vpx/vp9/common/blockd.h

518 lines
14 KiB
C
Raw Normal View History

2010-05-18 17:58:33 +02:00
/*
* Copyright (c) 2010 The WebM project authors. All Rights Reserved.
2010-05-18 17:58:33 +02:00
*
* 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 17:58:33 +02:00
*/
#ifndef __INC_BLOCKD_H
#define __INC_BLOCKD_H
void vpx_log(const char *format, ...);
#include "vpx_ports/config.h"
#include "vpx_scale/yv12config.h"
#include "mv.h"
#include "treecoder.h"
#include "subpixel.h"
#include "vpx_ports/mem.h"
#include "common.h"
2010-05-18 17:58:33 +02:00
#define TRUE 1
#define FALSE 0
// #define MODE_STATS
/*#define DCPRED 1*/
2010-05-18 17:58:33 +02:00
#define DCPREDSIMTHRESH 0
#define DCPREDCNTTHRESH 3
#define MB_FEATURE_TREE_PROBS 3
#define PREDICTION_PROBS 3
#define MBSKIP_CONTEXTS 3
2010-05-18 17:58:33 +02:00
#define MAX_MB_SEGMENTS 4
#define MAX_REF_LF_DELTAS 4
#define MAX_MODE_LF_DELTAS 4
/* Segment Feature Masks */
2010-05-18 17:58:33 +02:00
#define SEGMENT_DELTADATA 0
#define SEGMENT_ABSDATA 1
#define MAX_MV_REFS 19
2010-05-18 17:58:33 +02:00
typedef struct {
int r, c;
2010-05-18 17:58:33 +02:00
} POS;
typedef enum PlaneType {
PLANE_TYPE_Y_NO_DC = 0,
PLANE_TYPE_Y2,
PLANE_TYPE_UV,
PLANE_TYPE_Y_WITH_DC,
} PLANE_TYPE;
typedef char ENTROPY_CONTEXT;
typedef struct {
ENTROPY_CONTEXT y1[4];
ENTROPY_CONTEXT u[2];
ENTROPY_CONTEXT v[2];
ENTROPY_CONTEXT y2;
} ENTROPY_CONTEXT_PLANES;
2010-05-18 17:58:33 +02:00
extern const unsigned char vp9_block2left[25];
extern const unsigned char vp9_block2above[25];
extern const unsigned char vp9_block2left_8x8[25];
extern const unsigned char vp9_block2above_8x8[25];
#define VP9_COMBINEENTROPYCONTEXTS( Dest, A, B) \
Dest = ((A)!=0) + ((B)!=0);
2010-05-18 17:58:33 +02:00
typedef enum {
KEY_FRAME = 0,
INTER_FRAME = 1
2010-05-18 17:58:33 +02:00
} FRAME_TYPE;
typedef enum
{
SIXTAP = 0,
BILINEAR = 1,
EIGHTTAP = 2,
EIGHTTAP_SHARP = 3,
SWITCHABLE /* should be the last one */
} INTERPOLATIONFILTERTYPE;
typedef enum
{
DC_PRED, /* average of above and left pixels */
V_PRED, /* vertical prediction */
H_PRED, /* horizontal prediction */
D45_PRED, /* Directional 45 deg prediction [anti-clockwise from 0 deg hor] */
D135_PRED, /* Directional 135 deg prediction [anti-clockwise from 0 deg hor] */
D117_PRED, /* Directional 112 deg prediction [anti-clockwise from 0 deg hor] */
D153_PRED, /* Directional 157 deg prediction [anti-clockwise from 0 deg hor] */
D27_PRED, /* Directional 22 deg prediction [anti-clockwise from 0 deg hor] */
D63_PRED, /* Directional 67 deg prediction [anti-clockwise from 0 deg hor] */
TM_PRED, /* Truemotion prediction */
I8X8_PRED, /* 8x8 based prediction, each 8x8 has its own prediction mode */
B_PRED, /* block based prediction, each block has its own prediction mode */
2010-05-18 17:58:33 +02:00
NEARESTMV,
NEARMV,
ZEROMV,
NEWMV,
SPLITMV,
2010-05-18 17:58:33 +02:00
MB_MODE_COUNT
2010-05-18 17:58:33 +02:00
} MB_PREDICTION_MODE;
// Segment level features.
typedef enum {
SEG_LVL_ALT_Q = 0, // Use alternate Quantizer ....
SEG_LVL_ALT_LF = 1, // Use alternate loop filter value...
