vpx/vp9/common/vp9_reconintra.c

446 lines
16 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
*/
#include "./vpx_config.h"
#include "./vpx_dsp_rtcd.h"
#if CONFIG_VP9_HIGHBITDEPTH
#include "vpx_dsp/vpx_dsp_common.h"
#endif // CONFIG_VP9_HIGHBITDEPTH
#include "vpx_mem/vpx_mem.h"
#include "vpx_ports/mem.h"
#include "vpx_ports/vpx_once.h"
#include "vp9/common/vp9_reconintra.h"
#include "vp9/common/vp9_onyxc_int.h"
2010-05-18 17:58:33 +02:00
const TX_TYPE intra_mode_to_tx_type_lookup[INTRA_MODES] = {
DCT_DCT, // DC
ADST_DCT, // V
DCT_ADST, // H
DCT_DCT, // D45
ADST_ADST, // D135
ADST_DCT, // D117
DCT_ADST, // D153
DCT_ADST, // D207
ADST_DCT, // D63
ADST_ADST, // TM
};
enum {
NEED_LEFT = 1 << 1,
NEED_ABOVE = 1 << 2,
NEED_ABOVERIGHT = 1 << 3,
};
static const uint8_t extend_modes[INTRA_MODES] = {
NEED_ABOVE | NEED_LEFT, // DC
NEED_ABOVE, // V
NEED_LEFT, // H
NEED_ABOVERIGHT, // D45
NEED_LEFT | NEED_ABOVE, // D135
NEED_LEFT | NEED_ABOVE, // D117
NEED_LEFT | NEED_ABOVE, // D153
NEED_LEFT, // D207
NEED_ABOVERIGHT, // D63
NEED_LEFT | NEED_ABOVE, // TM
};
typedef void (*intra_pred_fn)(uint8_t *dst, ptrdiff_t stride,
const uint8_t *above, const uint8_t *left);
static intra_pred_fn pred[INTRA_MODES][TX_SIZES];
static intra_pred_fn dc_pred[2][2][TX_SIZES];
#if CONFIG_VP9_HIGHBITDEPTH
typedef void (*intra_high_pred_fn)(uint16_t *dst, ptrdiff_t stride,
const uint16_t *above, const uint16_t *left,
int bd);
static intra_high_pred_fn pred_high[INTRA_MODES][4];
static intra_high_pred_fn dc_pred_high[2][2][4];
#endif // CONFIG_VP9_HIGHBITDEPTH
static void vp9_init_intra_predictors_internal(void) {
#define INIT_ALL_SIZES(p, type) \
p[TX_4X4] = vpx_##type##_predictor_4x4; \
p[TX_8X8] = vpx_##type##_predictor_8x8; \
p[TX_16X16] = vpx_##type##_predictor_16x16; \
p[TX_32X32] = vpx_##type##_predictor_32x32
INIT_ALL_SIZES(pred[V_PRED], v);
INIT_ALL_SIZES(pred[H_PRED], h);
INIT_ALL_SIZES(pred[D207_PRED], d207);
INIT_ALL_SIZES(pred[D45_PRED], d45);
INIT_ALL_SIZES(pred[D63_PRED], d63);
INIT_ALL_SIZES(pred[D117_PRED], d117);
INIT_ALL_SIZES(pred[D135_PRED], d135);
INIT_ALL_SIZES(pred[D153_PRED], d153);
INIT_ALL_SIZES(pred[TM_PRED], tm);
INIT_ALL_SIZES(dc_pred[0][0], dc_128);
INIT_ALL_SIZES(dc_pred[0][1], dc_top);
INIT_ALL_SIZES(dc_pred[1][0], dc_left);
INIT_ALL_SIZES(dc_pred[1][1], dc);
#if CONFIG_VP9_HIGHBITDEPTH
INIT_ALL_SIZES(pred_high[V_PRED], highbd_v);
INIT_ALL_SIZES(pred_high[H_PRED], highbd_h);
INIT_ALL_SIZES(pred_high[D207_PRED], highbd_d207);
INIT_ALL_SIZES(pred_high[D45_PRED], highbd_d45);
INIT_ALL_SIZES(pred_high[D63_PRED], highbd_d63);
INIT_ALL_SIZES(pred_high[D117_PRED], highbd_d117);
INIT_ALL_SIZES(pred_high[D135_PRED], highbd_d135);
INIT_ALL_SIZES(pred_high[D153_PRED], highbd_d153);
INIT_ALL_SIZES(pred_high[TM_PRED], highbd_tm);
INIT_ALL_SIZES(dc_pred_high[0][0], highbd_dc_128);
INIT_ALL_SIZES(dc_pred_high[0][1], highbd_dc_top);
INIT_ALL_SIZES(dc_pred_high[1][0], highbd_dc_left);
