3119 lines
114 KiB
C
3119 lines
114 KiB
C
/*
|
|
* HEVC video Decoder
|
|
*
|
|
* Copyright (C) 2012 - 2013 Guillaume Martres
|
|
* Copyright (C) 2012 - 2013 Mickael Raulet
|
|
* Copyright (C) 2012 - 2013 Gildas Cocherel
|
|
* Copyright (C) 2012 - 2013 Wassim Hamidouche
|
|
*
|
|
* This file is part of FFmpeg.
|
|
*
|
|
* FFmpeg is free software; you can redistribute it and/or
|
|
* modify it under the terms of the GNU Lesser General Public
|
|
* License as published by the Free Software Foundation; either
|
|
* version 2.1 of the License, or (at your option) any later version.
|
|
*
|
|
* FFmpeg is distributed in the hope that it will be useful,
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|
* Lesser General Public License for more details.
|
|
*
|
|
* You should have received a copy of the GNU Lesser General Public
|
|
* License along with FFmpeg; if not, write to the Free Software
|
|
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
|
|
*/
|
|
|
|
#include "libavutil/attributes.h"
|
|
#include "libavutil/common.h"
|
|
#include "libavutil/internal.h"
|
|
#include "libavutil/md5.h"
|
|
#include "libavutil/opt.h"
|
|
#include "libavutil/pixdesc.h"
|
|
|
|
#include "bytestream.h"
|
|
#include "cabac_functions.h"
|
|
#include "dsputil.h"
|
|
#include "golomb.h"
|
|
#include "hevc.h"
|
|
|
|
const uint8_t ff_hevc_qpel_extra_before[4] = { 0, 3, 3, 2 };
|
|
const uint8_t ff_hevc_qpel_extra_after[4] = { 0, 3, 4, 4 };
|
|
const uint8_t ff_hevc_qpel_extra[4] = { 0, 6, 7, 6 };
|
|
|
|
static const uint8_t scan_1x1[1] = {
|
|
0,
|
|
};
|
|
|
|
static const uint8_t horiz_scan2x2_x[4] = {
|
|
0, 1, 0, 1,
|
|
};
|
|
|
|
static const uint8_t horiz_scan2x2_y[4] = {
|
|
0, 0, 1, 1
|
|
};
|
|
|
|
static const uint8_t horiz_scan4x4_x[16] = {
|
|
0, 1, 2, 3,
|
|
0, 1, 2, 3,
|
|
0, 1, 2, 3,
|
|
0, 1, 2, 3,
|
|
};
|
|
|
|
static const uint8_t horiz_scan4x4_y[16] = {
|
|
0, 0, 0, 0,
|
|
1, 1, 1, 1,
|
|
2, 2, 2, 2,
|
|
3, 3, 3, 3,
|
|
};
|
|
|
|
static const uint8_t horiz_scan8x8_inv[8][8] = {
|
|
{ 0, 1, 2, 3, 16, 17, 18, 19, },
|
|
{ 4, 5, 6, 7, 20, 21, 22, 23, },
|
|
{ 8, 9, 10, 11, 24, 25, 26, 27, },
|
|
{ 12, 13, 14, 15, 28, 29, 30, 31, },
|
|
{ 32, 33, 34, 35, 48, 49, 50, 51, },
|
|
{ 36, 37, 38, 39, 52, 53, 54, 55, },
|
|
{ 40, 41, 42, 43, 56, 57, 58, 59, },
|
|
{ 44, 45, 46, 47, 60, 61, 62, 63, },
|
|
};
|
|
|
|
static const uint8_t diag_scan4x1_x[4] = {
|
|
0, 1, 2, 3,
|
|
};
|
|
|
|
static const uint8_t diag_scan1x4_y[4] = {
|
|
0, 1, 2, 3,
|
|
};
|
|
|
|
static const uint8_t diag_scan2x2_x[4] = {
|
|
0, 0, 1, 1,
|
|
};
|
|
|
|
static const uint8_t diag_scan2x2_y[4] = {
|
|
0, 1, 0, 1,
|
|
};
|
|
|
|
static const uint8_t diag_scan2x2_inv[2][2] = {
|
|
{ 0, 2, },
|
|
{ 1, 3, },
|
|
};
|
|
|
|
static const uint8_t diag_scan8x2_x[16] = {
|
|
0, 0, 1, 1,
|
|
2, 2, 3, 3,
|
|
4, 4, 5, 5,
|
|
6, 6, 7, 7,
|
|
};
|
|
|
|
static const uint8_t diag_scan8x2_y[16] = {
|
|
0, 1, 0, 1,
|
|
0, 1, 0, 1,
|
|
0, 1, 0, 1,
|
|
0, 1, 0, 1,
|
|
};
|
|
|
|
static const uint8_t diag_scan8x2_inv[2][8] = {
|
|
{ 0, 2, 4, 6, 8, 10, 12, 14, },
|
|
{ 1, 3, 5, 7, 9, 11, 13, 15, },
|
|
};
|
|
|
|
static const uint8_t diag_scan2x8_x[16] = {
|
|
0, 0, 1, 0,
|
|
1, 0, 1, 0,
|
|
1, 0, 1, 0,
|
|
1, 0, 1, 1,
|
|
};
|
|
|
|
static const uint8_t diag_scan2x8_y[16] = {
|
|
0, 1, 0, 2,
|
|
1, 3, 2, 4,
|
|
3, 5, 4, 6,
|
|
5, 7, 6, 7,
|
|
};
|
|
|
|
static const uint8_t diag_scan2x8_inv[8][2] = {
|
|
{ 0, 2, },
|
|
{ 1, 4, },
|
|
{ 3, 6, },
|
|
{ 5, 8, },
|
|
{ 7, 10, },
|
|
{ 9, 12, },
|
|
{ 11, 14, },
|
|
{ 13, 15, },
|
|
};
|
|
|
|
const uint8_t ff_hevc_diag_scan4x4_x[16] = {
|
|
0, 0, 1, 0,
|
|
1, 2, 0, 1,
|
|
2, 3, 1, 2,
|
|
3, 2, 3, 3,
|
|
};
|
|
|
|
const uint8_t ff_hevc_diag_scan4x4_y[16] = {
|
|
0, 1, 0, 2,
|
|
1, 0, 3, 2,
|
|
1, 0, 3, 2,
|
|
1, 3, 2, 3,
|
|
};
|
|
|
|
static const uint8_t diag_scan4x4_inv[4][4] = {
|
|
{ 0, 2, 5, 9, },
|
|
{ 1, 4, 8, 12, },
|
|
{ 3, 7, 11, 14, },
|
|
{ 6, 10, 13, 15, },
|
|
};
|
|
|
|
const uint8_t ff_hevc_diag_scan8x8_x[64] = {
|
|
0, 0, 1, 0,
|
|
1, 2, 0, 1,
|
|
2, 3, 0, 1,
|
|
2, 3, 4, 0,
|
|
1, 2, 3, 4,
|
|
5, 0, 1, 2,
|
|
3, 4, 5, 6,
|
|
0, 1, 2, 3,
|
|
4, 5, 6, 7,
|
|
1, 2, 3, 4,
|
|
5, 6, 7, 2,
|
|
3, 4, 5, 6,
|
|
7, 3, 4, 5,
|
|
6, 7, 4, 5,
|
|
6, 7, 5, 6,
|
|
7, 6, 7, 7,
|
|
};
|
|
|
|
const uint8_t ff_hevc_diag_scan8x8_y[64] = {
|
|
0, 1, 0, 2,
|
|
1, 0, 3, 2,
|
|
1, 0, 4, 3,
|
|
2, 1, 0, 5,
|
|
4, 3, 2, 1,
|
|
0, 6, 5, 4,
|
|
3, 2, 1, 0,
|
|
7, 6, 5, 4,
|
|
3, 2, 1, 0,
|
|
7, 6, 5, 4,
|
|
3, 2, 1, 7,
|
|
6, 5, 4, 3,
|
|
2, 7, 6, 5,
|
|
4, 3, 7, 6,
|
|
5, 4, 7, 6,
|
|
5, 7, 6, 7,
|
|
};
|
|
|
|
static const uint8_t diag_scan8x8_inv[8][8] = {
|
|
{ 0, 2, 5, 9, 14, 20, 27, 35, },
|
|
{ 1, 4, 8, 13, 19, 26, 34, 42, },
|
|
{ 3, 7, 12, 18, 25, 33, 41, 48, },
|
|
{ 6, 11, 17, 24, 32, 40, 47, 53, },
|
|
{ 10, 16, 23, 31, 39, 46, 52, 57, },
|
|
{ 15, 22, 30, 38, 45, 51, 56, 60, },
|
|
{ 21, 29, 37, 44, 50, 55, 59, 62, },
|
|
{ 28, 36, 43, 49, 54, 58, 61, 63, },
|
|
};
|
|
|
|
/**
|
|
* NOTE: Each function hls_foo correspond to the function foo in the
|
|
* specification (HLS stands for High Level Syntax).
|
|
*/
|
|
|
|
/**
|
|
* Section 5.7
|
|
*/
|
|
|
|
/* free everything allocated by pic_arrays_init() */
|
|
static void pic_arrays_free(HEVCContext *s)
|
|
{
|
|
av_freep(&s->sao);
|
|
av_freep(&s->deblock);
|
|
av_freep(&s->split_cu_flag);
|
|
|
|
av_freep(&s->skip_flag);
|
|
av_freep(&s->tab_ct_depth);
|
|
|
|
av_freep(&s->tab_ipm);
|
|
av_freep(&s->cbf_luma);
|
|
av_freep(&s->is_pcm);
|
|
|
|
av_freep(&s->qp_y_tab);
|
|
av_freep(&s->tab_slice_address);
|
|
av_freep(&s->filter_slice_edges);
|
|
|
|
av_freep(&s->horizontal_bs);
|
|
av_freep(&s->vertical_bs);
|
|
|
|
av_buffer_pool_uninit(&s->tab_mvf_pool);
|
|
av_buffer_pool_uninit(&s->rpl_tab_pool);
|
|
}
|
|
|
|
/* allocate arrays that depend on frame dimensions */
|
|
static int pic_arrays_init(HEVCContext *s)
|
|
{
|
|
int log2_min_cb_size = s->sps->log2_min_coding_block_size;
|
|
int width = s->sps->width;
|
|
int height = s->sps->height;
|
|
int pic_size = width * height;
|
|
int pic_size_in_ctb = ((width >> log2_min_cb_size) + 1) *
|
|
((height >> log2_min_cb_size) + 1);
|
|
int ctb_count = s->sps->ctb_width * s->sps->ctb_height;
|
|
int pic_width_in_min_pu = width >> s->sps->log2_min_pu_size;
|
|
int pic_height_in_min_pu = height >> s->sps->log2_min_pu_size;
|
|
int pic_size_in_min_pu = pic_width_in_min_pu * pic_height_in_min_pu;
|
|
int pic_width_in_min_tu = width >> s->sps->log2_min_transform_block_size;
|
|
int pic_height_in_min_tu = height >> s->sps->log2_min_transform_block_size;
|
|
|
|
s->bs_width = width >> 3;
|
|
s->bs_height = height >> 3;
|
|
|
|
s->sao = av_mallocz_array(ctb_count, sizeof(*s->sao));
|
|
s->deblock = av_mallocz_array(ctb_count, sizeof(*s->deblock));
|
|
s->split_cu_flag = av_malloc(pic_size);
|
|
if (!s->sao || !s->deblock || !s->split_cu_flag)
|
|
goto fail;
|
|
|
|
s->skip_flag = av_malloc(pic_size_in_ctb);
|
|
s->tab_ct_depth = av_malloc(s->sps->min_cb_height * s->sps->min_cb_width);
|
|
if (!s->skip_flag || !s->tab_ct_depth)
|
|
goto fail;
|
|
|
|
s->tab_ipm = av_malloc(pic_size_in_min_pu);
|
|
s->cbf_luma = av_malloc(pic_width_in_min_tu * pic_height_in_min_tu);
|
|
s->is_pcm = av_malloc(pic_size_in_min_pu);
|
|
if (!s->tab_ipm || !s->cbf_luma || !s->is_pcm)
|
|
goto fail;
|
|
|
|
s->filter_slice_edges = av_malloc(ctb_count);
|
|
s->tab_slice_address = av_malloc(pic_size_in_ctb * sizeof(*s->tab_slice_address));
|
|
s->qp_y_tab = av_malloc(pic_size_in_ctb * sizeof(*s->qp_y_tab));
|
|
if (!s->qp_y_tab || !s->filter_slice_edges || !s->tab_slice_address)
|
|
goto fail;
|
|
|
|
s->horizontal_bs = av_mallocz(2 * s->bs_width * (s->bs_height + 1));
|
|
s->vertical_bs = av_mallocz(2 * s->bs_width * (s->bs_height + 1));
|
|
if (!s->horizontal_bs || !s->vertical_bs)
|
|
goto fail;
|
|
|
|
s->tab_mvf_pool = av_buffer_pool_init(pic_size_in_min_pu * sizeof(MvField),
|
|
av_buffer_alloc);
|
|
if (!s->tab_mvf_pool)
|
|
goto fail;
|
|
|
|
s->rpl_tab_pool = av_buffer_pool_init(ctb_count * sizeof(RefPicListTab),
|
|
av_buffer_allocz);
|
|
if (!s->rpl_tab_pool)
|
|
goto fail;
|
|
|
|
return 0;
|
|
fail:
|
|
pic_arrays_free(s);
|
|
return AVERROR(ENOMEM);
|
|
}
|
|
|
|
static void pred_weight_table(HEVCContext *s, GetBitContext *gb)
|
|
{
|
|
int i = 0;
|
|
int j = 0;
|
|
uint8_t luma_weight_l0_flag[16];
|
|
uint8_t chroma_weight_l0_flag[16];
|
|
uint8_t luma_weight_l1_flag[16];
|
|
uint8_t chroma_weight_l1_flag[16];
|
|
|
|
s->sh.luma_log2_weight_denom = get_ue_golomb(gb);
|
|
if (s->sps->chroma_format_idc != 0) {
|
|
int delta = get_se_golomb(gb);
|
|
s->sh.chroma_log2_weight_denom = av_clip_c(s->sh.luma_log2_weight_denom + delta, 0, 7);
|
|
}
|
|
|
|
for (i = 0; i < s->sh.nb_refs[L0]; i++) {
|
|
luma_weight_l0_flag[i] = get_bits1(gb);
|
|
if (!luma_weight_l0_flag[i]) {
|
|
s->sh.luma_weight_l0[i] = 1 << s->sh.luma_log2_weight_denom;
|
|
s->sh.luma_offset_l0[i] = 0;
|
|
}
|
|
}
|
|
if (s->sps->chroma_format_idc != 0) { //fix me ! invert "if" and "for"
|
|
for (i = 0; i < s->sh.nb_refs[L0]; i++) {
|
|
chroma_weight_l0_flag[i] = get_bits1(gb);
|
|
}
|
|
} else {
|
|
for (i = 0; i < s->sh.nb_refs[L0]; i++) {
|
|
chroma_weight_l0_flag[i] = 0;
|
|
}
|
|
}
|
|
for (i = 0; i < s->sh.nb_refs[L0]; i++) {
|
|
if (luma_weight_l0_flag[i]) {
|
|
int delta_luma_weight_l0 = get_se_golomb(gb);
|
|
s->sh.luma_weight_l0[i] = (1 << s->sh.luma_log2_weight_denom) + delta_luma_weight_l0;
|
|
s->sh.luma_offset_l0[i] = get_se_golomb(gb);
|
|
}
|
|
if (chroma_weight_l0_flag[i]) {
|
|
for (j = 0; j < 2; j++) {
|
|
int delta_chroma_weight_l0 = get_se_golomb(gb);
|
|
int delta_chroma_offset_l0 = get_se_golomb(gb);
|
|
s->sh.chroma_weight_l0[i][j] = (1 << s->sh.chroma_log2_weight_denom) + delta_chroma_weight_l0;
|
|
s->sh.chroma_offset_l0[i][j] = av_clip_c((delta_chroma_offset_l0 - ((128 * s->sh.chroma_weight_l0[i][j])
|
|
>> s->sh.chroma_log2_weight_denom) + 128), -128, 127);
|
|
}
|
|
} else {
|
|
s->sh.chroma_weight_l0[i][0] = 1 << s->sh.chroma_log2_weight_denom;
|
|
s->sh.chroma_offset_l0[i][0] = 0;
|
|
s->sh.chroma_weight_l0[i][1] = 1 << s->sh.chroma_log2_weight_denom;
|
|
s->sh.chroma_offset_l0[i][1] = 0;
|
|
}
|
|
}
|
|
if (s->sh.slice_type == B_SLICE) {
|
|
for (i = 0; i < s->sh.nb_refs[L1]; i++) {
|
|
luma_weight_l1_flag[i] = get_bits1(gb);
|
|
if (!luma_weight_l1_flag[i]) {
|
|
s->sh.luma_weight_l1[i] = 1 << s->sh.luma_log2_weight_denom;
|
|
s->sh.luma_offset_l1[i] = 0;
|
|
}
|
|
}
|
|
if (s->sps->chroma_format_idc != 0) {
|
|
for (i = 0; i < s->sh.nb_refs[L1]; i++) {
|
|
chroma_weight_l1_flag[i] = get_bits1(gb);
|
|
}
|
|
} else {
|
|
for (i = 0; i < s->sh.nb_refs[L1]; i++) {
|
|
chroma_weight_l1_flag[i] = 0;
|
|
}
|
|
}
|
|
for (i = 0; i < s->sh.nb_refs[L1]; i++) {
|
|
if (luma_weight_l1_flag[i]) {
|
|
int delta_luma_weight_l1 = get_se_golomb(gb);
|
|
s->sh.luma_weight_l1[i] = (1 << s->sh.luma_log2_weight_denom) + delta_luma_weight_l1;
|
|
s->sh.luma_offset_l1[i] = get_se_golomb(gb);
|
|
}
|
|
if (chroma_weight_l1_flag[i]) {
|
|
for (j = 0; j < 2; j++) {
|
|
int delta_chroma_weight_l1 = get_se_golomb(gb);
|
|
int delta_chroma_offset_l1 = get_se_golomb(gb);
|
|
s->sh.chroma_weight_l1[i][j] = (1 << s->sh.chroma_log2_weight_denom) + delta_chroma_weight_l1;
|
|
s->sh.chroma_offset_l1[i][j] = av_clip_c((delta_chroma_offset_l1 - ((128 * s->sh.chroma_weight_l1[i][j])
|
|
>> s->sh.chroma_log2_weight_denom) + 128), -128, 127);
|
|
}
|
|
} else {
|
|
s->sh.chroma_weight_l1[i][0] = 1 << s->sh.chroma_log2_weight_denom;
|
|
s->sh.chroma_offset_l1[i][0] = 0;
|
|
s->sh.chroma_weight_l1[i][1] = 1 << s->sh.chroma_log2_weight_denom;
|
|
s->sh.chroma_offset_l1[i][1] = 0;
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
static int decode_lt_rps(HEVCContext *s, LongTermRPS *rps, GetBitContext *gb)
|
|
{
|
|
const HEVCSPS *sps = s->sps;
|
|
int max_poc_lsb = 1 << sps->log2_max_poc_lsb;
|
|
int prev_delta_msb = 0;
|
|
int nb_sps = 0, nb_sh;
|
|
int i;
|
|
|
|
rps->nb_refs = 0;
|
|
if (!sps->long_term_ref_pics_present_flag)
|
|
return 0;
|
|
|
|
if (sps->num_long_term_ref_pics_sps > 0)
|
|
nb_sps = get_ue_golomb(gb);
|
|
nb_sh = get_ue_golomb(gb);
|
|
|
|
if (nb_sh + nb_sps > FF_ARRAY_ELEMS(rps->poc))
|
|
return AVERROR_INVALIDDATA;
|
|
|
|
rps->nb_refs = nb_sh + nb_sps;
|
|
|
|
for (i = 0; i < rps->nb_refs; i++) {
|
|
uint8_t delta_poc_msb_present;
|
|
|
|
if (i < nb_sps) {
|
|
uint8_t lt_idx_sps = 0;
|
|
|
|
if (sps->num_long_term_ref_pics_sps > 1)
|
|
lt_idx_sps = get_bits(gb, av_ceil_log2(sps->num_long_term_ref_pics_sps));
|
|
|
|
rps->poc[i] = sps->lt_ref_pic_poc_lsb_sps[lt_idx_sps];
|
|
rps->used[i] = sps->used_by_curr_pic_lt_sps_flag[lt_idx_sps];
|
|
} else {
|
|
rps->poc[i] = get_bits(gb, sps->log2_max_poc_lsb);
|
|
rps->used[i] = get_bits1(gb);
|
|
}
|
|
|
|
delta_poc_msb_present = get_bits1(gb);
|
|
if (delta_poc_msb_present) {
|
|
int delta = get_ue_golomb(gb);
|
|
|
|
if (i && i != nb_sps)
|
|
delta += prev_delta_msb;
|
|
|
|
rps->poc[i] += s->poc - delta * max_poc_lsb - s->sh.pic_order_cnt_lsb;
|
|
prev_delta_msb = delta;
|
|
}
|
|
}
|
|
|
|
return 0;
|
|
}
|
|
|
|
static int hls_slice_header(HEVCContext *s)
|
|
{
|
|
GetBitContext *gb = &s->HEVClc.gb;
|
|
SliceHeader *sh = &s->sh;
|
|
int i, ret;
|
|
|
|
// Coded parameters
|
|
sh->first_slice_in_pic_flag = get_bits1(gb);
|
|
if ((IS_IDR(s) || IS_BLA(s)) && sh->first_slice_in_pic_flag) {
|
|
s->seq_decode = (s->seq_decode + 1) & 0xff;
|
|
s->max_ra = INT_MAX;
|
|
if (IS_IDR(s))
|
|
ff_hevc_clear_refs(s);
|
|
}
|
|
if (s->nal_unit_type >= 16 && s->nal_unit_type <= 23)
|
|
sh->no_output_of_prior_pics_flag = get_bits1(gb);
|
|
|
|
sh->pps_id = get_ue_golomb(gb);
|
|
if (sh->pps_id >= MAX_PPS_COUNT || !s->pps_list[sh->pps_id]) {
|
|
av_log(s->avctx, AV_LOG_ERROR, "PPS id out of range: %d\n", sh->pps_id);
|
|
return AVERROR_INVALIDDATA;
|
|
}
|
|
s->pps = (HEVCPPS*)s->pps_list[sh->pps_id]->data;
|
|
|
|
if (s->sps != (HEVCSPS*)s->sps_list[s->pps->sps_id]->data) {
|
|
s->sps = (HEVCSPS*)s->sps_list[s->pps->sps_id]->data;
|
|
s->vps = s->vps_list[s->sps->vps_id];
|
|
|
|
pic_arrays_free(s);
|
|
ret = pic_arrays_init(s);
|
|
if (ret < 0) {
|
|
s->sps = NULL;
|
|
return AVERROR(ENOMEM);
|
|
}
|
|
|
|
s->width = s->sps->width;
|
|
s->height = s->sps->height;
|
|
|
|
s->avctx->coded_width = s->sps->width;
|
|
