lossless compression mode by command line option --lossless, and the decoder automatically recognizes a losslessly encoded clip and decodes accordingly. To achieve the lossless coding, this commit has changed the following: 1. To encode at lossless mode, encoder forces the use of unit quantizer, i.e, Q 0, where effective quantization is 1. Encoder also disables the usage of 8x8 transform and allows only 4x4 transform; 2. At Q 0, the first order 4x4 DCT/IDCT have been switched over to a pair of forward and inverse Walsh-Hadamard Transform (http://goo.gl/EIsfy), with proper scaling applied to match the range of the original 4x4 DCT/IDCT pair; 3. At Q 0, the second order remains to use the previous walsh-hadamard transform pair. However, to maintain the reversibility in second order transform at Q 0, scaling down is applied to first order DC coefficients prior to forward transform, and scaling up is applied to the second order output prior to quantization. Symmetric upscaling and downscaling are added around inverse second order transform; 4. At lossless mode, encoder also disables a number of minor features to ensure no loss is introduced, these features includes: a. Trellis quantization optimization b. Loop filtering c. Aggressive zero-binning, rounding and zero-bin boosting d. Mode based zero-bin boosting Lossless coding test was performed on all clips within the derf set, to verify that the commit has achieved lossless compression for all clips. The average compression ratio is around 2.57 to 1. (http://goo.gl/dEShs) Change-Id: Ia3aba7dd09df40dd590f93b9aba134defbc64e34
2012-06-14 06:03:31 +04:00
#endif
if (cpi->gfu_boost <= 400) {
cpi->zbin_mode_boost_enabled = FALSE;
}
// 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 based loop filter update flags
xd->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;
2010-05-18 19:58:33 +04:00
// Reset the loop filter deltas and segmentation map
setup_features(cpi);
2010-05-18 19:58:33 +04:00
// If segmentation is enabled force a map update for key frames
if (xd->segmentation_enabled) {
xd->update_mb_segmentation_map = 1;
xd->update_mb_segmentation_data = 1;
2010-05-18 19:58:33 +04:00
}
// The alternate reference frame cannot be active for a key frame
cpi->source_alt_ref_active = FALSE;
2011-10-05 14:26:00 +04:00
// Reset the RD threshold multipliers to default of * 1 (128)
for (i = 0; i < MAX_MODES; i++) {
cpi->rd_thresh_mult[i] = 128;
}
cm->error_resilient_mode = (cpi->oxcf.error_resilient_mode != 0);
cm->frame_parallel_decoding_mode =
(cpi->oxcf.frame_parallel_decoding_mode != 0);
if (cm->error_resilient_mode) {
cm->frame_parallel_decoding_mode = 1;
cm->refresh_entropy_probs = 0;
}
}
2010-05-18 19:58:33 +04:00
// Test code for new segment features
init_seg_features(cpi);
2010-05-18 19:58:33 +04:00
// Decide how big to make the frame
vp9_pick_frame_size(cpi);
vp9_clear_system_state();
// Set an active best quality and if necessary active worst quality
Q = cpi->active_worst_quality;
2010-05-18 19:58:33 +04:00
if (cm->frame_type == KEY_FRAME) {
int high = 2000;
int low = 400;
if (cpi->kf_boost > high)
cpi->active_best_quality = kf_low_motion_minq[Q];
else if (cpi->kf_boost < low)
cpi->active_best_quality = kf_high_motion_minq[Q];
else {
int gap = high - low;
int offset = high - cpi->kf_boost;
int qdiff = kf_high_motion_minq[Q] - kf_low_motion_minq[Q];
int adjustment = ((offset * qdiff) + (gap >> 1)) / gap;
cpi->active_best_quality = kf_low_motion_minq[Q] + adjustment;
}
2010-05-18 19:58:33 +04:00
// Make an adjustment based on the %s static
// The main impact of this is at lower Q to prevent overly large key
// frames unless a lot of the image is static.
if (cpi->kf_zeromotion_pct < 64)
cpi->active_best_quality += 4 - (cpi->kf_zeromotion_pct >> 4);
// Special case for key frames forced because we have reached
// the maximum key frame interval. Here force the Q to a range
// based on the ambient Q to reduce the risk of popping
if (cpi->this_key_frame_forced) {
int delta_qindex;
int qindex = cpi->last_boosted_qindex;
delta_qindex = compute_qdelta(cpi, qindex,
(qindex * 0.75));
2010-05-18 19:58:33 +04:00
cpi->active_best_quality = qindex + delta_qindex;
if (cpi->active_best_quality < cpi->best_quality)
2010-05-18 19:58:33 +04:00
cpi->active_best_quality = cpi->best_quality;
}
} else if (cpi->refresh_golden_frame || cpi->refresh_alt_ref_frame) {
int high = 2000;
int low = 400;
// Use the lower of cpi->active_worst_quality and recent
// average Q as basis for GF/ARF Q limit unless last frame was
// a key frame.
if ((cpi->frames_since_key > 1) &&
(cpi->avg_frame_qindex < cpi->active_worst_quality)) {
Q = cpi->avg_frame_qindex;
}
// For constrained quality dont allow Q less than the cq level
if ((cpi->oxcf.end_usage == USAGE_CONSTRAINED_QUALITY) &&
(Q < cpi->cq_target_quality)) {
Q = cpi->cq_target_quality;
}
if (cpi->gfu_boost > high)
cpi->active_best_quality = gf_low_motion_minq[Q];
else if (cpi->gfu_boost < low)
cpi->active_best_quality = gf_high_motion_minq[Q];
else {
int gap = high - low;
int offset = high - cpi->gfu_boost;
int qdiff = gf_high_motion_minq[Q] - gf_low_motion_minq[Q];
int adjustment = ((offset * qdiff) + (gap >> 1)) / gap;
cpi->active_best_quality = gf_low_motion_minq[Q] + adjustment;
}
// Constrained quality use slightly lower active best.
if (cpi->oxcf.end_usage == USAGE_CONSTRAINED_QUALITY) {
cpi->active_best_quality =
cpi->active_best_quality * 15 / 16;
}
} else {
cpi->active_best_quality = inter_minq[Q];
// For the constant/constrained quality mode we dont want
// q to fall below the cq level.
if ((cpi->oxcf.end_usage == USAGE_CONSTRAINED_QUALITY) &&
(cpi->active_best_quality < cpi->cq_target_quality)) {
// If we are strongly undershooting the target rate in the last
// frames then use the user passed in cq value not the auto
// cq value.
if (cpi->rolling_actual_bits < cpi->min_frame_bandwidth)
cpi->active_best_quality = cpi->oxcf.cq_level;
else
cpi->active_best_quality = cpi->cq_target_quality;
}
}
// 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;
if (cpi->active_best_quality > cpi->worst_quality)
cpi->active_best_quality = cpi->worst_quality;
if (cpi->active_worst_quality < cpi->active_best_quality)
cpi->active_worst_quality = cpi->active_best_quality;
// Special case code to try and match quality with forced key frames
if ((cm->frame_type == KEY_FRAME) && cpi->this_key_frame_forced) {
Q = cpi->last_boosted_qindex;
} else {
// Determine initial Q to try
Q = vp9_regulate_q(cpi, cpi->this_frame_target);
}
#if RESET_FOREACH_FILTER
last_zbin_oq = cpi->zbin_over_quant;
#endif
// Set highest allowed value for Zbin over quant
if (cm->frame_type == KEY_FRAME)
zbin_oq_high = 0; // ZBIN_OQ_MAX/16
else if (cpi->refresh_alt_ref_frame
|| (cpi->refresh_golden_frame && !cpi->source_alt_ref_active))
zbin_oq_high = 16;
else
zbin_oq_high = ZBIN_OQ_MAX;
vp9_compute_frame_size_bounds(cpi, &frame_under_shoot_limit,
&frame_over_shoot_limit);
// Limit Q range for the adaptive loop.