INIT_ALL_SIZES(dc_pred_high[1][1], highbd_dc);
#endif // CONFIG_VP9_HIGHBITDEPTH
#undef intra_pred_allsizes
}
#if CONFIG_VP9_HIGHBITDEPTH
static void build_intra_predictors_high(const MACROBLOCKD *xd,
const uint8_t *ref8,
int ref_stride,
uint8_t *dst8,
int dst_stride,
PREDICTION_MODE mode,
TX_SIZE tx_size,
int up_available,
int left_available,
int right_available,
int x, int y,
int plane, int bd) {
int i;
uint16_t *dst = CONVERT_TO_SHORTPTR(dst8);
uint16_t *ref = CONVERT_TO_SHORTPTR(ref8);
DECLARE_ALIGNED(16, uint16_t, left_col[32]);
DECLARE_ALIGNED(16, uint16_t, above_data[64 + 16]);
uint16_t *above_row = above_data + 16;
const uint16_t *const_above_row = above_row;
const int bs = 4 << tx_size;
int frame_width, frame_height;
int x0, y0;
const struct macroblockd_plane *const pd = &xd->plane[plane];
const int need_left = extend_modes[mode] & NEED_LEFT;
const int need_above = extend_modes[mode] & NEED_ABOVE;
const int need_aboveright = extend_modes[mode] & NEED_ABOVERIGHT;
int base = 128 << (bd - 8);
// 127 127 127 .. 127 127 127 127 127 127
// 129 A B .. Y Z
// 129 C D .. W X
// 129 E F .. U V
// 129 G H .. S T T T T T
// For 10 bit and 12 bit, 127 and 129 are replaced by base -1 and base + 1.
// Get current frame pointer, width and height.
if (plane == 0) {
frame_width = xd->cur_buf->y_width;
frame_height = xd->cur_buf->y_height;
} else {
frame_width = xd->cur_buf->uv_width;
frame_height = xd->cur_buf->uv_height;
}
// Get block position in current frame.
x0 = (-xd->mb_to_left_edge >> (3 + pd->subsampling_x)) + x;
y0 = (-xd->mb_to_top_edge >> (3 + pd->subsampling_y)) + y;
// NEED_LEFT
if (need_left) {
if (left_available) {
if (xd->mb_to_bottom_edge < 0) {
/* slower path if the block needs border extension */
if (y0 + bs <= frame_height) {
for (i = 0; i < bs; ++i)
left_col[i] = ref[i * ref_stride - 1];
} else {
const int extend_bottom = frame_height - y0;
for (i = 0; i < extend_bottom; ++i)
left_col[i] = ref[i * ref_stride - 1];
for (; i < bs; ++i)
left_col[i] = ref[(extend_bottom - 1) * ref_stride - 1];
}
} else {
/* faster path if the block does not need extension */
for (i = 0; i < bs; ++i)
left_col[i] = ref[i * ref_stride - 1];
}
} else {
vpx_memset16(left_col, base + 1, bs);
}
}
// NEED_ABOVE
if (need_above) {
if (up_available) {
const uint16_t *above_ref = ref - ref_stride;
if (xd->mb_to_right_edge < 0) {
/* slower path if the block needs border extension */
if (x0 + bs <= frame_width) {
memcpy(above_row, above_ref, bs * sizeof(above_row[0]));
} else if (x0 <= frame_width) {
const int r = frame_width - x0;
memcpy(above_row, above_ref, r * sizeof(above_row[0]));
vpx_memset16(above_row + r, above_row[r - 1], x0 + bs - frame_width);
}
} else {
/* faster path if the block does not need extension */
if (bs == 4 && right_available && left_available) {
const_above_row = above_ref;
} else {
memcpy(above_row, above_ref, bs * sizeof(above_row[0]));
}
}
