falkTX
/
FFmpeg
mirror of https://github.com/falkTX/FFmpeg.git
1
0
Fork 0
Code Issues 0 Releases 338 Wiki Activity
Browse Source

avcodec/hevc: Adjust white-spaces to reduce difference to 064698d381 

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
tags/n2.2-rc1
Michael Niedermayer 11 years ago
parent
38612379bf
commit
f578e5d937
12 changed files with 1125 additions and 1016 deletions
Split View
Diff Options
Show Stats Download Patch File Download Diff File
  1. +85
    -61
      libavcodec/hevc.c
  2. +285
    -288
      libavcodec/hevc_cabac.c
  3. +183
    -120
      libavcodec/hevc_filter.c
  4. +69
    -66
      libavcodec/hevc_mvs.c
  5. +7
    -6
      libavcodec/hevc_parser.c
  6. +31
    -29
      libavcodec/hevc_ps.c
  7. +29
    -27
      libavcodec/hevc_refs.c
  8. +10
    -9
      libavcodec/hevc_sei.c
  9. +126
    -127
      libavcodec/hevcdsp.c
  10. +222
    -210
      libavcodec/hevcdsp_template.c
  11. +6
    -6
      libavcodec/hevcpred.c
  12. +72
    -67
      libavcodec/hevcpred_template.c

+ 85
- 61
libavcodec/hevc.c View File

@@ -112,8 +112,10 @@ static int pic_arrays_init(HEVCContext *s, const HEVCSPS *sps)
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));
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;

@@ -130,6 +132,7 @@ static int pic_arrays_init(HEVCContext *s, const HEVCSPS *sps)
goto fail;

return 0;

fail:
pic_arrays_free(s);
return AVERROR(ENOMEM);
@@ -308,6 +311,7 @@ static int set_sps(HEVCContext *s, const HEVCSPS *sps)
s->sps = sps;
s->vps = s->vps_list[s->sps->vps_id];
return 0;

fail:
pic_arrays_free(s);
s->sps = NULL;
@@ -366,7 +370,8 @@ static int hls_slice_header(HEVCContext *s)
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",
av_log(s->avctx, AV_LOG_ERROR,
"Invalid slice segment address: %u.\n",
sh->slice_segment_addr);
return AVERROR_INVALIDDATA;
}
@@ -388,7 +393,8 @@ static int hls_slice_header(HEVCContext *s)
skip_bits(gb, 1); // slice_reserved_undetermined_flag[]

sh->slice_type = get_ue_golomb_long(gb);
if (!(sh->slice_type == I_SLICE || sh->slice_type == P_SLICE ||
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);
@@ -453,7 +459,7 @@ static int hls_slice_header(HEVCContext *s)
sh->slice_temporal_mvp_enabled_flag = 0;
} else {
s->sh.short_term_rps = NULL;
s->poc = 0;
s->poc = 0;
}

/* 8.3.1 */
@@ -537,7 +543,8 @@ static int hls_slice_header(HEVCContext *s)
sh->collocated_ref_idx = get_ue_golomb_long(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);
"Invalid collocated_ref_idx: %d.\n",
sh->collocated_ref_idx);
return AVERROR_INVALIDDATA;
}
}
@@ -580,13 +587,13 @@ static int hls_slice_header(HEVCContext *s)
}
} else {
sh->disable_deblocking_filter_flag = s->pps->disable_dbf;
sh->beta_offset = s->pps->beta_offset;
sh->tc_offset = s->pps->tc_offset;
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;
sh->beta_offset = 0;
sh->tc_offset = 0;
}

if (s->pps->seq_loop_filter_across_slices_enabled_flag &&
@@ -714,7 +721,8 @@ static void hls_sao_param(HEVCContext *s, int rx, int ry)
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));
SET_SAO(offset_sign[c_idx][i],
ff_hevc_sao_offset_sign_decode(s));
} else {
sao->offset_sign[c_idx][i] = 0;
}
@@ -741,7 +749,6 @@ static void hls_sao_param(HEVCContext *s, int rx, int ry)
#undef SET_SAO
#undef CTB


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,
@@ -761,7 +768,8 @@ static void hls_transform_unit(HEVCContext *s, int x0, int y0,
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);
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);
}
@@ -791,7 +799,7 @@ static void hls_transform_unit(HEVCContext *s, int x0, int y0,
scan_idx = SCAN_HORIZ;
}

if (lc->pu.intra_pred_mode_c >= 6 &&
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 &&
@@ -821,7 +829,7 @@ static void set_deblocking_bypass(HEVCContext *s, int x0, int y0, int log2_cb_si
int cb_size = 1 << log2_cb_size;
int log2_min_pu_size = s->sps->log2_min_pu_size;

int min_pu_width = s->sps->min_pu_width;
int min_pu_width = s->sps->min_pu_width;
int x_end = FFMIN(x0 + cb_size, s->sps->width);
int y_end = FFMIN(y0 + cb_size, s->sps->height);
int i, j;
@@ -859,18 +867,19 @@ static void hls_transform_tree(HEVCContext *s, int x0, int y0,
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);
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_tb_size &&
trafo_depth < lc->cu.max_trafo_depth &&
log2_trafo_size > s->sps->log2_min_tb_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);
(lc->cu.intra_split_flag && (trafo_depth == 0)) ||
lc->tt.inter_split_flag);
}

if (log2_trafo_size > 2) {
@@ -880,7 +889,8 @@ static void hls_transform_tree(HEVCContext *s, int x0, int 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)) {
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);
}
@@ -926,7 +936,8 @@ static void hls_transform_tree(HEVCContext *s, int x0, int y0,
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)
if (s->pps->transquant_bypass_enable_flag &&
lc->cu.cu_transquant_bypass_flag)
set_deblocking_bypass(s, x0, y0, log2_trafo_size);
}
}
@@ -957,7 +968,7 @@ static int hls_pcm_sample(HEVCContext *s, int x0, int y0, int log2_cb_size)
if (ret < 0)
return ret;

s->hevcdsp.put_pcm(dst0, stride0, cb_size, &gb, s->sps->pcm.bit_depth);
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_chroma);
s->hevcdsp.put_pcm(dst2, stride2, cb_size / 2, &gb, s->sps->pcm.bit_depth_chroma);
return 0;
@@ -1001,7 +1012,8 @@ static void luma_mc(HEVCContext *s, int16_t *dst, ptrdiff_t dststride,
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],
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;
@@ -1024,8 +1036,9 @@ static void luma_mc(HEVCContext *s, int16_t *dst, ptrdiff_t dststride,
* @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)
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];
@@ -1124,8 +1137,11 @@ static void hls_prediction_unit(HEVCContext *s, int x0, int y0,
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, &current_mv);
ff_hevc_luma_mv_merge_mode(s, x0, y0,
1 << log2_cb_size,
1 << log2_cb_size,
log2_cb_size, partIdx,
merge_idx, &current_mv);
x_pu = x0 >> s->sps->log2_min_pu_size;
y_pu = y0 >> s->sps->log2_min_pu_size;

@@ -1163,7 +1179,8 @@ static void hls_prediction_unit(HEVCContext *s, int x0, int y0,
ff_hevc_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, &current_mv, mvp_flag[0], 0);
partIdx, merge_idx, &current_mv,
mvp_flag[0], 0);
current_mv.mv[0].x += lc->pu.mvd.x;
current_mv.mv[0].y += lc->pu.mvd.y;
}
@@ -1184,7 +1201,8 @@ static void hls_prediction_unit(HEVCContext *s, int x0, int y0,
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, &current_mv, mvp_flag[1], 1);
partIdx, merge_idx, &current_mv,
mvp_flag[1], 1);
current_mv.mv[1].x += lc->pu.mvd.x;
current_mv.mv[1].y += lc->pu.mvd.y;
}
@@ -1212,7 +1230,7 @@ static void hls_prediction_unit(HEVCContext *s, int x0, int 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, 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,
@@ -1520,14 +1538,14 @@ static int hls_coding_unit(HEVCContext *s, int x0, int y0, int log2_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.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;

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;
@@ -1593,33 +1611,33 @@ static int hls_coding_unit(HEVCContext *s, int x0, int y0, int log2_cb_size)
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);
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, 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, 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);
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, 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);
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, 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;
}
@@ -1725,7 +1743,8 @@ static int hls_coding_quadtree(HEVCContext *s, int x0, int y0,
return 0;
}

static void hls_decode_neighbour(HEVCContext *s, int x_ctb, int y_ctb, int ctb_addr_ts)
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;
@@ -2009,7 +2028,7 @@ static int hls_nal_unit(HEVCContext *s)

static void restore_tqb_pixels(HEVCContext *s)
{
int min_pu_size = 1 << 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++) {
@@ -2076,6 +2095,7 @@ static int hevc_frame_start(HEVCContext *s)
ff_thread_finish_setup(s->avctx);

return 0;

fail:
if (s->ref && s->threads_type == FF_THREAD_FRAME)
ff_thread_report_progress(&s->ref->tf, INT_MAX, 0);
@@ -2316,8 +2336,8 @@ int ff_hevc_extract_rbsp(HEVCContext *s, const uint8_t *src, int length,
}
while (si < length)
dst[di++] = src[si++];
nsc:

nsc:
memset(dst + di, 0, FF_INPUT_BUFFER_PADDING_SIZE);

nal->data = dst;
@@ -2359,8 +2379,8 @@ static int decode_nal_units(HEVCContext *s, const uint8_t *buf, int length)
if (length < 4) {
av_log(s->avctx, AV_LOG_ERROR, "No start code is found.\n");
ret = AVERROR_INVALIDDATA;
goto fail;
}
goto fail;
}
}

buf += 3;
@@ -2378,7 +2398,8 @@ static int decode_nal_units(HEVCContext *s, const uint8_t *buf, int length)
goto fail;
}
s->nals = tmp;
memset(s->nals + s->nals_allocated, 0, (new_size - s->nals_allocated) * sizeof(*tmp));
memset(s->nals + s->nals_allocated, 0,
(new_size - s->nals_allocated) * sizeof(*tmp));
av_reallocp_array(&s->skipped_bytes_nal, new_size, sizeof(*s->skipped_bytes_nal));
av_reallocp_array(&s->skipped_bytes_pos_size_nal, new_size, sizeof(*s->skipped_bytes_pos_size_nal));
av_reallocp_array(&s->skipped_bytes_pos_nal, new_size, sizeof(*s->skipped_bytes_pos_nal));
@@ -2437,7 +2458,7 @@ fail:
return ret;
}

static void print_md5(void *log_ctx, int level, uint8_t md5[16])
static void print_md5(void *log_ctx, int level, uint8_t md5[16])
{
int i;
for (i = 0; i < 16; i++)
@@ -2526,7 +2547,7 @@ static int hevc_decode_frame(AVCodecContext *avctx, void *data, int *got_output,
}

s->ref = NULL;
ret = decode_nal_units(s, avpkt->data, avpkt->size);
ret = decode_nal_units(s, avpkt->data, avpkt->size);
if (ret < 0)
return ret;

@@ -2691,6 +2712,7 @@ static av_cold int hevc_init_context(AVCodecContext *avctx)
s->context_initialized = 1;

return 0;

fail:
hevc_decode_free(avctx);
return AVERROR(ENOMEM);
@@ -2803,7 +2825,8 @@ static int hevc_decode_extradata(HEVCContext *s)
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);
"Decoding nal unit %d %d from hvcC failed\n",
type, i);
return ret;
}
bytestream2_skip(&gb, nalsize);
@@ -2907,5 +2930,6 @@ AVCodec ff_hevc_decoder = {
.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_SLICE_THREADS | CODEC_CAP_FRAME_THREADS,
.capabilities = CODEC_CAP_DR1 | CODEC_CAP_DELAY |
CODEC_CAP_SLICE_THREADS | CODEC_CAP_FRAME_THREADS,
};