SEG_LVL_REF_FRAME = 2, // Optional Segment reference frame
SEG_LVL_MODE = 3, // Optional Segment mode
SEG_LVL_EOB = 4, // EOB end stop marker.
SEG_LVL_TRANSFORM = 5, // Block transform size.
SEG_LVL_MAX = 6 // Number of MB level features supported
2010-05-18 17:58:33 +02:00
} SEG_LVL_FEATURES;
2010-05-18 17:58:33 +02:00
// Segment level features.
typedef enum {
TX_4X4, // 4x4 dct transform
TX_8X8, // 8x8 dct transform
TX_16X16, // 16x16 dct transform
TX_SIZE_MAX // Number of different transforms available
} TX_SIZE;
typedef enum {
DCT_DCT = 0, // DCT in both horizontal and vertical
ADST_DCT = 1, // ADST in vertical, DCT in horizontal
DCT_ADST = 2, // DCT in vertical, ADST in horizontal
ADST_ADST = 3 // ADST in both directions
} TX_TYPE;
#define VP9_YMODES (B_PRED + 1)
#define VP9_UV_MODES (TM_PRED + 1)
#define VP9_I8X8_MODES (TM_PRED + 1)
#define VP9_I32X32_MODES (TM_PRED + 1)
2010-05-18 17:58:33 +02:00
#define VP9_MVREFS (1 + SPLITMV - NEARESTMV)
2010-05-18 17:58:33 +02:00
typedef enum {
B_DC_PRED, /* average of above and left pixels */
B_TM_PRED,
2010-05-18 17:58:33 +02:00
B_VE_PRED, /* vertical prediction */
B_HE_PRED, /* horizontal prediction */
2010-05-18 17:58:33 +02:00
B_LD_PRED,
B_RD_PRED,
2010-05-18 17:58:33 +02:00
B_VR_PRED,
B_VL_PRED,
B_HD_PRED,
B_HU_PRED,
2010-05-18 17:58:33 +02:00
LEFT4X4,
ABOVE4X4,
ZERO4X4,
NEW4X4,
2010-05-18 17:58:33 +02:00
B_MODE_COUNT
2010-05-18 17:58:33 +02:00
} B_PREDICTION_MODE;
#define VP9_BINTRAMODES (B_HU_PRED + 1) /* 10 */
#define VP9_SUBMVREFS (1 + NEW4X4 - LEFT4X4)
2010-05-18 17:58:33 +02:00
typedef enum {
PARTITIONING_16X8 = 0,
PARTITIONING_8X16,
PARTITIONING_8X8,
PARTITIONING_4X4,
NB_PARTITIONINGS,
} SPLITMV_PARTITIONING_TYPE;
2010-05-18 17:58:33 +02:00
/* For keyframes, intra block modes are predicted by the (already decoded)
modes for the Y blocks to the left and above us; for interframes, there
is a single probability table. */
union b_mode_info {
struct {
B_PREDICTION_MODE first;
TX_TYPE tx_type;
#if CONFIG_COMP_INTRA_PRED
B_PREDICTION_MODE second;
#endif
} as_mode;
struct {
int_mv first;
int_mv second;
} as_mv;
};
2010-05-18 17:58:33 +02:00
typedef enum {
INTRA_FRAME = 0,
LAST_FRAME = 1,
GOLDEN_FRAME = 2,
ALTREF_FRAME = 3,
MAX_REF_FRAMES = 4
2010-05-18 17:58:33 +02:00
} MV_REFERENCE_FRAME;
typedef struct {
MB_PREDICTION_MODE mode, uv_mode;
#if CONFIG_COMP_INTRA_PRED
MB_PREDICTION_MODE second_mode, second_uv_mode;
#endif
MV_REFERENCE_FRAME ref_frame, second_ref_frame;
TX_SIZE txfm_size;
int_mv mv[2]; // for each reference frame used
int_mv ref_mvs[MAX_REF_FRAMES][MAX_MV_REFS];
SPLITMV_PARTITIONING_TYPE partitioning;
unsigned char mb_skip_coeff; /* does this mb has coefficients at all, 1=no coefficients, 0=need decode tokens */
unsigned char need_to_clamp_mvs;
unsigned char need_to_clamp_secondmv;
unsigned char segment_id; /* Which set of segmentation parameters should be used for this MB */
// Flags used for prediction status of various bistream signals
unsigned char seg_id_predicted;
unsigned char ref_predicted;
// Indicates if the mb is part of the image (1) vs border (0)
// This can be useful in determining whether the MB provides
// a valid predictor
unsigned char mb_in_image;
#if CONFIG_PRED_FILTER
// Flag to turn prediction signal filter on(1)/off(0 ) at the MB level
unsigned int pred_filter_enabled;
#endif
INTERPOLATIONFILTERTYPE interp_filter;
#if CONFIG_SUPERBLOCKS
// FIXME need a SB array of 4 MB_MODE_INFOs that
// only needs one encoded_as_sb.
