4e692bbee2
This change has been imported from VP9 and alters the nature and use of exhaustive motion search. Firstly any exhaustive search is preceded by a normal step search. The exhaustive search is only carried out if the distortion resulting from the step search is above a threshold value. Secondly the simple +/- 64 exhaustive search is replaced by a multi stage mesh based search where each stage has a range and step/interval size. Subsequent stages use the best position from the previous stage as the center of the search but use a reduced range and interval size. For example: stage 1: Range +/- 64 interval 4 stage 2: Range +/- 32 interval 2 stage 3: Range +/- 15 interval 1 This process, especially when it follows on from a normal step search, has shown itself to be almost as effective as a full range exhaustive search with step 1 but greatly lowers the computational complexity such that it can be used in some cases for speeds 0-2. This patch also removes a double exhaustive search for sub 8x8 blocks which also contained a bug (the two searches used different distortion metrics). For best quality in my test animation sequence this patch has almost no impact on quality but improves encode speed by more than 5X. Restricted use in good quality speeds 0-2 yields significant quality gains on the animation test of 0.2 - 0.5 db with only a small impact on encode speed. On most natural video clips, however, where the step search is performing well, the quality gain and speed impact are small. Change-Id: Iac24152ae239f42a246f39ee5f00fe62d193cb98
144 lines
3.5 KiB
C
144 lines
3.5 KiB
C
/*
|
|
* Copyright (c) 2010 The WebM project authors. All Rights Reserved.
|
|
*
|
|
* Use of this source code is governed by a BSD-style license
|
|
* that can be found in the LICENSE file in the root of the source
|
|
* tree. An additional intellectual property rights grant can be found
|
|
* in the file PATENTS. All contributing project authors may
|
|
* be found in the AUTHORS file in the root of the source tree.
|
|
*/
|
|
|
|
#ifndef VP10_ENCODER_BLOCK_H_
|
|
#define VP10_ENCODER_BLOCK_H_
|
|
|
|
#include "vp10/common/entropymv.h"
|
|
#include "vp10/common/entropy.h"
|
|
|
|
#ifdef __cplusplus
|
|
extern "C" {
|
|
#endif
|
|
|
|
typedef struct {
|
|
unsigned int sse;
|
|
int sum;
|
|
unsigned int var;
|
|
} diff;
|
|
|
|
struct macroblock_plane {
|
|
DECLARE_ALIGNED(16, int16_t, src_diff[64 * 64]);
|
|
tran_low_t *qcoeff;
|
|
tran_low_t *coeff;
|
|
uint16_t *eobs;
|
|
struct buf_2d src;
|
|
|
|
// Quantizer setings
|
|
int16_t *quant_fp;
|
|
int16_t *round_fp;
|
|
int16_t *quant;
|
|
int16_t *quant_shift;
|
|
int16_t *zbin;
|
|
int16_t *round;
|
|
|
|
int64_t quant_thred[2];
|
|
};
|
|
|
|
/* The [2] dimension is for whether we skip the EOB node (i.e. if previous
|
|
* coefficient in this block was zero) or not. */
|
|
typedef unsigned int vp10_coeff_cost[PLANE_TYPES][REF_TYPES][COEF_BANDS][2]
|
|
[COEFF_CONTEXTS][ENTROPY_TOKENS];
|
|
|
|
typedef struct {
|
|
int_mv ref_mvs[MAX_REF_FRAMES][MAX_MV_REF_CANDIDATES];
|
|
uint8_t mode_context[MAX_REF_FRAMES];
|
|
} MB_MODE_INFO_EXT;
|
|
|
|
typedef struct macroblock MACROBLOCK;
|
|
struct macroblock {
|
|
struct macroblock_plane plane[MAX_MB_PLANE];
|
|
|
|
MACROBLOCKD e_mbd;
|
|
MB_MODE_INFO_EXT *mbmi_ext;
|
|
int skip_block;
|
|
int select_tx_size;
|
|
int skip_recode;
|
|
int skip_optimize;
|
|
int q_index;
|
|
|
|
int errorperbit;
|
|
int sadperbit16;
|
|
int sadperbit4;
|
|
int rddiv;
|
|
int rdmult;
|
|
int mb_energy;
|
|
int * m_search_count_ptr;
|
|
int * ex_search_count_ptr;
|
|
|
|
// These are set to their default values at the beginning, and then adjusted
|
|
// further in the encoding process.
|
|
BLOCK_SIZE min_partition_size;
|
|
BLOCK_SIZE max_partition_size;
|
|
|
|
int mv_best_ref_index[MAX_REF_FRAMES];
|
|
unsigned int max_mv_context[MAX_REF_FRAMES];
|
|
unsigned int source_variance;
|
|
unsigned int pred_sse[MAX_REF_FRAMES];
|
|
int pred_mv_sad[MAX_REF_FRAMES];
|
|
|
|
int nmvjointcost[MV_JOINTS];
|
|
int *nmvcost[2];
|
|
int *nmvcost_hp[2];
|
|
int **mvcost;
|
|
|
|
int nmvjointsadcost[MV_JOINTS];
|
|
int *nmvsadcost[2];
|
|
int *nmvsadcost_hp[2];
|
|
int **mvsadcost;
|
|
|
|
// These define limits to motion vector components to prevent them
|
|
// from extending outside the UMV borders
|
|
int mv_col_min;
|
|
int mv_col_max;
|
|
int mv_row_min;
|
|
int mv_row_max;
|
|
|
|
// Notes transform blocks where no coefficents are coded.
|
|
// Set during mode selection. Read during block encoding.
|
|
uint8_t zcoeff_blk[TX_SIZES][256];
|
|
|
|
int skip;
|
|
|
|
int encode_breakout;
|
|
|
|
// note that token_costs is the cost when eob node is skipped
|
|
vp10_coeff_cost token_costs[TX_SIZES];
|
|
|
|
int optimize;
|
|
|
|
// indicate if it is in the rd search loop or encoding process
|
|
int use_lp32x32fdct;
|
|
|
|
// use fast quantization process
|
|
int quant_fp;
|
|
|
|
// skip forward transform and quantization
|
|
uint8_t skip_txfm[MAX_MB_PLANE << 2];
|
|
#define SKIP_TXFM_NONE 0
|
|
#define SKIP_TXFM_AC_DC 1
|
|
#define SKIP_TXFM_AC_ONLY 2
|
|
|
|
int64_t bsse[MAX_MB_PLANE << 2];
|
|
|
|
// Used to store sub partition's choices.
|
|
MV pred_mv[MAX_REF_FRAMES];
|
|
|
|
// Strong color activity detection. Used in RTC coding mode to enhance
|
|
// the visual quality at the boundary of moving color objects.
|
|
uint8_t color_sensitivity[2];
|
|
};
|
|
|
|
#ifdef __cplusplus
|
|
} // extern "C"
|
|
#endif
|
|
|
|
#endif // VP10_ENCODER_BLOCK_H_
|