+ 285
- 288
libavcodec/hevc_cabac.c View File

@@ -33,94 +33,94 @@
* number of bin by SyntaxElement.
*/
static const int8_t num_bins_in_se[] = {
1, // sao_merge_flag
1, // sao_type_idx
0, // sao_eo_class
0, // sao_band_position
0, // sao_offset_abs
0, // sao_offset_sign
0, // end_of_slice_flag
3, // split_coding_unit_flag
1, // cu_transquant_bypass_flag
3, // skip_flag
3, // cu_qp_delta
1, // pred_mode
4, // part_mode
0, // pcm_flag
1, // prev_intra_luma_pred_mode
0, // mpm_idx
0, // rem_intra_luma_pred_mode
2, // intra_chroma_pred_mode
1, // merge_flag
1, // merge_idx
5, // inter_pred_idc
2, // ref_idx_l0
2, // ref_idx_l1
2, // abs_mvd_greater0_flag
2, // abs_mvd_greater1_flag
0, // abs_mvd_minus2
0, // mvd_sign_flag
1, // mvp_lx_flag
1, // no_residual_data_flag
3, // split_transform_flag
2, // cbf_luma
4, // cbf_cb, cbf_cr
2, // transform_skip_flag[][]
18, // last_significant_coeff_x_prefix
18, // last_significant_coeff_y_prefix
0, // last_significant_coeff_x_suffix
0, // last_significant_coeff_y_suffix
4, // significant_coeff_group_flag
42, // significant_coeff_flag
24, // coeff_abs_level_greater1_flag
6, // coeff_abs_level_greater2_flag
0, // coeff_abs_level_remaining
0, // coeff_sign_flag
1, // sao_merge_flag
1, // sao_type_idx
0, // sao_eo_class
0, // sao_band_position
0, // sao_offset_abs
0, // sao_offset_sign
0, // end_of_slice_flag
3, // split_coding_unit_flag
1, // cu_transquant_bypass_flag
3, // skip_flag
3, // cu_qp_delta
1, // pred_mode
4, // part_mode
0, // pcm_flag
1, // prev_intra_luma_pred_mode
0, // mpm_idx
0, // rem_intra_luma_pred_mode
2, // intra_chroma_pred_mode
1, // merge_flag
1, // merge_idx
5, // inter_pred_idc
2, // ref_idx_l0
2, // ref_idx_l1
2, // abs_mvd_greater0_flag
2, // abs_mvd_greater1_flag
0, // abs_mvd_minus2
0, // mvd_sign_flag
1, // mvp_lx_flag
1, // no_residual_data_flag
3, // split_transform_flag
2, // cbf_luma
4, // cbf_cb, cbf_cr
2, // transform_skip_flag[][]
18, // last_significant_coeff_x_prefix
18, // last_significant_coeff_y_prefix
0, // last_significant_coeff_x_suffix
0, // last_significant_coeff_y_suffix
4, // significant_coeff_group_flag
42, // significant_coeff_flag
24, // coeff_abs_level_greater1_flag
6, // coeff_abs_level_greater2_flag
0, // coeff_abs_level_remaining
0, // coeff_sign_flag
};

/**
* Offset to ctxIdx 0 in init_values and states, indexed by SyntaxElement.
*/
static const int elem_offset[sizeof(num_bins_in_se)] = {
0,
1,
2,
2,
2,
2,
2,
2,
5,
6,
9,
12,
13,
17,
17,
18,
18,
18,
20,
21,
22,
27,
29,
31,
33,
35,
35,
35,
36,
37,
40,
42,
46,
48,
66,
84,
84,
84,
88,
0,
1,
2,
2,
2,
2,
2,
2,
5,
6,
9,
12,
13,
17,
17,
18,
18,
18,
20,
21,
22,
27,
29,
31,
33,
35,
35,
35,
36,
37,
40,
42,
46,
48,
66,
84,
84,
84,
88,
130,
154,
160,
@@ -132,200 +132,195 @@ static const int elem_offset[sizeof(num_bins_in_se)] = {
* Indexed by init_type
*/
static const uint8_t init_values[3][HEVC_CONTEXTS] = {
{
// sao_merge_flag
153,
// sao_type_idx
200,
// split_coding_unit_flag
139, 141, 157,
// cu_transquant_bypass_flag
154,
// skip_flag
CNU, CNU, CNU,
// cu_qp_delta
154, 154, 154,
// pred_mode
CNU,
// part_mode
184, CNU, CNU, CNU,
// prev_intra_luma_pred_mode
184,
// intra_chroma_pred_mode
63, 139,
// merge_flag
CNU,
// merge_idx
CNU,
// inter_pred_idc
CNU, CNU, CNU, CNU, CNU,
// ref_idx_l0
CNU, CNU,
// ref_idx_l1
CNU, CNU,
// abs_mvd_greater1_flag
CNU, CNU,
// abs_mvd_greater1_flag
CNU, CNU,
// mvp_lx_flag
CNU,
// no_residual_data_flag
CNU,
// split_transform_flag
153, 138, 138,
// cbf_luma
111, 141,
// cbf_cb, cbf_cr
94, 138, 182, 154,
// transform_skip_flag
139, 139,
// last_significant_coeff_x_prefix
110, 110, 124, 125, 140, 153, 125, 127, 140, 109, 111, 143, 127, 111,
79, 108, 123, 63,
// last_significant_coeff_y_prefix
110, 110, 124, 125, 140, 153, 125, 127, 140, 109, 111, 143, 127, 111,
79, 108, 123, 63,
// significant_coeff_group_flag
91, 171, 134, 141,
// significant_coeff_flag
111, 111, 125, 110, 110, 94, 124, 108, 124, 107, 125, 141, 179, 153,
125, 107, 125, 141, 179, 153, 125, 107, 125, 141, 179, 153, 125, 140,
139, 182, 182, 152, 136, 152, 136, 153, 136, 139, 111, 136, 139, 111,
// coeff_abs_level_greater1_flag
140, 92, 137, 138, 140, 152, 138, 139, 153, 74, 149, 92, 139, 107,
122, 152, 140, 179, 166, 182, 140, 227, 122, 197,
// coeff_abs_level_greater2_flag
138, 153, 136, 167, 152, 152,
},
{
// sao_merge_flag
153,
// sao_type_idx
185,
// split_coding_unit_flag
107, 139, 126,
// cu_transquant_bypass_flag
154,
// skip_flag
197, 185, 201,
// cu_qp_delta
154, 154, 154,
// pred_mode
149,
// part_mode
154, 139, 154, 154,
// prev_intra_luma_pred_mode
154,
// intra_chroma_pred_mode
152, 139,
// merge_flag
110,
// merge_idx
122,
// inter_pred_idc
95, 79, 63, 31, 31,
// ref_idx_l0
153, 153,
// ref_idx_l1
153, 153,
// abs_mvd_greater1_flag
140, 198,
// abs_mvd_greater1_flag
140, 198,
// mvp_lx_flag
168,
// no_residual_data_flag
79,
// split_transform_flag
124, 138, 94,
// cbf_luma
153, 111,
// cbf_cb, cbf_cr
149, 107, 167, 154,
// transform_skip_flag
139, 139,
// last_significant_coeff_x_prefix
125, 110, 94, 110, 95, 79, 125, 111, 110, 78, 110, 111, 111, 95,
94, 108, 123, 108,
// last_significant_coeff_y_prefix
125, 110, 94, 110, 95, 79, 125, 111, 110, 78, 110, 111, 111, 95,
94, 108, 123, 108,
// significant_coeff_group_flag
121, 140, 61, 154,
// significant_coeff_flag
155, 154, 139, 153, 139, 123, 123, 63, 153, 166, 183, 140, 136, 153,
154, 166, 183, 140, 136, 153, 154, 166, 183, 140, 136, 153, 154, 170,
153, 123, 123, 107, 121, 107, 121, 167, 151, 183, 140, 151, 183, 140,
// coeff_abs_level_greater1_flag
154, 196, 196, 167, 154, 152, 167, 182, 182, 134, 149, 136, 153, 121,
136, 137, 169, 194, 166, 167, 154, 167, 137, 182,
// coeff_abs_level_greater2_flag
107, 167, 91, 122, 107, 167,
},
{
// sao_merge_flag
153,
// sao_type_idx
160,
// split_coding_unit_flag
107, 139, 126,
// cu_transquant_bypass_flag
154,
// skip_flag
197, 185, 201,
// cu_qp_delta
154, 154, 154,
// pred_mode
134,
// part_mode
154, 139, 154, 154,
// prev_intra_luma_pred_mode
183,
// intra_chroma_pred_mode
152, 139,
// merge_flag
154,
// merge_idx
137,
// inter_pred_idc
95, 79, 63, 31, 31,
// ref_idx_l0
153, 153,
// ref_idx_l1
153, 153,
// abs_mvd_greater1_flag
169, 198,
// abs_mvd_greater1_flag
169, 198,
// mvp_lx_flag
168,
// no_residual_data_flag
79,
// split_transform_flag
224, 167, 122,
// cbf_luma
153, 111,
// cbf_cb, cbf_cr
149, 92, 167, 154,
// transform_skip_flag
139, 139,
// last_significant_coeff_x_prefix
125, 110, 124, 110, 95, 94, 125, 111, 111, 79, 125, 126, 111, 111,
79, 108, 123, 93,
// last_significant_coeff_y_prefix
125, 110, 124, 110, 95, 94, 125, 111, 111, 79, 125, 126, 111, 111,
79, 108, 123, 93,
// significant_coeff_group_flag
121, 140, 61, 154,
// significant_coeff_flag
170, 154, 139, 153, 139, 123, 123, 63, 124, 166, 183, 140, 136, 153,
154, 166, 183, 140, 136, 153, 154, 166, 183, 140, 136, 153, 154, 170,
153, 138, 138, 122, 121, 122, 121, 167, 151, 183, 140, 151, 183, 140,
// coeff_abs_level_greater1_flag
154, 196, 167, 167, 154, 152, 167, 182, 182, 134, 149, 136, 153, 121,
136, 122, 169, 208, 166, 167, 154, 152, 167, 182,
// coeff_abs_level_greater2_flag
107, 167, 91, 107, 107, 167,
{ // sao_merge_flag
153,
// sao_type_idx
200,
// split_coding_unit_flag
139, 141, 157,
// cu_transquant_bypass_flag
154,
// skip_flag
CNU, CNU, CNU,
// cu_qp_delta
154, 154, 154,
// pred_mode
CNU,
// part_mode
184, CNU, CNU, CNU,
// prev_intra_luma_pred_mode
184,
// intra_chroma_pred_mode
63, 139,
// merge_flag
CNU,
// merge_idx
CNU,
// inter_pred_idc
CNU, CNU, CNU, CNU, CNU,
// ref_idx_l0
CNU, CNU,
// ref_idx_l1
CNU, CNU,
// abs_mvd_greater1_flag
CNU, CNU,
// abs_mvd_greater1_flag
CNU, CNU,
// mvp_lx_flag
CNU,
// no_residual_data_flag
CNU,
// split_transform_flag
153, 138, 138,
// cbf_luma
111, 141,
// cbf_cb, cbf_cr
94, 138, 182, 154,
// transform_skip_flag
139, 139,
// last_significant_coeff_x_prefix
110, 110, 124, 125, 140, 153, 125, 127, 140, 109, 111, 143, 127, 111,
79, 108, 123, 63,
// last_significant_coeff_y_prefix
110, 110, 124, 125, 140, 153, 125, 127, 140, 109, 111, 143, 127, 111,
79, 108, 123, 63,
// significant_coeff_group_flag
91, 171, 134, 141,
// significant_coeff_flag
111, 111, 125, 110, 110, 94, 124, 108, 124, 107, 125, 141, 179, 153,
125, 107, 125, 141, 179, 153, 125, 107, 125, 141, 179, 153, 125, 140,
139, 182, 182, 152, 136, 152, 136, 153, 136, 139, 111, 136, 139, 111,
// coeff_abs_level_greater1_flag
140, 92, 137, 138, 140, 152, 138, 139, 153, 74, 149, 92, 139, 107,
122, 152, 140, 179, 166, 182, 140, 227, 122, 197,
// coeff_abs_level_greater2_flag
138, 153, 136, 167, 152, 152, },
{ // sao_merge_flag
153,
// sao_type_idx
185,
// split_coding_unit_flag
107, 139, 126,
// cu_transquant_bypass_flag
154,
// skip_flag
197, 185, 201,
// cu_qp_delta
154, 154, 154,
// pred_mode
149,
// part_mode
154, 139, 154, 154,
// prev_intra_luma_pred_mode
154,
// intra_chroma_pred_mode
152, 139,
// merge_flag
110,
// merge_idx
122,
// inter_pred_idc
95, 79, 63, 31, 31,
// ref_idx_l0
153, 153,
// ref_idx_l1
153, 153,
// abs_mvd_greater1_flag
140, 198,
// abs_mvd_greater1_flag
140, 198,
// mvp_lx_flag
168,
// no_residual_data_flag
79,
// split_transform_flag
124, 138, 94,
// cbf_luma
153, 111,
// cbf_cb, cbf_cr
149, 107, 167, 154,
// transform_skip_flag
139, 139,
// last_significant_coeff_x_prefix
125, 110, 94, 110, 95, 79, 125, 111, 110, 78, 110, 111, 111, 95,
94, 108, 123, 108,
// last_significant_coeff_y_prefix
125, 110, 94, 110, 95, 79, 125, 111, 110, 78, 110, 111, 111, 95,
94, 108, 123, 108,
// significant_coeff_group_flag
121, 140, 61, 154,
// significant_coeff_flag
155, 154, 139, 153, 139, 123, 123, 63, 153, 166, 183, 140, 136, 153,
154, 166, 183, 140, 136, 153, 154, 166, 183, 140, 136, 153, 154, 170,
153, 123, 123, 107, 121, 107, 121, 167, 151, 183, 140, 151, 183, 140,
// coeff_abs_level_greater1_flag
154, 196, 196, 167, 154, 152, 167, 182, 182, 134, 149, 136, 153, 121,
136, 137, 169, 194, 166, 167, 154, 167, 137, 182,
// coeff_abs_level_greater2_flag
107, 167, 91, 122, 107, 167, },
{ // sao_merge_flag
153,
// sao_type_idx
160,
// split_coding_unit_flag
107, 139, 126,
// cu_transquant_bypass_flag
154,
// skip_flag
197, 185, 201,
// cu_qp_delta
154, 154, 154,
// pred_mode
134,
// part_mode
154, 139, 154, 154,
// prev_intra_luma_pred_mode
183,
// intra_chroma_pred_mode
152, 139,
// merge_flag
154,
// merge_idx
137,
// inter_pred_idc
95, 79, 63, 31, 31,
// ref_idx_l0
153, 153,
// ref_idx_l1
153, 153,
// abs_mvd_greater1_flag
169, 198,
// abs_mvd_greater1_flag
169, 198,
// mvp_lx_flag
168,
// no_residual_data_flag
79,
// split_transform_flag
224, 167, 122,
// cbf_luma
153, 111,
// cbf_cb, cbf_cr
149, 92, 167, 154,
// transform_skip_flag
139, 139,
// last_significant_coeff_x_prefix
125, 110, 124, 110, 95, 94, 125, 111, 111, 79, 125, 126, 111, 111,
79, 108, 123, 93,
// last_significant_coeff_y_prefix
125, 110, 124, 110, 95, 94, 125, 111, 111, 79, 125, 126, 111, 111,
79, 108, 123, 93,
// significant_coeff_group_flag
121, 140, 61, 154,
// significant_coeff_flag
170, 154, 139, 153, 139, 123, 123, 63, 124, 166, 183, 140, 136, 153,
154, 166, 183, 140, 136, 153, 154, 166, 183, 140, 136, 153, 154, 170,
153, 138, 138, 122, 121, 122, 121, 167, 151, 183, 140, 151, 183, 140,
// coeff_abs_level_greater1_flag
154, 196, 167, 167, 154, 152, 167, 182, 182, 134, 149, 136, 153, 121,
136, 122, 169, 208, 166, 167, 154, 152, 167, 182,
// coeff_abs_level_greater2_flag
107, 167, 91, 107, 107, 167,
},
};