s->avctx->coded_height = s->sps->height;
|
|
s->avctx->width = s->sps->output_width;
|
|
s->avctx->height = s->sps->output_height;
|
|
s->avctx->pix_fmt = s->sps->pix_fmt;
|
|
s->avctx->sample_aspect_ratio = s->sps->vui.sar;
|
|
s->avctx->has_b_frames = s->sps->temporal_layer[s->sps->max_sub_layers - 1].num_reorder_pics;
|
|
|
|
if (s->sps->chroma_format_idc == 0 || s->sps->separate_colour_plane_flag) {
|
|
av_log(s->avctx, AV_LOG_ERROR,
|
|
"TODO: s->sps->chroma_format_idc == 0 || "
|
|
"s->sps->separate_colour_plane_flag\n");
|
|
return AVERROR_PATCHWELCOME;
|
|
}
|
|
|
|
ff_hevc_pred_init(&s->hpc, s->sps->bit_depth);
|
|
ff_hevc_dsp_init (&s->hevcdsp, s->sps->bit_depth);
|
|
ff_videodsp_init (&s->vdsp, s->sps->bit_depth);
|
|
|
|
if (s->sps->sao_enabled) {
|
|
av_frame_unref(s->tmp_frame);
|
|
ret = ff_get_buffer(s->avctx, s->tmp_frame, 0);
|
|
if (ret < 0)
|
|
return ret;
|
|
s->frame = s->tmp_frame;
|
|
}
|
|
}
|
|
|
|
sh->dependent_slice_segment_flag = 0;
|
|
if (!sh->first_slice_in_pic_flag) {
|
|
int slice_address_length;
|
|
|
|
if (s->pps->dependent_slice_segments_enabled_flag)
|
|
sh->dependent_slice_segment_flag = get_bits1(gb);
|
|
|
|
slice_address_length = av_ceil_log2(s->sps->ctb_width *
|
|
s->sps->ctb_height);
|
|
sh->slice_segment_addr = get_bits(gb, slice_address_length);
|
|
if (sh->slice_segment_addr >= s->sps->ctb_width * s->sps->ctb_height) {
|
|
av_log(s->avctx, AV_LOG_ERROR, "Invalid slice segment address: %u.\n",
|
|
sh->slice_segment_addr);
|
|
return AVERROR_INVALIDDATA;
|
|
}
|
|
|
|
if (!sh->dependent_slice_segment_flag) {
|
|
sh->slice_addr = sh->slice_segment_addr;
|
|
s->slice_idx++;
|
|
}
|
|
} else {
|
|
sh->slice_segment_addr = sh->slice_addr = 0;
|
|
s->slice_idx = 0;
|
|
s->slice_initialized = 0;
|
|
}
|
|
|
|
if (!sh->dependent_slice_segment_flag) {
|
|
s->slice_initialized = 0;
|
|
|
|
for (i = 0; i < s->pps->num_extra_slice_header_bits; i++)
|
|
skip_bits(gb, 1); // slice_reserved_undetermined_flag[]
|
|
|
|
sh->slice_type = get_ue_golomb(gb);
|
|
if (!(sh->slice_type == I_SLICE || sh->slice_type == P_SLICE ||
|
|
sh->slice_type == B_SLICE)) {
|
|
av_log(s->avctx, AV_LOG_ERROR, "Unknown slice type: %d.\n",
|
|
sh->slice_type);
|
|
return AVERROR_INVALIDDATA;
|
|
}
|
|
|
|
if (s->pps->output_flag_present_flag)
|
|
sh->pic_output_flag = get_bits1(gb);
|
|
|
|
if (s->sps->separate_colour_plane_flag)
|
|
sh->colour_plane_id = get_bits(gb, 2);
|
|
|
|
if (!IS_IDR(s)) {
|
|
int short_term_ref_pic_set_sps_flag;
|
|
int poc;
|
|
|
|
sh->pic_order_cnt_lsb = get_bits(gb, s->sps->log2_max_poc_lsb);
|
|
poc = ff_hevc_compute_poc(s, sh->pic_order_cnt_lsb);
|
|
if (!sh->first_slice_in_pic_flag && poc != s->poc) {
|
|
av_log(s->avctx, AV_LOG_WARNING,
|
|
"Ignoring POC change between slices: %d -> %d\n", s->poc, poc);
|
|
if (s->avctx->err_recognition & AV_EF_EXPLODE)
|
|
return AVERROR_INVALIDDATA;
|
|
poc = s->poc;
|
|
}
|
|
s->poc = poc;
|
|
|
|
short_term_ref_pic_set_sps_flag = get_bits1(gb);
|
|
if (!short_term_ref_pic_set_sps_flag) {
|
|
ret = ff_hevc_decode_short_term_rps(s, &sh->slice_rps, s->sps, 1);
|
|
if (ret < 0)
|
|
return ret;
|
|
|
|
sh->short_term_rps = &sh->slice_rps;
|
|
} else {
|
|
int numbits, rps_idx;
|
|
|
|
if (!s->sps->nb_st_rps) {
|
|
av_log(s->avctx, AV_LOG_ERROR, "No ref lists in the SPS.\n");
|
|
return AVERROR_INVALIDDATA;
|
|
}
|
|
|
|
numbits = av_ceil_log2(s->sps->nb_st_rps);
|
|
rps_idx = (numbits > 0) ? get_bits(gb, numbits) : 0;
|
|
sh->short_term_rps = &s->sps->st_rps[rps_idx];
|
|
}
|
|
|
|
ret = decode_lt_rps(s, &sh->long_term_rps, gb);
|
|
if (ret < 0) {
|
|
av_log(s->avctx, AV_LOG_WARNING, "Invalid long term RPS.\n");
|
|
if (s->avctx->err_recognition & AV_EF_EXPLODE)
|
|
return AVERROR_INVALIDDATA;
|
|
}
|
|
|
|
if (s->sps->sps_temporal_mvp_enabled_flag)
|
|
sh->slice_temporal_mvp_enabled_flag = get_bits1(gb);
|
|
else
|
|
sh->slice_temporal_mvp_enabled_flag = 0;
|
|
} else {
|
|
s->sh.short_term_rps = NULL;
|
|
s->poc = 0;
|
|
}
|
|
|
|
if (s->temporal_id == 0 &&
|
|
s->nal_unit_type != NAL_TRAIL_N &&
|
|
s->nal_unit_type != NAL_TSA_N &&
|
|
s->nal_unit_type != NAL_STSA_N &&
|
|
s->nal_unit_type != NAL_TRAIL_N &&
|
|
s->nal_unit_type != NAL_RADL_N &&
|
|
s->nal_unit_type != NAL_RADL_R &&
|
|
s->nal_unit_type != NAL_RASL_R)
|
|
s->pocTid0 = s->poc;
|
|
|
|
if (s->sps->sao_enabled) {
|
|
sh->slice_sample_adaptive_offset_flag[0] = get_bits1(gb);
|
|
sh->slice_sample_adaptive_offset_flag[2] =
|
|
sh->slice_sample_adaptive_offset_flag[1] = get_bits1(gb);
|
|
} else {
|
|
sh->slice_sample_adaptive_offset_flag[0] = 0;
|
|
sh->slice_sample_adaptive_offset_flag[1] = 0;
|
|
sh->slice_sample_adaptive_offset_flag[2] = 0;
|
|
}
|
|
|
|
sh->nb_refs[L0] = sh->nb_refs[L1] = 0;
|
|
if (sh->slice_type == P_SLICE || sh->slice_type == B_SLICE) {
|
|
int nb_refs;
|
|
|
|
sh->nb_refs[L0] = s->pps->num_ref_idx_l0_default_active;
|
|
if (sh->slice_type == B_SLICE)
|
|
sh->nb_refs[L1] = s->pps->num_ref_idx_l1_default_active;
|
|
|
|
if (get_bits1(gb)) { // num_ref_idx_active_override_flag
|
|
sh->nb_refs[L0] = get_ue_golomb(gb) + 1;
|
|
if (sh->slice_type == B_SLICE)
|
|
sh->nb_refs[L1] = get_ue_golomb(gb) + 1;
|
|
}
|
|
if (sh->nb_refs[L0] > MAX_REFS || sh->nb_refs[L1] > MAX_REFS) {
|
|
av_log(s->avctx, AV_LOG_ERROR, "Too many refs: %d/%d.\n",
|
|
sh->nb_refs[L0], sh->nb_refs[L1]);
|
|
return AVERROR_INVALIDDATA;
|
|
}
|
|
|
|
sh->rpl_modification_flag[0] = 0;
|
|
sh->rpl_modification_flag[1] = 0;
|
|
nb_refs = ff_hevc_frame_nb_refs(s);
|
|
if (!nb_refs) {
|
|
av_log(s->avctx, AV_LOG_ERROR, "Zero refs for a frame with P or B slices.\n");
|
|
return AVERROR_INVALIDDATA;
|
|
}
|
|
|
|
if (s->pps->lists_modification_present_flag && nb_refs > 1) {
|
|
sh->rpl_modification_flag[0] = get_bits1(gb);
|
|
if (sh->rpl_modification_flag[0]) {
|
|
for (i = 0; i < sh->nb_refs[L0]; i++)
|
|
sh->list_entry_lx[0][i] = get_bits(gb, av_ceil_log2(nb_refs));
|
|
}
|
|
|
|
if (sh->slice_type == B_SLICE) {
|
|
sh->rpl_modification_flag[1] = get_bits1(gb);
|
|
if (sh->rpl_modification_flag[1] == 1)
|
|
for (i = 0; i < sh->nb_refs[L1]; i++)
|
|
sh->list_entry_lx[1][i] = get_bits(gb, av_ceil_log2(nb_refs));
|
|
}
|
|
}
|
|
|
|
if (sh->slice_type == B_SLICE)
|
|
sh->mvd_l1_zero_flag = get_bits1(gb);
|
|
|
|
if (s->pps->cabac_init_present_flag)
|
|
sh->cabac_init_flag = get_bits1(gb);
|
|
else
|
|
sh->cabac_init_flag = 0;
|
|
|
|
sh->collocated_ref_idx = 0;
|
|
if (sh->slice_temporal_mvp_enabled_flag) {
|
|
sh->collocated_list = L0;
|
|
if (sh->slice_type == B_SLICE)
|
|
sh->collocated_list = !get_bits1(gb);
|
|
|
|
if (sh->nb_refs[sh->collocated_list] > 1) {
|
|
sh->collocated_ref_idx = get_ue_golomb(gb);
|
|
if (sh->collocated_ref_idx >= sh->nb_refs[sh->collocated_list]) {
|
|
av_log(s->avctx, AV_LOG_ERROR,
|
|
"Invalid collocated_ref_idx: %d.\n", sh->collocated_ref_idx);
|
|
return AVERROR_INVALIDDATA;
|
|
}
|
|
}
|
|
}
|
|
|
|
if ((s->pps->weighted_pred_flag && sh->slice_type == P_SLICE) ||
|
|
(s->pps->weighted_bipred_flag && sh->slice_type == B_SLICE)) {
|
|
pred_weight_table(s, gb);
|
|
}
|
|
|
|
sh->max_num_merge_cand = 5 - get_ue_golomb(gb);
|
|
}
|
|
|
|
sh->slice_qp_delta = get_se_golomb(gb);
|
|
if (s->pps->pic_slice_level_chroma_qp_offsets_present_flag) {
|
|
sh->slice_cb_qp_offset = get_se_golomb(gb);
|
|
sh->slice_cr_qp_offset = get_se_golomb(gb);
|
|
} else {
|
|
sh->slice_cb_qp_offset = 0;
|
|
sh->slice_cr_qp_offset = 0;
|
|
}
|
|
|
|
if (s->pps->deblocking_filter_control_present_flag) {
|
|
int deblocking_filter_override_flag = 0;
|
|
|
|
if (s->pps->deblocking_filter_override_enabled_flag)
|
|
deblocking_filter_override_flag = get_bits1(gb);
|
|
|
|
if (deblocking_filter_override_flag) {
|
|
sh->disable_deblocking_filter_flag = get_bits1(gb);
|
|
if (!sh->disable_deblocking_filter_flag) {
|
|
sh->beta_offset = get_se_golomb(gb) * 2;
|
|
sh->tc_offset = get_se_golomb(gb) * 2;
|
|
}
|
|
} else {
|
|
sh->disable_deblocking_filter_flag = s->pps->pps_disable_deblocking_filter_flag;
|
|
sh->beta_offset = s->pps->beta_offset;
|
|
sh->tc_offset = s->pps->tc_offset;
|
|
}
|
|
} else {
|
|
sh->disable_deblocking_filter_flag = 0;
|
|
sh->beta_offset = 0;
|
|
sh->tc_offset = 0;
|
|
}
|
|
|
|
|
|
if (s->pps->seq_loop_filter_across_slices_enabled_flag &&
|
|
(sh->slice_sample_adaptive_offset_flag[0] ||
|
|
sh->slice_sample_adaptive_offset_flag[1] ||
|
|
!sh->disable_deblocking_filter_flag)) {
|
|
sh->slice_loop_filter_across_slices_enabled_flag = get_bits1(gb);
|
|
} else {
|
|
sh->slice_loop_filter_across_slices_enabled_flag = s->pps->seq_loop_filter_across_slices_enabled_flag;
|
|
}
|
|
} else if (!s->slice_initialized) {
|
|
av_log(s->avctx, AV_LOG_ERROR, "Independent slice segment missing.\n");
|
|
return AVERROR_INVALIDDATA;
|
|
}
|
|
|
|
sh->num_entry_point_offsets = 0;
|
|
if (s->pps->tiles_enabled_flag || s->pps->entropy_coding_sync_enabled_flag) {
|
|
sh->num_entry_point_offsets = get_ue_golomb(gb);
|
|
if (sh->num_entry_point_offsets > 0) {
|
|
int offset_len = get_ue_golomb(gb) + 1;
|
|
|
|
for (i = 0; i < sh->num_entry_point_offsets; i++)
|
|
skip_bits(gb, offset_len);
|
|
}
|
|
}
|
|
|
|
if (s->pps->slice_header_extension_present_flag) {
|
|
int length = get_ue_golomb(gb);
|
|
for (i = 0; i < length; i++)
|
|
skip_bits(gb, 8); // slice_header_extension_data_byte
|
|
}
|
|
|
|
// Inferred parameters
|
|
sh->slice_qp = 26 + s->pps->pic_init_qp_minus26 + sh->slice_qp_delta;
|
|
sh->slice_ctb_addr_rs = sh->slice_segment_addr;
|
|
|
|
s->HEVClc.first_qp_group = !s->sh.dependent_slice_segment_flag;
|
|
|
|
if (!s->pps->cu_qp_delta_enabled_flag)
|
|
s->HEVClc.qp_y = ((s->sh.slice_qp + 52 + 2 * s->sps->qp_bd_offset) %
|
|
(52 + s->sps->qp_bd_offset)) - s->sps->qp_bd_offset;
|
|
|
|
s->slice_initialized = 1;
|
|
|
|
return 0;
|
|
}
|
|
|
|
#define CTB(tab, x, y) ((tab)[(y) * s->sps->ctb_width + (x)])
|
|
|
|
#define SET_SAO(elem, value) \
|
|
do { \
|
|
if (!sao_merge_up_flag && !sao_merge_left_flag) \
|
|
sao->elem = value; \
|
|
else if (sao_merge_left_flag) \
|
|
sao->elem = CTB(s->sao, rx-1, ry).elem; \
|
|
else if (sao_merge_up_flag) \
|
|
sao->elem = CTB(s->sao, rx, ry-1).elem; \
|
|
else \
|
|
sao->elem = 0; \
|
|
} while (0)
|
|
|
|
static void hls_sao_param(HEVCContext *s, int rx, int ry)
|
|
{
|
|
HEVCLocalContext *lc = &s->HEVClc;
|
|
int sao_merge_left_flag = 0;
|
|
int sao_merge_up_flag = 0;
|
|
int shift = s->sps->bit_depth - FFMIN(s->sps->bit_depth, 10);
|
|
SAOParams *sao = &CTB(s->sao, rx, ry);
|
|
int c_idx, i;
|
|
|
|
if (s->sh.slice_sample_adaptive_offset_flag[0] ||
|
|
s->sh.slice_sample_adaptive_offset_flag[1]) {
|
|
if (rx > 0) {
|
|
if (lc->ctb_left_flag)
|
|
sao_merge_left_flag = ff_hevc_sao_merge_flag_decode(s);
|
|
}
|
|
if (ry > 0 && !sao_merge_left_flag) {
|
|
if (lc->ctb_up_flag)
|
|
sao_merge_up_flag = ff_hevc_sao_merge_flag_decode(s);
|
|
}
|
|
}
|
|
|
|
for (c_idx = 0; c_idx < 3; c_idx++) {
|
|
if (!s->sh.slice_sample_adaptive_offset_flag[c_idx]) {
|
|
sao->type_idx[c_idx] = SAO_NOT_APPLIED;
|
|
continue;
|
|
}
|
|
|
|
if (c_idx == 2) {
|
|
sao->type_idx[2] = sao->type_idx[1];
|
|
sao->eo_class[2] = sao->eo_class[1];
|
|
} else {
|
|
SET_SAO(type_idx[c_idx], ff_hevc_sao_type_idx_decode(s));
|
|
}
|
|
|
|
if (sao->type_idx[c_idx] == SAO_NOT_APPLIED)
|
|
continue;
|
|
|
|
for (i = 0; i < 4; i++)
|
|
SET_SAO(offset_abs[c_idx][i], ff_hevc_sao_offset_abs_decode(s));
|
|
|
|
if (sao->type_idx[c_idx] == SAO_BAND) {
|
|
for (i = 0; i < 4; i++) {
|
|
if (sao->offset_abs[c_idx][i]) {
|
|
SET_SAO(offset_sign[c_idx][i], ff_hevc_sao_offset_sign_decode(s));
|
|
} else {
|
|
sao->offset_sign[c_idx][i] = 0;
|
|
}
|
|
}
|
|
SET_SAO(band_position[c_idx], ff_hevc_sao_band_position_decode(s));
|
|
} else if (c_idx != 2) {
|
|
SET_SAO(eo_class[c_idx], ff_hevc_sao_eo_class_decode(s));
|
|
}
|
|
|
|
// Inferred parameters
|
|
sao->offset_val[c_idx][0] = 0; //avoid undefined values
|
|
for (i = 0; i < 4; i++) {
|
|
sao->offset_val[c_idx][i + 1] = sao->offset_abs[c_idx][i] << shift;
|
|
if (sao->type_idx[c_idx] == SAO_EDGE) {
|
|
if (i > 1)
|
|
sao->offset_val[c_idx][i + 1] = -sao->offset_val[c_idx][i + 1];
|
|
} else if (sao->offset_sign[c_idx][i]) {
|
|
sao->offset_val[c_idx][i + 1] = -sao->offset_val[c_idx][i + 1];
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
#undef SET_SAO
|
|
#undef CTB
|
|
|
|
static void hls_residual_coding(HEVCContext *s, int x0, int y0,
|
|
int log2_trafo_size, enum ScanType scan_idx,
|
|
int c_idx)
|
|
{
|
|
#define GET_COORD(offset, n) \
|
|
do { \
|
|
x_c = (scan_x_cg[offset >> 4] << 2) + scan_x_off[n]; \
|
|
y_c = (scan_y_cg[offset >> 4] << 2) + scan_y_off[n]; \
|
|
} while (0)
|
|
HEVCLocalContext *lc = &s->HEVClc;
|
|
int transform_skip_flag = 0;
|
|
|
|
int last_significant_coeff_x, last_significant_coeff_y;
|
|
int last_scan_pos;
|
|
int n_end;
|
|
int num_coeff = 0;
|
|
int num_last_subset;
|
|
int x_cg_last_sig, y_cg_last_sig;
|
|
|
|
const uint8_t *scan_x_cg, *scan_y_cg, *scan_x_off, *scan_y_off;
|
|
|
|
ptrdiff_t stride = s->frame->linesize[c_idx];
|
|
int hshift = s->sps->hshift[c_idx];
|
|
int vshift = s->sps->vshift[c_idx];
|
|
uint8_t *dst = &s->frame->data[c_idx][(y0 >> vshift) * stride +
|
|
((x0 >> hshift) << s->sps->pixel_shift)];
|
|
DECLARE_ALIGNED( 16, int16_t, coeffs[MAX_TB_SIZE * MAX_TB_SIZE] ) = {0};
|
|
|
|
int trafo_size = 1 << log2_trafo_size;
|
|
int i;
|
|
int qp,shift,add,scale,scale_m;
|
|
const uint8_t level_scale[] = { 40, 45, 51, 57, 64, 72 };
|
|
const uint8_t *scale_matrix;
|
|
uint8_t dc_scale;
|
|
|
|
// Derive QP for dequant
|
|
if (!lc->cu.cu_transquant_bypass_flag) {
|
|
static const int qp_c[] = { 29, 30, 31, 32, 33, 33, 34, 34, 35, 35, 36, 36, 37, 37 };
|
|
int qp_y = lc->qp_y;
|
|
|
|
if (c_idx == 0) {
|
|
qp = qp_y + s->sps->qp_bd_offset;
|
|
} else {
|
|
int qp_i, offset;
|
|
|
|
if (c_idx == 1)
|
|
offset = s->pps->cb_qp_offset + s->sh.slice_cb_qp_offset;
|
|
else
|
|
offset = s->pps->cr_qp_offset + s->sh.slice_cr_qp_offset;
|
|
|
|
qp_i = av_clip_c(qp_y + offset, - s->sps->qp_bd_offset, 57);
|
|
if (qp_i < 30)
|
|
qp = qp_i;
|
|
else if (qp_i > 43)
|
|
qp = qp_i - 6;
|
|
else
|
|
qp = qp_c[qp_i - 30];
|
|
|
|
qp += s->sps->qp_bd_offset;
|
|
}
|
|
|
|
shift = s->sps->bit_depth + log2_trafo_size - 5;
|
|
add = 1 << (shift-1);
|
|
scale = level_scale[qp%6] << (qp/6);
|
|
scale_m = 16; // default when no custom scaling lists.