bottom_index = cpi->active_best_quality;
top_index = cpi->active_worst_quality;
q_low = cpi->active_best_quality;
q_high = cpi->active_worst_quality;
loop_count = 0;
2010-05-18 19:58:33 +04:00
if (cm->frame_type != KEY_FRAME) {
/* TODO: Decide this more intelligently */
if (sf->search_best_filter) {
cm->mcomp_filter_type = mcomp_filters_to_search[0];
mcomp_filter_index = 0;
} else {
cm->mcomp_filter_type = DEFAULT_INTERP_FILTER;
}
/* TODO: Decide this more intelligently */
xd->allow_high_precision_mv = (Q < HIGH_PRECISION_MV_QTHRESH);
set_mvcost(&cpi->mb);
}
#if CONFIG_COMP_INTERINTRA_PRED
if (cm->current_video_frame == 0) {
cm->use_interintra = 1;
}
#endif
#if CONFIG_POSTPROC
2010-05-18 19:58:33 +04:00
if (cpi->oxcf.noise_sensitivity > 0) {
uint8_t *src;
int l = 0;
2010-05-18 19:58:33 +04:00
switch (cpi->oxcf.noise_sensitivity) {
case 1:
l = 20;
break;
case 2:
l = 40;
break;
case 3:
l = 60;
break;
case 4:
2010-05-18 19:58:33 +04:00
case 5:
l = 100;
break;
case 6:
l = 150;
break;
}
2010-05-18 19:58:33 +04:00
if (cm->frame_type == KEY_FRAME) {
vp9_de_noise(cpi->Source, cpi->Source, l, 1, 0);
} else {
vp9_de_noise(cpi->Source, cpi->Source, l, 1, 0);
2010-05-18 19:58:33 +04:00
src = cpi->Source->y_buffer;
if (cpi->Source->y_stride < 0) {
src += cpi->Source->y_stride * (cpi->Source->y_height - 1);
}
2010-05-18 19:58:33 +04:00
}
}
2010-05-18 19:58:33 +04:00
#endif
#ifdef OUTPUT_YUV_SRC
vp9_write_yuv_frame(cpi->Source);
2010-05-18 19:58:33 +04:00
#endif
#if RESET_FOREACH_FILTER
if (sf->search_best_filter) {
q_low0 = q_low;
q_high0 = q_high;
Q0 = Q;
zbin_oq_low0 = zbin_oq_low;
zbin_oq_high0 = zbin_oq_high;
last_zbin_oq0 = last_zbin_oq;
rate_correction_factor0 = cpi->rate_correction_factor;
gf_rate_correction_factor0 = cpi->gf_rate_correction_factor;
active_best_quality0 = cpi->active_best_quality;
active_worst_quality0 = cpi->active_worst_quality;
}
#endif
do {
vp9_clear_system_state(); // __asm emms;
2010-05-18 19:58:33 +04:00
vp9_set_quantizer(cpi, Q);
2010-05-18 19:58:33 +04:00
if (loop_count == 0) {
// setup skip prob for costing in mode/mv decision
if (cpi->common.mb_no_coeff_skip) {
int k;
for (k = 0; k < MBSKIP_CONTEXTS; k++)
cm->mbskip_pred_probs[k] = cpi->base_skip_false_prob[Q][k];
2010-05-18 19:58:33 +04:00
if (cm->frame_type != KEY_FRAME) {
if (cpi->refresh_alt_ref_frame) {
for (k = 0; k < MBSKIP_CONTEXTS; k++) {
if (cpi->last_skip_false_probs[2][k] != 0)
cm->mbskip_pred_probs[k] = cpi->last_skip_false_probs[2][k];
}
} else if (cpi->refresh_golden_frame) {
for (k = 0; k < MBSKIP_CONTEXTS; k++) {
if (cpi->last_skip_false_probs[1][k] != 0)
cm->mbskip_pred_probs[k] = cpi->last_skip_false_probs[1][k];
}
} else {
int k;
for (k = 0; k < MBSKIP_CONTEXTS; k++) {
if (cpi->last_skip_false_probs[0][k] != 0)
cm->mbskip_pred_probs[k] = cpi->last_skip_false_probs[0][k];
}
}
2010-05-18 19:58:33 +04:00
// as this is for cost estimate, let's make sure it does not
// get extreme either way
{
int k;
for (k = 0; k < MBSKIP_CONTEXTS; ++k) {
if (cm->mbskip_pred_probs[k] < 5)
cm->mbskip_pred_probs[k] = 5;
if (cm->mbskip_pred_probs[k] > 250)
cm->mbskip_pred_probs[k] = 250;
if (cpi->is_src_frame_alt_ref)
cm->mbskip_pred_probs[k] = 1;
}
}
}
}
// Set up entropy depending on frame type.
if (cm->frame_type == KEY_FRAME) {
/* Choose which entropy context to use. When using a forward reference
* frame, it immediately follows the keyframe, and thus benefits from
* using the same entropy context established by the keyframe. Otherwise,
* use the default context 0.
*/
cm->frame_context_idx = cpi->oxcf.play_alternate;
vp9_setup_key_frame(cpi);
} else {
/* Choose which entropy context to use. Currently there are only two
* contexts used, one for normal frames and one for alt ref frames.
*/
cpi->common.frame_context_idx = cpi->refresh_alt_ref_frame;
vp9_setup_inter_frame(cpi);
}
}
// transform / motion compensation build reconstruction frame
vp9_encode_frame(cpi);
// Update the skip mb flag probabilities based on the distribution
// seen in the last encoder iteration.
update_base_skip_probs(cpi);
vp9_clear_system_state(); // __asm emms;
2010-05-18 19:58:33 +04:00
// Dummy pack of the bitstream using up to date stats to get an
// accurate estimate of output frame size to determine if we need
// to recode.
vp9_save_coding_context(cpi);
cpi->dummy_packing = 1;
vp9_pack_bitstream(cpi, dest, size);
cpi->projected_frame_size = (*size) << 3;
vp9_restore_coding_context(cpi);
if (frame_over_shoot_limit == 0)
frame_over_shoot_limit = 1;
active_worst_qchanged = FALSE;
// Special case handling for forced key frames
if ((cm->frame_type == KEY_FRAME) && cpi->this_key_frame_forced) {
int last_q = Q;
int kf_err = vp9_calc_ss_err(cpi->Source,
&cm->yv12_fb[cm->new_fb_idx]);
int high_err_target = cpi->ambient_err;
int low_err_target = (cpi->ambient_err >> 1);
// Prevent possible divide by zero error below for perfect KF
kf_err += (!kf_err);
// The key frame is not good enough or we can afford
// to make it better without undue risk of popping.
if (((kf_err > high_err_target) &&
(cpi->projected_frame_size <= frame_over_shoot_limit)) ||
((kf_err > low_err_target) &&
(cpi->projected_frame_size <= frame_under_shoot_limit))) {
// Lower q_high
q_high = (Q > q_low) ? (Q - 1) : q_low;
// Adjust Q
Q = (Q * high_err_target) / kf_err;
if (Q < ((q_high + q_low) >> 1))
Q = (q_high + q_low) >> 1;
}
// The key frame is much better than the previous frame
else if ((kf_err < low_err_target) &&
(cpi->projected_frame_size >= frame_under_shoot_limit)) {
// Raise q_low
q_low = (Q < q_high) ? (Q + 1) : q_high;
// Adjust Q
Q = (Q * low_err_target) / kf_err;
if (Q > ((q_high + q_low + 1) >> 1))
Q = (q_high + q_low + 1) >> 1;
}
// Clamp Q to upper and lower limits:
if (Q > q_high)
Q = q_high;
else if (Q < q_low)
Q = q_low;
Loop = ((Q != last_q)) ? TRUE : FALSE;
}
// Is the projected frame size out of range and are we allowed to attempt to recode.
else if (recode_loop_test(cpi,
frame_over_shoot_limit, frame_under_shoot_limit,
Q, top_index, bottom_index)) {
int last_q = Q;
int Retries = 0;
// Frame size out of permitted range:
// Update correction factor & compute new Q to try...