@@ -548,7 +543,7 @@ static void cabac_init_state(HEVCContext *s)
pre ^= pre >> 31;
if (pre > 124)
pre = 124 + (pre & 1);
s->HEVClc->cabac_state[i] = pre;
s->HEVClc->cabac_state[i] = pre;
}
}

@@ -561,7 +556,8 @@ void ff_hevc_cabac_init(HEVCContext *s, int ctb_addr_ts)
(s->pps->tile_id[ctb_addr_ts] != s->pps->tile_id[ctb_addr_ts - 1])))
cabac_init_state(s);

if (!s->sh.first_slice_in_pic_flag && s->pps->entropy_coding_sync_enabled_flag) {
if (!s->sh.first_slice_in_pic_flag &&
s->pps->entropy_coding_sync_enabled_flag) {
if ((ctb_addr_ts % s->sps->ctb_width) == 0) {
if (s->sps->ctb_width == 1)
cabac_init_state(s);
@@ -640,7 +636,7 @@ int ff_hevc_sao_offset_sign_decode(HEVCContext *s)
int ff_hevc_sao_eo_class_decode(HEVCContext *s)
{
int ret = (get_cabac_bypass(&s->HEVClc->cc) << 1);
ret |= get_cabac_bypass(&s->HEVClc->cc);
ret |= get_cabac_bypass(&s->HEVClc->cc);
return ret;
}

@@ -662,9 +658,9 @@ int ff_hevc_skip_flag_decode(HEVCContext *s, int x0, int y0, int x_cb, int y_cb)
int y0b = y0 & ((1 << s->sps->log2_ctb_size) - 1);

if (s->HEVClc->ctb_left_flag || x0b)
inc = !!SAMPLE_CTB(s->skip_flag, x_cb-1, y_cb);
inc = !!SAMPLE_CTB(s->skip_flag, x_cb - 1, y_cb);
if (s->HEVClc->ctb_up_flag || y0b)
inc += !!SAMPLE_CTB(s->skip_flag, x_cb, y_cb-1);
inc += !!SAMPLE_CTB(s->skip_flag, x_cb, y_cb - 1);

return GET_CABAC(elem_offset[SKIP_FLAG] + inc);
}
@@ -707,18 +703,19 @@ int ff_hevc_pred_mode_decode(HEVCContext *s)
int ff_hevc_split_coding_unit_flag_decode(HEVCContext *s, int ct_depth, int x0, int y0)
{
int inc = 0, depth_left = 0, depth_top = 0;
int x0b = x0 & ((1 << s->sps->log2_ctb_size) - 1);
int y0b = y0 & ((1 << s->sps->log2_ctb_size) - 1);
int x0b = x0 & ((1 << s->sps->log2_ctb_size) - 1);
int y0b = y0 & ((1 << s->sps->log2_ctb_size) - 1);
int x_cb = x0 >> s->sps->log2_min_cb_size;
int y_cb = y0 >> s->sps->log2_min_cb_size;

if (s->HEVClc->ctb_left_flag || x0b)
depth_left = s->tab_ct_depth[(y_cb)*s->sps->min_cb_width + x_cb-1];
depth_left = s->tab_ct_depth[(y_cb) * s->sps->min_cb_width + x_cb - 1];
if (s->HEVClc->ctb_up_flag || y0b)
depth_top = s->tab_ct_depth[(y_cb-1)*s->sps->min_cb_width + x_cb];
depth_top = s->tab_ct_depth[(y_cb - 1) * s->sps->min_cb_width + x_cb];

inc += (depth_left > ct_depth);
inc += (depth_top > ct_depth);
inc += (depth_top > ct_depth);

return GET_CABAC(elem_offset[SPLIT_CODING_UNIT_FLAG] + inc);
}

@@ -756,7 +753,7 @@ int ff_hevc_part_mode_decode(HEVCContext *s, int log2_cb_size)
return PART_Nx2N;
if (get_cabac_bypass(&s->HEVClc->cc)) // 0001
return PART_nRx2N;
return PART_nLx2N; // 0000
return PART_nLx2N; // 0000
}

int ff_hevc_pcm_flag_decode(HEVCContext *s)
@@ -794,7 +791,7 @@ int ff_hevc_intra_chroma_pred_mode_decode(HEVCContext *s)
return 4;

ret = (get_cabac_bypass(&s->HEVClc->cc) << 1);
ret |= get_cabac_bypass(&s->HEVClc->cc);
ret |= get_cabac_bypass(&s->HEVClc->cc);
return ret;
}



+ 183
- 120
libavcodec/hevc_filter.c View File

@@ -35,20 +35,22 @@
#define CR 2

static const uint8_t tctable[54] = {
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, // QP 0...18
1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 4, // QP 19...37
5, 5, 6, 6, 7, 8, 9,10,11,13,14,16,18,20,22,24 // QP 38...53
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, // QP 0...18
1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 4, // QP 19...37
5, 5, 6, 6, 7, 8, 9, 10, 11, 13, 14, 16, 18, 20, 22, 24 // QP 38...53
};

static const uint8_t betatable[52] = {
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 6, 7, 8, // QP 0...18
9,10,11,12,13,14,15,16,17,18,20,22,24,26,28,30,32,34,36, // QP 19...37
38,40,42,44,46,48,50,52,54,56,58,60,62,64 // QP 38...51
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 6, 7, 8, // QP 0...18
9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 20, 22, 24, 26, 28, 30, 32, 34, 36, // QP 19...37
38, 40, 42, 44, 46, 48, 50, 52, 54, 56, 58, 60, 62, 64 // QP 38...51
};

static int chroma_tc(HEVCContext *s, int qp_y, int c_idx, int tc_offset)
{
static const int qp_c[] = { 29, 30, 31, 32, 33, 33, 34, 34, 35, 35, 36, 36, 37, 37 };
static const int qp_c[] = {
29, 30, 31, 32, 33, 33, 34, 34, 35, 35, 36, 36, 37, 37
};
int qp_i, offset;
int qp;
int idxt;
@@ -71,19 +73,23 @@ static int chroma_tc(HEVCContext *s, int qp_y, int c_idx, int tc_offset)
return tctable[idxt];
}