above_row[-1] = left_available ? above_ref[-1] : (base + 1);
} else {
vpx_memset16(above_row, base - 1, bs);
above_row[-1] = base - 1;
}
}
// NEED_ABOVERIGHT
if (need_aboveright) {
if (up_available) {
const uint16_t *above_ref = ref - ref_stride;
if (xd->mb_to_right_edge < 0) {
/* slower path if the block needs border extension */
if (x0 + 2 * bs <= frame_width) {
if (right_available && bs == 4) {
memcpy(above_row, above_ref, 2 * bs * sizeof(above_row[0]));
} else {
memcpy(above_row, above_ref, bs * sizeof(above_row[0]));
vpx_memset16(above_row + bs, above_row[bs - 1], bs);
}
} else if (x0 + bs <= frame_width) {
const int r = frame_width - x0;
if (right_available && bs == 4) {
memcpy(above_row, above_ref, r * sizeof(above_row[0]));
vpx_memset16(above_row + r, above_row[r - 1],
x0 + 2 * bs - frame_width);
} else {
memcpy(above_row, above_ref, bs * sizeof(above_row[0]));
vpx_memset16(above_row + bs, above_row[bs - 1], bs);
}
} else if (x0 <= frame_width) {
const int r = frame_width - x0;
memcpy(above_row, above_ref, r * sizeof(above_row[0]));
vpx_memset16(above_row + r, above_row[r - 1],
x0 + 2 * bs - frame_width);
}
above_row[-1] = left_available ? above_ref[-1] : (base + 1);
} else {
/* faster path if the block does not need extension */
if (bs == 4 && right_available && left_available) {
const_above_row = above_ref;
} else {
memcpy(above_row, above_ref, bs * sizeof(above_row[0]));
if (bs == 4 && right_available)
memcpy(above_row + bs, above_ref + bs, bs * sizeof(above_row[0]));
else
vpx_memset16(above_row + bs, above_row[bs - 1], bs);
above_row[-1] = left_available ? above_ref[-1] : (base + 1);
}
}
} else {
vpx_memset16(above_row, base - 1, bs * 2);
above_row[-1] = base - 1;
}
}
// predict
if (mode == DC_PRED) {
dc_pred_high[left_available][up_available][tx_size](dst, dst_stride,
const_above_row,
left_col, xd->bd);
} else {
pred_high[mode][tx_size](dst, dst_stride, const_above_row, left_col,
xd->bd);
}
}
#endif // CONFIG_VP9_HIGHBITDEPTH
static void build_intra_predictors(const MACROBLOCKD *xd, const uint8_t *ref,
int ref_stride, uint8_t *dst, int dst_stride,
PREDICTION_MODE mode, TX_SIZE tx_size,
int up_available, int left_available,
int right_available, int x, int y,
int plane) {
int i;
DECLARE_ALIGNED(16, uint8_t, left_col[32]);
DECLARE_ALIGNED(16, uint8_t, above_data[64 + 16]);
uint8_t *above_row = above_data + 16;
const uint8_t *const_above_row = above_row;
const int bs = 4 << tx_size;
int frame_width, frame_height;
int x0, y0;
const struct macroblockd_plane *const pd = &xd->plane[plane];
// 127 127 127 .. 127 127 127 127 127 127
// 129 A B .. Y Z
// 129 C D .. W X
// 129 E F .. U V
// 129 G H .. S T T T T T
// ..
// Get current frame pointer, width and height.
if (plane == 0) {
frame_width = xd->cur_buf->y_width;
frame_height = xd->cur_buf->y_height;
} else {
frame_width = xd->cur_buf->uv_width;
frame_height = xd->cur_buf->uv_height;
}
// Get block position in current frame.