|
|
dc_scale = 16;
|
|
|
|
if (s->sps->scaling_list_enable_flag) {
|
|
const ScalingList *sl = s->pps->pps_scaling_list_data_present_flag ?
|
|
&s->pps->scaling_list : &s->sps->scaling_list;
|
|
int matrix_id = lc->cu.pred_mode != MODE_INTRA;
|
|
|
|
if (log2_trafo_size != 5)
|
|
matrix_id = 3 * matrix_id + c_idx;
|
|
|
|
scale_matrix = sl->sl[log2_trafo_size - 2][matrix_id];
|
|
if (log2_trafo_size >= 4)
|
|
dc_scale = sl->sl_dc[log2_trafo_size - 4][matrix_id];
|
|
}
|
|
}
|
|
|
|
memset(lc->rc.significant_coeff_group_flag, 0, 8 * 8);
|
|
|
|
if (s->pps->transform_skip_enabled_flag && !lc->cu.cu_transquant_bypass_flag &&
|
|
log2_trafo_size == 2) {
|
|
transform_skip_flag = ff_hevc_transform_skip_flag_decode(s, c_idx);
|
|
}
|
|
|
|
last_significant_coeff_x =
|
|
ff_hevc_last_significant_coeff_x_prefix_decode(s, c_idx, log2_trafo_size);
|
|
last_significant_coeff_y =
|
|
ff_hevc_last_significant_coeff_y_prefix_decode(s, c_idx, log2_trafo_size);
|
|
|
|
if (last_significant_coeff_x > 3) {
|
|
int suffix = ff_hevc_last_significant_coeff_suffix_decode(s, last_significant_coeff_x);
|
|
last_significant_coeff_x = (1 << ((last_significant_coeff_x >> 1) - 1)) *
|
|
(2 + (last_significant_coeff_x & 1)) +
|
|
suffix;
|
|
}
|
|
|
|
if (last_significant_coeff_y > 3) {
|
|
int suffix = ff_hevc_last_significant_coeff_suffix_decode(s, last_significant_coeff_y);
|
|
last_significant_coeff_y = (1 << ((last_significant_coeff_y >> 1) - 1)) *
|
|
(2 + (last_significant_coeff_y & 1)) +
|
|
suffix;
|
|
}
|
|
|
|
if (scan_idx == SCAN_VERT)
|
|
FFSWAP(int, last_significant_coeff_x, last_significant_coeff_y);
|
|
|
|
x_cg_last_sig = last_significant_coeff_x >> 2;
|
|
y_cg_last_sig = last_significant_coeff_y >> 2;
|
|
|
|
switch (scan_idx) {
|
|
case SCAN_DIAG: {
|
|
int last_x_c = last_significant_coeff_x & 3;
|
|
int last_y_c = last_significant_coeff_y & 3;
|
|
|
|
scan_x_off = ff_hevc_diag_scan4x4_x;
|
|
scan_y_off = ff_hevc_diag_scan4x4_y;
|
|
num_coeff = diag_scan4x4_inv[last_y_c][last_x_c];
|
|
if (trafo_size == 4) {
|
|
scan_x_cg = scan_1x1;
|
|
scan_y_cg = scan_1x1;
|
|
} else if (trafo_size == 8) {
|
|
num_coeff += diag_scan2x2_inv[y_cg_last_sig][x_cg_last_sig] << 4;
|
|
scan_x_cg = diag_scan2x2_x;
|
|
scan_y_cg = diag_scan2x2_y;
|
|
} else if (trafo_size == 16) {
|
|
num_coeff += diag_scan4x4_inv[y_cg_last_sig][x_cg_last_sig] << 4;
|
|
scan_x_cg = ff_hevc_diag_scan4x4_x;
|
|
scan_y_cg = ff_hevc_diag_scan4x4_y;
|
|
} else { // trafo_size == 32
|
|
num_coeff += diag_scan8x8_inv[y_cg_last_sig][x_cg_last_sig] << 4;
|
|
scan_x_cg = ff_hevc_diag_scan8x8_x;
|
|
scan_y_cg = ff_hevc_diag_scan8x8_y;
|
|
}
|
|
break;
|
|
}
|
|
case SCAN_HORIZ:
|
|
scan_x_cg = horiz_scan2x2_x;
|
|
scan_y_cg = horiz_scan2x2_y;
|
|
scan_x_off = horiz_scan4x4_x;
|
|
scan_y_off = horiz_scan4x4_y;
|
|
num_coeff = horiz_scan8x8_inv[last_significant_coeff_y][last_significant_coeff_x];
|
|
break;
|
|
default: //SCAN_VERT
|
|
scan_x_cg = horiz_scan2x2_y;
|
|
scan_y_cg = horiz_scan2x2_x;
|
|
scan_x_off = horiz_scan4x4_y;
|
|
scan_y_off = horiz_scan4x4_x;
|
|
num_coeff = horiz_scan8x8_inv[last_significant_coeff_x][last_significant_coeff_y];
|
|
break;
|
|
}
|
|
num_coeff++;
|
|
|
|
num_last_subset = (num_coeff - 1) >> 4;
|
|
|
|
for (i = num_last_subset; i >= 0; i--) {
|
|
int n, m;
|
|
int first_nz_pos_in_cg, last_nz_pos_in_cg, num_sig_coeff, first_greater1_coeff_idx;
|
|
int sign_hidden;
|
|
int sum_abs;
|
|
int x_cg, y_cg, x_c, y_c, pos;
|
|
int implicit_non_zero_coeff = 0;
|
|
int64_t trans_coeff_level;
|
|
|
|
int offset = i << 4;
|
|
|
|
uint8_t significant_coeff_flag_idx[16] = {0};
|
|
uint8_t coeff_abs_level_greater1_flag[16] = {0};
|
|
uint8_t coeff_abs_level_greater2_flag[16] = {0};
|
|
uint16_t coeff_sign_flag;
|
|
uint8_t nb_significant_coeff_flag = 0;
|
|
|
|
int first_elem;
|
|
|
|
x_cg = scan_x_cg[i];
|
|
y_cg = scan_y_cg[i];
|
|
|
|
if ((i < num_last_subset) && (i > 0)) {
|
|
lc->rc.significant_coeff_group_flag[x_cg][y_cg] =
|
|
ff_hevc_significant_coeff_group_flag_decode(s, c_idx, x_cg, y_cg,
|
|
log2_trafo_size);
|
|
implicit_non_zero_coeff = 1;
|
|
} else {
|
|
lc->rc.significant_coeff_group_flag[x_cg][y_cg] =
|
|
((x_cg == x_cg_last_sig && y_cg == y_cg_last_sig) ||
|
|
(x_cg == 0 && y_cg == 0));
|
|
}
|
|
|
|
last_scan_pos = num_coeff - offset - 1;
|
|
|
|
if (i == num_last_subset) {
|
|
n_end = last_scan_pos - 1;
|
|
significant_coeff_flag_idx[0] = last_scan_pos;
|
|
nb_significant_coeff_flag = 1;
|
|
} else {
|
|
n_end = 15;
|
|
}
|
|
for (n = n_end; n >= 0; n--) {
|
|
GET_COORD(offset, n);
|
|
|
|
if (lc->rc.significant_coeff_group_flag[x_cg][y_cg] &&
|
|
(n > 0 || implicit_non_zero_coeff == 0)) {
|
|
if (ff_hevc_significant_coeff_flag_decode(s, c_idx, x_c, y_c, log2_trafo_size, scan_idx) == 1) {
|
|
significant_coeff_flag_idx[nb_significant_coeff_flag] = n;
|
|
nb_significant_coeff_flag = nb_significant_coeff_flag + 1;
|
|
implicit_non_zero_coeff = 0;
|
|
}
|
|
} else {
|
|
int last_cg = (x_c == (x_cg << 2) && y_c == (y_cg << 2));
|
|
if (last_cg && implicit_non_zero_coeff && lc->rc.significant_coeff_group_flag[x_cg][y_cg]) {
|
|
significant_coeff_flag_idx[nb_significant_coeff_flag] = n;
|
|
nb_significant_coeff_flag = nb_significant_coeff_flag + 1;
|
|
}
|
|
}
|
|
}
|
|
|
|
n_end = nb_significant_coeff_flag;
|
|
|
|
first_nz_pos_in_cg = 16;
|
|
last_nz_pos_in_cg = -1;
|
|
num_sig_coeff = 0;
|
|
first_greater1_coeff_idx = -1;
|
|
for (m = 0; m < n_end; m++) {
|
|
n = significant_coeff_flag_idx[m];
|
|
if (num_sig_coeff < 8) {
|
|
coeff_abs_level_greater1_flag[n] =
|
|
ff_hevc_coeff_abs_level_greater1_flag_decode(s, c_idx, i, n,
|
|
(num_sig_coeff == 0),
|
|
(i == num_last_subset));
|
|
num_sig_coeff++;
|
|
if (coeff_abs_level_greater1_flag[n] &&
|
|
first_greater1_coeff_idx == -1)
|
|
first_greater1_coeff_idx = n;
|
|
}
|
|
if (last_nz_pos_in_cg == -1)
|
|
last_nz_pos_in_cg = n;
|
|
first_nz_pos_in_cg = n;
|
|
}
|
|
|
|
sign_hidden = (last_nz_pos_in_cg - first_nz_pos_in_cg >= 4 &&
|
|
!lc->cu.cu_transquant_bypass_flag);
|
|
if (first_greater1_coeff_idx != -1) {
|
|
coeff_abs_level_greater2_flag[first_greater1_coeff_idx] =
|
|
ff_hevc_coeff_abs_level_greater2_flag_decode(s, c_idx, i, first_greater1_coeff_idx);
|
|
}
|
|
if (!s->pps->sign_data_hiding_flag || !sign_hidden ) {
|
|
coeff_sign_flag = ff_hevc_coeff_sign_flag(s, nb_significant_coeff_flag) << (16 - nb_significant_coeff_flag);
|
|
} else {
|
|
coeff_sign_flag = ff_hevc_coeff_sign_flag(s, nb_significant_coeff_flag-1) << (16 - (nb_significant_coeff_flag - 1));
|
|
}
|
|
|
|
num_sig_coeff = 0;
|
|
sum_abs = 0;
|
|
first_elem = 1;
|
|
for (m = 0; m < n_end; m++) {
|
|
n = significant_coeff_flag_idx[m];
|
|
GET_COORD(offset, n);
|
|
trans_coeff_level = 1 + coeff_abs_level_greater1_flag[n] +
|
|
coeff_abs_level_greater2_flag[n];
|
|
if (trans_coeff_level == ((num_sig_coeff < 8) ?