static int get_qPy_pred(HEVCContext *s, int xC, int yC, int xBase, int yBase, int log2_cb_size)
static int get_qPy_pred(HEVCContext *s, int xC, int yC,
int xBase, int yBase, int log2_cb_size)
{
HEVCLocalContext *lc = s->HEVClc;
int ctb_size_mask = (1 << s->sps->log2_ctb_size) - 1;
int MinCuQpDeltaSizeMask = (1 << (s->sps->log2_ctb_size - s->pps->diff_cu_qp_delta_depth)) - 1;
int xQgBase = xBase - ( xBase & MinCuQpDeltaSizeMask );
int yQgBase = yBase - ( yBase & MinCuQpDeltaSizeMask );
int MinCuQpDeltaSizeMask = (1 << (s->sps->log2_ctb_size -
s->pps->diff_cu_qp_delta_depth)) - 1;
int xQgBase = xBase - (xBase & MinCuQpDeltaSizeMask);
int yQgBase = yBase - (yBase & MinCuQpDeltaSizeMask);
int min_cb_width = s->sps->min_cb_width;
int min_cb_height = s->sps->min_cb_height;
int x_cb = xQgBase >> s->sps->log2_min_cb_size;
int y_cb = yQgBase >> s->sps->log2_min_cb_size;
int availableA = (xBase & ctb_size_mask) && (xQgBase & ctb_size_mask);
int availableB = (yBase & ctb_size_mask) && (yQgBase & ctb_size_mask);
int availableA = (xBase & ctb_size_mask) &&
(xQgBase & ctb_size_mask);
int availableB = (yBase & ctb_size_mask) &&
(yQgBase & ctb_size_mask);
int qPy_pred;
int qPy_a;
int qPy_b;
@@ -94,31 +100,32 @@ static int get_qPy_pred(HEVCContext *s, int xC, int yC, int xBase, int yBase, in
qPy_pred = s->sh.slice_qp;
} else {
qPy_pred = lc->qp_y;
if (log2_cb_size < s->sps->log2_ctb_size - s->pps->diff_cu_qp_delta_depth) {
if (log2_cb_size < s->sps->log2_ctb_size -
s->pps->diff_cu_qp_delta_depth) {
static const int offsetX[8][8] = {
{-1, 1, 3, 1, 7, 1, 3, 1},
{ 0, 0, 0, 0, 0, 0, 0, 0},
{ 1, 3, 1, 3, 1, 3, 1, 3},
{ 2, 2, 2, 2, 2, 2, 2, 2},
{ 3, 5, 7, 5, 3, 5, 7, 5},
{ 4, 4, 4, 4, 4, 4, 4, 4},
{ 5, 7, 5, 7, 5, 7, 5, 7},
{ 6, 6, 6, 6, 6, 6, 6, 6}
{ -1, 1, 3, 1, 7, 1, 3, 1 },
{ 0, 0, 0, 0, 0, 0, 0, 0 },
{ 1, 3, 1, 3, 1, 3, 1, 3 },
{ 2, 2, 2, 2, 2, 2, 2, 2 },
{ 3, 5, 7, 5, 3, 5, 7, 5 },
{ 4, 4, 4, 4, 4, 4, 4, 4 },
{ 5, 7, 5, 7, 5, 7, 5, 7 },
{ 6, 6, 6, 6, 6, 6, 6, 6 }
};
static const int offsetY[8][8] = {
{ 7, 0, 1, 2, 3, 4, 5, 6},
{ 0, 1, 2, 3, 4, 5, 6, 7},
{ 1, 0, 3, 2, 5, 4, 7, 6},
{ 0, 1, 2, 3, 4, 5, 6, 7},
{ 3, 0, 1, 2, 7, 4, 5, 6},
{ 0, 1, 2, 3, 4, 5, 6, 7},
{ 1, 0, 3, 2, 5, 4, 7, 6},
{ 0, 1, 2, 3, 4, 5, 6, 7}
{ 7, 0, 1, 2, 3, 4, 5, 6 },
{ 0, 1, 2, 3, 4, 5, 6, 7 },
{ 1, 0, 3, 2, 5, 4, 7, 6 },
{ 0, 1, 2, 3, 4, 5, 6, 7 },
{ 3, 0, 1, 2, 7, 4, 5, 6 },
{ 0, 1, 2, 3, 4, 5, 6, 7 },
{ 1, 0, 3, 2, 5, 4, 7, 6 },
{ 0, 1, 2, 3, 4, 5, 6, 7 }
};
int xC0b = (xC - (xC & ctb_size_mask)) >> s->sps->log2_min_cb_size;
int yC0b = (yC - (yC & ctb_size_mask)) >> s->sps->log2_min_cb_size;
int idxX = (xQgBase & ctb_size_mask) >> s->sps->log2_min_cb_size;
int idxY = (yQgBase & ctb_size_mask) >> s->sps->log2_min_cb_size;
int idxX = (xQgBase & ctb_size_mask) >> s->sps->log2_min_cb_size;
int idxY = (yQgBase & ctb_size_mask) >> s->sps->log2_min_cb_size;
int idx_mask = ctb_size_mask >> s->sps->log2_min_cb_size;
int x, y;

@@ -149,13 +156,15 @@ static int get_qPy_pred(HEVCContext *s, int xC, int yC, int xBase, int yBase, in
return (qPy_a + qPy_b + 1) >> 1;
}

void ff_hevc_set_qPy(HEVCContext *s, int xC, int yC, int xBase, int yBase, int log2_cb_size)
void ff_hevc_set_qPy(HEVCContext *s, int xC, int yC,
int xBase, int yBase, int log2_cb_size)
{
int qp_y = get_qPy_pred(s, xC, yC, xBase, yBase, log2_cb_size);

if (s->HEVClc->tu.cu_qp_delta != 0) {
int off = s->sps->qp_bd_offset;
s->HEVClc->qp_y = ((qp_y + s->HEVClc->tu.cu_qp_delta + 52 + 2 * off) % (52 + off)) - off;
s->HEVClc->qp_y = ((qp_y + s->HEVClc->tu.cu_qp_delta + 52 + 2 * off) %
(52 + off)) - off;
} else
s->HEVClc->qp_y = qp_y;
}
@@ -168,11 +177,12 @@ static int get_qPy(HEVCContext *s, int xC, int yC)
return s->qp_y_tab[x + y * s->sps->min_cb_width];
}

static void copy_CTB(uint8_t *dst, uint8_t *src, int width, int height, int stride)
static void copy_CTB(uint8_t *dst, uint8_t *src,
int width, int height, int stride)
{
int i;

for(i=0; i< height; i++){
for (i = 0; i < height; i++) {
memcpy(dst, src, width);
dst += stride;
src += stride;
@@ -187,28 +197,29 @@ static void sao_filter_CTB(HEVCContext *s, int x, int y)
// TODO: skip CBs when (cu_transquant_bypass_flag || (pcm_loop_filter_disable_flag && pcm_flag))
int c_idx = 0;
int class = 1, class_index;
int edges[4]; // 0 left 1 top 2 right 3 bottom
int edges[4]; // 0 left 1 top 2 right 3 bottom
SAOParams *sao[4];
int classes[4];
int x_shift = 0, y_shift = 0;
int x_ctb = x>>s->sps->log2_ctb_size;
int y_ctb = y>>s->sps->log2_ctb_size;
int x_ctb = x >> s->sps->log2_ctb_size;
int y_ctb = y >> s->sps->log2_ctb_size;
int ctb_addr_rs = y_ctb * s->sps->ctb_width + x_ctb;
int ctb_addr_ts = s->pps->ctb_addr_rs_to_ts[ctb_addr_rs];

// flags indicating unfilterable edges
uint8_t vert_edge[] = {0,0,0,0};
uint8_t horiz_edge[] = {0,0,0,0};
uint8_t diag_edge[] = {0,0,0,0};
uint8_t vert_edge[] = { 0, 0, 0, 0 };
uint8_t horiz_edge[] = { 0, 0, 0, 0 };
uint8_t diag_edge[] = { 0, 0, 0, 0 };
uint8_t lfase[3]; // current, above, left
uint8_t no_tile_filter = s->pps->tiles_enabled_flag && !s->pps->loop_filter_across_tiles_enabled_flag;
uint8_t no_tile_filter = s->pps->tiles_enabled_flag &&
!s->pps->loop_filter_across_tiles_enabled_flag;
uint8_t left_tile_edge = 0;
uint8_t up_tile_edge = 0;

sao[0] = &CTB(s->sao, x_ctb, y_ctb);
edges[0] = x_ctb == 0;
edges[1] = y_ctb == 0;
edges[2] = x_ctb == (s->sps->ctb_width - 1);
edges[2] = x_ctb == (s->sps->ctb_width - 1);
edges[3] = y_ctb == (s->sps->ctb_height - 1);
lfase[0] = CTB(s->filter_slice_edges, x_ctb, y_ctb);
classes[0] = 0;
@@ -249,10 +260,12 @@ static void sao_filter_CTB(HEVCContext *s, int x, int y)
diag_edge[3] = diag_edge[0];

// Does left CTB comes after above CTB?
if(CTB(s->tab_slice_address, x_ctb - 1, y_ctb) > CTB(s->tab_slice_address, x_ctb, y_ctb - 1)) {
if (CTB(s->tab_slice_address, x_ctb - 1, y_ctb) >
CTB(s->tab_slice_address, x_ctb, y_ctb - 1)) {
diag_edge[2] = !lfase[2] || left_tile_edge || up_tile_edge;
diag_edge[1] = diag_edge[2];
} else if(CTB(s->tab_slice_address, x_ctb - 1, y_ctb) < CTB(s->tab_slice_address, x_ctb, y_ctb - 1)) {
} else if (CTB(s->tab_slice_address, x_ctb - 1, y_ctb) <
CTB(s->tab_slice_address, x_ctb, y_ctb - 1)) {
diag_edge[1] = !lfase[1] || left_tile_edge || up_tile_edge;
diag_edge[2] = diag_edge[1];
} else {
@@ -286,10 +299,21 @@ static void sao_filter_CTB(HEVCContext *s, int x, int y)

switch (sao[class_index]->type_idx[c_idx]) {
case SAO_BAND:
s->hevcdsp.sao_band_filter[classes[class_index]](dst, src, stride, sao[class_index], edges, width, height, c_idx);
break;
s->hevcdsp.sao_band_filter[classes[class_index]](dst, src,
stride,
sao[class_index],
edges, width,
height, c_idx);
break;
case SAO_EDGE:
s->hevcdsp.sao_edge_filter[classes[class_index]](dst, src, stride, sao[class_index], edges, width, height, c_idx, vert_edge[classes[class_index]], horiz_edge[classes[class_index]], diag_edge[classes[class_index]]);
s->hevcdsp.sao_edge_filter[classes[class_index]](dst, src,
stride,
sao[class_index],
edges, width,
height, c_idx,
vert_edge[classes[class_index]],
horiz_edge[classes[class_index]],
diag_edge[classes[class_index]]);
break;
}
}
@@ -318,33 +342,35 @@ static void deblocking_filter_CTB(HEVCContext *s, int x0, int y0)
int c_tc[2];
int beta[2];
int tc[2];
uint8_t no_p[2] = {0};
uint8_t no_q[2] = {0};
uint8_t no_p[2] = { 0 };
uint8_t no_q[2] = { 0 };

int log2_ctb_size = s->sps->log2_ctb_size;
int log2_ctb_size = s->sps->log2_ctb_size;
int x_end, y_end;
int ctb_size = 1<<log2_ctb_size;
int ctb = (x0 >> log2_ctb_size) + (y0 >> log2_ctb_size) * s->sps->ctb_width;
int ctb_size = 1 << log2_ctb_size;
int ctb = (x0 >> log2_ctb_size) +
(y0 >> log2_ctb_size) * s->sps->ctb_width;
int cur_tc_offset = s->deblock[ctb].tc_offset;
int cur_beta_offset = s->deblock[ctb].beta_offset;
int left_tc_offset, left_beta_offset;
int tc_offset, beta_offset;
int pcmf = (s->sps->pcm_enabled_flag && s->sps->pcm.loop_filter_disable_flag) ||
s->pps->transquant_bypass_enable_flag;
int pcmf = (s->sps->pcm_enabled_flag &&
s->sps->pcm.loop_filter_disable_flag) ||
s->pps->transquant_bypass_enable_flag;

if (x0) {
left_tc_offset = s->deblock[ctb-1].tc_offset;
left_beta_offset = s->deblock[ctb-1].beta_offset;
left_tc_offset = s->deblock[ctb - 1].tc_offset;
left_beta_offset = s->deblock[ctb - 1].beta_offset;
}

x_end = x0+ctb_size;
x_end = x0 + ctb_size;
if (x_end > s->sps->width)
x_end = s->sps->width;
y_end = y0+ctb_size;
y_end = y0 + ctb_size;
if (y_end > s->sps->height)
y_end = s->sps->height;

tc_offset = cur_tc_offset;
tc_offset = cur_tc_offset;
beta_offset = cur_beta_offset;