unsigned char encoded_as_sb;
#endif
2010-05-18 17:58:33 +02:00
} MB_MODE_INFO;
typedef struct {
MB_MODE_INFO mbmi;
union b_mode_info bmi[16];
2010-05-18 17:58:33 +02:00
} MODE_INFO;
typedef struct blockd {
short *qcoeff;
short *dqcoeff;
unsigned char *predictor;
short *diff;
short *dequant;
2010-05-18 17:58:33 +02:00
/* 16 Y blocks, 4 U blocks, 4 V blocks each with 16 entries */
unsigned char **base_pre;
unsigned char **base_second_pre;
int pre;
int pre_stride;
2010-05-18 17:58:33 +02:00
unsigned char **base_dst;
int dst;
int dst_stride;
2010-05-18 17:58:33 +02:00
int eob;
2010-05-18 17:58:33 +02:00
union b_mode_info bmi;
2010-05-18 17:58:33 +02:00
} BLOCKD;
typedef struct macroblockd {
DECLARE_ALIGNED(16, short, diff[400]); /* from idct diff */
DECLARE_ALIGNED(16, unsigned char, predictor[384]);
DECLARE_ALIGNED(16, short, qcoeff[400]);
DECLARE_ALIGNED(16, short, dqcoeff[400]);
DECLARE_ALIGNED(16, unsigned short, eobs[25]);
2010-05-18 17:58:33 +02:00
/* 16 Y blocks, 4 U, 4 V, 1 DC 2nd order block, each with 16 entries. */
BLOCKD block[25];
int fullpixel_mask;
2010-05-18 17:58:33 +02:00
YV12_BUFFER_CONFIG pre; /* Filtered copy of previous frame reconstruction */
struct {
uint8_t *y_buffer, *u_buffer, *v_buffer;
} second_pre;
YV12_BUFFER_CONFIG dst;
2010-05-18 17:58:33 +02:00
MODE_INFO *prev_mode_info_context;
MODE_INFO *mode_info_context;
int mode_info_stride;
2010-05-18 17:58:33 +02:00
FRAME_TYPE frame_type;
2010-05-18 17:58:33 +02:00
int up_available;
int left_available;
2010-05-18 17:58:33 +02:00
/* Y,U,V,Y2 */
ENTROPY_CONTEXT_PLANES *above_context;
ENTROPY_CONTEXT_PLANES *left_context;
2010-05-18 17:58:33 +02:00
/* 0 indicates segmentation at MB level is not enabled. Otherwise the individual bits indicate which features are active. */
unsigned char segmentation_enabled;
2010-05-18 17:58:33 +02:00
/* 0 (do not update) 1 (update) the macroblock segmentation map. */
unsigned char update_mb_segmentation_map;
2010-05-18 17:58:33 +02:00
/* 0 (do not update) 1 (update) the macroblock segmentation feature data. */
unsigned char update_mb_segmentation_data;
2010-05-18 17:58:33 +02:00
/* 0 (do not update) 1 (update) the macroblock segmentation feature data. */
unsigned char mb_segment_abs_delta;
2010-05-18 17:58:33 +02:00
/* Per frame flags that define which MB level features (such as quantizer or loop filter level) */
/* are enabled and when enabled the proabilities used to decode the per MB flags in MB_MODE_INFO */
Further work on Segmentation Experiment: This check in includes quite a lot of clean up and refactoring. Most of the analysis and set up for the different coding options for the segment map (currently simple distribution based coding or temporaly predicted coding), has been moved to one location (the function choose_segmap_coding_method() in segmenation.c). This code was previously scattered around in various locations making integration with other experiments and modification / debug more difficult. Currently the functionality is as it was with the exception that the prediction probabilities are now only transmitted when the temporal prediction mode is selected. There is still quite a bit more clean up work that will be possible when the #ifdef is removed. Also at that time I may rename and alter the sense of macroblock based variable "segment_flag" which indicates (1 that the segmnet id is not predicted vs 0 that it is predicted). I also intend to experiment with a spatial prediction mode that can be used when coding a key frame segment map or in cases where temporal prediction does not work well but there is spatial correlation. In a later check in when the ifdefs have gone I may also move the call to choose_segmap_coding_method() to just before where the bitsream is packed (currently it is in vp8_encode_frame()) to further reduce the possibility of clashes with other experiments and prevent it being called on each itteration of the recode loop. Change-Id: I3d4aba2a2826ec21f367678d5b07c1d1c36db168