|
|
((n == first_greater1_coeff_idx) ? 3 : 2) : 1)) {
|
|
trans_coeff_level += ff_hevc_coeff_abs_level_remaining(s, first_elem, trans_coeff_level);
|
|
first_elem = 0;
|
|
}
|
|
if (s->pps->sign_data_hiding_flag && sign_hidden) {
|
|
sum_abs += trans_coeff_level;
|
|
if (n == first_nz_pos_in_cg && ((sum_abs&1) == 1))
|
|
trans_coeff_level = -trans_coeff_level;
|
|
}
|
|
if (coeff_sign_flag >> 15)
|
|
trans_coeff_level = -trans_coeff_level;
|
|
coeff_sign_flag <<= 1;
|
|
num_sig_coeff++;
|
|
|
|
if (!lc->cu.cu_transquant_bypass_flag) {
|
|
if(s->sps->scaling_list_enable_flag) {
|
|
if(y_c || x_c || log2_trafo_size < 4) {
|
|
switch(log2_trafo_size) {
|
|
case 3: pos = (y_c << 3) + x_c; break;
|
|
case 4: pos = ((y_c >> 1) << 3) + (x_c >> 1); break;
|
|
case 5: pos = ((y_c >> 2) << 3) + (x_c >> 2); break;
|
|
default: pos = (y_c << 2) + x_c;
|
|
}
|
|
scale_m = scale_matrix[pos];
|
|
} else
|
|
scale_m = dc_scale;
|
|
}
|
|
trans_coeff_level = (trans_coeff_level * (int64_t)scale * (int64_t)scale_m + add) >> shift;
|
|
if (trans_coeff_level < 0) {
|
|
if((~trans_coeff_level) & 0xFffffffffff8000)
|
|
trans_coeff_level = -32768;
|
|
} else {
|
|
if(trans_coeff_level & 0xffffffffffff8000)
|
|
trans_coeff_level = 32767;
|
|
}
|
|
}
|
|
|
|
coeffs[y_c * trafo_size + x_c] = trans_coeff_level;
|
|
}
|
|
}
|
|
|
|
if (lc->cu.cu_transquant_bypass_flag) {
|
|
s->hevcdsp.transquant_bypass[log2_trafo_size-2](dst, coeffs, stride);
|
|
} else {
|
|
if (transform_skip_flag)
|
|
s->hevcdsp.transform_skip(dst, coeffs, stride);
|
|
else if (lc->cu.pred_mode == MODE_INTRA && c_idx == 0 && log2_trafo_size == 2)
|
|
s->hevcdsp.transform_4x4_luma_add(dst, coeffs, stride);
|
|
else
|
|
s->hevcdsp.transform_add[log2_trafo_size-2](dst, coeffs, stride);
|
|
}
|
|
}
|
|
|
|
static void hls_transform_unit(HEVCContext *s, int x0, int y0, int xBase, int yBase, int cb_xBase, int cb_yBase,
|
|
int log2_cb_size, int log2_trafo_size, int trafo_depth, int blk_idx)
|
|
{
|
|
HEVCLocalContext *lc = &s->HEVClc;
|
|
int scan_idx = SCAN_DIAG;
|
|
int scan_idx_c = SCAN_DIAG;
|
|
|
|
if (lc->cu.pred_mode == MODE_INTRA) {
|
|
int trafo_size = 1 << log2_trafo_size;
|
|
ff_hevc_set_neighbour_available(s, x0, y0, trafo_size, trafo_size);
|
|
|
|
s->hpc.intra_pred(s, x0, y0, log2_trafo_size, 0);
|
|
if (log2_trafo_size > 2) {
|
|
trafo_size = trafo_size<<(s->sps->hshift[1]-1);
|
|
ff_hevc_set_neighbour_available(s, x0, y0, trafo_size, trafo_size);
|
|
s->hpc.intra_pred(s, x0, y0, log2_trafo_size - 1, 1);
|
|
s->hpc.intra_pred(s, x0, y0, log2_trafo_size - 1, 2);
|
|
} else if (blk_idx == 3) {
|
|
trafo_size = trafo_size<<(s->sps->hshift[1]);
|
|
ff_hevc_set_neighbour_available(s, xBase, yBase, trafo_size, trafo_size);
|
|
s->hpc.intra_pred(s, xBase, yBase, log2_trafo_size, 1);
|
|
s->hpc.intra_pred(s, xBase, yBase, log2_trafo_size, 2);
|
|
}
|
|
}
|
|
|
|
if (lc->tt.cbf_luma ||
|
|
SAMPLE_CBF(lc->tt.cbf_cb[trafo_depth], x0, y0) ||
|
|
SAMPLE_CBF(lc->tt.cbf_cr[trafo_depth], x0, y0)) {
|
|
if (s->pps->cu_qp_delta_enabled_flag && !lc->tu.is_cu_qp_delta_coded) {
|
|
lc->tu.cu_qp_delta = ff_hevc_cu_qp_delta_abs(s);
|
|
if (lc->tu.cu_qp_delta != 0)
|
|
if (ff_hevc_cu_qp_delta_sign_flag(s) == 1)
|
|
lc->tu.cu_qp_delta = -lc->tu.cu_qp_delta;
|
|
lc->tu.is_cu_qp_delta_coded = 1;
|
|
ff_hevc_set_qPy(s, x0, y0, cb_xBase, cb_yBase, log2_cb_size);
|
|
}
|
|
|
|
if (lc->cu.pred_mode == MODE_INTRA && log2_trafo_size < 4) {
|
|
if (lc->tu.cur_intra_pred_mode >= 6 &&
|
|
lc->tu.cur_intra_pred_mode <= 14) {
|
|
scan_idx = SCAN_VERT;
|
|
} else if (lc->tu.cur_intra_pred_mode >= 22 &&
|
|
lc->tu.cur_intra_pred_mode <= 30) {
|
|
scan_idx = SCAN_HORIZ;
|
|
}
|
|
|
|
if (lc->pu.intra_pred_mode_c >= 6 &&
|
|
lc->pu.intra_pred_mode_c <= 14) {
|
|
scan_idx_c = SCAN_VERT;
|
|
} else if (lc->pu.intra_pred_mode_c >= 22 &&
|
|
lc->pu.intra_pred_mode_c <= 30) {
|
|
scan_idx_c = SCAN_HORIZ;
|
|
}
|
|
}
|
|
|
|
if (lc->tt.cbf_luma)
|
|
hls_residual_coding(s, x0, y0, log2_trafo_size, scan_idx, 0);
|
|
if (log2_trafo_size > 2) {
|
|
if (SAMPLE_CBF(lc->tt.cbf_cb[trafo_depth], x0, y0))
|
|
hls_residual_coding(s, x0, y0, log2_trafo_size - 1, scan_idx_c, 1);
|
|
if (SAMPLE_CBF(lc->tt.cbf_cr[trafo_depth], x0, y0))
|
|
hls_residual_coding(s, x0, y0, log2_trafo_size - 1, scan_idx_c, 2);
|
|
} else if (blk_idx == 3) {
|
|
if (SAMPLE_CBF(lc->tt.cbf_cb[trafo_depth], xBase, yBase))
|
|
hls_residual_coding(s, xBase, yBase, log2_trafo_size, scan_idx_c, 1);
|
|
if (SAMPLE_CBF(lc->tt.cbf_cr[trafo_depth], xBase, yBase))
|
|
hls_residual_coding(s, xBase, yBase, log2_trafo_size, scan_idx_c, 2);
|
|
}
|
|
}
|
|
}
|
|
|
|
static void set_deblocking_bypass(HEVCContext *s, int x0, int y0, int log2_cb_size)
|
|
{
|
|
int cb_size = 1 << log2_cb_size;
|
|
int log2_min_pu_size = s->sps->log2_min_pu_size;
|
|
|
|
int pic_width_in_min_pu = s->sps->width >> s->sps->log2_min_pu_size;
|
|
int x_end = FFMIN(x0 + cb_size, s->sps->width);
|
|
int y_end = FFMIN(y0 + cb_size, s->sps->height);
|
|
int i, j;
|
|
|
|
for (j = (y0 >> log2_min_pu_size); j < (y_end >> log2_min_pu_size); j++)
|
|
for (i = (x0 >> log2_min_pu_size); i < (x_end >> log2_min_pu_size); i++)
|
|
s->is_pcm[i + j * pic_width_in_min_pu] = 2;
|
|
}
|
|
|
|
static void hls_transform_tree(HEVCContext *s, int x0, int y0, int xBase, int yBase, int cb_xBase, int cb_yBase,
|
|
int log2_cb_size, int log2_trafo_size, int trafo_depth, int blk_idx)
|
|
{
|
|
HEVCLocalContext *lc = &s->HEVClc;
|
|
uint8_t split_transform_flag;
|
|
|
|
if (trafo_depth > 0 && log2_trafo_size == 2) {
|
|
SAMPLE_CBF(lc->tt.cbf_cb[trafo_depth], x0, y0) =
|
|
SAMPLE_CBF(lc->tt.cbf_cb[trafo_depth - 1], xBase, yBase);
|
|
SAMPLE_CBF(lc->tt.cbf_cr[trafo_depth], x0, y0) =
|
|
SAMPLE_CBF(lc->tt.cbf_cr[trafo_depth - 1], xBase, yBase);
|
|
} else {
|
|
SAMPLE_CBF(lc->tt.cbf_cb[trafo_depth], x0, y0) =
|
|
SAMPLE_CBF(lc->tt.cbf_cr[trafo_depth], x0, y0) = 0;
|
|
}
|
|
|
|
if (lc->cu.intra_split_flag) {
|
|
if (trafo_depth == 1)
|
|
lc->tu.cur_intra_pred_mode = lc->pu.intra_pred_mode[blk_idx];
|
|
} else {
|
|
lc->tu.cur_intra_pred_mode = lc->pu.intra_pred_mode[0];
|
|
}
|
|
|
|
lc->tt.cbf_luma = 1;
|
|
|
|
lc->tt.inter_split_flag = (s->sps->max_transform_hierarchy_depth_inter == 0 &&
|
|
lc->cu.pred_mode == MODE_INTER &&
|
|
lc->cu.part_mode != PART_2Nx2N && trafo_depth == 0);
|
|
|
|
if (log2_trafo_size <= s->sps->log2_max_trafo_size &&
|
|
log2_trafo_size > s->sps->log2_min_transform_block_size &&
|
|
trafo_depth < lc->cu.max_trafo_depth &&
|
|
!(lc->cu.intra_split_flag && trafo_depth == 0)) {
|
|
split_transform_flag = ff_hevc_split_transform_flag_decode(s, log2_trafo_size);
|
|
} else {
|
|
split_transform_flag = (log2_trafo_size > s->sps->log2_max_trafo_size ||
|
|
(lc->cu.intra_split_flag && (trafo_depth == 0)) ||
|
|
lc->tt.inter_split_flag);
|
|
}
|
|
|
|
if (log2_trafo_size > 2) {
|
|
if (trafo_depth == 0 ||
|
|
SAMPLE_CBF(lc->tt.cbf_cb[trafo_depth - 1], xBase, yBase)) {
|
|
SAMPLE_CBF(lc->tt.cbf_cb[trafo_depth], x0, y0) =
|
|
ff_hevc_cbf_cb_cr_decode(s, trafo_depth);
|
|
}
|
|
|
|
if (trafo_depth == 0 || SAMPLE_CBF(lc->tt.cbf_cr[trafo_depth - 1], xBase, yBase)) {
|
|
SAMPLE_CBF(lc->tt.cbf_cr[trafo_depth], x0, y0) =
|
|
ff_hevc_cbf_cb_cr_decode(s, trafo_depth);
|
|
}
|
|
}
|
|
|
|
if (split_transform_flag) {
|
|
int x1 = x0 + ((1 << log2_trafo_size) >> 1);
|
|
int y1 = y0 + ((1 << log2_trafo_size) >> 1);
|
|
|
|
hls_transform_tree(s, x0, y0, x0, y0, cb_xBase, cb_yBase, log2_cb_size,
|
|
log2_trafo_size - 1, trafo_depth + 1, 0);
|
|
hls_transform_tree(s, x1, y0, x0, y0, cb_xBase, cb_yBase, log2_cb_size,
|
|
log2_trafo_size - 1, trafo_depth + 1, 1);
|
|
hls_transform_tree(s, x0, y1, x0, y0, cb_xBase, cb_yBase, log2_cb_size,
|
|
log2_trafo_size - 1, trafo_depth + 1, 2);
|
|
hls_transform_tree(s, x1, y1, x0, y0, cb_xBase, cb_yBase, log2_cb_size,
|
|
log2_trafo_size - 1, trafo_depth + 1, 3);
|
|
} else {
|
|
int min_tu_size = 1 << s->sps->log2_min_transform_block_size;
|
|
int log2_min_tu_size = s->sps->log2_min_transform_block_size;
|
|
int pic_width_in_min_tu = s->sps->width >> log2_min_tu_size;
|
|
int i, j;
|
|
|
|
if (lc->cu.pred_mode == MODE_INTRA || trafo_depth != 0 ||
|
|
SAMPLE_CBF(lc->tt.cbf_cb[trafo_depth], x0, y0) ||
|
|
SAMPLE_CBF(lc->tt.cbf_cr[trafo_depth], x0, y0)) {
|
|
lc->tt.cbf_luma = ff_hevc_cbf_luma_decode(s, trafo_depth);
|
|
}
|
|
|
|
hls_transform_unit(s, x0, y0, xBase, yBase, cb_xBase, cb_yBase,
|
|
log2_cb_size, log2_trafo_size, trafo_depth, blk_idx);
|
|
|
|
// TODO: store cbf_luma somewhere else
|
|
if (lc->tt.cbf_luma)
|
|
for (i = 0; i < (1 << log2_trafo_size); i += min_tu_size)
|
|
for (j = 0; j < (1 << log2_trafo_size); j += min_tu_size) {
|
|
int x_tu = (x0 + j) >> log2_min_tu_size;
|
|
int y_tu = (y0 + i) >> log2_min_tu_size;
|
|
s->cbf_luma[y_tu * pic_width_in_min_tu + x_tu] = 1;
|
|
}
|
|
if (!s->sh.disable_deblocking_filter_flag) {
|
|
ff_hevc_deblocking_boundary_strengths(s, x0, y0, log2_trafo_size,
|
|
lc->slice_or_tiles_up_boundary,
|
|
lc->slice_or_tiles_left_boundary);
|
|
if (s->pps->transquant_bypass_enable_flag && lc->cu.cu_transquant_bypass_flag)
|
|
set_deblocking_bypass(s, x0, y0, log2_trafo_size);
|
|
}
|
|
}
|
|
}
|
|
|
|
static int hls_pcm_sample(HEVCContext *s, int x0, int y0, int log2_cb_size)
|
|
{
|
|
//TODO: non-4:2:0 support
|
|
HEVCLocalContext *lc = &s->HEVClc;
|
|
GetBitContext gb;
|
|
int cb_size = 1 << log2_cb_size;
|
|
int stride0 = s->frame->linesize[0];
|
|
uint8_t *dst0 = &s->frame->data[0][y0 * stride0 + (x0 << s->sps->pixel_shift)];
|
|
int stride1 = s->frame->linesize[1];
|
|
uint8_t *dst1 = &s->frame->data[1][(y0 >> s->sps->vshift[1]) * stride1 + ((x0 >> s->sps->hshift[1]) << s->sps->pixel_shift)];
|
|
int stride2 = s->frame->linesize[2];
|
|
uint8_t *dst2 = &s->frame->data[2][(y0 >> s->sps->vshift[2]) * stride2 + ((x0 >> s->sps->hshift[2]) << s->sps->pixel_shift)];
|
|
|
|
int length = cb_size * cb_size * s->sps->pcm.bit_depth + ((cb_size * cb_size) >> 1) * s->sps->pcm.bit_depth;
|
|
const uint8_t *pcm = skip_bytes(&s->HEVClc.cc, (length + 7) >> 3);
|
|
int ret;
|
|
|
|
ff_hevc_deblocking_boundary_strengths(s, x0, y0, log2_cb_size,
|
|
lc->slice_or_tiles_up_boundary,
|
|
lc->slice_or_tiles_left_boundary);
|
|
|
|
ret = init_get_bits(&gb, pcm, length);
|
|
if (ret < 0)
|
|
return ret;
|
|
|
|
s->hevcdsp.put_pcm(dst0, stride0, cb_size, &gb, s->sps->pcm.bit_depth);
|
|
s->hevcdsp.put_pcm(dst1, stride1, cb_size / 2, &gb, s->sps->pcm.bit_depth);
|
|
s->hevcdsp.put_pcm(dst2, stride2, cb_size / 2, &gb, s->sps->pcm.bit_depth);
|
|
return 0;
|
|
}
|
|
|
|
static void hls_mvd_coding(HEVCContext *s, int x0, int y0, int log2_cb_size)
|
|
{
|
|
HEVCLocalContext *lc = &s->HEVClc;
|
|
int x = ff_hevc_abs_mvd_greater0_flag_decode(s);
|
|
int y = ff_hevc_abs_mvd_greater0_flag_decode(s);
|
|
|
|
if (x)
|
|
x += ff_hevc_abs_mvd_greater1_flag_decode(s);
|
|
if (y)
|
|
y += ff_hevc_abs_mvd_greater1_flag_decode(s);
|
|
|
|
switch (x) {
|
|
case 2: lc->pu.mvd.x = ff_hevc_mvd_decode(s); break;
|
|
case 1: lc->pu.mvd.x = ff_hevc_mvd_sign_flag_decode(s); break;
|
|
case 0: lc->pu.mvd.x = 0; break;
|
|
}
|
|
|
|
switch (y) {
|
|
case 2: lc->pu.mvd.y = ff_hevc_mvd_decode(s); break;
|
|
case 1: lc->pu.mvd.y = ff_hevc_mvd_sign_flag_decode(s); break;
|
|
case 0: lc->pu.mvd.y = 0; break;
|
|
}
|
|
}
|
|
|
|
/**
|
|
* 8.5.3.2.2.1 Luma sample interpolation process
|
|
*
|
|
* @param s HEVC decoding context
|
|
* @param dst target buffer for block data at block position
|
|
* @param dststride stride of the dst buffer
|
|
* @param ref reference picture buffer at origin (0, 0)
|
|
* @param mv motion vector (relative to block position) to get pixel data from
|
|
* @param x_off horizontal position of block from origin (0, 0)
|
|
* @param y_off vertical position of block from origin (0, 0)
|
|
* @param block_w width of block
|
|
* @param block_h height of block
|
|
*/
|
|
static void luma_mc(HEVCContext *s, int16_t *dst, ptrdiff_t dststride, AVFrame *ref,
|
|
const Mv *mv, int x_off, int y_off, int block_w, int block_h)
|
|
{
|
|
HEVCLocalContext *lc = &s->HEVClc;
|
|
uint8_t *src = ref->data[0];
|
|
ptrdiff_t srcstride = ref->linesize[0];
|
|
int pic_width = s->sps->width;
|
|
int pic_height = s->sps->height;
|
|
|
|
int mx = mv->x & 3;
|
|
int my = mv->y & 3;
|
|
int extra_left = ff_hevc_qpel_extra_before[mx];
|
|
int extra_top = ff_hevc_qpel_extra_before[my];
|
|
|
|
x_off += mv->x >> 2;
|
|
y_off += mv->y >> 2;
|
|
src += y_off * srcstride + (x_off << s->sps->pixel_shift);
|
|
|
|
if (x_off < extra_left || x_off >= pic_width - block_w - ff_hevc_qpel_extra_after[mx] ||
|
|
y_off < extra_top || y_off >= pic_height - block_h - ff_hevc_qpel_extra_after[my]) {
|
|
int offset = extra_top * srcstride + (extra_left << s->sps->pixel_shift);
|
|
|
|
s->vdsp.emulated_edge_mc(lc->edge_emu_buffer, srcstride, src - offset, srcstride,
|
|
block_w + ff_hevc_qpel_extra[mx], block_h + ff_hevc_qpel_extra[my],
|
|
x_off - extra_left, y_off - extra_top,
|
|
pic_width, pic_height);
|
|
src = lc->edge_emu_buffer + offset;
|
|
}
|
|
s->hevcdsp.put_hevc_qpel[my][mx](dst, dststride, src, srcstride, block_w,
|
|
block_h, lc->mc_buffer);
|
|
}
|
|
|
|
/**
|
|
* 8.5.3.2.2.2 Chroma sample interpolation process
|
|
*
|
|
* @param s HEVC decoding context
|
|
* @param dst1 target buffer for block data at block position (U plane)
|
|
* @param dst2 target buffer for block data at block position (V plane)