// vertical filtering luma
@@ -353,22 +379,26 @@ static void deblocking_filter_CTB(HEVCContext *s, int x0, int y0)
const int bs0 = s->vertical_bs[(x >> 3) + (y >> 2) * s->bs_width];
const int bs1 = s->vertical_bs[(x >> 3) + ((y + 4) >> 2) * s->bs_width];
if (bs0 || bs1) {
const int qp0 = (get_qPy(s, x - 1, y) + get_qPy(s, x, y) + 1) >> 1;
const int qp0 = (get_qPy(s, x - 1, y) + get_qPy(s, x, y) + 1) >> 1;
const int qp1 = (get_qPy(s, x - 1, y + 4) + get_qPy(s, x, y + 4) + 1) >> 1;

beta[0] = betatable[av_clip(qp0 + ((beta_offset >> 1) << 1), 0, MAX_QP)];
beta[1] = betatable[av_clip(qp1 + ((beta_offset >> 1) << 1), 0, MAX_QP)];
tc[0] = bs0 ? TC_CALC(qp0, bs0) : 0;
tc[1] = bs1 ? TC_CALC(qp1, bs1) : 0;
src = &s->frame->data[LUMA][y * s->frame->linesize[LUMA] + (x << s->sps->pixel_shift)];
tc[0] = bs0 ? TC_CALC(qp0, bs0) : 0;
tc[1] = bs1 ? TC_CALC(qp1, bs1) : 0;
src = &s->frame->data[LUMA][y * s->frame->linesize[LUMA] + (x << s->sps->pixel_shift)];
if (pcmf) {
no_p[0] = get_pcm(s, x - 1, y);
no_p[1] = get_pcm(s, x - 1, y + 4);
no_q[0] = get_pcm(s, x, y);
no_q[1] = get_pcm(s, x, y + 4);
s->hevcdsp.hevc_v_loop_filter_luma_c(src, s->frame->linesize[LUMA], beta, tc, no_p, no_q);
s->hevcdsp.hevc_v_loop_filter_luma_c(src,
s->frame->linesize[LUMA],
beta, tc, no_p, no_q);
} else
s->hevcdsp.hevc_v_loop_filter_luma(src, s->frame->linesize[LUMA], beta, tc, no_p, no_q);
s->hevcdsp.hevc_v_loop_filter_luma(src,
s->frame->linesize[LUMA],
beta, tc, no_p, no_q);
}
}
}
@@ -376,24 +406,28 @@ static void deblocking_filter_CTB(HEVCContext *s, int x0, int y0)
// vertical filtering chroma
for (chroma = 1; chroma <= 2; chroma++) {
for (y = y0; y < y_end; y += 16) {
for (x = x0 ? x0:16; x < x_end; x += 16) {
const int bs0 = s->vertical_bs[(x >> 3) + (y >> 2) * s->bs_width];
for (x = x0 ? x0 : 16; x < x_end; x += 16) {
const int bs0 = s->vertical_bs[(x >> 3) + (y >> 2) * s->bs_width];
const int bs1 = s->vertical_bs[(x >> 3) + ((y + 8) >> 2) * s->bs_width];
if ((bs0 == 2) || (bs1 == 2)) {
const int qp0 = (get_qPy(s, x - 1, y) + get_qPy(s, x, y) + 1) >> 1;
const int qp0 = (get_qPy(s, x - 1, y) + get_qPy(s, x, y) + 1) >> 1;
const int qp1 = (get_qPy(s, x - 1, y + 8) + get_qPy(s, x, y + 8) + 1) >> 1;

c_tc[0] = (bs0 == 2) ? chroma_tc(s, qp0, chroma, tc_offset) : 0;
c_tc[1] = (bs1 == 2) ? chroma_tc(s, qp1, chroma, tc_offset) : 0;
src = &s->frame->data[chroma][(y / 2) * s->frame->linesize[chroma] + ((x / 2) << s->sps->pixel_shift)];
src = &s->frame->data[chroma][(y / 2) * s->frame->linesize[chroma] + ((x / 2) << s->sps->pixel_shift)];
if (pcmf) {
no_p[0] = get_pcm(s, x - 1, y);
no_p[1] = get_pcm(s, x - 1, y + 8);
no_q[0] = get_pcm(s, x, y);
no_q[1] = get_pcm(s, x, y + 8);
s->hevcdsp.hevc_v_loop_filter_chroma_c(src, s->frame->linesize[chroma], c_tc, no_p, no_q);
s->hevcdsp.hevc_v_loop_filter_chroma_c(src,
s->frame->linesize[chroma],
c_tc, no_p, no_q);
} else
s->hevcdsp.hevc_v_loop_filter_chroma(src, s->frame->linesize[chroma], c_tc, no_p, no_q);
s->hevcdsp.hevc_v_loop_filter_chroma(src,
s->frame->linesize[chroma],
c_tc, no_p, no_q);
}
}
}
@@ -410,22 +444,26 @@ static void deblocking_filter_CTB(HEVCContext *s, int x0, int y0)
const int qp0 = (get_qPy(s, x, y - 1) + get_qPy(s, x, y) + 1) >> 1;
const int qp1 = (get_qPy(s, x + 4, y - 1) + get_qPy(s, x + 4, y) + 1) >> 1;

tc_offset = x >= x0 ? cur_tc_offset : left_tc_offset;
tc_offset = x >= x0 ? cur_tc_offset : left_tc_offset;
beta_offset = x >= x0 ? cur_beta_offset : left_beta_offset;

beta[0] = betatable[av_clip(qp0 + ((beta_offset >> 1) << 1), 0, MAX_QP)];
beta[1] = betatable[av_clip(qp1 + ((beta_offset >> 1) << 1), 0, MAX_QP)];
tc[0] = bs0 ? TC_CALC(qp0, bs0) : 0;
tc[1] = bs1 ? TC_CALC(qp1, bs1) : 0;
src = &s->frame->data[LUMA][y * s->frame->linesize[LUMA] + (x << s->sps->pixel_shift)];
beta[0] = betatable[av_clip(qp0 + ((beta_offset >> 1) << 1), 0, MAX_QP)];
beta[1] = betatable[av_clip(qp1 + ((beta_offset >> 1) << 1), 0, MAX_QP)];
tc[0] = bs0 ? TC_CALC(qp0, bs0) : 0;
tc[1] = bs1 ? TC_CALC(qp1, bs1) : 0;
src = &s->frame->data[LUMA][y * s->frame->linesize[LUMA] + (x << s->sps->pixel_shift)];
if (pcmf) {
no_p[0] = get_pcm(s, x, y - 1);
no_p[1] = get_pcm(s, x + 4, y - 1);
no_q[0] = get_pcm(s, x, y);
no_q[1] = get_pcm(s, x + 4, y);
s->hevcdsp.hevc_h_loop_filter_luma_c(src, s->frame->linesize[LUMA], beta, tc, no_p, no_q);
s->hevcdsp.hevc_h_loop_filter_luma_c(src,
s->frame->linesize[LUMA],
beta, tc, no_p, no_q);
} else
s->hevcdsp.hevc_h_loop_filter_luma(src, s->frame->linesize[LUMA], beta, tc, no_p, no_q);
s->hevcdsp.hevc_h_loop_filter_luma(src,
s->frame->linesize[LUMA],
beta, tc, no_p, no_q);
}
}
}
@@ -441,29 +479,33 @@ static void deblocking_filter_CTB(HEVCContext *s, int x0, int y0)
bs0 = 0;
bs1 = s->horizontal_bs[(x + 8 + y * s->bs_width) >> 2];
} else if (x >= x_end - 8) {
bs0 = s->horizontal_bs[(x + y * s->bs_width) >> 2];
bs0 = s->horizontal_bs[(x + y * s->bs_width) >> 2];
bs1 = 0;
} else {
bs0 = s->horizontal_bs[(x + y * s->bs_width) >> 2];
bs0 = s->horizontal_bs[(x + y * s->bs_width) >> 2];
bs1 = s->horizontal_bs[(x + 8 + y * s->bs_width) >> 2];
}

if ((bs0 == 2) || (bs1 == 2)) {
const int qp0 = (bs0 == 2) ? ((get_qPy(s, x, y - 1) + get_qPy(s, x, y) + 1) >> 1) : 0;
const int qp0 = (bs0 == 2) ? ((get_qPy(s, x, y - 1) + get_qPy(s, x, y) + 1) >> 1) : 0;
const int qp1 = (bs1 == 2) ? ((get_qPy(s, x + 8, y - 1) + get_qPy(s, x + 8, y) + 1) >> 1) : 0;

tc_offset = x >= x0 ? cur_tc_offset : left_tc_offset;
c_tc[0] = (bs0 == 2) ? chroma_tc(s, qp0, chroma, tc_offset) : 0;
c_tc[0] = (bs0 == 2) ? chroma_tc(s, qp0, chroma, tc_offset) : 0;
c_tc[1] = (bs1 == 2) ? chroma_tc(s, qp1, chroma, cur_tc_offset) : 0;
src = &s->frame->data[chroma][(y / 2) * s->frame->linesize[chroma] + ((x / 2) << s->sps->pixel_shift)];
src = &s->frame->data[chroma][(y / 2) * s->frame->linesize[chroma] + ((x / 2) << s->sps->pixel_shift)];
if (pcmf) {
no_p[0] = get_pcm(s, x, y - 1);
no_p[1] = get_pcm(s, x + 8, y - 1);
no_q[0] = get_pcm(s, x, y);
no_q[1] = get_pcm(s, x + 8, y);
s->hevcdsp.hevc_h_loop_filter_chroma_c(src, s->frame->linesize[chroma], c_tc, no_p, no_q);
s->hevcdsp.hevc_h_loop_filter_chroma_c(src,
s->frame->linesize[chroma],
c_tc, no_p, no_q);
} else
s->hevcdsp.hevc_h_loop_filter_chroma(src, s->frame->linesize[chroma], c_tc, no_p, no_q);
s->hevcdsp.hevc_h_loop_filter_chroma(src,
s->frame->linesize[chroma],
c_tc, no_p, no_q);
}
}
}
@@ -472,7 +514,8 @@ static void deblocking_filter_CTB(HEVCContext *s, int x0, int y0)

static int boundary_strength(HEVCContext *s, MvField *curr,
uint8_t curr_cbf_luma, MvField *neigh,
uint8_t neigh_cbf_luma, RefPicList *neigh_refPicList,
uint8_t neigh_cbf_luma,
RefPicList *neigh_refPicList,
int tu_border)
{
int mvs = curr->pred_flag[0] + curr->pred_flag[1];
@@ -487,7 +530,7 @@ static int boundary_strength(HEVCContext *s, MvField *curr,
if (mvs == neigh->pred_flag[0] + neigh->pred_flag[1]) {
if (mvs == 2) {
// same L0 and L1
if (s->ref->refPicList[0].list[curr->ref_idx[0]] == neigh_refPicList[0].list[neigh->ref_idx[0]] &&
if (s->ref->refPicList[0].list[curr->ref_idx[0]] == neigh_refPicList[0].list[neigh->ref_idx[0]] &&
s->ref->refPicList[0].list[curr->ref_idx[0]] == s->ref->refPicList[1].list[curr->ref_idx[1]] &&
neigh_refPicList[0].list[neigh->ref_idx[0]] == neigh_refPicList[1].list[neigh->ref_idx[1]]) {
if ((abs(neigh->mv[0].x - curr->mv[0].x) >= 4 || abs(neigh->mv[0].y - curr->mv[0].y) >= 4 ||
@@ -520,18 +563,18 @@ static int boundary_strength(HEVCContext *s, MvField *curr,
int ref_B;

if (curr->pred_flag[0]) {
A = curr->mv[0];
A = curr->mv[0];
ref_A = s->ref->refPicList[0].list[curr->ref_idx[0]];
} else {
A = curr->mv[1];
A = curr->mv[1];
ref_A = s->ref->refPicList[1].list[curr->ref_idx[1]];
}

if (neigh->pred_flag[0]) {
B = neigh->mv[0];
B = neigh->mv[0];
ref_B = neigh_refPicList[0].list[neigh->ref_idx[0]];
} else {
B = neigh->mv[1];
B = neigh->mv[1];
ref_B = neigh_refPicList[1].list[neigh->ref_idx[1]];
}