2011-11-15 12:13:33 +01:00
// Probability Tree used to code Segment number
vp9_prob mb_segment_tree_probs[MB_FEATURE_TREE_PROBS];
Further work on Segmentation Experiment: This check in includes quite a lot of clean up and refactoring. Most of the analysis and set up for the different coding options for the segment map (currently simple distribution based coding or temporaly predicted coding), has been moved to one location (the function choose_segmap_coding_method() in segmenation.c). This code was previously scattered around in various locations making integration with other experiments and modification / debug more difficult. Currently the functionality is as it was with the exception that the prediction probabilities are now only transmitted when the temporal prediction mode is selected. There is still quite a bit more clean up work that will be possible when the #ifdef is removed. Also at that time I may rename and alter the sense of macroblock based variable "segment_flag" which indicates (1 that the segmnet id is not predicted vs 0 that it is predicted). I also intend to experiment with a spatial prediction mode that can be used when coding a key frame segment map or in cases where temporal prediction does not work well but there is spatial correlation. In a later check in when the ifdefs have gone I may also move the call to choose_segmap_coding_method() to just before where the bitsream is packed (currently it is in vp8_encode_frame()) to further reduce the possibility of clashes with other experiments and prevent it being called on each itteration of the recode loop. Change-Id: I3d4aba2a2826ec21f367678d5b07c1d1c36db168
2011-11-15 12:13:33 +01:00
#if CONFIG_NEW_MVREF
vp9_prob mb_mv_ref_id_probs[MAX_REF_FRAMES][3];
#endif
// Segment features
signed char segment_feature_data[MAX_MB_SEGMENTS][SEG_LVL_MAX];
unsigned int segment_feature_mask[MAX_MB_SEGMENTS];
2010-05-18 17:58:33 +02:00
/* mode_based Loop filter adjustment */
unsigned char mode_ref_lf_delta_enabled;
unsigned char mode_ref_lf_delta_update;
/* Delta values have the range +/- MAX_LOOP_FILTER */
signed char last_ref_lf_deltas[MAX_REF_LF_DELTAS]; /* 0 = Intra, Last, GF, ARF */
signed char ref_lf_deltas[MAX_REF_LF_DELTAS]; /* 0 = Intra, Last, GF, ARF */
signed char last_mode_lf_deltas[MAX_MODE_LF_DELTAS]; /* 0 = BPRED, ZERO_MV, MV, SPLIT */
signed char mode_lf_deltas[MAX_MODE_LF_DELTAS]; /* 0 = BPRED, ZERO_MV, MV, SPLIT */
/* Distance of MB away from frame edges */
int mb_to_left_edge;
int mb_to_right_edge;
int mb_to_top_edge;
int mb_to_bottom_edge;
unsigned int frames_since_golden;
unsigned int frames_till_alt_ref_frame;
vp9_subpix_fn_t subpixel_predict;
vp9_subpix_fn_t subpixel_predict8x4;
vp9_subpix_fn_t subpixel_predict8x8;
vp9_subpix_fn_t subpixel_predict16x16;
vp9_subpix_fn_t subpixel_predict_avg;
vp9_subpix_fn_t subpixel_predict_avg8x4;
vp9_subpix_fn_t subpixel_predict_avg8x8;
vp9_subpix_fn_t subpixel_predict_avg16x16;
int allow_high_precision_mv;
2010-05-18 17:58:33 +02:00
int corrupted;
#if !CONFIG_SUPERBLOCKS && (ARCH_X86 || ARCH_X86_64)
/* This is an intermediate buffer currently used in sub-pixel motion search
* to keep a copy of the reference area. This buffer can be used for other
* purpose.