// Frame is too large
if (cpi->projected_frame_size > cpi->this_frame_target) {
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 || (loop_count > 1)) {
// Update rate_correction_factor unless cpi->active_worst_quality has changed.
if (!active_worst_qchanged)
vp9_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)
vp9_update_rate_correction_factors(cpi, 0);
Q = vp9_regulate_q(cpi, cpi->this_frame_target);
while (((Q < q_low) || (cpi->zbin_over_quant < zbin_oq_low)) && (Retries < 10)) {
vp9_update_rate_correction_factors(cpi, 0);
Q = vp9_regulate_q(cpi, cpi->this_frame_target);
Retries++;
}
}
2010-05-18 19:58:33 +04:00
overshoot_seen = TRUE;
}
// Frame is too small
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 || (loop_count > 1)) {
// Update rate_correction_factor unless cpi->active_worst_quality has changed.
if (!active_worst_qchanged)
vp9_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)
vp9_update_rate_correction_factors(cpi, 0);
Q = vp9_regulate_q(cpi, cpi->this_frame_target);
// Special case reset for qlow for constrained quality.
// This should only trigger where there is very substantial
// undershoot on a frame and the auto cq level is above
// the user passsed in value.
if ((cpi->oxcf.end_usage == USAGE_CONSTRAINED_QUALITY) &&
(Q < q_low)) {
q_low = Q;
}
while (((Q > q_high) || (cpi->zbin_over_quant > zbin_oq_high)) && (Retries < 10)) {
vp9_update_rate_correction_factors(cpi, 0);
Q = vp9_regulate_q(cpi, cpi->this_frame_target);
Retries++;
}
}
undershoot_seen = TRUE;
}
2010-05-18 19:58:33 +04:00
// Clamp Q to upper and lower limits:
if (Q > q_high)
Q = q_high;
else if (Q < q_low)
Q = q_low;
2010-05-18 19:58:33 +04:00
// 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;
2010-05-18 19:58:33 +04:00
// Loop = ((Q != last_q) || (last_zbin_oq != cpi->zbin_over_quant)) ? TRUE : FALSE;
Loop = ((Q != last_q)) ? TRUE : FALSE;
#if RESET_FOREACH_FILTER
last_zbin_oq = cpi->zbin_over_quant;
#endif
} else
Loop = FALSE;
2010-05-18 19:58:33 +04:00
if (cpi->is_src_frame_alt_ref)
Loop = FALSE;
2010-05-18 19:58:33 +04:00
if (Loop == FALSE && cm->frame_type != KEY_FRAME && sf->search_best_filter) {
if (mcomp_filter_index < mcomp_filters) {
int64_t err = vp9_calc_ss_err(cpi->Source,
&cm->yv12_fb[cm->new_fb_idx]);
int64_t rate = cpi->projected_frame_size << 8;
mcomp_filter_cost[mcomp_filter_index] =
(RDCOST(cpi->RDMULT, cpi->RDDIV, rate, err));
mcomp_filter_index++;
if (mcomp_filter_index < mcomp_filters) {
cm->mcomp_filter_type = mcomp_filters_to_search[mcomp_filter_index];
loop_count = -1;
Loop = TRUE;
} else {
int f;
int64_t best_cost = mcomp_filter_cost[0];
int mcomp_best_filter = mcomp_filters_to_search[0];
for (f = 1; f < mcomp_filters; f++) {
if (mcomp_filter_cost[f] < best_cost) {
mcomp_best_filter = mcomp_filters_to_search[f];
best_cost = mcomp_filter_cost[f];
2010-05-18 19:58:33 +04:00
}
}
if (mcomp_best_filter != mcomp_filters_to_search[mcomp_filters - 1]) {
loop_count = -1;
Loop = TRUE;
cm->mcomp_filter_type = mcomp_best_filter;
}
/*
printf(" best filter = %d, ( ", mcomp_best_filter);
for (f=0;f<mcomp_filters; f++) printf("%d ", mcomp_filter_cost[f]);
printf(")\n");
*/
2010-05-18 19:58:33 +04:00
}
#if RESET_FOREACH_FILTER
if (Loop == TRUE) {
overshoot_seen = FALSE;
undershoot_seen = FALSE;
zbin_oq_low = zbin_oq_low0;
zbin_oq_high = zbin_oq_high0;
q_low = q_low0;
q_high = q_high0;
Q = Q0;
cpi->zbin_over_quant = last_zbin_oq = last_zbin_oq0;
cpi->rate_correction_factor = rate_correction_factor0;
cpi->gf_rate_correction_factor = gf_rate_correction_factor0;
cpi->active_best_quality = active_best_quality0;
cpi->active_worst_quality = active_worst_quality0;
}
#endif
}
}
if (Loop == TRUE) {
loop_count++;
#if CONFIG_INTERNAL_STATS
cpi->tot_recode_hits++;
2010-05-18 19:58:33 +04:00
#endif
}
} while (Loop == TRUE);
// Special case code to reduce pulsing when key frames are forced at a
// fixed interval. Note the reconstruction error if it is the frame before
// the force key frame
if (cpi->next_key_frame_forced && (cpi->twopass.frames_to_key == 0)) {
cpi->ambient_err = vp9_calc_ss_err(cpi->Source,
&cm->yv12_fb[cm->new_fb_idx]);
}
// This frame's MVs are saved and will be used in next frame's MV
// prediction. Last frame has one more line(add to bottom) and one
// more column(add to right) than cm->mip. The edge elements are
// initialized to 0.
if (cm->show_frame) { // do not save for altref frame
int mb_row;
int mb_col;
MODE_INFO *tmp = cm->mip;
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 + 1)].as_int = tmp->mbmi.mv[0].as_int;
cpi->lf_ref_frame_sign_bias[mb_col + mb_row * (cm->mode_info_stride + 1)] = cm->ref_frame_sign_bias[tmp->mbmi.ref_frame];
cpi->lf_ref_frame[mb_col + mb_row * (cm->mode_info_stride + 1)] = tmp->mbmi.ref_frame;
tmp++;
}
}
}
}
// 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
vp9_update_gf_useage_maps(cpi, cm, &cpi->mb);
2010-05-18 19:58:33 +04:00
if (cm->frame_type == KEY_FRAME)
cpi->refresh_last_frame = 1;
2010-05-18 19:58:33 +04:00
#if 0
{
FILE *f = fopen("gfactive.stt", "a");
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,
cpi->refresh_golden_frame);
fclose(f);
}
#endif
2010-05-18 19:58:33 +04:00
cm->frame_to_show = &cm->yv12_fb[cm->new_fb_idx];
#if WRITE_RECON_BUFFER
if (cm->show_frame)
write_cx_frame_to_file(cm->frame_to_show,
cm->current_video_frame);
else
write_cx_frame_to_file(cm->frame_to_show,
cm->current_video_frame + 1000);
#endif
// Pick the loop filter level for the frame.