@@ -548,24 +591,26 @@ static int boundary_strength(HEVCContext *s, MvField *curr,
return 1;
}

void ff_hevc_deblocking_boundary_strengths(HEVCContext *s, int x0, int y0, int log2_trafo_size,
int slice_or_tiles_up_boundary, int slice_or_tiles_left_boundary)
void ff_hevc_deblocking_boundary_strengths(HEVCContext *s, int x0, int y0,
int log2_trafo_size,
int slice_or_tiles_up_boundary,
int slice_or_tiles_left_boundary)
{
MvField *tab_mvf = s->ref->tab_mvf;
int log2_min_pu_size = s->sps->log2_min_pu_size;
int log2_min_tu_size = s->sps->log2_min_tb_size;
int min_pu_width = s->sps->min_pu_width;
int min_tu_width = s->sps->min_tb_width;
int is_intra = tab_mvf[(y0 >> log2_min_pu_size) * min_pu_width + (x0 >> log2_min_pu_size)].is_intra;
int is_intra = tab_mvf[(y0 >> log2_min_pu_size) * min_pu_width +
(x0 >> log2_min_pu_size)].is_intra;
int i, j;
int bs;

if (y0 > 0 && (y0 & 7) == 0) {
int yp_pu = (y0 - 1) >> log2_min_pu_size;
int yq_pu = y0 >> log2_min_pu_size;
int yq_pu = y0 >> log2_min_pu_size;
int yp_tu = (y0 - 1) >> log2_min_tu_size;
int yq_tu = y0 >> log2_min_tu_size;
int yq_tu = y0 >> log2_min_tu_size;

for (i = 0; i < (1 << log2_trafo_size); i += 4) {
int x_pu = (x0 + i) >> log2_min_pu_size;
@@ -574,12 +619,18 @@ void ff_hevc_deblocking_boundary_strengths(HEVCContext *s, int x0, int y0, int l
MvField *curr = &tab_mvf[yq_pu * min_pu_width + x_pu];
uint8_t top_cbf_luma = s->cbf_luma[yp_tu * min_tu_width + x_tu];
uint8_t curr_cbf_luma = s->cbf_luma[yq_tu * min_tu_width + x_tu];
RefPicList* top_refPicList = ff_hevc_get_ref_list(s, s->ref, x0 + i, y0 - 1);

bs = boundary_strength(s, curr, curr_cbf_luma, top, top_cbf_luma, top_refPicList, 1);
if (!s->sh.slice_loop_filter_across_slices_enabled_flag && (slice_or_tiles_up_boundary & 1) && (y0 % (1 << s->sps->log2_ctb_size)) == 0)
RefPicList *top_refPicList = ff_hevc_get_ref_list(s, s->ref,
x0 + i, y0 - 1);

bs = boundary_strength(s, curr, curr_cbf_luma,
top, top_cbf_luma, top_refPicList, 1);
if (!s->sh.slice_loop_filter_across_slices_enabled_flag &&
(slice_or_tiles_up_boundary & 1) &&
(y0 % (1 << s->sps->log2_ctb_size)) == 0)
bs = 0;
else if (!s->pps->loop_filter_across_tiles_enabled_flag && (slice_or_tiles_up_boundary & 2) && (y0 % (1 << s->sps->log2_ctb_size)) == 0)
else if (!s->pps->loop_filter_across_tiles_enabled_flag &&
(slice_or_tiles_up_boundary & 2) &&
(y0 % (1 << s->sps->log2_ctb_size)) == 0)
bs = 0;
if (y0 == 0 || s->sh.disable_deblocking_filter_flag == 1)
bs = 0;
@@ -596,17 +647,19 @@ void ff_hevc_deblocking_boundary_strengths(HEVCContext *s, int x0, int y0, int l
int yp_tu = (y0 + j - 1) >> log2_min_tu_size;
int yq_tu = (y0 + j) >> log2_min_tu_size;


for (i = 0; i < (1<<log2_trafo_size); i += 4) {
for (i = 0; i < (1 << log2_trafo_size); i += 4) {
int x_pu = (x0 + i) >> log2_min_pu_size;
int x_tu = (x0 + i) >> log2_min_tu_size;
MvField *top = &tab_mvf[yp_pu * min_pu_width + x_pu];
MvField *curr = &tab_mvf[yq_pu * min_pu_width + x_pu];
uint8_t top_cbf_luma = s->cbf_luma[yp_tu * min_tu_width + x_tu];
uint8_t curr_cbf_luma = s->cbf_luma[yq_tu * min_tu_width + x_tu];
RefPicList* top_refPicList = ff_hevc_get_ref_list(s, s->ref, x0 + i, y0 + j - 1);
RefPicList *top_refPicList = ff_hevc_get_ref_list(s, s->ref,
x0 + i,
y0 + j - 1);

bs = boundary_strength(s, curr, curr_cbf_luma, top, top_cbf_luma, top_refPicList, 0);
bs = boundary_strength(s, curr, curr_cbf_luma,
top, top_cbf_luma, top_refPicList, 0);
if (s->sh.disable_deblocking_filter_flag == 1)
bs = 0;
if (bs)
@@ -622,19 +675,25 @@ void ff_hevc_deblocking_boundary_strengths(HEVCContext *s, int x0, int y0, int l
int xq_tu = x0 >> log2_min_tu_size;

for (i = 0; i < (1 << log2_trafo_size); i += 4) {
int y_pu = (y0 + i) >> log2_min_pu_size;
int y_tu = (y0 + i) >> log2_min_tu_size;
int y_pu = (y0 + i) >> log2_min_pu_size;
int y_tu = (y0 + i) >> log2_min_tu_size;
MvField *left = &tab_mvf[y_pu * min_pu_width + xp_pu];
MvField *curr = &tab_mvf[y_pu * min_pu_width + xq_pu];

uint8_t left_cbf_luma = s->cbf_luma[y_tu * min_tu_width + xp_tu];
uint8_t curr_cbf_luma = s->cbf_luma[y_tu * min_tu_width + xq_tu];
RefPicList* left_refPicList = ff_hevc_get_ref_list(s, s->ref, x0 - 1, y0 + i);

bs = boundary_strength(s, curr, curr_cbf_luma, left, left_cbf_luma, left_refPicList, 1);
if (!s->sh.slice_loop_filter_across_slices_enabled_flag && (slice_or_tiles_left_boundary & 1) && (x0 % (1 << s->sps->log2_ctb_size)) == 0)
RefPicList *left_refPicList = ff_hevc_get_ref_list(s, s->ref,
x0 - 1, y0 + i);

bs = boundary_strength(s, curr, curr_cbf_luma,
left, left_cbf_luma, left_refPicList, 1);
if (!s->sh.slice_loop_filter_across_slices_enabled_flag &&
(slice_or_tiles_left_boundary & 1) &&
(x0 % (1 << s->sps->log2_ctb_size)) == 0)
bs = 0;
else if (!s->pps->loop_filter_across_tiles_enabled_flag && (slice_or_tiles_left_boundary & 2) && (x0 % (1 << s->sps->log2_ctb_size)) == 0)
else if (!s->pps->loop_filter_across_tiles_enabled_flag &&
(slice_or_tiles_left_boundary & 2) &&
(x0 % (1 << s->sps->log2_ctb_size)) == 0)
bs = 0;
if (x0 == 0 || s->sh.disable_deblocking_filter_flag == 1)
bs = 0;
@@ -651,16 +710,19 @@ void ff_hevc_deblocking_boundary_strengths(HEVCContext *s, int x0, int y0, int l

for (i = 8; i < (1 << log2_trafo_size); i += 8) {
int xp_pu = (x0 + i - 1) >> log2_min_pu_size;
int xq_pu = (x0 + i) >> log2_min_pu_size;
int xq_pu = (x0 + i) >> log2_min_pu_size;
int xp_tu = (x0 + i - 1) >> log2_min_tu_size;
int xq_tu = (x0 + i) >> log2_min_tu_size;
int xq_tu = (x0 + i) >> log2_min_tu_size;
MvField *left = &tab_mvf[y_pu * min_pu_width + xp_pu];
MvField *curr = &tab_mvf[y_pu * min_pu_width + xq_pu];
uint8_t left_cbf_luma = s->cbf_luma[y_tu * min_tu_width + xp_tu];
uint8_t curr_cbf_luma = s->cbf_luma[y_tu * min_tu_width + xq_tu];
RefPicList* left_refPicList = ff_hevc_get_ref_list(s, s->ref, x0 + i - 1, y0 + j);
RefPicList *left_refPicList = ff_hevc_get_ref_list(s, s->ref,
x0 + i - 1,
y0 + j);

bs = boundary_strength(s, curr, curr_cbf_luma, left, left_cbf_luma, left_refPicList, 0);
bs = boundary_strength(s, curr, curr_cbf_luma,
left, left_cbf_luma, left_refPicList, 0);
if (s->sh.disable_deblocking_filter_flag == 1)
bs = 0;
if (bs)
@@ -668,6 +730,7 @@ void ff_hevc_deblocking_boundary_strengths(HEVCContext *s, int x0, int y0, int l
}
}
}

#undef LUMA
#undef CB
#undef CR


+ 69
- 66
libavcodec/hevc_mvs.c View File

@@ -38,7 +38,8 @@ static const uint8_t l0_l1_cand_idx[12][2] = {
{ 3, 2, },
};

void ff_hevc_set_neighbour_available(HEVCContext *s, int x0, int y0, int nPbW, int nPbH)
void ff_hevc_set_neighbour_available(HEVCContext *s, int x0, int y0,
int nPbW, int nPbH)
{
HEVCLocalContext *lc = s->HEVClc;
int x0b = x0 & ((1 << s->sps->log2_ctb_size) - 1);
@@ -65,8 +66,8 @@ static int z_scan_block_avail(HEVCContext *s, int xCurr, int yCurr,
{
#define MIN_TB_ADDR_ZS(x, y) \
s->pps->min_tb_addr_zs[(y) * s->sps->min_tb_width + (x)]
int Curr = MIN_TB_ADDR_ZS(xCurr >> s->sps->log2_min_tb_size,
yCurr >> s->sps->log2_min_tb_size);
int Curr = MIN_TB_ADDR_ZS(xCurr >> s->sps->log2_min_tb_size,
yCurr >> s->sps->log2_min_tb_size);
int N;

if ((xN < 0) || (yN < 0) ||
@@ -80,7 +81,6 @@ static int z_scan_block_avail(HEVCContext *s, int xCurr, int yCurr,
return N <= Curr;
}


static int same_prediction_block(HEVCLocalContext *lc, int log2_cb_size,
int x0, int y0, int nPbW, int nPbH,
int xA1, int yA1, int partIdx)
@@ -169,7 +169,7 @@ static int check_mvset(Mv *mvLXCol, Mv *mvCol,
cur_poc_diff = poc - refPicList[X].list[refIdxLx];

if (!col_poc_diff)
col_poc_diff = 1; // error resilience
col_poc_diff = 1; // error resilience

if (cur_lt || col_poc_diff == cur_poc_diff) {
mvLXCol->x = mvCol->x;
@@ -180,16 +180,16 @@ static int check_mvset(Mv *mvLXCol, Mv *mvCol,
return 1;
}

#define CHECK_MVSET(l) \
check_mvset(mvLXCol, temp_col.mv + l, \
colPic, s->poc, \
refPicList, X, refIdxLx, \
refPicList_col, L##l, temp_col.ref_idx[l])
#define CHECK_MVSET(l) \
check_mvset(mvLXCol, temp_col.mv + l, \
colPic, s->poc, \
refPicList, X, refIdxLx, \
refPicList_col, L ## l, temp_col.ref_idx[l])

// derive the motion vectors section 8.5.3.1.8
static int derive_temporal_colocated_mvs(HEVCContext *s, MvField temp_col,
int refIdxLx, Mv* mvLXCol, int X,
int colPic, RefPicList* refPicList_col)
int refIdxLx, Mv *mvLXCol, int X,
int colPic, RefPicList *refPicList_col)
{
RefPicList *refPicList = s->ref->refPicList;

@@ -229,15 +229,15 @@ static int derive_temporal_colocated_mvs(HEVCContext *s, MvField temp_col,
return 0;
}

#define TAB_MVF(x, y) \
#define TAB_MVF(x, y) \
tab_mvf[(y) * min_pu_width + x]

#define TAB_MVF_PU(v) \
TAB_MVF(x##v##_pu, y##v##_pu)
#define TAB_MVF_PU(v) \
TAB_MVF(x ## v ## _pu, y ## v ## _pu)

#define DERIVE_TEMPORAL_COLOCATED_MVS \
derive_temporal_colocated_mvs(s, temp_col, \
refIdxLx, mvLXCol, X, colPic, \
#define DERIVE_TEMPORAL_COLOCATED_MVS \
derive_temporal_colocated_mvs(s, temp_col, \
refIdxLx, mvLXCol, X, colPic, \
ff_hevc_get_ref_list(s, ref, x, y))