*/
DECLARE_ALIGNED(32, unsigned char, y_buf[22 * 32]);
#endif
2010-05-18 17:58:33 +02:00
#if CONFIG_RUNTIME_CPU_DETECT
struct VP9_COMMON_RTCD *rtcd;
2010-05-18 17:58:33 +02:00
#endif
int mb_index; // Index of the MB in the SB (0..3)
int q_index;
2010-05-18 17:58:33 +02:00
} MACROBLOCKD;
#define ACTIVE_HT 110 // quantization stepsize threshold
#define ACTIVE_HT8 300
#define ACTIVE_HT16 300
// convert MB_PREDICTION_MODE to B_PREDICTION_MODE
static B_PREDICTION_MODE pred_mode_conv(MB_PREDICTION_MODE mode) {
B_PREDICTION_MODE b_mode;
switch (mode) {
case DC_PRED:
b_mode = B_DC_PRED;
break;
case V_PRED:
b_mode = B_VE_PRED;
break;
case H_PRED:
b_mode = B_HE_PRED;
break;
case TM_PRED:
b_mode = B_TM_PRED;
break;
case D45_PRED:
b_mode = B_LD_PRED;
break;
case D135_PRED:
b_mode = B_RD_PRED;
break;
case D117_PRED:
b_mode = B_VR_PRED;
break;
case D153_PRED:
b_mode = B_HD_PRED;
break;
case D27_PRED:
b_mode = B_HU_PRED;
break;
case D63_PRED:
b_mode = B_VL_PRED;
break;
default :
// for debug purpose, to be removed after full testing
assert(0);
break;
}
return b_mode;
}
// transform mapping
static TX_TYPE txfm_map(B_PREDICTION_MODE bmode) {
// map transform type
TX_TYPE tx_type;
switch (bmode) {
case B_TM_PRED :
case B_RD_PRED :
tx_type = ADST_ADST;
break;
case B_VE_PRED :
case B_VR_PRED :
tx_type = ADST_DCT;
break;
case B_HE_PRED :
case B_HD_PRED :
case B_HU_PRED :
tx_type = DCT_ADST;
break;
default :
tx_type = DCT_DCT;
break;
}
return tx_type;
}
static TX_TYPE get_tx_type_4x4(const MACROBLOCKD *xd, const BLOCKD *b) {
TX_TYPE tx_type = DCT_DCT;
if (xd->mode_info_context->mbmi.mode == B_PRED &&
xd->q_index < ACTIVE_HT) {
tx_type = txfm_map(b->bmi.as_mode.first);
}
return tx_type;
}
static TX_TYPE get_tx_type_8x8(const MACROBLOCKD *xd, const BLOCKD *b) {
TX_TYPE tx_type = DCT_DCT;
if (xd->mode_info_context->mbmi.mode == I8X8_PRED &&
xd->q_index < ACTIVE_HT8) {
// TODO(rbultje): MB_PREDICTION_MODE / B_PREDICTION_MODE should be merged
// or the relationship otherwise modified to address this type conversion.
tx_type = txfm_map(pred_mode_conv(
(MB_PREDICTION_MODE)b->bmi.as_mode.first));
}
return tx_type;
}
static TX_TYPE get_tx_type_16x16(const MACROBLOCKD *xd, const BLOCKD *b) {
TX_TYPE tx_type = DCT_DCT;
if (xd->mode_info_context->mbmi.mode < I8X8_PRED &&
xd->q_index < ACTIVE_HT16) {
tx_type = txfm_map(pred_mode_conv(xd->mode_info_context->mbmi.mode));
}
return tx_type;
}
static TX_TYPE get_tx_type(const MACROBLOCKD *xd, const BLOCKD *b) {
TX_TYPE tx_type = DCT_DCT;
int ib = (int)(b - xd->block);
if (ib >= 16)
return tx_type;
if (xd->mode_info_context->mbmi.txfm_size == TX_16X16) {
tx_type = get_tx_type_16x16(xd, b);
}
if (xd->mode_info_context->mbmi.txfm_size == TX_8X8) {
ib = (ib & 8) + ((ib & 4) >> 1);
tx_type = get_tx_type_8x8(xd, &xd->block[ib]);
}
if (xd->mode_info_context->mbmi.txfm_size == TX_4X4) {
tx_type = get_tx_type_4x4(xd, b);
}
return tx_type;
}
2010-05-18 17:58:33 +02:00
extern void vp9_build_block_doffsets(MACROBLOCKD *xd);
extern void vp9_setup_block_dptrs(MACROBLOCKD *xd);
2010-05-18 17:58:33 +02:00
static void update_blockd_bmi(MACROBLOCKD *xd) {
int i;
int is_4x4;
is_4x4 = (xd->mode_info_context->mbmi.mode == SPLITMV) ||
(xd->mode_info_context->mbmi.mode == I8X8_PRED) ||
(xd->mode_info_context->mbmi.mode == B_PRED);
if (is_4x4) {
for (i = 0; i < 16; i++) {
xd->block[i].bmi = xd->mode_info_context->bmi[i];
}
}
}
2010-05-18 17:58:33 +02:00
#endif /* __INC_BLOCKD_H */