loopfilter_frame(cpi, cm);
// build the bitstream
cpi->dummy_packing = 0;
vp9_pack_bitstream(cpi, dest, size);
if (cpi->mb.e_mbd.update_mb_segmentation_map) {
update_reference_segmentation_map(cpi);
}
update_reference_frames(cpi);
vp9_copy(cpi->common.fc.coef_counts_4x4, cpi->coef_counts_4x4);
vp9_copy(cpi->common.fc.hybrid_coef_counts_4x4,
cpi->hybrid_coef_counts_4x4);
vp9_copy(cpi->common.fc.coef_counts_8x8, cpi->coef_counts_8x8);
vp9_copy(cpi->common.fc.hybrid_coef_counts_8x8,
cpi->hybrid_coef_counts_8x8);
vp9_copy(cpi->common.fc.coef_counts_16x16, cpi->coef_counts_16x16);
vp9_copy(cpi->common.fc.hybrid_coef_counts_16x16,
cpi->hybrid_coef_counts_16x16);
32x32 transform for superblocks. This adds Debargha's DCT/DWT hybrid and a regular 32x32 DCT, and adds code all over the place to wrap that in the bitstream/encoder/decoder/RD. Some implementation notes (these probably need careful review): - token range is extended by 1 bit, since the value range out of this transform is [-16384,16383]. - the coefficients coming out of the FDCT are manually scaled back by 1 bit, or else they won't fit in int16_t (they are 17 bits). Because of this, the RD error scoring does not right-shift the MSE score by two (unlike for 4x4/8x8/16x16). - to compensate for this loss in precision, the quantizer is halved also. This is currently a little hacky. - FDCT and IDCT is double-only right now. Needs a fixed-point impl. - There are no default probabilities for the 32x32 transform yet; I'm simply using the 16x16 luma ones. A future commit will add newly generated probabilities for all transforms. - No ADST version. I don't think we'll add one for this level; if an ADST is desired, transform-size selection can scale back to 16x16 or lower, and use an ADST at that level. Additional notes specific to Debargha's DWT/DCT hybrid: - coefficient scale is different for the top/left 16x16 (DCT-over-DWT) block than for the rest (DWT pixel differences) of the block. Therefore, RD error scoring isn't easily scalable between coefficient and pixel domain. Thus, unfortunately, we need to compute the RD distortion in the pixel domain until we figure out how to scale these appropriately. Change-Id: I00386f20f35d7fabb19aba94c8162f8aee64ef2b
2012-12-08 02:45:05 +04:00
vp9_copy(cpi->common.fc.coef_counts_32x32, cpi->coef_counts_32x32);
if (!cpi->common.error_resilient_mode &&
!cpi->common.frame_parallel_decoding_mode)
vp9_adapt_coef_probs(&cpi->common);
if (cpi->common.frame_type != KEY_FRAME) {
vp9_copy(cpi->common.fc.sb_ymode_counts, cpi->sb_ymode_count);
vp9_copy(cpi->common.fc.ymode_counts, cpi->ymode_count);
vp9_copy(cpi->common.fc.uv_mode_counts, cpi->y_uv_mode_count);
vp9_copy(cpi->common.fc.bmode_counts, cpi->bmode_count);
vp9_copy(cpi->common.fc.i8x8_mode_counts, cpi->i8x8_mode_count);
vp9_copy(cpi->common.fc.sub_mv_ref_counts, cpi->sub_mv_ref_count);
vp9_copy(cpi->common.fc.mbsplit_counts, cpi->mbsplit_count);
#if CONFIG_COMP_INTERINTRA_PRED
vp9_copy(cpi->common.fc.interintra_counts, cpi->interintra_count);
#endif
cpi->common.fc.NMVcount = cpi->NMVcount;
if (!cpi->common.error_resilient_mode &&
!cpi->common.frame_parallel_decoding_mode) {
vp9_adapt_mode_probs(&cpi->common);
vp9_adapt_mode_context(&cpi->common);
vp9_adapt_nmv_probs(&cpi->common, cpi->mb.e_mbd.allow_high_precision_mv);
}
}
#if CONFIG_COMP_INTERINTRA_PRED
if (cm->frame_type != KEY_FRAME)
select_interintra_mode(cpi);
#endif
/* Move storing frame_type out of the above loop since it is also
* needed in motion search besides loopfilter */
cm->last_frame_type = cm->frame_type;
// Update rate control heuristics
cpi->total_byte_count += (*size);
cpi->projected_frame_size = (*size) << 3;
if (!active_worst_qchanged)
vp9_update_rate_correction_factors(cpi, 2);
cpi->last_q[cm->frame_type] = cm->base_qindex;
// Keep record of last boosted (KF/KF/ARF) Q value.
// If the current frame is coded at a lower Q then we also update it.
// If all mbs in this group are skipped only update if the Q value is
// better than that already stored.
// This is used to help set quality in forced key frames to reduce popping
if ((cm->base_qindex < cpi->last_boosted_qindex) ||
((cpi->static_mb_pct < 100) &&
((cm->frame_type == KEY_FRAME) ||
cpi->refresh_alt_ref_frame ||
(cpi->refresh_golden_frame && !cpi->is_src_frame_alt_ref)))) {
cpi->last_boosted_qindex = cm->base_qindex;
}
if (cm->frame_type == KEY_FRAME) {
vp9_adjust_key_frame_context(cpi);
}
// Keep a record of ambient average Q.
if (cm->frame_type != KEY_FRAME)
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)
&& !cpi->refresh_golden_frame && !cpi->refresh_alt_ref_frame) {
cpi->ni_frames++;
cpi->tot_q += vp9_convert_qindex_to_q(Q);
cpi->avg_q = cpi->tot_q / (double)cpi->ni_frames;
// Calculate the average Q for normal inter frames (not key or GFU
// frames).
cpi->ni_tot_qi += Q;
cpi->ni_av_qi = (cpi->ni_tot_qi / cpi->ni_frames);
}
// Update the buffer level variable.
// Non-viewable frames are a special case and are treated as pure overhead.
if (!cm->show_frame)
cpi->bits_off_target -= cpi->projected_frame_size;
else
cpi->bits_off_target += cpi->av_per_frame_bandwidth - cpi->projected_frame_size;
// Clip the buffer level at the maximum buffer size
if (cpi->bits_off_target > cpi->oxcf.maximum_buffer_size)
cpi->bits_off_target = cpi->oxcf.maximum_buffer_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->twopass.kf_group_bits += cpi->this_frame_target - cpi->projected_frame_size;
if (cpi->twopass.kf_group_bits < 0)
cpi->twopass.kf_group_bits = 0;
} else if (cpi->refresh_golden_frame || cpi->refresh_alt_ref_frame) {
cpi->twopass.gf_group_bits += cpi->this_frame_target - cpi->projected_frame_size;
if (cpi->twopass.gf_group_bits < 0)
cpi->twopass.gf_group_bits = 0;
}
// Update the skip mb flag probabilities based on the distribution seen
// in this frame.