/*
@@ -245,7 +245,7 @@ static int derive_temporal_colocated_mvs(HEVCContext *s, MvField temp_col,
*/
static int temporal_luma_motion_vector(HEVCContext *s, int x0, int y0,
int nPbW, int nPbH, int refIdxLx,
Mv* mvLXCol, int X)
Mv *mvLXCol, int X)
{
MvField *tab_mvf;
MvField temp_col;
@@ -273,44 +273,45 @@ static int temporal_luma_motion_vector(HEVCContext *s, int x0, int y0,
(y0 >> s->sps->log2_ctb_size) == (y >> s->sps->log2_ctb_size) &&
y < s->sps->height &&
x < s->sps->width) {
x = ((x >> 4) << 4);
y = ((y >> 4) << 4);
x_pu = x >> s->sps->log2_min_pu_size;
y_pu = y >> s->sps->log2_min_pu_size;
temp_col = TAB_MVF(x_pu, y_pu);
x = ((x >> 4) << 4);
y = ((y >> 4) << 4);
x_pu = x >> s->sps->log2_min_pu_size;
y_pu = y >> s->sps->log2_min_pu_size;
temp_col = TAB_MVF(x_pu, y_pu);
availableFlagLXCol = DERIVE_TEMPORAL_COLOCATED_MVS;
}

// derive center collocated motion vector
if (tab_mvf && !availableFlagLXCol) {
x = x0 + (nPbW >> 1);
y = y0 + (nPbH >> 1);
x = ((x >> 4) << 4);
y = ((y >> 4) << 4);
x_pu = x >> s->sps->log2_min_pu_size;
y_pu = y >> s->sps->log2_min_pu_size;
temp_col = TAB_MVF(x_pu, y_pu);
x = x0 + (nPbW >> 1);
y = y0 + (nPbH >> 1);
x = ((x >> 4) << 4);
y = ((y >> 4) << 4);
x_pu = x >> s->sps->log2_min_pu_size;
y_pu = y >> s->sps->log2_min_pu_size;
temp_col = TAB_MVF(x_pu, y_pu);
availableFlagLXCol = DERIVE_TEMPORAL_COLOCATED_MVS;
}
return availableFlagLXCol;
}

#define AVAILABLE(cand, v) \
#define AVAILABLE(cand, v) \
(cand && !TAB_MVF_PU(v).is_intra)

#define PRED_BLOCK_AVAILABLE(v) \
check_prediction_block_available(s, log2_cb_size, \
x0, y0, nPbW, nPbH, \
x##v, y##v, part_idx)
#define PRED_BLOCK_AVAILABLE(v) \
check_prediction_block_available(s, log2_cb_size, \
x0, y0, nPbW, nPbH, \
x ## v, y ## v, part_idx)

#define COMPARE_MV_REFIDX(a, b) \
#define COMPARE_MV_REFIDX(a, b) \
compareMVrefidx(TAB_MVF_PU(a), TAB_MVF_PU(b))

/*
* 8.5.3.1.2 Derivation process for spatial merging candidates
*/
static void derive_spatial_merge_candidates(HEVCContext *s, int x0, int y0,
int nPbW, int nPbH, int log2_cb_size,
int nPbW, int nPbH,
int log2_cb_size,
int singleMCLFlag, int part_idx,
struct MvField mergecandlist[])
{
@@ -318,7 +319,7 @@ static void derive_spatial_merge_candidates(HEVCContext *s, int x0, int y0,
RefPicList *refPicList = s->ref->refPicList;
MvField *tab_mvf = s->ref->tab_mvf;

const int min_pu_width = s->sps->min_pu_width;
const int min_pu_width = s->sps->min_pu_width;

const int cand_bottom_left = lc->na.cand_bottom_left;
const int cand_left = lc->na.cand_left;
@@ -402,7 +403,7 @@ static void derive_spatial_merge_candidates(HEVCContext *s, int x0, int y0,

// above right spatial merge candidate
check_MER = 1;
check_B0 = PRED_BLOCK_AVAILABLE(B0);
check_B0 = PRED_BLOCK_AVAILABLE(B0);

is_available_b0 = check_B0 && AVAILABLE(cand_up_right, B0);

@@ -417,7 +418,7 @@ static void derive_spatial_merge_candidates(HEVCContext *s, int x0, int y0,

// left bottom spatial merge candidate
check_MER = 1;
check_A0 = PRED_BLOCK_AVAILABLE(A0);
check_A0 = PRED_BLOCK_AVAILABLE(A0);

is_available_a0 = check_A0 && AVAILABLE(cand_bottom_left, A0);

@@ -539,11 +540,11 @@ void ff_hevc_luma_mv_merge_mode(HEVCContext *s, int x0, int y0, int nPbW,

if (s->pps->log2_parallel_merge_level > 2 && nCS == 8) {
singleMCLFlag = 1;
x0 = lc->cu.x;
y0 = lc->cu.y;
nPbW = nCS;
nPbH = nCS;
part_idx = 0;
x0 = lc->cu.x;
y0 = lc->cu.y;
nPbW = nCS;
nPbH = nCS;
part_idx = 0;
}

ff_hevc_set_neighbour_available(s, x0, y0, nPbW, nPbH);
@@ -553,21 +554,21 @@ void ff_hevc_luma_mv_merge_mode(HEVCContext *s, int x0, int y0, int nPbW,
if (mergecand_list[merge_idx].pred_flag[0] == 1 &&
mergecand_list[merge_idx].pred_flag[1] == 1 &&
(nPbW2 + nPbH2) == 12) {
mergecand_list[merge_idx].ref_idx[1] = -1;
mergecand_list[merge_idx].ref_idx[1] = -1;
mergecand_list[merge_idx].pred_flag[1] = 0;
}

*mv = mergecand_list[merge_idx];
}

static av_always_inline void dist_scale(HEVCContext *s, Mv * mv,
static av_always_inline void dist_scale(HEVCContext *s, Mv *mv,
int min_pu_width, int x, int y,
int elist, int ref_idx_curr, int ref_idx)
{
RefPicList *refPicList = s->ref->refPicList;
MvField *tab_mvf = s->ref->tab_mvf;
int ref_pic_elist = refPicList[elist].list[TAB_MVF(x, y).ref_idx[elist]];
int ref_pic_curr = refPicList[ref_idx_curr].list[ref_idx];
MvField *tab_mvf = s->ref->tab_mvf;
int ref_pic_elist = refPicList[elist].list[TAB_MVF(x, y).ref_idx[elist]];
int ref_pic_curr = refPicList[ref_idx_curr].list[ref_idx];

if (ref_pic_elist != ref_pic_curr)
mv_scale(mv, mv, s->poc - ref_pic_elist, s->poc - ref_pic_curr);
@@ -589,7 +590,6 @@ static int mv_mp_mode_mx(HEVCContext *s, int x, int y, int pred_flag_index,
return 0;
}


static int mv_mp_mode_mx_lt(HEVCContext *s, int x, int y, int pred_flag_index,
Mv *mv, int ref_idx_curr, int ref_idx)
{
@@ -597,25 +597,27 @@ static int mv_mp_mode_mx_lt(HEVCContext *s, int x, int y, int pred_flag_index,
int min_pu_width = s->sps->min_pu_width;

RefPicList *refPicList = s->ref->refPicList;
int currIsLongTerm = refPicList[ref_idx_curr].isLongTerm[ref_idx];
int currIsLongTerm = refPicList[ref_idx_curr].isLongTerm[ref_idx];

int colIsLongTerm =
refPicList[pred_flag_index].isLongTerm[(TAB_MVF(x, y).ref_idx[pred_flag_index])];

if (TAB_MVF(x, y).pred_flag[pred_flag_index] && colIsLongTerm == currIsLongTerm) {
if (TAB_MVF(x, y).pred_flag[pred_flag_index] &&
colIsLongTerm == currIsLongTerm) {
*mv = TAB_MVF(x, y).mv[pred_flag_index];
if (!currIsLongTerm)
dist_scale(s, mv, min_pu_width, x, y, pred_flag_index, ref_idx_curr, ref_idx);
dist_scale(s, mv, min_pu_width, x, y,
pred_flag_index, ref_idx_curr, ref_idx);
return 1;
}
return 0;
}

#define MP_MX(v, pred, mx) \
mv_mp_mode_mx(s, x##v##_pu, y##v##_pu, pred, &mx, ref_idx_curr, ref_idx)
#define MP_MX(v, pred, mx) \
mv_mp_mode_mx(s, x ## v ## _pu, y ## v ## _pu, pred, &mx, ref_idx_curr, ref_idx)

#define MP_MX_LT(v, pred, mx) \
mv_mp_mode_mx_lt(s, x##v##_pu, y##v##_pu, pred, &mx, ref_idx_curr, ref_idx)
#define MP_MX_LT(v, pred, mx) \
mv_mp_mode_mx_lt(s, x ## v ## _pu, y ## v ## _pu, pred, &mx, ref_idx_curr, ref_idx)

void ff_hevc_luma_mv_mvp_mode(HEVCContext *s, int x0, int y0, int nPbW,
int nPbH, int log2_cb_size, int part_idx,
@@ -683,8 +685,8 @@ void ff_hevc_luma_mv_mvp_mode(HEVCContext *s, int x0, int y0, int nPbW,
is_available_a0 = PRED_BLOCK_AVAILABLE(A0) && AVAILABLE(cand_bottom_left, A0);

//left spatial merge candidate
xA1 = x0 - 1;
yA1 = y0 + nPbH - 1;
xA1 = x0 - 1;
yA1 = y0 + nPbH - 1;
xA1_pu = xA1 >> s->sps->log2_min_pu_size;
yA1_pu = yA1 >> s->sps->log2_min_pu_size;

@@ -719,8 +721,8 @@ void ff_hevc_luma_mv_mvp_mode(HEVCContext *s, int x0, int y0, int nPbW,

// B candidates
// above right spatial merge candidate
xB0 = x0 + nPbW;
yB0 = y0 - 1;
xB0 = x0 + nPbW;
yB0 = y0 - 1;
xB0_pu = xB0 >> s->sps->log2_min_pu_size;
yB0_pu = yB0 >> s->sps->log2_min_pu_size;

@@ -734,8 +736,8 @@ void ff_hevc_luma_mv_mvp_mode(HEVCContext *s, int x0, int y0, int nPbW,

if (!availableFlagLXB0) {
// above spatial merge candidate
xB1 = x0 + nPbW - 1;
yB1 = y0 - 1;
xB1 = x0 + nPbW - 1;
yB1 = y0 - 1;
xB1_pu = xB1 >> s->sps->log2_min_pu_size;
yB1_pu = yB1 >> s->sps->log2_min_pu_size;

@@ -800,7 +802,8 @@ void ff_hevc_luma_mv_mvp_mode(HEVCContext *s, int x0, int y0, int nPbW,
if (numMVPCandLX < 2 && s->sh.slice_temporal_mvp_enabled_flag) {
Mv mv_col;
int available_col = temporal_luma_motion_vector(s, x0, y0, nPbW,
nPbH, ref_idx, &mv_col, LX);
nPbH, ref_idx,
&mv_col, LX);
if (available_col)
mvpcand_list[numMVPCandLX++] = mv_col;
}


+ 7
- 6
libavcodec/hevc_parser.c View File

@@ -21,6 +21,7 @@
*/

#include "libavutil/common.h"

#include "parser.h"
#include "hevc.h"
#include "golomb.h"
@@ -36,7 +37,8 @@ typedef struct HEVCParseContext {
* Find the end of the current frame in the bitstream.
* @return the position of the first byte of the next frame, or END_NOT_FOUND
*/
static int hevc_find_frame_end(AVCodecParserContext *s, const uint8_t *buf, int buf_size)
static int hevc_find_frame_end(AVCodecParserContext *s, const uint8_t *buf,
int buf_size)
{
int i;
ParseContext *pc = &((HEVCParseContext *)s->priv_data)->pc;
@@ -82,9 +84,8 @@ static int hevc_find_frame_end(AVCodecParserContext *s, const uint8_t *buf, int
* @param buf buffer with field/frame data.
* @param buf_size size of the buffer.
*/
static inline int parse_nal_units(AVCodecParserContext *s,
AVCodecContext *avctx,
const uint8_t *buf, int buf_size)
static inline int parse_nal_units(AVCodecParserContext *s, AVCodecContext *avctx,
const uint8_t *buf, int buf_size)
{
HEVCContext *h = &((HEVCParseContext *)s->priv_data)->h;
GetBitContext *gb = &h->HEVClc->gb;
@@ -269,7 +270,7 @@ static int hevc_parse(AVCodecParserContext *s,
} else {
next = hevc_find_frame_end(s, buf, buf_size);
if (ff_combine_frame(pc, next, &buf, &buf_size) < 0) {
*poutbuf = NULL;
*poutbuf = NULL;
*poutbuf_size = 0;
return buf_size;
}
@@ -299,9 +300,9 @@ static int hevc_split(AVCodecContext *avctx, const uint8_t *buf, int buf_size)
return i - 3;
} else { // no parameter set at the beginning of the stream
return 0;
}
}
}
}
return 0;
}