|
|
* @param dststride stride of the dst1 and dst2 buffers
|
|
* @param ref reference picture buffer at origin (0, 0)
|
|
* @param mv motion vector (relative to block position) to get pixel data from
|
|
* @param x_off horizontal position of block from origin (0, 0)
|
|
* @param y_off vertical position of block from origin (0, 0)
|
|
* @param block_w width of block
|
|
* @param block_h height of block
|
|
*/
|
|
static void chroma_mc(HEVCContext *s, int16_t *dst1, int16_t *dst2, ptrdiff_t dststride, AVFrame *ref,
|
|
const Mv *mv, int x_off, int y_off, int block_w, int block_h)
|
|
{
|
|
HEVCLocalContext *lc = &s->HEVClc;
|
|
uint8_t *src1 = ref->data[1];
|
|
uint8_t *src2 = ref->data[2];
|
|
ptrdiff_t src1stride = ref->linesize[1];
|
|
ptrdiff_t src2stride = ref->linesize[2];
|
|
int pic_width = s->sps->width >> 1;
|
|
int pic_height = s->sps->height >> 1;
|
|
|
|
int mx = mv->x & 7;
|
|
int my = mv->y & 7;
|
|
|
|
x_off += mv->x >> 3;
|
|
y_off += mv->y >> 3;
|
|
src1 += y_off * src1stride + (x_off << s->sps->pixel_shift);
|
|
src2 += y_off * src2stride + (x_off << s->sps->pixel_shift);
|
|
|
|
if (x_off < EPEL_EXTRA_BEFORE || x_off >= pic_width - block_w - EPEL_EXTRA_AFTER ||
|
|
y_off < EPEL_EXTRA_AFTER || y_off >= pic_height - block_h - EPEL_EXTRA_AFTER) {
|
|
int offset1 = EPEL_EXTRA_BEFORE * (src1stride + (1 << s->sps->pixel_shift));
|
|
int offset2 = EPEL_EXTRA_BEFORE * (src2stride + (1 << s->sps->pixel_shift));
|
|
|
|
s->vdsp.emulated_edge_mc(lc->edge_emu_buffer, src1stride, src1 - offset1, src1stride,
|
|
block_w + EPEL_EXTRA, block_h + EPEL_EXTRA,
|
|
x_off - EPEL_EXTRA_BEFORE, y_off - EPEL_EXTRA_BEFORE,
|
|
pic_width, pic_height);
|
|
|
|
src1 = lc->edge_emu_buffer + offset1;
|
|
s->hevcdsp.put_hevc_epel[!!my][!!mx](dst1, dststride, src1, src1stride,
|
|
block_w, block_h, mx, my, lc->mc_buffer);
|
|
|
|
s->vdsp.emulated_edge_mc(lc->edge_emu_buffer, src2stride, src2 - offset2, src2stride,
|
|
block_w + EPEL_EXTRA, block_h + EPEL_EXTRA,
|
|
x_off - EPEL_EXTRA_BEFORE, y_off - EPEL_EXTRA_BEFORE,
|
|
pic_width, pic_height);
|
|
src2 = lc->edge_emu_buffer + offset2;
|
|
s->hevcdsp.put_hevc_epel[!!my][!!mx](dst2, dststride, src2, src2stride,
|
|
block_w, block_h, mx, my, lc->mc_buffer);
|
|
} else {
|
|
s->hevcdsp.put_hevc_epel[!!my][!!mx](dst1, dststride, src1, src1stride,
|
|
block_w, block_h, mx, my, lc->mc_buffer);
|
|
s->hevcdsp.put_hevc_epel[!!my][!!mx](dst2, dststride, src2, src2stride,
|
|
block_w, block_h, mx, my, lc->mc_buffer);
|
|
}
|
|
}
|
|
|
|
static void hevc_await_progress(HEVCContext *s, HEVCFrame *ref,
|
|
const Mv *mv, int y0)
|
|
{
|
|
int y = (mv->y >> 2) + y0;
|
|
|
|
//ff_thread_await_progress(&ref->tf, FFMIN(s->height, y), 0);
|
|
ff_thread_await_progress(&ref->tf, INT_MAX, 0);
|
|
}
|
|
|
|
static void hls_prediction_unit(HEVCContext *s, int x0, int y0, int nPbW, int nPbH, int log2_cb_size, int partIdx)
|
|
{
|
|
#define POS(c_idx, x, y) \
|
|
&s->frame->data[c_idx][((y) >> s->sps->vshift[c_idx]) * s->frame->linesize[c_idx] + \
|
|
(((x) >> s->sps->hshift[c_idx]) << s->sps->pixel_shift)]
|
|
HEVCLocalContext *lc = &s->HEVClc;
|
|
int merge_idx = 0;
|
|
enum InterPredIdc inter_pred_idc = PRED_L0;
|
|
struct MvField current_mv = {{{ 0 }}};
|
|
|
|
int pic_width_in_min_pu = s->sps->width >> s->sps->log2_min_pu_size;
|
|
|
|
MvField *tab_mvf = s->ref->tab_mvf;
|
|
RefPicList *refPicList = s->ref->refPicList;
|
|
HEVCFrame *ref0, *ref1;
|
|
|
|
int tmpstride = MAX_PB_SIZE;
|
|
|
|
uint8_t *dst0 = POS(0, x0, y0);
|
|
uint8_t *dst1 = POS(1, x0, y0);
|
|
uint8_t *dst2 = POS(2, x0, y0);
|
|
int log2_min_cb_size = s->sps->log2_min_coding_block_size;
|
|
int pic_width_in_ctb = s->sps->width>>log2_min_cb_size;
|
|
int x_cb = x0 >> log2_min_cb_size;
|
|
int y_cb = y0 >> log2_min_cb_size;
|
|
int ref_idx[2];
|
|
int mvp_flag[2];
|
|
int x_pu, y_pu;
|
|
int i, j;
|
|
|
|
if (SAMPLE_CTB(s->skip_flag, x_cb, y_cb)) {
|
|
if (s->sh.max_num_merge_cand > 1)
|
|
merge_idx = ff_hevc_merge_idx_decode(s);
|
|
else
|
|
merge_idx = 0;
|
|
|
|
ff_hevc_luma_mv_merge_mode(s, x0, y0, 1 << log2_cb_size, 1 << log2_cb_size,
|
|
log2_cb_size, partIdx, merge_idx, ¤t_mv);
|
|
x_pu = x0 >> s->sps->log2_min_pu_size;
|
|
y_pu = y0 >> s->sps->log2_min_pu_size;
|
|
|
|
for (i = 0; i < nPbW >> s->sps->log2_min_pu_size; i++)
|
|
for (j = 0; j < nPbH >> s->sps->log2_min_pu_size; j++)
|
|
tab_mvf[(y_pu + j) * pic_width_in_min_pu + x_pu + i] = current_mv;
|
|
} else { /* MODE_INTER */
|
|
lc->pu.merge_flag = ff_hevc_merge_flag_decode(s);
|
|
if (lc->pu.merge_flag) {
|
|
if (s->sh.max_num_merge_cand > 1)
|
|
merge_idx = ff_hevc_merge_idx_decode(s);
|
|
else
|
|
merge_idx = 0;
|
|
|
|
ff_hevc_luma_mv_merge_mode(s, x0, y0, nPbW, nPbH, log2_cb_size,
|
|
partIdx, merge_idx, ¤t_mv);
|
|
x_pu = x0 >> s->sps->log2_min_pu_size;
|
|
y_pu = y0 >> s->sps->log2_min_pu_size;
|
|
|
|
for (i = 0; i < nPbW >> s->sps->log2_min_pu_size; i++)
|
|
for (j = 0; j < nPbH >> s->sps->log2_min_pu_size; j++)
|
|
tab_mvf[(y_pu + j) * pic_width_in_min_pu + x_pu + i] = current_mv;
|
|
} else {
|
|
ff_hevc_set_neighbour_available(s, x0, y0, nPbW, nPbH);
|
|
if (s->sh.slice_type == B_SLICE)
|
|
inter_pred_idc = ff_hevc_inter_pred_idc_decode(s, nPbW, nPbH);
|
|
|
|
if (inter_pred_idc != PRED_L1) {
|
|
if (s->sh.nb_refs[L0]) {
|
|
ref_idx[0] = ff_hevc_ref_idx_lx_decode(s, s->sh.nb_refs[L0]);
|
|
current_mv.ref_idx[0] = ref_idx[0];
|
|
}
|
|
current_mv.pred_flag[0] = 1;
|
|
hls_mvd_coding(s, x0, y0, 0);
|
|
mvp_flag[0] = ff_hevc_mvp_lx_flag_decode(s);
|
|
ff_hevc_luma_mv_mvp_mode(s, x0, y0, nPbW, nPbH, log2_cb_size,
|
|
partIdx, merge_idx, ¤t_mv, mvp_flag[0], 0);
|
|
current_mv.mv[0].x += lc->pu.mvd.x;
|
|
current_mv.mv[0].y += lc->pu.mvd.y;
|
|
}
|
|
|
|
if (inter_pred_idc != PRED_L0) {
|
|
if (s->sh.nb_refs[L1]) {
|
|
ref_idx[1] = ff_hevc_ref_idx_lx_decode(s, s->sh.nb_refs[L1]);
|
|
current_mv.ref_idx[1] = ref_idx[1];
|
|
}
|
|
|
|
if (s->sh.mvd_l1_zero_flag == 1 && inter_pred_idc == PRED_BI) {
|
|
lc->pu.mvd.x = 0;
|
|
lc->pu.mvd.y = 0;
|
|
} else {
|
|
hls_mvd_coding(s, x0, y0, 1);
|
|
}
|
|
|
|
current_mv.pred_flag[1] = 1;
|
|
mvp_flag[1] = ff_hevc_mvp_lx_flag_decode(s);
|
|
ff_hevc_luma_mv_mvp_mode(s, x0, y0, nPbW, nPbH, log2_cb_size,
|
|
partIdx, merge_idx, ¤t_mv, mvp_flag[1], 1);
|
|
current_mv.mv[1].x += lc->pu.mvd.x;
|
|
current_mv.mv[1].y += lc->pu.mvd.y;
|
|
}
|
|
|
|
x_pu = x0 >> s->sps->log2_min_pu_size;
|
|
y_pu = y0 >> s->sps->log2_min_pu_size;
|
|
|
|
for (i = 0; i < nPbW >> s->sps->log2_min_pu_size; i++)
|
|
for(j = 0; j < nPbH >> s->sps->log2_min_pu_size; j++)
|
|
tab_mvf[(y_pu + j) * pic_width_in_min_pu + x_pu + i] = current_mv;
|
|
}
|
|
}
|
|
|
|
if (current_mv.pred_flag[0]) {
|
|
ref0 = refPicList[0].ref[current_mv.ref_idx[0]];
|
|
if (!ref0)
|
|
return;
|
|
hevc_await_progress(s, ref0, ¤t_mv.mv[0], y0);
|
|
}
|
|
if (current_mv.pred_flag[1]) {
|
|
ref1 = refPicList[1].ref[current_mv.ref_idx[1]];
|
|
if (!ref1)
|
|
return;
|
|
hevc_await_progress(s, ref1, ¤t_mv.mv[1], y0);
|
|
}
|
|
|
|
if (current_mv.pred_flag[0] && !current_mv.pred_flag[1]) {
|
|
DECLARE_ALIGNED(16, int16_t, tmp [MAX_PB_SIZE * MAX_PB_SIZE]);
|
|
DECLARE_ALIGNED(16, int16_t, tmp2[MAX_PB_SIZE * MAX_PB_SIZE]);
|
|
|
|
luma_mc(s, tmp, tmpstride, ref0->frame,
|
|
¤t_mv.mv[0], x0, y0, nPbW, nPbH);
|
|
|
|
if ((s->sh.slice_type == P_SLICE && s->pps->weighted_pred_flag) ||
|
|
(s->sh.slice_type == B_SLICE && s->pps->weighted_bipred_flag)) {
|
|
s->hevcdsp.weighted_pred(s->sh.luma_log2_weight_denom,
|
|
s->sh.luma_weight_l0[current_mv.ref_idx[0]],
|
|
s->sh.luma_offset_l0[current_mv.ref_idx[0]],
|
|
dst0, s->frame->linesize[0], tmp, tmpstride, nPbW, nPbH);
|
|
} else {
|
|
s->hevcdsp.put_unweighted_pred(dst0, s->frame->linesize[0], tmp, tmpstride, nPbW, nPbH);
|
|
}
|
|
chroma_mc(s, tmp, tmp2, tmpstride, ref0->frame,
|
|
¤t_mv.mv[0], x0 / 2, y0 / 2, nPbW / 2, nPbH / 2);
|
|
|
|
if ((s->sh.slice_type == P_SLICE && s->pps->weighted_pred_flag) ||
|
|
(s->sh.slice_type == B_SLICE && s->pps->weighted_bipred_flag)) {
|
|
s->hevcdsp.weighted_pred(s->sh.chroma_log2_weight_denom,
|
|
s->sh.chroma_weight_l0[current_mv.ref_idx[0]][0],
|
|
s->sh.chroma_offset_l0[current_mv.ref_idx[0]][0],
|
|
dst1, s->frame->linesize[1], tmp, tmpstride,
|
|
nPbW / 2, nPbH / 2);
|
|
s->hevcdsp.weighted_pred(s->sh.chroma_log2_weight_denom,
|
|
s->sh.chroma_weight_l0[current_mv.ref_idx[0]][1],
|
|
s->sh.chroma_offset_l0[current_mv.ref_idx[0]][1],
|
|
dst2, s->frame->linesize[2], tmp2, tmpstride,
|
|
nPbW / 2, nPbH / 2);
|
|
} else {
|
|
s->hevcdsp.put_unweighted_pred(dst1, s->frame->linesize[1], tmp, tmpstride, nPbW/2, nPbH/2);
|
|
s->hevcdsp.put_unweighted_pred(dst2, s->frame->linesize[2], tmp2, tmpstride, nPbW/2, nPbH/2);
|
|
}
|
|
} else if (!current_mv.pred_flag[0] && current_mv.pred_flag[1]) {
|
|
DECLARE_ALIGNED(16, int16_t, tmp [MAX_PB_SIZE * MAX_PB_SIZE]);
|
|
DECLARE_ALIGNED(16, int16_t, tmp2[MAX_PB_SIZE * MAX_PB_SIZE]);
|
|
|
|
if (!ref1)
|
|
return;
|
|
|
|
luma_mc(s, tmp, tmpstride, ref1->frame,
|
|
¤t_mv.mv[1], x0, y0, nPbW, nPbH);
|
|
|
|
if ((s->sh.slice_type == P_SLICE && s->pps->weighted_pred_flag) ||
|
|
(s->sh.slice_type == B_SLICE && s->pps->weighted_bipred_flag)) {
|
|
s->hevcdsp.weighted_pred(s->sh.luma_log2_weight_denom,
|
|
s->sh.luma_weight_l1[current_mv.ref_idx[1]],
|
|
s->sh.luma_offset_l1[current_mv.ref_idx[1]],
|
|
dst0, s->frame->linesize[0], tmp, tmpstride,
|
|
nPbW, nPbH);
|
|
} else {
|
|
s->hevcdsp.put_unweighted_pred(dst0, s->frame->linesize[0], tmp, tmpstride, nPbW, nPbH);
|
|
}
|
|
|
|
chroma_mc(s, tmp, tmp2, tmpstride, ref1->frame,
|
|
¤t_mv.mv[1], x0/2, y0/2, nPbW/2, nPbH/2);
|
|
|
|
if ((s->sh.slice_type == P_SLICE && s->pps->weighted_pred_flag) ||
|
|
(s->sh.slice_type == B_SLICE && s->pps->weighted_bipred_flag)) {
|
|
s->hevcdsp.weighted_pred(s->sh.chroma_log2_weight_denom,
|
|
s->sh.chroma_weight_l1[current_mv.ref_idx[1]][0],
|
|
s->sh.chroma_offset_l1[current_mv.ref_idx[1]][0],
|
|
dst1, s->frame->linesize[1], tmp, tmpstride, nPbW/2, nPbH/2);
|
|
s->hevcdsp.weighted_pred(s->sh.chroma_log2_weight_denom,
|
|
s->sh.chroma_weight_l1[current_mv.ref_idx[1]][1],
|
|
s->sh.chroma_offset_l1[current_mv.ref_idx[1]][1],
|
|
dst2, s->frame->linesize[2], tmp2, tmpstride, nPbW/2, nPbH/2);
|
|
} else {
|
|
s->hevcdsp.put_unweighted_pred(dst1, s->frame->linesize[1], tmp, tmpstride, nPbW/2, nPbH/2);
|
|
s->hevcdsp.put_unweighted_pred(dst2, s->frame->linesize[2], tmp2, tmpstride, nPbW/2, nPbH/2);
|
|
}
|
|
} else if (current_mv.pred_flag[0] && current_mv.pred_flag[1]) {
|
|
DECLARE_ALIGNED(16, int16_t, tmp [MAX_PB_SIZE * MAX_PB_SIZE]);
|
|
DECLARE_ALIGNED(16, int16_t, tmp2[MAX_PB_SIZE * MAX_PB_SIZE]);
|
|
DECLARE_ALIGNED(16, int16_t, tmp3[MAX_PB_SIZE * MAX_PB_SIZE]);
|
|
DECLARE_ALIGNED(16, int16_t, tmp4[MAX_PB_SIZE * MAX_PB_SIZE]);
|
|
HEVCFrame *ref0 = refPicList[0].ref[current_mv.ref_idx[0]];
|
|
HEVCFrame *ref1 = refPicList[1].ref[current_mv.ref_idx[1]];
|
|
|
|
if (!ref0 || !ref1)
|
|
return;
|
|
|
|
luma_mc(s, tmp, tmpstride, ref0->frame,
|
|
¤t_mv.mv[0], x0, y0, nPbW, nPbH);
|
|
luma_mc(s, tmp2, tmpstride, ref1->frame,
|
|
¤t_mv.mv[1], x0, y0, nPbW, nPbH);
|
|
|
|
if ((s->sh.slice_type == P_SLICE && s->pps->weighted_pred_flag) ||
|
|
(s->sh.slice_type == B_SLICE && s->pps->weighted_bipred_flag)){
|
|
s->hevcdsp.weighted_pred_avg(s->sh.luma_log2_weight_denom,
|
|
s->sh.luma_weight_l0[current_mv.ref_idx[0]],
|
|
s->sh.luma_weight_l1[current_mv.ref_idx[1]],
|
|
s->sh.luma_offset_l0[current_mv.ref_idx[0]],
|
|
s->sh.luma_offset_l1[current_mv.ref_idx[1]],
|
|
dst0, s->frame->linesize[0], tmp, tmp2, tmpstride, nPbW, nPbH);
|
|
} else {
|
|
s->hevcdsp.put_weighted_pred_avg(dst0, s->frame->linesize[0], tmp, tmp2, tmpstride, nPbW, nPbH);
|
|
}
|
|
|
|
chroma_mc(s, tmp, tmp2, tmpstride, ref0->frame,
|
|
¤t_mv.mv[0], x0/2, y0/2, nPbW/2, nPbH/2);
|
|
chroma_mc(s, tmp3, tmp4, tmpstride, ref1->frame,
|
|
¤t_mv.mv[1], x0/2, y0/2, nPbW/2, nPbH/2);
|
|
|
|
if ((s->sh.slice_type == P_SLICE && s->pps->weighted_pred_flag) ||
|
|
(s->sh.slice_type == B_SLICE && s->pps->weighted_bipred_flag)) {
|
|
s->hevcdsp.weighted_pred_avg(s->sh.chroma_log2_weight_denom ,
|
|
s->sh.chroma_weight_l0[current_mv.ref_idx[0]][0],
|
|
s->sh.chroma_weight_l1[current_mv.ref_idx[1]][0],
|
|
s->sh.chroma_offset_l0[current_mv.ref_idx[0]][0],
|
|
s->sh.chroma_offset_l1[current_mv.ref_idx[1]][0],
|
|
dst1, s->frame->linesize[1], tmp, tmp3, tmpstride, nPbW/2, nPbH/2);
|
|
s->hevcdsp.weighted_pred_avg(s->sh.chroma_log2_weight_denom ,
|
|
s->sh.chroma_weight_l0[current_mv.ref_idx[0]][1],
|
|
s->sh.chroma_weight_l1[current_mv.ref_idx[1]][1],
|
|
s->sh.chroma_offset_l0[current_mv.ref_idx[0]][1],
|
|
s->sh.chroma_offset_l1[current_mv.ref_idx[1]][1],
|
|
dst2, s->frame->linesize[2], tmp2, tmp4, tmpstride, nPbW/2, nPbH/2);
|
|
} else {
|
|
s->hevcdsp.put_weighted_pred_avg(dst1, s->frame->linesize[1], tmp, tmp3, tmpstride, nPbW/2, nPbH/2);
|
|
s->hevcdsp.put_weighted_pred_avg(dst2, s->frame->linesize[2], tmp2, tmp4, tmpstride, nPbW/2, nPbH/2);
|
|
}
|
|
}
|
|
}
|
|
|
|
/**
|
|
* 8.4.1
|
|
*/
|
|
static int luma_intra_pred_mode(HEVCContext *s, int x0, int y0, int pu_size,
|
|
int prev_intra_luma_pred_flag)
|
|
{
|
|
HEVCLocalContext *lc = &s->HEVClc;
|
|
int x_pu = x0 >> s->sps->log2_min_pu_size;
|
|
int y_pu = y0 >> s->sps->log2_min_pu_size;
|
|
int pic_width_in_min_pu = s->sps->width >> s->sps->log2_min_pu_size;