update_base_skip_probs(cpi);
#if 0// 1 && CONFIG_INTERNAL_STATS
{
FILE *f = fopen("tmp.stt", "a");
int recon_err;
vp9_clear_system_state(); // __asm emms;
recon_err = vp9_calc_ss_err(cpi->Source,
&cm->yv12_fb[cm->new_fb_idx]);
if (cpi->twopass.total_left_stats->coded_error != 0.0)
fprintf(f, "%10d %10d %10d %10d %10d %10d %10d %10d"
"%7.2f %7.2f %7.2f %7.2f %7.2f %7.2f %7.2f"
"%6d %5d %5d %5d %8d %8.2f %10d %10.3f"
"%10.3f %8d %10d %10d %10d\n",
cpi->common.current_video_frame, cpi->this_frame_target,
cpi->projected_frame_size, 0, //loop_size_estimate,
(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,
vp9_convert_qindex_to_q(cm->base_qindex),
(double)vp9_dc_quant(cm->base_qindex, 0) / 4.0,
vp9_convert_qindex_to_q(cpi->active_best_quality),
vp9_convert_qindex_to_q(cpi->active_worst_quality),
cpi->avg_q,
vp9_convert_qindex_to_q(cpi->ni_av_qi),
vp9_convert_qindex_to_q(cpi->cq_target_quality),
cpi->zbin_over_quant,
// cpi->avg_frame_qindex, cpi->zbin_over_quant,
cpi->refresh_golden_frame, cpi->refresh_alt_ref_frame,
cm->frame_type, cpi->gfu_boost,
cpi->twopass.est_max_qcorrection_factor,
(int)cpi->twopass.bits_left,
cpi->twopass.total_left_stats->coded_error,
(double)cpi->twopass.bits_left /
cpi->twopass.total_left_stats->coded_error,
cpi->tot_recode_hits, recon_err, cpi->kf_boost,
cpi->kf_zeromotion_pct);
else
fprintf(f, "%10d %10d %10d %10d %10d %10d %10d %10d"
"%7.2f %7.2f %7.2f %7.2f %7.2f %7.2f %7.2f"
"%6d %5d %5d %5d %8d %8.2f %10d %10.3f"
"%8d %10d %10d %10d\n",
cpi->common.current_video_frame,
cpi->this_frame_target, cpi->projected_frame_size,
0, //loop_size_estimate,
(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,
vp9_convert_qindex_to_q(cm->base_qindex),
(double)vp9_dc_quant(cm->base_qindex, 0) / 4.0,
vp9_convert_qindex_to_q(cpi->active_best_quality),
vp9_convert_qindex_to_q(cpi->active_worst_quality),
cpi->avg_q,
vp9_convert_qindex_to_q(cpi->ni_av_qi),
vp9_convert_qindex_to_q(cpi->cq_target_quality),
cpi->zbin_over_quant,
// cpi->avg_frame_qindex, cpi->zbin_over_quant,
cpi->refresh_golden_frame, cpi->refresh_alt_ref_frame,
cm->frame_type, cpi->gfu_boost,
cpi->twopass.est_max_qcorrection_factor,
(int)cpi->twopass.bits_left,
cpi->twopass.total_left_stats->coded_error,
cpi->tot_recode_hits, recon_err, cpi->kf_boost,
cpi->kf_zeromotion_pct);
fclose(f);
if (0) {
FILE *fmodes = fopen("Modes.stt", "a");
int i;
fprintf(fmodes, "%6d:%1d:%1d:%1d ",
cpi->common.current_video_frame,
cm->frame_type, cpi->refresh_golden_frame,
cpi->refresh_alt_ref_frame);
for (i = 0; i < MAX_MODES; i++)
fprintf(fmodes, "%5d ", cpi->mode_chosen_counts[i]);
fprintf(fmodes, "\n");
fclose(fmodes);
}
}
2010-05-18 19:58:33 +04:00
#endif
#if 0
// Debug stats for segment feature experiments.
print_seg_map(cpi);
2010-05-18 19:58:33 +04:00
#endif
// If this was a kf or Gf note the Q
if ((cm->frame_type == KEY_FRAME)
|| cpi->refresh_golden_frame || cpi->refresh_alt_ref_frame)
cm->last_kf_gf_q = cm->base_qindex;
2010-05-18 19:58:33 +04:00
if (cpi->refresh_golden_frame == 1)
cm->frame_flags = cm->frame_flags | FRAMEFLAGS_GOLDEN;
else
cm->frame_flags = cm->frame_flags&~FRAMEFLAGS_GOLDEN;
2010-05-18 19:58:33 +04:00
if (cpi->refresh_alt_ref_frame == 1)
cm->frame_flags = cm->frame_flags | FRAMEFLAGS_ALTREF;
else
cm->frame_flags = cm->frame_flags&~FRAMEFLAGS_ALTREF;
2010-05-18 19:58:33 +04:00
if (cpi->refresh_last_frame & cpi->refresh_golden_frame)
cpi->gold_is_last = 1;
else if (cpi->refresh_last_frame ^ cpi->refresh_golden_frame)
cpi->gold_is_last = 0;
2010-05-18 19:58:33 +04:00
if (cpi->refresh_last_frame & cpi->refresh_alt_ref_frame)
cpi->alt_is_last = 1;
else if (cpi->refresh_last_frame ^ cpi->refresh_alt_ref_frame)
cpi->alt_is_last = 0;
2010-05-18 19:58:33 +04:00
if (cpi->refresh_alt_ref_frame & cpi->refresh_golden_frame)
cpi->gold_is_alt = 1;
else if (cpi->refresh_alt_ref_frame ^ cpi->refresh_golden_frame)
cpi->gold_is_alt = 0;
2010-05-18 19:58:33 +04:00
cpi->ref_frame_flags = VP9_ALT_FLAG | VP9_GOLD_FLAG | VP9_LAST_FLAG;
2010-05-18 19:58:33 +04:00
if (cpi->gold_is_last)
cpi->ref_frame_flags &= ~VP9_GOLD_FLAG;
2010-05-18 19:58:33 +04:00
if (cpi->alt_is_last)
cpi->ref_frame_flags &= ~VP9_ALT_FLAG;
2010-05-18 19:58:33 +04:00
if (cpi->gold_is_alt)
cpi->ref_frame_flags &= ~VP9_ALT_FLAG;
2010-05-18 19:58:33 +04:00
if (cpi->oxcf.play_alternate && cpi->refresh_alt_ref_frame
&& (cm->frame_type != KEY_FRAME))
// Update the alternate reference frame stats as appropriate.
update_alt_ref_frame_stats(cpi);
else
// Update the Golden frame stats as appropriate.
update_golden_frame_stats(cpi);
2010-05-18 19:58:33 +04:00
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;
2010-05-18 19:58:33 +04:00
// As this frame is a key frame the next defaults to an inter frame.
cm->frame_type = INTER_FRAME;
} else {
*frame_flags = cm->frame_flags&~FRAMEFLAGS_KEY;
}
2010-05-18 19:58:33 +04:00
// Clear the one shot update flags for segmentation map and mode/ref loop filter deltas.
xd->update_mb_segmentation_map = 0;
xd->update_mb_segmentation_data = 0;
xd->mode_ref_lf_delta_update = 0;
2010-05-18 19:58:33 +04:00
// 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++;
}
2010-05-18 19:58:33 +04:00
// reset to normal state now that we are done.