x0 = (-xd->mb_to_left_edge >> (3 + pd->subsampling_x)) + x;
y0 = (-xd->mb_to_top_edge >> (3 + pd->subsampling_y)) + y;
// NEED_LEFT
if (extend_modes[mode] & NEED_LEFT) {
if (left_available) {
if (xd->mb_to_bottom_edge < 0) {
/* slower path if the block needs border extension */
if (y0 + bs <= frame_height) {
for (i = 0; i < bs; ++i)
left_col[i] = ref[i * ref_stride - 1];
} else {
const int extend_bottom = frame_height - y0;
for (i = 0; i < extend_bottom; ++i)
left_col[i] = ref[i * ref_stride - 1];
for (; i < bs; ++i)
left_col[i] = ref[(extend_bottom - 1) * ref_stride - 1];
}
} else {
/* faster path if the block does not need extension */
for (i = 0; i < bs; ++i)
left_col[i] = ref[i * ref_stride - 1];
}
} else {
memset(left_col, 129, bs);
}
[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 18:35:28 +01:00
}
// NEED_ABOVE
if (extend_modes[mode] & NEED_ABOVE) {
if (up_available) {
const uint8_t *above_ref = ref - ref_stride;
if (xd->mb_to_right_edge < 0) {
/* slower path if the block needs border extension */
if (x0 + bs <= frame_width) {
memcpy(above_row, above_ref, bs);
} else if (x0 <= frame_width) {
const int r = frame_width - x0;
memcpy(above_row, above_ref, r);
memset(above_row + r, above_row[r - 1], x0 + bs - frame_width);
}
} else {
/* faster path if the block does not need extension */
if (bs == 4 && right_available && left_available) {
const_above_row = above_ref;
} else {
memcpy(above_row, above_ref, bs);
}
}
above_row[-1] = left_available ? above_ref[-1] : 129;
} else {
memset(above_row, 127, bs);
above_row[-1] = 127;
}
}
// NEED_ABOVERIGHT
if (extend_modes[mode] & NEED_ABOVERIGHT) {
if (up_available) {
const uint8_t *above_ref = ref - ref_stride;
if (xd->mb_to_right_edge < 0) {
/* slower path if the block needs border extension */
if (x0 + 2 * bs <= frame_width) {
if (right_available && bs == 4) {
memcpy(above_row, above_ref, 2 * bs);
} else {
memcpy(above_row, above_ref, bs);
memset(above_row + bs, above_row[bs - 1], bs);
}
} else if (x0 + bs <= frame_width) {
const int r = frame_width - x0;
if (right_available && bs == 4) {
memcpy(above_row, above_ref, r);
memset(above_row + r, above_row[r - 1], x0 + 2 * bs - frame_width);
} else {
memcpy(above_row, above_ref, bs);
memset(above_row + bs, above_row[bs - 1], bs);
}
} else if (x0 <= frame_width) {
const int r = frame_width - x0;
memcpy(above_row, above_ref, r);
memset(above_row + r, above_row[r - 1], x0 + 2 * bs - frame_width);
}
} else {
/* faster path if the block does not need extension */
if (bs == 4 && right_available && left_available) {
const_above_row = above_ref;
} else {
memcpy(above_row, above_ref, bs);
if (bs == 4 && right_available)
memcpy(above_row + bs, above_ref + bs, bs);
else
memset(above_row + bs, above_row[bs - 1], bs);
}
}
above_row[-1] = left_available ? above_ref[-1] : 129;
} else {
memset(above_row, 127, bs * 2);
above_row[-1] = 127;
}
}
// predict
if (mode == DC_PRED) {
dc_pred[left_available][up_available][tx_size](dst, dst_stride,
const_above_row, left_col);
} else {
pred[mode][tx_size](dst, dst_stride, const_above_row, left_col);
}
2010-05-18 17:58:33 +02:00
}
void vp9_predict_intra_block(const MACROBLOCKD *xd, int bwl_in,
TX_SIZE tx_size, PREDICTION_MODE mode,
const uint8_t *ref, int ref_stride,
uint8_t *dst, int dst_stride,
int aoff, int loff, int plane) {
const int bw = (1 << bwl_in);
const int txw = (1 << tx_size);
const int have_top = loff || (xd->above_mi != NULL);
const int have_left = aoff || (xd->left_mi != NULL);
const int have_right = (aoff + txw) < bw;
const int x = aoff * 4;
const int y = loff * 4;
#if CONFIG_VP9_HIGHBITDEPTH
if (xd->cur_buf->flags & YV12_FLAG_HIGHBITDEPTH) {
build_intra_predictors_high(xd, ref, ref_stride, dst, dst_stride, mode,
tx_size, have_top, have_left, have_right,
x, y, plane, xd->bd);
return;
}
#endif
build_intra_predictors(xd, ref, ref_stride, dst, dst_stride, mode, tx_size,
have_top, have_left, have_right, x, y, plane);
}
void vp9_init_intra_predictors(void) {
once(vp9_init_intra_predictors_internal);
}