|
|
int size_in_pus = pu_size >> s->sps->log2_min_pu_size;
|
|
int x0b = x0 & ((1 << s->sps->log2_ctb_size) - 1);
|
|
int y0b = y0 & ((1 << s->sps->log2_ctb_size) - 1);
|
|
|
|
int cand_up = (lc->ctb_up_flag || y0b) ? s->tab_ipm[(y_pu-1)*pic_width_in_min_pu+x_pu] : INTRA_DC ;
|
|
int cand_left = (lc->ctb_left_flag || x0b) ? s->tab_ipm[y_pu*pic_width_in_min_pu+x_pu-1] : INTRA_DC ;
|
|
|
|
int y_ctb = (y0 >> (s->sps->log2_ctb_size)) << (s->sps->log2_ctb_size);
|
|
MvField *tab_mvf = s->ref->tab_mvf;
|
|
int intra_pred_mode;
|
|
int candidate[3];
|
|
int i, j;
|
|
|
|
// intra_pred_mode prediction does not cross vertical CTB boundaries
|
|
if ((y0 - 1) < y_ctb)
|
|
cand_up = INTRA_DC;
|
|
|
|
if (cand_left == cand_up) {
|
|
if (cand_left < 2) {
|
|
candidate[0] = INTRA_PLANAR;
|
|
candidate[1] = INTRA_DC;
|
|
candidate[2] = INTRA_ANGULAR_26;
|
|
} else {
|
|
candidate[0] = cand_left;
|
|
candidate[1] = 2 + ((cand_left - 2 - 1 + 32) & 31);
|
|
candidate[2] = 2 + ((cand_left - 2 + 1) & 31);
|
|
}
|
|
} else {
|
|
candidate[0] = cand_left;
|
|
candidate[1] = cand_up;
|
|
if (candidate[0] != INTRA_PLANAR && candidate[1] != INTRA_PLANAR) {
|
|
candidate[2] = INTRA_PLANAR;
|
|
} else if (candidate[0] != INTRA_DC && candidate[1] != INTRA_DC) {
|
|
candidate[2] = INTRA_DC;
|
|
} else {
|
|
candidate[2] = INTRA_ANGULAR_26;
|
|
}
|
|
}
|
|
|
|
if (prev_intra_luma_pred_flag) {
|
|
intra_pred_mode = candidate[lc->pu.mpm_idx];
|
|
} else {
|
|
if (candidate[0] > candidate[1])
|
|
FFSWAP(uint8_t, candidate[0], candidate[1]);
|
|
if (candidate[0] > candidate[2])
|
|
FFSWAP(uint8_t, candidate[0], candidate[2]);
|
|
if (candidate[1] > candidate[2])
|
|
FFSWAP(uint8_t, candidate[1], candidate[2]);
|
|
|
|
intra_pred_mode = lc->pu.rem_intra_luma_pred_mode;
|
|
for (i = 0; i < 3; i++) {
|
|
if (intra_pred_mode >= candidate[i])
|
|
intra_pred_mode++;
|
|
}
|
|
}
|
|
|
|
/* write the intra prediction units into the mv array */
|
|
if(!size_in_pus)
|
|
size_in_pus = 1;
|
|
for (i = 0; i < size_in_pus; i++) {
|
|
memset(&s->tab_ipm[(y_pu + i) * pic_width_in_min_pu + x_pu],
|
|
intra_pred_mode, size_in_pus);
|
|
|
|
for (j = 0; j < size_in_pus; j++) {
|
|
tab_mvf[(y_pu + j) * pic_width_in_min_pu + x_pu + i].is_intra = 1;
|
|
tab_mvf[(y_pu + j) * pic_width_in_min_pu + x_pu + i].pred_flag[0] = 0;
|
|
tab_mvf[(y_pu + j) * pic_width_in_min_pu + x_pu + i].pred_flag[1] = 0;
|
|
tab_mvf[(y_pu + j) * pic_width_in_min_pu + x_pu + i].ref_idx[0] = 0;
|
|
tab_mvf[(y_pu + j) * pic_width_in_min_pu + x_pu + i].ref_idx[1] = 0;
|
|
tab_mvf[(y_pu + j) * pic_width_in_min_pu + x_pu + i].mv[0].x = 0;
|
|
tab_mvf[(y_pu + j) * pic_width_in_min_pu + x_pu + i].mv[0].y = 0;
|
|
tab_mvf[(y_pu + j) * pic_width_in_min_pu + x_pu + i].mv[1].x = 0;
|
|
tab_mvf[(y_pu + j) * pic_width_in_min_pu + x_pu + i].mv[1].y = 0;
|
|
}
|
|
}
|
|
|
|
return intra_pred_mode;
|
|
}
|
|
|
|
static av_always_inline void set_ct_depth(HEVCContext *s, int x0, int y0,
|
|
int log2_cb_size, int ct_depth)
|
|
{
|
|
int length = (1 << log2_cb_size) >> s->sps->log2_min_coding_block_size;
|
|
int x_cb = x0 >> s->sps->log2_min_coding_block_size;
|
|
int y_cb = y0 >> s->sps->log2_min_coding_block_size;
|
|
int y;
|
|
|
|
for (y = 0; y < length; y++)
|
|
memset(&s->tab_ct_depth[(y_cb + y) * s->sps->min_cb_width + x_cb],
|
|
ct_depth, length);
|
|
}
|
|
|
|
static void intra_prediction_unit(HEVCContext *s, int x0, int y0, int log2_cb_size)
|
|
{
|
|
HEVCLocalContext *lc = &s->HEVClc;
|
|
static const uint8_t intra_chroma_table[4] = {0, 26, 10, 1};
|
|
uint8_t prev_intra_luma_pred_flag[4];
|
|
int split = lc->cu.part_mode == PART_NxN;
|
|
int pb_size = (1 << log2_cb_size) >> split;
|
|
int side = split + 1;
|
|
int chroma_mode;
|
|
int i, j;
|
|
|
|
for (i = 0; i < side; i++)
|
|
for (j = 0; j < side; j++)
|
|
prev_intra_luma_pred_flag[2 * i + j] = ff_hevc_prev_intra_luma_pred_flag_decode(s);
|
|
|
|
for (i = 0; i < side; i++) {
|
|
for (j = 0; j < side; j++) {
|
|
if (prev_intra_luma_pred_flag[2*i+j])
|
|
lc->pu.mpm_idx = ff_hevc_mpm_idx_decode(s);
|
|
else
|
|
lc->pu.rem_intra_luma_pred_mode = ff_hevc_rem_intra_luma_pred_mode_decode(s);
|
|
|
|
lc->pu.intra_pred_mode[2 * i + j] =
|
|
luma_intra_pred_mode(s, x0 + pb_size * j, y0 + pb_size * i, pb_size,
|
|
prev_intra_luma_pred_flag[2 * i + j]);
|
|
}
|
|
}
|
|
|
|
chroma_mode = ff_hevc_intra_chroma_pred_mode_decode(s);
|
|
if (chroma_mode != 4) {
|
|
if (lc->pu.intra_pred_mode[0] == intra_chroma_table[chroma_mode])
|
|
lc->pu.intra_pred_mode_c = 34;
|
|
else
|
|
lc->pu.intra_pred_mode_c = intra_chroma_table[chroma_mode];
|
|
} else {
|
|
lc->pu.intra_pred_mode_c = lc->pu.intra_pred_mode[0];
|
|
}
|
|
}
|
|
|
|
static void intra_prediction_unit_default_value(HEVCContext *s, int x0, int y0, int log2_cb_size)
|
|
{
|
|
HEVCLocalContext *lc = &s->HEVClc;
|
|
int pb_size = 1 << log2_cb_size;
|
|
int size_in_pus = pb_size >> s->sps->log2_min_pu_size;
|
|
int pic_width_in_min_pu = s->sps->width >> s->sps->log2_min_pu_size;
|
|
MvField *tab_mvf = s->ref->tab_mvf;
|
|
int x_pu = x0 >> s->sps->log2_min_pu_size;
|
|
int y_pu = y0 >> s->sps->log2_min_pu_size;
|
|
int j, k;
|
|
|
|
if (size_in_pus == 0)
|
|
size_in_pus = 1;
|
|
for (j = 0; j < size_in_pus; j++) {
|
|
memset(&s->tab_ipm[(y_pu + j) * pic_width_in_min_pu + x_pu], INTRA_DC, size_in_pus);
|
|
for (k = 0; k <size_in_pus; k++)
|
|
tab_mvf[(y_pu + j) * pic_width_in_min_pu + x_pu + k].is_intra = lc->cu.pred_mode == MODE_INTRA;
|
|
}
|
|
}
|
|
|
|
static int hls_coding_unit(HEVCContext *s, int x0, int y0, int log2_cb_size)
|
|
{
|
|
int cb_size = 1 << log2_cb_size;
|
|
HEVCLocalContext *lc = &s->HEVClc;
|
|
int log2_min_cb_size = s->sps->log2_min_coding_block_size;
|
|
int length = cb_size >> log2_min_cb_size;
|
|
int pic_width_in_ctb = s->sps->width >> log2_min_cb_size;
|
|
int x_cb = x0 >> log2_min_cb_size;
|
|
int y_cb = y0 >> log2_min_cb_size;
|
|
int x, y;
|
|
|
|
lc->cu.x = x0;
|
|
lc->cu.y = y0;
|
|
lc->cu.rqt_root_cbf = 1;
|
|
|
|
lc->cu.pred_mode = MODE_INTRA;
|
|
lc->cu.part_mode = PART_2Nx2N;
|
|
lc->cu.intra_split_flag = 0;
|
|
lc->cu.pcm_flag = 0;
|
|
SAMPLE_CTB(s->skip_flag, x_cb, y_cb) = 0;
|
|
for (x = 0; x < 4; x++)
|
|
lc->pu.intra_pred_mode[x] = 1;
|
|
if (s->pps->transquant_bypass_enable_flag) {
|
|
lc->cu.cu_transquant_bypass_flag = ff_hevc_cu_transquant_bypass_flag_decode(s);
|
|
if (lc->cu.cu_transquant_bypass_flag)
|
|
set_deblocking_bypass(s, x0, y0, log2_cb_size);
|
|
} else
|
|
lc->cu.cu_transquant_bypass_flag = 0;
|
|
|
|
|
|
if (s->sh.slice_type != I_SLICE) {
|
|
uint8_t skip_flag = ff_hevc_skip_flag_decode(s, x0, y0, x_cb, y_cb);
|
|
|
|
lc->cu.pred_mode = MODE_SKIP;
|
|
x = y_cb * pic_width_in_ctb + x_cb;
|
|
for (y = 0; y < length; y++) {
|
|
memset(&s->skip_flag[x], skip_flag, length);
|
|
x += pic_width_in_ctb;
|
|
}
|
|
lc->cu.pred_mode = skip_flag ? MODE_SKIP : MODE_INTER;
|
|
}
|
|
|
|
if (SAMPLE_CTB(s->skip_flag, x_cb, y_cb)) {
|
|
hls_prediction_unit(s, x0, y0, cb_size, cb_size, log2_cb_size, 0);
|
|
intra_prediction_unit_default_value(s, x0, y0, log2_cb_size);
|
|
|
|
if (!s->sh.disable_deblocking_filter_flag)
|
|
ff_hevc_deblocking_boundary_strengths(s, x0, y0, log2_cb_size,
|
|
lc->slice_or_tiles_up_boundary,
|
|
lc->slice_or_tiles_left_boundary);
|
|
} else {
|
|
if (s->sh.slice_type != I_SLICE)
|
|
lc->cu.pred_mode = ff_hevc_pred_mode_decode(s);
|
|
if (lc->cu.pred_mode != MODE_INTRA ||
|
|
log2_cb_size == s->sps->log2_min_coding_block_size) {
|
|
lc->cu.part_mode = ff_hevc_part_mode_decode(s, log2_cb_size);
|
|
lc->cu.intra_split_flag = lc->cu.part_mode == PART_NxN &&
|
|
lc->cu.pred_mode == MODE_INTRA;
|
|
}
|
|
|
|
if (lc->cu.pred_mode == MODE_INTRA) {
|
|
if (lc->cu.part_mode == PART_2Nx2N && s->sps->pcm_enabled_flag &&
|
|
log2_cb_size >= s->sps->pcm.log2_min_pcm_cb_size &&
|
|
log2_cb_size <= s->sps->pcm.log2_max_pcm_cb_size) {
|
|
lc->cu.pcm_flag = ff_hevc_pcm_flag_decode(s);
|
|
}
|
|
if (lc->cu.pcm_flag) {
|
|
int ret;
|
|
intra_prediction_unit_default_value(s, x0, y0, log2_cb_size);
|
|
ret = hls_pcm_sample(s, x0, y0, log2_cb_size);
|
|
if(s->sps->pcm.loop_filter_disable_flag)
|
|
set_deblocking_bypass(s, x0, y0, log2_cb_size);
|
|
|
|
if (ret < 0)
|
|
return ret;
|
|
} else {
|
|
intra_prediction_unit(s, x0, y0, log2_cb_size);
|
|
}
|
|
} else {
|
|
intra_prediction_unit_default_value(s, x0, y0, log2_cb_size);
|
|
switch (lc->cu.part_mode) {
|
|
case PART_2Nx2N:
|
|
hls_prediction_unit(s, x0, y0, cb_size, cb_size, log2_cb_size, 0);
|
|
break;
|
|
case PART_2NxN:
|
|
hls_prediction_unit(s, x0, y0, cb_size, cb_size / 2, log2_cb_size, 0);
|
|
hls_prediction_unit(s, x0, y0 + cb_size / 2, cb_size, cb_size/2, log2_cb_size, 1);
|
|
break;
|
|
case PART_Nx2N:
|
|
hls_prediction_unit(s, x0, y0, cb_size / 2, cb_size, log2_cb_size, 0);
|
|
hls_prediction_unit(s, x0 + cb_size / 2, y0, cb_size / 2, cb_size, log2_cb_size, 1);
|
|
break;
|
|
case PART_2NxnU:
|
|
hls_prediction_unit(s, x0, y0, cb_size, cb_size / 4, log2_cb_size, 0);
|
|
hls_prediction_unit(s, x0, y0 + cb_size / 4, cb_size, cb_size * 3 / 4, log2_cb_size, 1);
|
|
break;
|
|
case PART_2NxnD:
|
|
hls_prediction_unit(s, x0, y0, cb_size, cb_size * 3 / 4, log2_cb_size, 0);
|
|
hls_prediction_unit(s, x0, y0 + cb_size * 3 / 4, cb_size, cb_size / 4, log2_cb_size, 1);
|
|
break;
|
|
case PART_nLx2N:
|
|
hls_prediction_unit(s, x0, y0, cb_size / 4, cb_size, log2_cb_size,0);
|
|
hls_prediction_unit(s, x0 + cb_size / 4, y0, cb_size * 3 / 4, cb_size, log2_cb_size, 1);
|
|
break;
|
|
case PART_nRx2N:
|
|
hls_prediction_unit(s, x0, y0, cb_size * 3 / 4, cb_size, log2_cb_size,0);
|
|
hls_prediction_unit(s, x0 + cb_size * 3 / 4, y0, cb_size/4, cb_size, log2_cb_size, 1);
|
|
break;
|
|
case PART_NxN:
|
|
hls_prediction_unit(s, x0, y0, cb_size / 2, cb_size / 2, log2_cb_size, 0);
|
|
hls_prediction_unit(s, x0 + cb_size / 2, y0, cb_size / 2, cb_size / 2, log2_cb_size, 1);
|
|
hls_prediction_unit(s, x0, y0 + cb_size / 2, cb_size / 2, cb_size / 2, log2_cb_size, 2);
|
|
hls_prediction_unit(s, x0 + cb_size / 2, y0 + cb_size / 2, cb_size / 2, cb_size / 2, log2_cb_size, 3);
|
|
break;
|
|
}
|
|
}
|
|
|
|
if (!lc->cu.pcm_flag) {
|
|
if (lc->cu.pred_mode != MODE_INTRA &&
|
|
!(lc->cu.part_mode == PART_2Nx2N && lc->pu.merge_flag)) {
|
|
lc->cu.rqt_root_cbf = ff_hevc_no_residual_syntax_flag_decode(s);
|
|
}
|
|
if (lc->cu.rqt_root_cbf) {
|
|
lc->cu.max_trafo_depth = lc->cu.pred_mode == MODE_INTRA ?
|
|
s->sps->max_transform_hierarchy_depth_intra + lc->cu.intra_split_flag :
|
|
s->sps->max_transform_hierarchy_depth_inter;
|
|
hls_transform_tree(s, x0, y0, x0, y0, x0, y0, log2_cb_size,
|
|
log2_cb_size, 0, 0);
|
|
} else {
|
|
if (!s->sh.disable_deblocking_filter_flag)
|
|
ff_hevc_deblocking_boundary_strengths(s, x0, y0, log2_cb_size,
|
|
lc->slice_or_tiles_up_boundary,
|
|
lc->slice_or_tiles_left_boundary);
|
|
}
|
|
}
|
|
}
|
|
|
|
if (s->pps->cu_qp_delta_enabled_flag && lc->tu.is_cu_qp_delta_coded == 0)
|
|
ff_hevc_set_qPy(s, x0, y0, x0, y0, log2_cb_size);
|
|
|
|
x = y_cb * pic_width_in_ctb + x_cb;
|
|
for (y = 0; y < length; y++) {
|
|
memset(&s->qp_y_tab[x], lc->qp_y, length);
|
|
x += pic_width_in_ctb;
|
|
}
|
|
|
|
set_ct_depth(s, x0, y0, log2_cb_size, lc->ct.depth);
|
|
|
|
return 0;
|
|
}
|
|
|
|
static int hls_coding_quadtree(HEVCContext *s, int x0, int y0, int log2_cb_size, int cb_depth)
|
|
{
|
|
HEVCLocalContext *lc = &s->HEVClc;
|
|
int ret;
|
|
|
|
lc->ct.depth = cb_depth;
|
|
if ((x0 + (1 << log2_cb_size) <= s->sps->width) &&
|
|
(y0 + (1 << log2_cb_size) <= s->sps->height) &&
|
|
log2_cb_size > s->sps->log2_min_coding_block_size) {
|
|
SAMPLE(s->split_cu_flag, x0, y0) =
|
|
ff_hevc_split_coding_unit_flag_decode(s, cb_depth, x0, y0);
|
|
} else {
|
|
SAMPLE(s->split_cu_flag, x0, y0) =
|
|
(log2_cb_size > s->sps->log2_min_coding_block_size);
|
|
}
|
|
if (s->pps->cu_qp_delta_enabled_flag &&
|
|
log2_cb_size >= s->sps->log2_ctb_size - s->pps->diff_cu_qp_delta_depth) {
|
|
lc->tu.is_cu_qp_delta_coded = 0;
|
|
lc->tu.cu_qp_delta = 0;
|
|
}
|
|
|
|
if (SAMPLE(s->split_cu_flag, x0, y0)) {
|
|
int more_data = 0;
|
|
int cb_size = (1 << (log2_cb_size)) >> 1;
|
|
int x1 = x0 + cb_size;
|
|
int y1 = y0 + cb_size;
|
|
|
|
more_data = hls_coding_quadtree(s, x0, y0, log2_cb_size - 1, cb_depth + 1);
|
|
if (more_data < 0)
|
|
return more_data;
|
|
|
|
if (more_data && x1 < s->sps->width)
|
|
more_data = hls_coding_quadtree(s, x1, y0, log2_cb_size - 1, cb_depth + 1);
|
|
if (more_data && y1 < s->sps->height)
|
|
more_data = hls_coding_quadtree(s, x0, y1, log2_cb_size - 1, cb_depth + 1);
|
|
if (more_data && x1 < s->sps->width &&
|
|
y1 < s->sps->height) {
|
|
return hls_coding_quadtree(s, x1, y1, log2_cb_size - 1, cb_depth + 1);
|
|
}
|
|
if (more_data)
|
|
return ((x1 + cb_size) < s->sps->width ||
|
|
(y1 + cb_size) < s->sps->height);
|
|
else
|
|
return 0;
|
|
} else {
|
|
ret = hls_coding_unit(s, x0, y0, log2_cb_size);
|
|
if (ret < 0)
|
|
return ret;
|
|
if ((!((x0 + (1 << log2_cb_size)) %
|
|
(1 << (s->sps->log2_ctb_size))) ||
|
|
(x0 + (1 << log2_cb_size) >= s->sps->width)) &&
|
|
(!((y0 + (1 << log2_cb_size)) %
|
|
(1 << (s->sps->log2_ctb_size))) ||
|
|
(y0 + (1 << log2_cb_size) >= s->sps->height))) {
|
|
int end_of_slice_flag = ff_hevc_end_of_slice_flag_decode(s);
|
|
return !end_of_slice_flag;
|
|
} else {
|
|
return 1;
|
|
}
|
|
}
|
|
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* 7.3.4