2010-05-18 19:58:33 +04:00
#if 0
{
char filename[512];
FILE *recon_file;
sprintf(filename, "enc%04d.yuv", (int) cm->current_video_frame);
recon_file = fopen(filename, "wb");
fwrite(cm->yv12_fb[cm->active_ref_idx[cpi->lst_fb_idx]].buffer_alloc,
cm->yv12_fb[cm->active_ref_idx[cpi->lst_fb_idx]].frame_size,
1, recon_file);
fclose(recon_file);
}
#endif
#ifdef OUTPUT_YUV_REC
vp9_write_yuv_rec_frame(cm);
experiment extending the quantizer range Prior to this change, VP8 min quantizer is 4, which caps the highest quality around 51DB. This experimental change extends the min quantizer to 1, removes the cap and allows the highest quality to be around ~73DB, consistent with the fdct/idct round trip error. To test this change, at configure time use options: --enable-experimental --enable-extend_qrange The following is a brief log of changes in each of the patch sets patch set 1: In this commit, the quantization/dequantization constants are kept unchanged, instead scaling factor 4 is rolled into fdct/idct. Fixed Q0 encoding tests on mobile: Before: 9560.567kbps Overall PSNR:50.255DB VPXSSIM:98.288 Now: 18035.774kbps Overall PSNR:73.022DB VPXSSIM:99.991 patch set 2: regenerated dc/ac quantizer lookup tables based on the scaling factor rolled in the fdct/idct. Also slightly extended the range towards the high quantizer end. patch set 3: slightly tweaked the quantizer tables and generated bits_per_mb table based on Paul's suggestions. patch set 4: fix a typo in idct, re-calculated tables relating active max Q to active min Q patch set 5: added rdmult lookup table based on Q patch set 6: fix rdmult scale: dct coefficient has scaled up by 4 patch set 7: make transform coefficients to be within 16bits patch set 8: normalize 2nd order quantizers patch set 9: fix mis-spellings patch set 10: change the configure script and macros to allow experimental code to be enabled at configure time with --enable-extend_qrange patch set 11: rebase for merge Change-Id: Ib50641ddd44aba2a52ed890222c309faa31cc59c
2010-12-02 02:50:14 +03:00
#endif
2010-05-18 19:58:33 +04:00
if (cm->show_frame) {
vpx_memcpy(cm->prev_mip, cm->mip,
(cm->mb_cols + 1) * (cm->mb_rows + 1)* sizeof(MODE_INFO));
} else {
vpx_memset(cm->prev_mip, 0,
(cm->mb_cols + 1) * (cm->mb_rows + 1)* sizeof(MODE_INFO));
}
2010-05-18 19:58:33 +04:00
}
static void Pass2Encode(VP9_COMP *cpi, unsigned long *size,
unsigned char *dest, unsigned int *frame_flags) {
2010-05-18 19:58:33 +04:00
if (!cpi->refresh_alt_ref_frame)
vp9_second_pass(cpi);
2010-05-18 19:58:33 +04:00
encode_frame_to_data_rate(cpi, size, dest, frame_flags);
cpi->twopass.bits_left -= 8 * *size;
2010-05-18 19:58:33 +04:00
if (!cpi->refresh_alt_ref_frame) {
double lower_bounds_min_rate = FRAME_OVERHEAD_BITS * cpi->oxcf.frame_rate;
double two_pass_min_rate = (double)(cpi->oxcf.target_bandwidth
* cpi->oxcf.two_pass_vbrmin_section / 100);
if (two_pass_min_rate < lower_bounds_min_rate)
two_pass_min_rate = lower_bounds_min_rate;
cpi->twopass.bits_left += (int64_t)(two_pass_min_rate / cpi->oxcf.frame_rate);
}
2010-05-18 19:58:33 +04:00
}
int vp9_receive_raw_frame(VP9_PTR ptr, unsigned int frame_flags,
YV12_BUFFER_CONFIG *sd, int64_t time_stamp,
int64_t end_time) {
VP9_COMP *cpi = (VP9_COMP *) ptr;
VP9_COMMON *cm = &cpi->common;
struct vpx_usec_timer timer;
int res = 0;
2010-05-18 19:58:33 +04:00
vpx_usec_timer_start(&timer);
if (vp9_lookahead_push(cpi->lookahead, sd, time_stamp, end_time, frame_flags,
cpi->active_map_enabled ? cpi->active_map : NULL))
res = -1;
cm->clr_type = sd->clrtype;
vpx_usec_timer_mark(&timer);
cpi->time_receive_data += vpx_usec_timer_elapsed(&timer);
2010-05-18 19:58:33 +04:00
return res;
2010-05-18 19:58:33 +04:00
}
static int frame_is_reference(const VP9_COMP *cpi) {
const VP9_COMMON *cm = &cpi->common;
const MACROBLOCKD *xd = &cpi->mb.e_mbd;
return cm->frame_type == KEY_FRAME || cpi->refresh_last_frame
|| cpi->refresh_golden_frame || cpi->refresh_alt_ref_frame
|| cm->refresh_entropy_probs
|| xd->mode_ref_lf_delta_update
|| xd->update_mb_segmentation_map || xd->update_mb_segmentation_data;
}
int vp9_get_compressed_data(VP9_PTR ptr, unsigned int *frame_flags,
unsigned long *size, unsigned char *dest,
int64_t *time_stamp, int64_t *time_end, int flush) {
VP9_COMP *cpi = (VP9_COMP *) ptr;
VP9_COMMON *cm = &cpi->common;
struct vpx_usec_timer cmptimer;
YV12_BUFFER_CONFIG *force_src_buffer = NULL;
2010-05-18 19:58:33 +04:00
if (!cpi)
return -1;
2010-05-18 19:58:33 +04:00
vpx_usec_timer_start(&cmptimer);
2010-05-18 19:58:33 +04:00
cpi->source = NULL;
2010-05-18 19:58:33 +04:00
cpi->mb.e_mbd.allow_high_precision_mv = ALTREF_HIGH_PRECISION_MV;
set_mvcost(&cpi->mb);
// Should we code an alternate reference frame
if (cpi->oxcf.play_alternate &&
cpi->source_alt_ref_pending) {
if ((cpi->source = vp9_lookahead_peek(cpi->lookahead,
cpi->frames_till_gf_update_due))) {
cpi->alt_ref_source = cpi->source;
if (cpi->oxcf.arnr_max_frames > 0) {
vp9_temporal_filter_prepare(cpi, cpi->frames_till_gf_update_due);
force_src_buffer = &cpi->alt_ref_buffer;
}
cm->frames_till_alt_ref_frame = cpi->frames_till_gf_update_due;
cpi->refresh_alt_ref_frame = 1;
cpi->refresh_golden_frame = 0;
cpi->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;
}
}
if (!cpi->source) {
if ((cpi->source = vp9_lookahead_pop(cpi->lookahead, flush))) {
cm->show_frame = 1;
cpi->is_src_frame_alt_ref = cpi->alt_ref_source
&& (cpi->source == cpi->alt_ref_source);
if (cpi->is_src_frame_alt_ref)
cpi->alt_ref_source = NULL;
}
}
if (cpi->source) {
cpi->un_scaled_source =
cpi->Source = force_src_buffer ? force_src_buffer : &cpi->source->img;
*time_stamp = cpi->source->ts_start;
*time_end = cpi->source->ts_end;
*frame_flags = cpi->source->flags;
} else {
*size = 0;
if (flush && cpi->pass == 1 && !cpi->twopass.first_pass_done) {
vp9_end_first_pass(cpi); /* get last stats packet */
cpi->twopass.first_pass_done = 1;
2010-05-18 19:58:33 +04:00
}
return -1;
}
2010-05-18 19:58:33 +04:00
if (cpi->source->ts_start < cpi->first_time_stamp_ever) {
cpi->first_time_stamp_ever = cpi->source->ts_start;
cpi->last_end_time_stamp_seen = cpi->source->ts_start;
}
// adjust frame rates based on timestamps given
if (!cpi->refresh_alt_ref_frame) {
int64_t this_duration;
int step = 0;
2010-05-18 19:58:33 +04:00
if (cpi->source->ts_start == cpi->first_time_stamp_ever) {
this_duration = cpi->source->ts_end - cpi->source->ts_start;
step = 1;
} else {
int64_t last_duration;
2010-05-18 19:58:33 +04:00
this_duration = cpi->source->ts_end - cpi->last_end_time_stamp_seen;
last_duration = cpi->last_end_time_stamp_seen
- cpi->last_time_stamp_seen;
// do a step update if the duration changes by 10%
if (last_duration)
step = (int)((this_duration - last_duration) * 10 / last_duration);
}
2010-05-18 19:58:33 +04:00
if (this_duration) {
if (step)
vp9_new_frame_rate(cpi, 10000000.0 / this_duration);
else {
double avg_duration, interval;
/* Average this frame's rate into the last second's average
* frame rate. If we haven't seen 1 second yet, then average
* over the whole interval seen.
*/
interval = (double)(cpi->source->ts_end
- cpi->first_time_stamp_ever);
if (interval > 10000000.0)
interval = 10000000;
avg_duration = 10000000.0 / cpi->oxcf.frame_rate;
avg_duration *= (interval - avg_duration + this_duration);
avg_duration /= interval;
2010-05-18 19:58:33 +04:00
vp9_new_frame_rate(cpi, 10000000.0 / avg_duration);
}
2010-05-18 19:58:33 +04:00
}
cpi->last_time_stamp_seen = cpi->source->ts_start;
cpi->last_end_time_stamp_seen = cpi->source->ts_end;
}
2010-05-18 19:58:33 +04:00
// start with a 0 size frame
*size = 0;
2010-05-18 19:58:33 +04:00
// Clear down mmx registers
vp9_clear_system_state(); // __asm emms;
cm->frame_type = INTER_FRAME;
cm->frame_flags = *frame_flags;
2010-05-18 19:58:33 +04:00
#if 0
if (cpi->refresh_alt_ref_frame) {
// cpi->refresh_golden_frame = 1;
cpi->refresh_golden_frame = 0;
cpi->refresh_last_frame = 0;
} else {
cpi->refresh_golden_frame = 0;
cpi->refresh_last_frame = 1;
}
2010-05-18 19:58:33 +04:00
#endif
/* find a free buffer for the new frame, releasing the reference previously
* held.
*/
cm->fb_idx_ref_cnt[cm->new_fb_idx]--;
cm->new_fb_idx = get_free_fb(cm);
/* Reset the frame pointers to the current frame size */
vp8_yv12_realloc_frame_buffer(&cm->yv12_fb[cm->new_fb_idx],
cm->mb_cols * 16, cm->mb_rows * 16,
VP9BORDERINPIXELS);
/* Disable any references that have different size */
if ((cm->yv12_fb[cm->active_ref_idx[cpi->lst_fb_idx]].y_width !=
cm->yv12_fb[cm->new_fb_idx].y_width) ||
(cm->yv12_fb[cm->active_ref_idx[cpi->lst_fb_idx]].y_height !=
cm->yv12_fb[cm->new_fb_idx].y_height))
cpi->ref_frame_flags &= ~VP9_LAST_FLAG;
if ((cm->yv12_fb[cm->active_ref_idx[cpi->gld_fb_idx]].y_width !=
cm->yv12_fb[cm->new_fb_idx].y_width) ||
(cm->yv12_fb[cm->active_ref_idx[cpi->gld_fb_idx]].y_height !=
cm->yv12_fb[cm->new_fb_idx].y_height))
cpi->ref_frame_flags &= ~VP9_GOLD_FLAG;
if ((cm->yv12_fb[cm->active_ref_idx[cpi->alt_fb_idx]].y_width !=
cm->yv12_fb[cm->new_fb_idx].y_width) ||
(cm->yv12_fb[cm->active_ref_idx[cpi->alt_fb_idx]].y_height !=
cm->yv12_fb[cm->new_fb_idx].y_height))
cpi->ref_frame_flags &= ~VP9_ALT_FLAG;
vp9_setup_interp_filters(&cpi->mb.e_mbd, DEFAULT_INTERP_FILTER, cm);
if (cpi->pass == 1) {
Pass1Encode(cpi, size, dest, frame_flags);
} else if (cpi->pass == 2) {
Pass2Encode(cpi, size, dest, frame_flags);
} else {
encode_frame_to_data_rate(cpi, size, dest, frame_flags);
}
2010-05-18 19:58:33 +04:00
if (cm->refresh_entropy_probs) {
vpx_memcpy(&cm->frame_contexts[cm->frame_context_idx], &cm->fc,
sizeof(cm->fc));
}
2010-05-18 19:58:33 +04:00
// if its a dropped frame honor the requests on subsequent frames
if (*size > 0) {
cpi->droppable = !frame_is_reference(cpi);
2010-05-18 19:58:33 +04:00
// return to normal state
cm->refresh_entropy_probs = 1;
cpi->refresh_alt_ref_frame = 0;
cpi->refresh_golden_frame = 0;
cpi->refresh_last_frame = 1;
cm->frame_type = INTER_FRAME;
2010-05-18 19:58:33 +04:00
}
2010-05-18 19:58:33 +04:00
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);
}
2010-05-18 19:58:33 +04:00
#if CONFIG_INTERNAL_STATS
2010-05-18 19:58:33 +04:00
if (cpi->pass != 1) {
cpi->bytes += *size;
2010-05-18 19:58:33 +04:00
if (cm->show_frame) {
2010-05-18 19:58:33 +04:00
cpi->count++;
2010-05-18 19:58:33 +04:00
if (cpi->b_calculate_psnr) {
double ye, ue, ve;
double frame_psnr;
YV12_BUFFER_CONFIG *orig = cpi->Source;
YV12_BUFFER_CONFIG *recon = cpi->common.frame_to_show;
YV12_BUFFER_CONFIG *pp = &cm->post_proc_buffer;
int y_samples = orig->y_height * orig->y_width;
int uv_samples = orig->uv_height * orig->uv_width;
int t_samples = y_samples + 2 * uv_samples;
double sq_error;
ye = (double)calc_plane_error(orig->y_buffer, orig->y_stride,
recon->y_buffer, recon->y_stride, orig->y_width,
orig->y_height);
ue = (double)calc_plane_error(orig->u_buffer, orig->uv_stride,
recon->u_buffer, recon->uv_stride, orig->uv_width,
orig->uv_height);
ve = (double)calc_plane_error(orig->v_buffer, orig->uv_stride,
recon->v_buffer, recon->uv_stride, orig->uv_width,
orig->uv_height);
2010-05-18 19:58:33 +04:00
sq_error = ye + ue + ve;
2010-05-18 19:58:33 +04:00
frame_psnr = vp9_mse2psnr(t_samples, 255.0, sq_error);
2010-05-18 19:58:33 +04:00
cpi->total_y += vp9_mse2psnr(y_samples, 255.0, ye);
cpi->total_u += vp9_mse2psnr(uv_samples, 255.0, ue);
cpi->total_v += vp9_mse2psnr(uv_samples, 255.0, ve);
cpi->total_sq_error += sq_error;
cpi->total += frame_psnr;
{
double frame_psnr2, frame_ssim2 = 0;
double weight = 0;
#if CONFIG_POSTPROC
vp9_deblock(cm->frame_to_show, &cm->post_proc_buffer,
cm->filter_level * 10 / 6, 1, 0);
#endif
vp9_clear_system_state();
ye = (double)calc_plane_error(orig->y_buffer, orig->y_stride,
pp->y_buffer, pp->y_stride, orig->y_width,
orig->y_height);
ue = (double)calc_plane_error(orig->u_buffer, orig->uv_stride,
pp->u_buffer, pp->uv_stride, orig->uv_width,
orig->uv_height);
ve = (double)calc_plane_error(orig->v_buffer, orig->uv_stride,
pp->v_buffer, pp->uv_stride, orig->uv_width,
orig->uv_height);
sq_error = ye + ue + ve;
frame_psnr2 = vp9_mse2psnr(t_samples, 255.0, sq_error);
cpi->totalp_y += vp9_mse2psnr(y_samples, 255.0, ye);
cpi->totalp_u += vp9_mse2psnr(uv_samples, 255.0, ue);
cpi->totalp_v += vp9_mse2psnr(uv_samples, 255.0, ve);
cpi->total_sq_error2 += sq_error;
cpi->totalp += frame_psnr2;
2010-05-18 19:58:33 +04:00
frame_ssim2 = vp9_calc_ssim(cpi->Source,
&cm->post_proc_buffer, 1, &weight);
cpi->summed_quality += frame_ssim2 * weight;
cpi->summed_weights += weight;
experiment extending the quantizer range Prior to this change, VP8 min quantizer is 4, which caps the highest quality around 51DB. This experimental change extends the min quantizer to 1, removes the cap and allows the highest quality to be around ~73DB, consistent with the fdct/idct round trip error. To test this change, at configure time use options: --enable-experimental --enable-extend_qrange The following is a brief log of changes in each of the patch sets patch set 1: In this commit, the quantization/dequantization constants are kept unchanged, instead scaling factor 4 is rolled into fdct/idct. Fixed Q0 encoding tests on mobile: Before: 9560.567kbps Overall PSNR:50.255DB VPXSSIM:98.288 Now: 18035.774kbps Overall PSNR:73.022DB VPXSSIM:99.991 patch set 2: regenerated dc/ac quantizer lookup tables based on the scaling factor rolled in the fdct/idct. Also slightly extended the range towards the high quantizer end. patch set 3: slightly tweaked the quantizer tables and generated bits_per_mb table based on Paul's suggestions. patch set 4: fix a typo in idct, re-calculated tables relating active max Q to active min Q patch set 5: added rdmult lookup table based on Q patch set 6: fix rdmult scale: dct coefficient has scaled up by 4 patch set 7: make transform coefficients to be within 16bits patch set 8: normalize 2nd order quantizers patch set 9: fix mis-spellings patch set 10: change the configure script and macros to allow experimental code to be enabled at configure time with --enable-extend_qrange patch set 11: rebase for merge Change-Id: Ib50641ddd44aba2a52ed890222c309faa31cc59c