|
|
*/
|
|
|
|
static void hls_decode_neighbour(HEVCContext *s, int x_ctb, int y_ctb, int ctb_addr_ts)
|
|
{
|
|
HEVCLocalContext *lc = &s->HEVClc;
|
|
int ctb_size = 1 << s->sps->log2_ctb_size;
|
|
int ctb_addr_rs = s->pps->ctb_addr_ts_to_rs[ctb_addr_ts];
|
|
int ctb_addr_in_slice = ctb_addr_rs - s->sh.slice_addr;
|
|
|
|
int tile_left_boundary;
|
|
int tile_up_boundary;
|
|
int slice_left_boundary;
|
|
int slice_up_boundary;
|
|
|
|
s->tab_slice_address[ctb_addr_rs] = s->sh.slice_addr;
|
|
|
|
|
|
if (s->pps->entropy_coding_sync_enabled_flag) {
|
|
if (x_ctb == 0 && (y_ctb & (ctb_size - 1)) == 0)
|
|
lc->first_qp_group = 1;
|
|
lc->end_of_tiles_x = s->sps->width;
|
|
} else if (s->pps->tiles_enabled_flag) {
|
|
if (ctb_addr_ts && s->pps->tile_id[ctb_addr_ts] != s->pps->tile_id[ctb_addr_ts - 1]) {
|
|
int idxX = s->pps->col_idxX[x_ctb >> s->sps->log2_ctb_size];
|
|
lc->start_of_tiles_x = x_ctb;
|
|
lc->end_of_tiles_x = x_ctb + (s->pps->column_width[idxX]<< s->sps->log2_ctb_size);
|
|
lc->first_qp_group = 1;
|
|
}
|
|
} else {
|
|
lc->end_of_tiles_x = s->sps->width;
|
|
}
|
|
|
|
lc->end_of_tiles_y = FFMIN(y_ctb + ctb_size, s->sps->height);
|
|
|
|
if (s->pps->tiles_enabled_flag) {
|
|
tile_left_boundary = ((x_ctb > 0) &&
|
|
(s->pps->tile_id[ctb_addr_ts] == s->pps->tile_id[s->pps->ctb_addr_rs_to_ts[ctb_addr_rs-1]]));
|
|
slice_left_boundary = ((x_ctb > 0) &&
|
|
(s->tab_slice_address[ctb_addr_rs] == s->tab_slice_address[ctb_addr_rs - 1]));
|
|
tile_up_boundary = ((y_ctb > 0) &&
|
|
(s->pps->tile_id[ctb_addr_ts] == s->pps->tile_id[s->pps->ctb_addr_rs_to_ts[ctb_addr_rs - s->sps->ctb_width]]));
|
|
slice_up_boundary = ((y_ctb > 0) &&
|
|
(s->tab_slice_address[ctb_addr_rs] == s->tab_slice_address[ctb_addr_rs - s->sps->ctb_width]));
|
|
} else {
|
|
tile_left_boundary =
|
|
tile_up_boundary = 1;
|
|
slice_left_boundary = ctb_addr_in_slice > 0;
|
|
slice_up_boundary = ctb_addr_in_slice >= s->sps->ctb_width;
|
|
}
|
|
lc->slice_or_tiles_left_boundary = (!slice_left_boundary) + (!tile_left_boundary << 1);
|
|
lc->slice_or_tiles_up_boundary = (!slice_up_boundary + (!tile_up_boundary << 1));
|
|
lc->ctb_left_flag = ((x_ctb > 0) && (ctb_addr_in_slice > 0) && tile_left_boundary);
|
|
lc->ctb_up_flag = ((y_ctb > 0) && (ctb_addr_in_slice >= s->sps->ctb_width) && tile_up_boundary);
|
|
lc->ctb_up_right_flag = ((y_ctb > 0) && (ctb_addr_in_slice+1 >= s->sps->ctb_width) && (s->pps->tile_id[ctb_addr_ts] == s->pps->tile_id[s->pps->ctb_addr_rs_to_ts[ctb_addr_rs+1 - s->sps->ctb_width]]));
|
|
lc->ctb_up_left_flag = ((x_ctb > 0) && (y_ctb > 0) && (ctb_addr_in_slice-1 >= s->sps->ctb_width) && (s->pps->tile_id[ctb_addr_ts] == s->pps->tile_id[s->pps->ctb_addr_rs_to_ts[ctb_addr_rs-1 - s->sps->ctb_width]]));
|
|
}
|
|
|
|
static int hls_slice_data(HEVCContext *s)
|
|
{
|
|
int ctb_size = 1 << s->sps->log2_ctb_size;
|
|
int more_data = 1;
|
|
int x_ctb = 0;
|
|
int y_ctb = 0;
|
|
int ctb_addr_ts = s->pps->ctb_addr_rs_to_ts[s->sh.slice_ctb_addr_rs];
|
|
|
|
while (more_data && ctb_addr_ts < s->sps->ctb_size) {
|
|
int ctb_addr_rs = s->pps->ctb_addr_ts_to_rs[ctb_addr_ts];
|
|
|
|
x_ctb = (ctb_addr_rs % ((s->sps->width + (ctb_size - 1)) >> s->sps->log2_ctb_size)) << s->sps->log2_ctb_size;
|
|
y_ctb = (ctb_addr_rs / ((s->sps->width + (ctb_size - 1)) >> s->sps->log2_ctb_size)) << s->sps->log2_ctb_size;
|
|
hls_decode_neighbour(s, x_ctb, y_ctb, ctb_addr_ts);
|
|
|
|
ff_hevc_cabac_init(s, ctb_addr_ts);
|
|
|
|
hls_sao_param(s, x_ctb >> s->sps->log2_ctb_size, y_ctb >> s->sps->log2_ctb_size);
|
|
|
|
s->deblock[ctb_addr_rs].disable = s->sh.disable_deblocking_filter_flag;
|
|
s->deblock[ctb_addr_rs].beta_offset = s->sh.beta_offset;
|
|
s->deblock[ctb_addr_rs].tc_offset = s->sh.tc_offset;
|
|
s->filter_slice_edges[ctb_addr_rs] = s->sh.slice_loop_filter_across_slices_enabled_flag;
|
|
|
|
more_data = hls_coding_quadtree(s, x_ctb, y_ctb, s->sps->log2_ctb_size, 0);
|
|
if (more_data < 0)
|
|
return more_data;
|
|
|
|
ctb_addr_ts++;
|
|
ff_hevc_save_states(s, ctb_addr_ts);
|
|
ff_hevc_hls_filters(s, x_ctb, y_ctb, ctb_size);
|
|
}
|
|
|
|
if (x_ctb + ctb_size >= s->sps->width &&
|
|
y_ctb + ctb_size >= s->sps->height)
|
|
ff_hevc_hls_filter(s, x_ctb, y_ctb);
|
|
|
|
return ctb_addr_ts;
|
|
}
|
|
|
|
/**
|
|
* @return AVERROR_INVALIDDATA if the packet is not a valid NAL unit,
|
|
* 0 if the unit should be skipped, 1 otherwise
|
|
*/
|
|
static int hls_nal_unit(HEVCContext *s)
|
|
{
|
|
GetBitContext *gb = &s->HEVClc.gb;
|
|
int nuh_layer_id;
|
|
|
|
if (get_bits1(gb) != 0)
|
|
return AVERROR_INVALIDDATA;
|
|
|
|
s->nal_unit_type = get_bits(gb, 6);
|
|
|
|
nuh_layer_id = get_bits(gb, 6);
|
|
s->temporal_id = get_bits(gb, 3) - 1;
|
|
if (s->temporal_id < 0)
|
|
return AVERROR_INVALIDDATA;
|
|
|
|
av_log(s->avctx, AV_LOG_DEBUG,
|
|
"nal_unit_type: %d, nuh_layer_id: %dtemporal_id: %d\n",
|
|
s->nal_unit_type, nuh_layer_id, s->temporal_id);
|
|
|
|
return (nuh_layer_id == 0);
|
|
}
|
|
|
|
static void restore_tqb_pixels(HEVCContext *s)
|
|
{
|
|
int pic_width_in_min_pu = s->sps->width >> s->sps->log2_min_pu_size;
|
|
int pic_height_in_min_pu = s->sps->height >> s->sps->log2_min_pu_size;
|
|
int min_pu_size = 1 << s->sps->log2_min_pu_size;
|
|
int x, y, c_idx;
|
|
|
|
for (c_idx = 0; c_idx < 3; c_idx++) {
|
|
ptrdiff_t stride = s->frame->linesize[c_idx];
|
|
int hshift = s->sps->hshift[c_idx];
|
|
int vshift = s->sps->vshift[c_idx];
|
|
for (y = 0; y < pic_height_in_min_pu; y++) {
|
|
for (x = 0; x < pic_width_in_min_pu; x++) {
|
|
if (s->is_pcm[y*pic_width_in_min_pu+x]) {
|
|
int n;
|
|
int len = min_pu_size >> hshift;
|
|
uint8_t *src = &s->frame->data[c_idx][((y << s->sps->log2_min_pu_size) >> vshift) * stride + (((x << s->sps->log2_min_pu_size) >> hshift) << s->sps->pixel_shift)];
|
|
uint8_t *dst = &s->sao_frame->data[c_idx][((y << s->sps->log2_min_pu_size) >> vshift) * stride + (((x << s->sps->log2_min_pu_size) >> hshift) << s->sps->pixel_shift)];
|
|
for (n = 0;n < (min_pu_size >> vshift); n++) {
|
|
memcpy(dst,src,len);
|
|
src += stride;
|
|
dst += stride;
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
static int hevc_frame_start(HEVCContext *s)
|
|
{
|
|
HEVCLocalContext *lc = &s->HEVClc;
|
|
int pic_width_in_min_pu = s->sps->width >> s->sps->log2_min_pu_size;
|
|
int pic_height_in_min_pu = s->sps->height >> s->sps->log2_min_pu_size;
|
|
int pic_width_in_min_tu = s->sps->width >> s->sps->log2_min_transform_block_size;
|
|
int pic_height_in_min_tu = s->sps->height >> s->sps->log2_min_transform_block_size;
|
|
int ret;
|
|
|
|
memset(s->horizontal_bs, 0, 2 * s->bs_width * (s->bs_height + 1));
|
|
memset(s->vertical_bs, 0, 2 * s->bs_width * (s->bs_height + 1));
|
|
memset(s->cbf_luma, 0, pic_width_in_min_tu * pic_height_in_min_tu);
|
|
memset(s->is_pcm, 0, pic_width_in_min_pu * pic_height_in_min_pu);
|
|
|
|
lc->start_of_tiles_x = 0;
|
|
s->is_decoded = 0;
|
|
|
|
if (s->pps->tiles_enabled_flag)
|
|
lc->end_of_tiles_x = s->pps->column_width[0] << s->sps->log2_ctb_size;
|
|
|
|
ret = ff_hevc_set_new_ref(s, s->sps->sao_enabled ? &s->sao_frame : &s->frame,
|
|
s->poc);
|
|
if (ret < 0)
|
|
goto fail;
|
|
|
|
av_fast_malloc(&lc->edge_emu_buffer, &lc->edge_emu_buffer_size,
|
|
(MAX_PB_SIZE + 7) * s->ref->frame->linesize[0]);
|
|
if (!lc->edge_emu_buffer) {
|
|
ret = AVERROR(ENOMEM);
|
|
goto fail;
|
|
}
|
|
|
|
ret = ff_hevc_frame_rps(s);
|
|
if (ret < 0) {
|
|
av_log(s->avctx, AV_LOG_ERROR, "Error constructing the frame RPS.\n");
|
|
goto fail;
|
|
}
|
|
|
|
av_frame_unref(s->output_frame);
|
|
ret = ff_hevc_output_frame(s, s->output_frame, 0);
|
|
if (ret < 0)
|
|
goto fail;
|
|
|
|
ff_thread_finish_setup(s->avctx);
|
|
|
|
return 0;
|
|
fail:
|
|
if (s->ref)
|
|
ff_thread_report_progress(&s->ref->tf, INT_MAX, 0);
|
|
s->ref = NULL;
|
|
return ret;
|
|
}
|
|
|
|
static int decode_nal_unit(HEVCContext *s, const uint8_t *nal, int length)
|
|
{
|
|
HEVCLocalContext *lc = &s->HEVClc;
|
|
GetBitContext *gb = &lc->gb;
|
|
int ctb_addr_ts;
|
|
int ret;
|
|
|
|
ret = init_get_bits8(gb, nal, length);
|
|
if (ret < 0)
|
|
return ret;
|
|
|
|
ret = hls_nal_unit(s);
|
|
if (ret < 0) {
|
|
av_log(s->avctx, AV_LOG_ERROR, "Invalid NAL unit %d, skipping.\n",
|
|
s->nal_unit_type);
|
|
if (s->avctx->err_recognition & AV_EF_EXPLODE)
|
|
return ret;
|
|
return 0;
|
|
} else if (!ret)
|
|
return 0;
|
|
|
|
switch (s->nal_unit_type) {
|
|
case NAL_VPS:
|
|
ret = ff_hevc_decode_nal_vps(s);
|
|
if (ret < 0)
|
|
return ret;
|
|
break;
|
|
case NAL_SPS:
|
|
ret = ff_hevc_decode_nal_sps(s);
|
|
if (ret < 0)
|
|
return ret;
|
|
break;
|
|
case NAL_PPS:
|
|
ret = ff_hevc_decode_nal_pps(s);
|
|
if (ret < 0)
|
|
return ret;
|
|
break;
|
|
case NAL_SEI_PREFIX:
|
|
case NAL_SEI_SUFFIX:
|
|
ret = ff_hevc_decode_nal_sei(s);
|
|
if (ret < 0)
|
|
return ret;
|
|
break;
|
|
case NAL_TRAIL_R:
|
|
case NAL_TRAIL_N:
|
|
case NAL_TSA_N:
|
|
case NAL_TSA_R:
|
|
case NAL_STSA_N:
|
|
case NAL_STSA_R:
|
|
case NAL_BLA_W_LP:
|
|
case NAL_BLA_W_RADL:
|
|
case NAL_BLA_N_LP:
|
|
case NAL_IDR_W_RADL:
|
|
case NAL_IDR_N_LP:
|
|
case NAL_CRA_NUT:
|
|
case NAL_RADL_N:
|
|
case NAL_RADL_R:
|
|
case NAL_RASL_N:
|
|
case NAL_RASL_R:
|
|
ret = hls_slice_header(s);
|
|
if (ret < 0)
|
|
return ret;
|
|
|
|
if (s->max_ra == INT_MAX) {
|
|
if (s->nal_unit_type == NAL_CRA_NUT ||
|
|
s->nal_unit_type == NAL_BLA_W_LP ||
|
|
s->nal_unit_type == NAL_BLA_N_LP ||
|
|
s->nal_unit_type == NAL_BLA_N_LP) {
|
|
s->max_ra = s->poc;
|
|
} else {
|
|
if (IS_IDR(s))
|
|
s->max_ra = INT_MIN;
|
|
}
|
|
}
|
|
|
|
if ((s->nal_unit_type == NAL_RASL_R || s->nal_unit_type == NAL_RASL_N) &&
|
|
s->poc <= s->max_ra) {
|
|
s->is_decoded = 0;
|
|
break;
|
|
} else {
|
|
if (s->nal_unit_type == NAL_RASL_R && s->poc > s->max_ra)
|
|
s->max_ra = INT_MIN;
|
|
}
|
|
|
|
if (s->sh.first_slice_in_pic_flag) {
|
|
ret = hevc_frame_start(s);
|
|
if (ret < 0)
|
|
return ret;
|
|
} else if (!s->ref) {
|
|
av_log(s->avctx, AV_LOG_ERROR, "First slice in a frame missing.\n");
|
|
return AVERROR_INVALIDDATA;
|
|
}
|
|
|
|
if (!s->sh.dependent_slice_segment_flag &&
|
|
s->sh.slice_type != I_SLICE) {
|
|
ret = ff_hevc_slice_rpl(s);
|
|
if (ret < 0) {
|
|
av_log(s->avctx, AV_LOG_WARNING, "Error constructing the reference "
|
|
"lists for the current slice.\n");
|
|
if (s->avctx->err_recognition & AV_EF_EXPLODE)
|
|
return ret;
|
|
}
|
|
}
|
|
|
|
ctb_addr_ts = hls_slice_data(s);
|
|
if (ctb_addr_ts >= (s->sps->ctb_width * s->sps->ctb_height)) {
|
|
s->is_decoded = 1;
|
|
if ((s->pps->transquant_bypass_enable_flag ||
|
|
(s->sps->pcm.loop_filter_disable_flag && s->sps->pcm_enabled_flag)) &&
|
|
s->sps->sao_enabled)
|
|
restore_tqb_pixels(s);
|
|
}
|
|
|
|
if (ctb_addr_ts < 0)
|
|
return ctb_addr_ts;
|
|
break;
|
|
case NAL_EOS_NUT:
|
|
case NAL_EOB_NUT:
|
|
s->seq_decode = (s->seq_decode + 1) & 0xff;
|
|
s->max_ra = INT_MAX;
|
|
break;
|
|
case NAL_AUD:
|
|
case NAL_FD_NUT:
|
|
break;
|
|
default:
|
|
av_log(s->avctx, AV_LOG_INFO, "Skipping NAL unit %d\n", s->nal_unit_type);
|
|
}
|
|
|
|
return 0;
|
|
}
|
|
|
|
/* FIXME: This is adapted from ff_h264_decode_nal, avoiding duplication
|
|
between these functions would be nice. */
|
|
static int extract_rbsp(const uint8_t *src, int length,
|
|
HEVCNAL *nal)
|
|
{
|
|
int i, si, di;
|
|
uint8_t *dst;
|
|
|
|
#define STARTCODE_TEST \
|
|
if (i + 2 < length && src[i + 1] == 0 && src[i + 2] <= 3) { \
|
|
if (src[i + 2] != 3) { \
|
|
/* startcode, so we must be past the end */ \
|
|
length = i; \
|
|
} \
|
|
break; \
|
|
}
|
|
#if HAVE_FAST_UNALIGNED
|
|
#define FIND_FIRST_ZERO \
|
|
if (i > 0 && !src[i]) \
|
|
i--; \
|
|
while (src[i]) \
|
|
i++
|
|
#if HAVE_FAST_64BIT
|
|
for (i = 0; i + 1 < length; i += 9) {
|
|
if (!((~AV_RN64A(src + i) &
|
|
(AV_RN64A(src + i) - 0x0100010001000101ULL)) &
|
|
0x8000800080008080ULL))
|
|
continue;
|
|
FIND_FIRST_ZERO;
|
|
STARTCODE_TEST;
|
|
i -= 7;
|
|
}
|
|
#else
|
|
for (i = 0; i + 1 < length; i += 5) {
|
|
if (!((~AV_RN32A(src + i) &
|
|
(AV_RN32A(src + i) - 0x01000101U)) &
|
|
0x80008080U))
|
|
continue;
|
|
FIND_FIRST_ZERO;
|
|
STARTCODE_TEST;
|
|
i -= 3;
|
|
}
|
|
#endif
|
|
#else
|
|
for (i = 0; i + 1 < length; i += 2) {
|
|
if (src[i])
|
|
continue;
|
|
if (i > 0 && src[i - 1] == 0)
|
|
i--;
|
|
STARTCODE_TEST;
|
|
}
|
|
#endif
|
|
|
|
if (i >= length - 1) { // no escaped 0
|
|
nal->data = src;
|
|
nal->size = length;
|
|
return length;
|
|
}
|
|
|
|
av_fast_malloc(&nal->rbsp_buffer, &nal->rbsp_buffer_size,
|
|
length + FF_INPUT_BUFFER_PADDING_SIZE);
|
|
if (!nal->rbsp_buffer)
|
|
return AVERROR(ENOMEM);
|
|
|
|
dst = nal->rbsp_buffer;
|
|
|
|
memcpy(dst, src, i);
|
|
si = di = i;
|
|
while (si + 2 < length) {
|
|
// remove escapes (very rare 1:2^22)
|
|
if (src[si + 2] > 3) {
|
|
dst[di++] = src[si++];
|
|
dst[di++] = src[si++];
|
|
} else if (src[si] == 0 && src[si + 1] == 0) {
|
|
if (src[si + 2] == 3) { // escape
|
|
dst[di++] = 0;
|
|
dst[di++] = 0;
|
|
si += 3;
|
|
|
|
continue;
|
|
} else // next start code
|
|
goto nsc;
|
|
}
|
|
|
|
dst[di++] = src[si++];
|
|
}
|
|
while (si < length)
|
|
dst[di++] = src[si++];
|
|
nsc:
|
|
|
|
memset(dst + di, 0, FF_INPUT_BUFFER_PADDING_SIZE);
|
|
|
|
nal->data = dst;
|
|
nal->size = di;
|
|
return si;
|
|
}
|
|
|
|
static int decode_nal_units(HEVCContext *s, const uint8_t *buf, int length)
|
|
{
|
|
int i, consumed, ret = 0;
|
|
|
|
s->ref = NULL;
|
|
s->eos = 0;
|
|
|
|