2010-12-02 02:50:14 +03:00
#if 0
{
FILE *f = fopen("q_used.stt", "a");
fprintf(f, "%5d : Y%f7.3:U%f7.3:V%f7.3:F%f7.3:S%7.3f\n",
cpi->common.current_video_frame, y2, u2, v2,
frame_psnr2, frame_ssim2);
fclose(f);
}
experiment extending the quantizer range Prior to this change, VP8 min quantizer is 4, which caps the highest quality around 51DB. This experimental change extends the min quantizer to 1, removes the cap and allows the highest quality to be around ~73DB, consistent with the fdct/idct round trip error. To test this change, at configure time use options: --enable-experimental --enable-extend_qrange The following is a brief log of changes in each of the patch sets patch set 1: In this commit, the quantization/dequantization constants are kept unchanged, instead scaling factor 4 is rolled into fdct/idct. Fixed Q0 encoding tests on mobile: Before: 9560.567kbps Overall PSNR:50.255DB VPXSSIM:98.288 Now: 18035.774kbps Overall PSNR:73.022DB VPXSSIM:99.991 patch set 2: regenerated dc/ac quantizer lookup tables based on the scaling factor rolled in the fdct/idct. Also slightly extended the range towards the high quantizer end. patch set 3: slightly tweaked the quantizer tables and generated bits_per_mb table based on Paul's suggestions. patch set 4: fix a typo in idct, re-calculated tables relating active max Q to active min Q patch set 5: added rdmult lookup table based on Q patch set 6: fix rdmult scale: dct coefficient has scaled up by 4 patch set 7: make transform coefficients to be within 16bits patch set 8: normalize 2nd order quantizers patch set 9: fix mis-spellings patch set 10: change the configure script and macros to allow experimental code to be enabled at configure time with --enable-extend_qrange patch set 11: rebase for merge Change-Id: Ib50641ddd44aba2a52ed890222c309faa31cc59c
2010-12-02 02:50:14 +03:00
#endif
}
}
2010-05-18 19:58:33 +04:00
if (cpi->b_calculate_ssimg) {
double y, u, v, frame_all;
frame_all = vp9_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;
}
2010-05-18 19:58:33 +04:00
}
}
2010-05-18 19:58:33 +04:00
#endif
return 0;
2010-05-18 19:58:33 +04:00
}
int vp9_get_preview_raw_frame(VP9_PTR comp, YV12_BUFFER_CONFIG *dest,
vp9_ppflags_t *flags) {
VP9_COMP *cpi = (VP9_COMP *) comp;
2010-05-18 19:58:33 +04:00
if (cpi->refresh_alt_ref_frame)
return -1;
else {
int ret;
2010-05-18 19:58:33 +04:00
#if CONFIG_POSTPROC
ret = vp9_post_proc_frame(&cpi->common, dest, flags);
2010-05-18 19:58:33 +04:00
#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;
2010-05-18 19:58:33 +04:00
}
#endif // !CONFIG_POSTPROC
vp9_clear_system_state();
return ret;
}
}
2010-05-18 19:58:33 +04:00
int vp9_set_roimap(VP9_PTR comp, unsigned char *map, unsigned int rows,
unsigned int cols, int delta_q[4], int delta_lf[4],
unsigned int threshold[4]) {
VP9_COMP *cpi = (VP9_COMP *) comp;
signed char feature_data[SEG_LVL_MAX][MAX_MB_SEGMENTS];
MACROBLOCKD *xd = &cpi->mb.e_mbd;
int i;
2010-05-18 19:58:33 +04:00
if (cpi->common.mb_rows != rows || cpi->common.mb_cols != cols)
return -1;
2010-05-18 19:58:33 +04:00
if (!map) {
vp9_disable_segmentation((VP9_PTR)cpi);
return 0;
}
// Set the segmentation Map
vp9_set_segmentation_map((VP9_PTR)cpi, map);
// Activate segmentation.
vp9_enable_segmentation((VP9_PTR)cpi);
// Set up the quant segment data
feature_data[SEG_LVL_ALT_Q][0] = delta_q[0];
feature_data[SEG_LVL_ALT_Q][1] = delta_q[1];
feature_data[SEG_LVL_ALT_Q][2] = delta_q[2];
feature_data[SEG_LVL_ALT_Q][3] = delta_q[3];
// Set up the loop segment data s
feature_data[SEG_LVL_ALT_LF][0] = delta_lf[0];
feature_data[SEG_LVL_ALT_LF][1] = delta_lf[1];
feature_data[SEG_LVL_ALT_LF][2] = delta_lf[2];
feature_data[SEG_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];
// Enable the loop and quant changes in the feature mask
for (i = 0; i < 4; i++) {
if (delta_q[i])
vp9_enable_segfeature(xd, i, SEG_LVL_ALT_Q);
else
vp9_disable_segfeature(xd, i, SEG_LVL_ALT_Q);
2010-05-18 19:58:33 +04:00
if (delta_lf[i])
vp9_enable_segfeature(xd, i, SEG_LVL_ALT_LF);
else
vp9_disable_segfeature(xd, i, SEG_LVL_ALT_LF);
}
2011-10-05 14:26:00 +04:00
// Initialise the feature data structure
// SEGMENT_DELTADATA 0, SEGMENT_ABSDATA 1
vp9_set_segment_data((VP9_PTR)cpi, &feature_data[0][0], SEGMENT_DELTADATA);
2010-05-18 19:58:33 +04:00
return 0;
2010-05-18 19:58:33 +04:00
}
int vp9_set_active_map(VP9_PTR comp, unsigned char *map,
unsigned int rows, unsigned int cols) {
VP9_COMP *cpi = (VP9_COMP *) comp;
2010-05-18 19:58:33 +04:00
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;
2010-05-18 19:58:33 +04:00
return 0;
} else {
// cpi->active_map_enabled = 0;
return -1;
}
2010-05-18 19:58:33 +04:00
}
int vp9_set_internal_size(VP9_PTR comp,
VPX_SCALING horiz_mode, VPX_SCALING vert_mode) {
VP9_COMP *cpi = (VP9_COMP *) comp;
2010-05-18 19:58:33 +04:00
if (horiz_mode <= ONETWO)
cpi->horiz_scale = horiz_mode;
else
return -1;
2010-05-18 19:58:33 +04:00
if (vert_mode <= ONETWO)
cpi->vert_scale = vert_mode;
else
return -1;
2010-05-18 19:58:33 +04:00
vp9_change_config(comp, &cpi->oxcf);
return 0;
2010-05-18 19:58:33 +04:00
}
int vp9_calc_ss_err(YV12_BUFFER_CONFIG *source, YV12_BUFFER_CONFIG *dest) {
int i, j;
int Total = 0;
2010-05-18 19:58:33 +04:00
uint8_t *src = source->y_buffer;
uint8_t *dst = dest->y_buffer;
2010-05-18 19:58:33 +04:00
// 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 += vp9_mse16x16(src + j, source->y_stride, dst + j, dest->y_stride,
&sse);
2010-05-18 19:58:33 +04:00
}
src += 16 * source->y_stride;
dst += 16 * dest->y_stride;
}
return Total;
2010-05-18 19:58:33 +04:00
}
int vp9_get_quantizer(VP9_PTR c) {
VP9_COMP *cpi = (VP9_COMP *) c;
return cpi->common.base_qindex;
2010-05-18 19:58:33 +04:00
}