/* split the input packet into NAL units, so we know the upper bound on the
|
|
* number of slices in the frame */
|
|
s->nb_nals = 0;
|
|
while (length >= 4) {
|
|
HEVCNAL *nal;
|
|
int extract_length = 0;
|
|
|
|
if (s->disable_au == 0) {
|
|
if (s->is_nalff) {
|
|
int i;
|
|
for (i = 0; i < s->nal_length_size; i++)
|
|
extract_length = (extract_length << 8) | buf[i];
|
|
buf += s->nal_length_size;
|
|
length -= s->nal_length_size;
|
|
|
|
if (extract_length > length) {
|
|
av_log(s->avctx, AV_LOG_ERROR, "Invalid NAL unit size.\n");
|
|
ret = AVERROR_INVALIDDATA;
|
|
goto fail;
|
|
}
|
|
} else {
|
|
if (buf[2] == 0) {
|
|
length--;
|
|
buf++;
|
|
continue;
|
|
}
|
|
if (buf[0] != 0 || buf[1] != 0 || buf[2] != 1) {
|
|
ret = AVERROR_INVALIDDATA;
|
|
goto fail;
|
|
}
|
|
|
|
buf += 3;
|
|
length -= 3;
|
|
}
|
|
}
|
|
if (!s->is_nalff || s->disable_au)
|
|
extract_length = length;
|
|
|
|
if (s->nals_allocated < s->nb_nals + 1) {
|
|
int new_size = s->nals_allocated + 1;
|
|
HEVCNAL *tmp = av_realloc_array(s->nals, new_size, sizeof(*tmp));
|
|
if (!tmp) {
|
|
ret = AVERROR(ENOMEM);
|
|
goto fail;
|
|
}
|
|
s->nals = tmp;
|
|
memset(s->nals + s->nals_allocated, 0, (new_size - s->nals_allocated) * sizeof(*tmp));
|
|
s->nals_allocated = new_size;
|
|
}
|
|
nal = &s->nals[s->nb_nals++];
|
|
|
|
consumed = extract_rbsp(buf, extract_length, nal);
|
|
if (consumed < 0) {
|
|
ret = consumed;
|
|
goto fail;
|
|
}
|
|
|
|
ret = init_get_bits8(&s->HEVClc.gb, nal->data, nal->size);
|
|
if (ret < 0)
|
|
goto fail;
|
|
hls_nal_unit(s);
|
|
|
|
if (s->nal_unit_type == NAL_EOS_NUT || s->nal_unit_type == NAL_EOS_NUT)
|
|
s->eos = 1;
|
|
|
|
buf += consumed;
|
|
length -= consumed;
|
|
}
|
|
|
|
/* parse the NAL units */
|
|
for (i = 0; i < s->nb_nals; i++) {
|
|
int ret = decode_nal_unit(s, s->nals[i].data, s->nals[i].size);
|
|
if (ret < 0) {
|
|
av_log(s->avctx, AV_LOG_WARNING, "Error parsing NAL unit #%d.\n", i);
|
|
if (s->avctx->err_recognition & AV_EF_EXPLODE)
|
|
goto fail;
|
|
}
|
|
}
|
|
|
|
fail:
|
|
if (s->ref)
|
|
ff_thread_report_progress(&s->ref->tf, INT_MAX, 0);
|
|
|
|
return ret;
|
|
}
|
|
|
|
static void print_md5(void *log_ctx, int level, uint8_t md5[16])
|
|
{
|
|
int i;
|
|
for (i = 0; i < 16; i++)
|
|
av_log(log_ctx, level, "%02"PRIx8, md5[i]);
|
|
}
|
|
|
|
static int verify_md5(HEVCContext *s, AVFrame *frame)
|
|
{
|
|
const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(frame->format);
|
|
int pixel_shift = desc->comp[0].depth_minus1 > 7;
|
|
int i, j;
|
|
|
|
if (!desc)
|
|
return AVERROR(EINVAL);
|
|
|
|
av_log(s->avctx, AV_LOG_DEBUG, "Verifying checksum for frame with POC %d: ",
|
|
s->poc);
|
|
|
|
/* the checksums are LE, so we have to byteswap for >8bpp formats
|
|
* on BE arches */
|
|
#if HAVE_BIGENDIAN
|
|
if (pixel_shift && !s->checksum_buf) {
|
|
av_fast_malloc(&s->checksum_buf, &s->checksum_buf_size,
|
|
FFMAX3(frame->linesize[0], frame->linesize[1],
|
|
frame->linesize[2]));
|
|
if (!s->checksum_buf)
|
|
return AVERROR(ENOMEM);
|
|
}
|
|
#endif
|
|
|
|
for (i = 0; frame->data[i]; i++) {
|
|
int width = s->avctx->coded_width;
|
|
int height = s->avctx->coded_height;
|
|
int w = (i == 1 || i == 2) ? (width >> desc->log2_chroma_w) : width;
|
|
int h = (i == 1 || i == 2) ? (height >> desc->log2_chroma_h) : height;
|
|
uint8_t md5[16];
|
|
|
|
av_md5_init(s->md5_ctx);
|
|
for (j = 0; j < h; j++) {
|
|
const uint8_t *src = frame->data[i] + j * frame->linesize[i];
|
|
#if HAVE_BIGENDIAN
|
|
if (pixel_shift) {
|
|
s->dsp.bswap16_buf((uint16_t*)s->checksum_buf,
|
|
(const uint16_t*)src, w);
|
|
src = s->checksum_buf;
|
|
}
|
|
#endif
|
|
av_md5_update(s->md5_ctx, src, w << pixel_shift);
|
|
}
|
|
av_md5_final(s->md5_ctx, md5);
|
|
|
|
if (!memcmp(md5, s->md5[i], 16)) {
|
|
av_log (s->avctx, AV_LOG_DEBUG, "plane %d - correct ", i);
|
|
print_md5(s->avctx, AV_LOG_DEBUG, md5);
|
|
av_log (s->avctx, AV_LOG_DEBUG, "; ");
|
|
} else {
|
|
av_log (s->avctx, AV_LOG_ERROR, "mismatching checksum of plane %d - ", i);
|
|
print_md5(s->avctx, AV_LOG_ERROR, md5);
|
|
av_log (s->avctx, AV_LOG_ERROR, " != ");
|
|
print_md5(s->avctx, AV_LOG_ERROR, s->md5[i]);
|
|
av_log (s->avctx, AV_LOG_ERROR, "\n");
|
|
return AVERROR_INVALIDDATA;
|
|
}
|
|
}
|
|
|
|
av_log(s->avctx, AV_LOG_DEBUG, "\n");
|
|
|
|
return 0;
|
|
}
|
|
|
|
static int hevc_decode_frame(AVCodecContext *avctx, void *data, int *got_output,
|
|
AVPacket *avpkt)
|
|
{
|
|
int ret;
|
|
HEVCContext *s = avctx->priv_data;
|
|
|
|
//av_log(avctx, AV_LOG_WARNING, "decode size %d\n", avpkt->size);
|
|
|
|
if (!avpkt->size) {
|
|
ret = ff_hevc_output_frame(s, data, 1);
|
|
if (ret < 0)
|
|
return ret;
|
|
|
|
*got_output = ret;
|
|
return 0;
|
|
}
|
|
|
|
s->ref = NULL;
|
|
ret = decode_nal_units(s, avpkt->data, avpkt->size);
|
|
if (ret < 0)
|
|
return ret;
|
|
|
|
/* verify the SEI checksum */
|
|
if (avctx->err_recognition & AV_EF_CRCCHECK && s->is_decoded &&
|
|
s->is_md5) {
|
|
ret = verify_md5(s, s->ref->frame);
|
|
if (ret < 0 && avctx->err_recognition & AV_EF_EXPLODE) {
|
|
ff_hevc_unref_frame(s, s->ref, ~0);
|
|
return ret;
|
|
}
|
|
}
|
|
s->is_md5 = 0;
|
|
|
|
if (s->is_decoded) {
|
|
av_log(avctx, AV_LOG_DEBUG, "Decoded frame with POC %d.\n", s->poc);
|
|
s->is_decoded = 0;
|
|
}
|
|
|
|
if (s->output_frame->buf[0]) {
|
|
av_frame_move_ref(data, s->output_frame);
|
|
*got_output = 1;
|
|
}
|
|
|
|
return avpkt->size;
|
|
}
|
|
|
|
static int hevc_ref_frame(HEVCContext *s, HEVCFrame *dst, HEVCFrame *src)
|
|
{
|
|
int ret;
|
|
|
|
ret = ff_thread_ref_frame(&dst->tf, &src->tf);
|
|
if (ret < 0)
|
|
return ret;
|
|
|
|
dst->tab_mvf_buf = av_buffer_ref(src->tab_mvf_buf);
|
|
if (!dst->tab_mvf_buf)
|
|
goto fail;
|
|
dst->tab_mvf = src->tab_mvf;
|
|
|
|
dst->rpl_tab_buf = av_buffer_ref(src->rpl_tab_buf);
|
|
if (!dst->rpl_tab_buf)
|
|
goto fail;
|
|
dst->rpl_tab = src->rpl_tab;
|
|
|
|
dst->rpl_buf = av_buffer_ref(src->rpl_buf);
|
|
if (!dst->rpl_buf)
|
|
goto fail;
|
|
|
|
dst->poc = src->poc;
|
|
dst->ctb_count = src->ctb_count;
|
|
dst->window = src->window;
|
|
dst->flags = src->flags;
|
|
dst->sequence = src->sequence;
|
|
|
|
return 0;
|
|
fail:
|
|
ff_hevc_unref_frame(s, dst, ~0);
|
|
return AVERROR(ENOMEM);
|
|
}
|
|
|
|
static av_cold int hevc_decode_free(AVCodecContext *avctx)
|
|
{
|
|
HEVCContext *s = avctx->priv_data;
|
|
HEVCLocalContext *lc = &s->HEVClc;
|
|
int i;
|
|
|
|
pic_arrays_free(s);
|
|
|
|
av_freep(&lc->edge_emu_buffer);
|
|
av_freep(&s->md5_ctx);
|
|
|
|
av_frame_free(&s->tmp_frame);
|
|
av_frame_free(&s->output_frame);
|
|
|
|
for (i = 0; i < FF_ARRAY_ELEMS(s->DPB); i++) {
|
|
ff_hevc_unref_frame(s, &s->DPB[i], ~0);
|
|
av_frame_free(&s->DPB[i].frame);
|
|
}
|
|
|
|
for (i = 0; i < FF_ARRAY_ELEMS(s->vps_list); i++)
|
|
av_freep(&s->vps_list[i]);
|
|
for (i = 0; i < FF_ARRAY_ELEMS(s->sps_list); i++)
|
|
av_buffer_unref(&s->sps_list[i]);
|
|
for (i = 0; i < FF_ARRAY_ELEMS(s->pps_list); i++)
|
|
av_buffer_unref(&s->pps_list[i]);
|
|
|
|
for (i = 0; i < s->nals_allocated; i++)
|
|
av_freep(&s->nals[i].rbsp_buffer);
|
|
av_freep(&s->nals);
|
|
s->nals_allocated = 0;
|
|
|
|
return 0;
|
|
}
|
|
|
|
static av_cold int hevc_init_context(AVCodecContext *avctx)
|
|
{
|
|
HEVCContext *s = avctx->priv_data;
|
|
int i;
|
|
|
|
s->avctx = avctx;
|
|
|
|
s->tmp_frame = av_frame_alloc();
|
|
if (!s->tmp_frame)
|
|
goto fail;
|
|
|
|
s->output_frame = av_frame_alloc();
|
|
if (!s->output_frame)
|
|
goto fail;
|
|
|
|
for (i = 0; i < FF_ARRAY_ELEMS(s->DPB); i++) {
|
|
s->DPB[i].frame = av_frame_alloc();
|
|
if (!s->DPB[i].frame)
|
|
goto fail;
|
|
s->DPB[i].tf.f = s->DPB[i].frame;
|
|
}
|
|
|
|
s->max_ra = INT_MAX;
|
|
|
|
s->md5_ctx = av_md5_alloc();
|
|
if (!s->md5_ctx)
|
|
goto fail;
|
|
|
|
ff_dsputil_init(&s->dsp, avctx);
|
|
|
|
s->context_initialized = 1;
|
|
|
|
return 0;
|
|
fail:
|
|
hevc_decode_free(avctx);
|
|
return AVERROR(ENOMEM);
|
|
}
|
|
|
|
static int hevc_update_thread_context(AVCodecContext *dst,
|
|
const AVCodecContext *src)
|
|
{
|
|
HEVCContext *s = dst->priv_data;
|
|
HEVCContext *s0 = src->priv_data;
|
|
int i, ret;
|
|
|
|
if (!s->context_initialized) {
|
|
ret = hevc_init_context(dst);
|
|
if (ret < 0)
|
|
return ret;
|
|
}
|
|
|
|
for (i = 0; i < FF_ARRAY_ELEMS(s->DPB); i++) {
|
|
ff_hevc_unref_frame(s, &s->DPB[i], ~0);
|
|
if (s0->DPB[i].frame->buf[0]) {
|
|
ret = hevc_ref_frame(s, &s->DPB[i], &s0->DPB[i]);
|
|
if (ret < 0)
|
|
return ret;
|
|
}
|
|
}
|
|
|
|
for (i = 0; i < FF_ARRAY_ELEMS(s->sps_list); i++) {
|
|
av_buffer_unref(&s->sps_list[i]);
|
|
if (s0->sps_list[i]) {
|
|
s->sps_list[i] = av_buffer_ref(s0->sps_list[i]);
|
|
if (!s->sps_list[i])
|
|
return AVERROR(ENOMEM);
|
|
}
|
|
}
|
|
|
|
for (i = 0; i < FF_ARRAY_ELEMS(s->pps_list); i++) {
|
|
av_buffer_unref(&s->pps_list[i]);
|
|
if (s0->pps_list[i]) {
|
|
s->pps_list[i] = av_buffer_ref(s0->pps_list[i]);
|
|
if (!s->pps_list[i])
|
|
return AVERROR(ENOMEM);
|
|
}
|
|
}
|
|
|
|
s->seq_decode = s0->seq_decode;
|
|
s->seq_output = s0->seq_output;
|
|
s->pocTid0 = s0->pocTid0;
|
|
s->max_ra = s0->max_ra;
|
|
|
|
s->is_nalff = s0->is_nalff;
|
|
s->nal_length_size = s0->nal_length_size;
|
|
|
|
|
|
if (s0->eos) {
|
|
s->seq_decode = (s->seq_decode + 1) & 0xff;
|
|
s->max_ra = INT_MAX;
|
|
}
|
|
|
|
return 0;
|
|
}
|
|
|
|
static int hevc_decode_extradata(HEVCContext *s)
|
|
{
|
|
AVCodecContext *avctx = s->avctx;
|
|
GetByteContext gb;
|
|
int ret;
|
|
|
|
bytestream2_init(&gb, avctx->extradata, avctx->extradata_size);
|
|
|
|
if (avctx->extradata_size > 3 &&
|
|
(avctx->extradata[0] || avctx->extradata[1] ||
|
|
avctx->extradata[2] > 1)) {
|
|
/* It seems the extradata is encoded as hvcC format.
|
|
* Temporarily, we support configurationVersion==0 until 14496-15 3rd finalized.
|
|
* When finalized, configurationVersion will be 1 and we can recognize hvcC by
|
|
* checking if avctx->extradata[0]==1 or not. */
|
|
int i, j, num_arrays;
|
|
int nal_len_size;
|
|
|
|
s->is_nalff = 1;
|
|
|
|
bytestream2_skip(&gb, 21);
|
|
nal_len_size = (bytestream2_get_byte(&gb) & 3) + 1;
|
|
num_arrays = bytestream2_get_byte(&gb);
|
|
|
|
/* nal units in the hvcC always have length coded with 2 bytes,
|
|
* so put a fake nal_length_size = 2 while parsing them */
|
|
s->nal_length_size = 2;
|
|
|
|
/* Decode nal units from hvcC. */
|
|
for (i = 0; i < num_arrays; i++) {
|
|
int type = bytestream2_get_byte(&gb) & 0x3f;
|
|
int cnt = bytestream2_get_be16(&gb);
|
|
|
|
for (j = 0; j < cnt; j++) {
|
|
// +2 for the nal size field
|
|
int nalsize = bytestream2_peek_be16(&gb) + 2;
|
|
if (bytestream2_get_bytes_left(&gb) < nalsize) {
|
|
av_log(s->avctx, AV_LOG_ERROR, "Invalid NAL unit size in extradata.\n");
|
|
return AVERROR_INVALIDDATA;
|
|
}
|
|
|
|
ret = decode_nal_units(s, gb.buffer, nalsize);
|
|
if (ret < 0) {
|
|
av_log(avctx, AV_LOG_ERROR,
|
|
"Decoding nal unit %d %d from hvcC failed\n", type, i);
|
|
return ret;
|
|
}
|
|
bytestream2_skip(&gb, nalsize);
|
|
}
|
|
}
|
|
|
|
/* Now store right nal length size, that will be used to parse all other nals */
|
|
s->nal_length_size = nal_len_size;
|
|
} else {
|
|
s->is_nalff = 0;
|
|
ret = decode_nal_units(s, avctx->extradata, avctx->extradata_size);
|
|
if (ret < 0)
|
|
return ret;
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
static av_cold int hevc_decode_init(AVCodecContext *avctx)
|
|
{
|
|
HEVCContext *s = avctx->priv_data;
|
|
int ret;
|
|
|
|
ff_init_cabac_states();
|
|
|
|
avctx->internal->allocate_progress = 1;
|
|
|
|
ret = hevc_init_context(avctx);
|
|
if (ret < 0)
|
|
return ret;
|
|
|
|
if (avctx->extradata_size > 0 && avctx->extradata) {
|
|
ret = hevc_decode_extradata(s);
|
|
if (ret < 0) {
|
|
hevc_decode_free(avctx);
|
|
return ret;
|
|
}
|
|
}
|
|
|
|
return 0;
|
|
}
|
|
|
|
static av_cold int hevc_init_thread_copy(AVCodecContext *avctx)
|
|
{
|
|
HEVCContext *s = avctx->priv_data;
|
|
int ret;
|
|
|
|
memset(s, 0, sizeof(*s));
|
|
|
|
ret = hevc_init_context(avctx);
|
|
if (ret < 0)
|
|
return ret;
|
|
|
|
return 0;
|
|
}
|
|
|
|
static void hevc_decode_flush(AVCodecContext *avctx)
|
|
{
|
|
HEVCContext *s = avctx->priv_data;
|
|
ff_hevc_flush_dpb(s);
|
|
s->max_ra = INT_MAX;
|
|
}
|
|
|
|
#define OFFSET(x) offsetof(HEVCContext, x)
|
|
#define PAR (AV_OPT_FLAG_DECODING_PARAM | AV_OPT_FLAG_VIDEO_PARAM)
|
|
static const AVOption options[] = {
|
|
{ "disable-au", "disable read frame AU by AU", OFFSET(disable_au),
|
|
AV_OPT_TYPE_INT, {.i64 = 0}, 0, 1, PAR },
|
|
{ "strict-displaywin", "stricly apply default display window size", OFFSET(strict_def_disp_win),
|
|
AV_OPT_TYPE_INT, {.i64 = 0}, 0, 1, PAR },
|
|
{ NULL },
|
|
};
|
|
|
|
static const AVClass hevc_decoder_class = {
|
|
.class_name = "HEVC decoder",
|
|
.item_name = av_default_item_name,
|
|
.option = options,
|
|
.version = LIBAVUTIL_VERSION_INT,
|
|
};
|
|
|
|
AVCodec ff_hevc_decoder = {
|
|
.name = "hevc",
|
|
.long_name = NULL_IF_CONFIG_SMALL("HEVC (High Efficiency Video Coding)"),
|
|
.type = AVMEDIA_TYPE_VIDEO,
|
|
.id = AV_CODEC_ID_HEVC,
|
|
.priv_data_size = sizeof(HEVCContext),
|
|
.priv_class = &hevc_decoder_class,
|
|
.init = hevc_decode_init,
|
|
.close = hevc_decode_free,
|
|
.decode = hevc_decode_frame,
|
|
.flush = hevc_decode_flush,
|
|
.update_thread_context = hevc_update_thread_context,
|
|
.init_thread_copy = hevc_init_thread_copy,
|
|
.capabilities = CODEC_CAP_DR1 | CODEC_CAP_DELAY | CODEC_CAP_FRAME_THREADS | CODEC_CAP_EXPERIMENTAL,
|
|
};
|