+ 31
- 29
libavcodec/hevc_ps.c View File

@@ -229,7 +229,7 @@ static int decode_profile_tier_level(HEVCContext *s, PTL *ptl,
}
if (max_num_sub_layers - 1 > 0)
for (i = max_num_sub_layers - 1; i < 8; i++)
skip_bits(gb, 2); // reserved_zero_2bits[i]
skip_bits(gb, 2); // reserved_zero_2bits[i]
for (i = 0; i < max_num_sub_layers - 1; i++) {
if (ptl->sub_layer_profile_present_flag[i]) {
ptl->sub_layer_profile_space[i] = get_bits(gb, 2);
@@ -273,7 +273,8 @@ static void decode_sublayer_hrd(HEVCContext *s, int nb_cpb,
}
}

static void decode_hrd(HEVCContext *s, int common_inf_present, int max_sublayers)
static void decode_hrd(HEVCContext *s, int common_inf_present,
int max_sublayers)
{
GetBitContext *gb = &s->HEVClc->gb;
int nal_params_present = 0, vcl_params_present = 0;
@@ -298,7 +299,7 @@ static void decode_hrd(HEVCContext *s, int common_inf_present, int max_sublayers
skip_bits(gb, 4); // cpb_size_scale

if (subpic_params_present)
skip_bits(gb, 4); // cpb_size_du_scale
skip_bits(gb, 4); // cpb_size_du_scale

skip_bits(gb, 5); // initial_cpb_removal_delay_length_minus1
skip_bits(gb, 5); // au_cpb_removal_delay_length_minus1
@@ -315,7 +316,7 @@ static void decode_hrd(HEVCContext *s, int common_inf_present, int max_sublayers
fixed_rate = get_bits1(gb);

if (fixed_rate)
get_ue_golomb_long(gb); // elemental_duration_in_tc_minus1
get_ue_golomb_long(gb); // elemental_duration_in_tc_minus1
else
low_delay = get_bits1(gb);

@@ -396,7 +397,7 @@ int ff_hevc_decode_nal_vps(HEVCContext *s)
vps->vps_num_layer_sets = get_ue_golomb_long(gb) + 1;
for (i = 1; i < vps->vps_num_layer_sets; i++)
for (j = 0; j <= vps->vps_max_layer_id; j++)
skip_bits(gb, 1); // layer_id_included_flag[i][j]
skip_bits(gb, 1); // layer_id_included_flag[i][j]

vps->vps_timing_info_present_flag = get_bits1(gb);
if (vps->vps_timing_info_present_flag) {
@@ -582,7 +583,7 @@ static int scaling_list_data(HEVCContext *s, ScalingList *sl)
int32_t scaling_list_delta_coef;

next_coef = 8;
coef_num = FFMIN(64, 1 << (4 + (size_id << 1)));
coef_num = FFMIN(64, 1 << (4 + (size_id << 1)));
if (size_id > 1) {
scaling_list_dc_coef[size_id - 2][matrix_id] = get_se_golomb(gb) + 8;
next_coef = scaling_list_dc_coef[size_id - 2][matrix_id];
@@ -643,7 +644,8 @@ int ff_hevc_decode_nal_sps(HEVCContext *s)
}

skip_bits1(gb); // temporal_id_nesting_flag
if (decode_profile_tier_level(s, &sps->ptl, sps->max_sub_layers) < 0) {
if (decode_profile_tier_level(s, &sps->ptl,
sps->max_sub_layers) < 0) {
av_log(s->avctx, AV_LOG_ERROR, "error decoding profile tier level\n");
ret = AVERROR_INVALIDDATA;
goto err;
@@ -820,11 +822,12 @@ int ff_hevc_decode_nal_sps(HEVCContext *s)

sps->nb_st_rps = get_ue_golomb_long(gb);
if (sps->nb_st_rps > MAX_SHORT_TERM_RPS_COUNT) {
av_log(s->avctx, AV_LOG_ERROR, "Too many short term RPS: %d.\n",
av_log(s->avctx, AV_LOG_ERROR,
"Too many short term RPS: %d.\n",
sps->nb_st_rps);
ret = AVERROR_INVALIDDATA;
goto err;
}
goto err;
}
for (i = 0; i < sps->nb_st_rps; i++) {
if ((ret = ff_hevc_decode_short_term_rps(s, &sps->st_rps[i],
sps, 0)) < 0)
@@ -870,10 +873,9 @@ int ff_hevc_decode_nal_sps(HEVCContext *s)
sps->output_width, sps->output_height);
if (s->avctx->err_recognition & AV_EF_EXPLODE) {
ret = AVERROR_INVALIDDATA;
goto err;
}
av_log(s->avctx, AV_LOG_WARNING,
"Displaying the whole video surface.\n");
goto err;
}
av_log(s->avctx, AV_LOG_WARNING, "Displaying the whole video surface.\n");
sps->pic_conf_win.left_offset =
sps->pic_conf_win.right_offset =
sps->pic_conf_win.top_offset =
@@ -1016,8 +1018,8 @@ int ff_hevc_decode_nal_pps(HEVCContext *s)
if (pps_id >= MAX_PPS_COUNT) {
av_log(s->avctx, AV_LOG_ERROR, "PPS id out of range: %d\n", pps_id);
ret = AVERROR_INVALIDDATA;
goto err;
}
goto err;
}
pps->sps_id = get_ue_golomb_long(gb);
if (pps->sps_id >= MAX_SPS_COUNT) {
av_log(s->avctx, AV_LOG_ERROR, "SPS id out of range: %d\n", pps->sps_id);
@@ -1080,11 +1082,11 @@ int ff_hevc_decode_nal_pps(HEVCContext *s)
pps->num_tile_rows = get_ue_golomb_long(gb) + 1;
if (pps->num_tile_columns == 0 ||
pps->num_tile_columns >= sps->width) {
av_log(s->avctx, AV_LOG_ERROR, "num_tile_columns_minus1 out of range: %d\n",
av_log(s->avctx, AV_LOG_ERROR, "num_tile_columns_minus1 out of range: %d\n",
pps->num_tile_columns - 1);
ret = AVERROR_INVALIDDATA;
goto err;
}
goto err;
}
if (pps->num_tile_rows == 0 ||
pps->num_tile_rows >= sps->height) {
av_log(s->avctx, AV_LOG_ERROR, "num_tile_rows_minus1 out of range: %d\n",
@@ -1110,8 +1112,8 @@ int ff_hevc_decode_nal_pps(HEVCContext *s)
if (sum >= sps->ctb_width) {
av_log(s->avctx, AV_LOG_ERROR, "Invalid tile widths.\n");
ret = AVERROR_INVALIDDATA;
goto err;
}
goto err;
}
pps->column_width[pps->num_tile_columns - 1] = sps->ctb_width - sum;

sum = 0;
@@ -1166,8 +1168,8 @@ int ff_hevc_decode_nal_pps(HEVCContext *s)
av_log(s->avctx, AV_LOG_ERROR, "log2_parallel_merge_level_minus2 out of range: %d\n",
pps->log2_parallel_merge_level - 2);
ret = AVERROR_INVALIDDATA;
goto err;
}
goto err;
}

pps->slice_header_extension_present_flag = get_bits1(gb);
pps->pps_extension_flag = get_bits1(gb);
@@ -1187,9 +1189,9 @@ int ff_hevc_decode_nal_pps(HEVCContext *s)
pps->row_height = av_malloc_array(pps->num_tile_rows, sizeof(*pps->row_height));
}
if (!pps->column_width || !pps->row_height) {
ret = AVERROR(ENOMEM);
goto err;
}
ret = AVERROR(ENOMEM);
goto err;
}

for (i = 0; i < pps->num_tile_columns; i++) {
pps->column_width[i] = ((i + 1) * sps->ctb_width) / pps->num_tile_columns -
@@ -1230,9 +1232,9 @@ int ff_hevc_decode_nal_pps(HEVCContext *s)
pps->min_tb_addr_zs = av_malloc_array(pic_area_in_min_tbs, sizeof(*pps->min_tb_addr_zs));
if (!pps->ctb_addr_rs_to_ts || !pps->ctb_addr_ts_to_rs ||
!pps->tile_id || !pps->min_cb_addr_zs || !pps->min_tb_addr_zs) {
ret = AVERROR(ENOMEM);
goto err;
}
ret = AVERROR(ENOMEM);
goto err;
}

for (ctb_addr_rs = 0; ctb_addr_rs < pic_area_in_ctbs; ctb_addr_rs++) {
int tb_x = ctb_addr_rs % sps->ctb_width;


+ 29
- 27
libavcodec/hevc_refs.c View File

@@ -49,16 +49,17 @@ void ff_hevc_unref_frame(HEVCContext *s, HEVCFrame *frame, int flags)
}
}

RefPicList* ff_hevc_get_ref_list(HEVCContext *s, HEVCFrame *ref, int x0, int y0)
RefPicList *ff_hevc_get_ref_list(HEVCContext *s, HEVCFrame *ref, int x0, int y0)
{
if (x0 < 0 || y0 < 0) {
return s->ref->refPicList;
} else {
int x_cb = x0 >> s->sps->log2_ctb_size;
int y_cb = y0 >> s->sps->log2_ctb_size;
int pic_width_cb = (s->sps->width + (1<<s->sps->log2_ctb_size)-1 ) >> s->sps->log2_ctb_size;
int pic_width_cb = (s->sps->width + (1 << s->sps->log2_ctb_size) - 1) >>
s->sps->log2_ctb_size;
int ctb_addr_ts = s->pps->ctb_addr_rs_to_ts[y_cb * pic_width_cb + x_cb];
return (RefPicList*) ref->rpl_tab[ctb_addr_ts];
return (RefPicList *)ref->rpl_tab[ctb_addr_ts];
}
}

@@ -67,7 +68,8 @@ void ff_hevc_clear_refs(HEVCContext *s)
int i;
for (i = 0; i < FF_ARRAY_ELEMS(s->DPB); i++)
ff_hevc_unref_frame(s, &s->DPB[i],
HEVC_FRAME_FLAG_SHORT_REF | HEVC_FRAME_FLAG_LONG_REF);
HEVC_FRAME_FLAG_SHORT_REF |
HEVC_FRAME_FLAG_LONG_REF);
}

void ff_hevc_flush_dpb(HEVCContext *s)
@@ -85,7 +87,8 @@ static HEVCFrame *alloc_frame(HEVCContext *s)
if (frame->frame->buf[0])
continue;

ret = ff_thread_get_buffer(s->avctx, &frame->tf, AV_GET_BUFFER_FLAG_REF);
ret = ff_thread_get_buffer(s->avctx, &frame->tf,
AV_GET_BUFFER_FLAG_REF);
if (ret < 0)
return NULL;

@@ -96,15 +99,15 @@ static HEVCFrame *alloc_frame(HEVCContext *s)
frame->tab_mvf_buf = av_buffer_pool_get(s->tab_mvf_pool);
if (!frame->tab_mvf_buf)
goto fail;
frame->tab_mvf = (MvField*)frame->tab_mvf_buf->data;
frame->tab_mvf = (MvField *)frame->tab_mvf_buf->data;

frame->rpl_tab_buf = av_buffer_pool_get(s->rpl_tab_pool);
if (!frame->rpl_tab_buf)
goto fail;
frame->rpl_tab = (RefPicListTab**)frame->rpl_tab_buf->data;
frame->rpl_tab = (RefPicListTab **)frame->rpl_tab_buf->data;
frame->ctb_count = s->sps->ctb_width * s->sps->ctb_height;
for (j = 0; j < frame->ctb_count; j++)
frame->rpl_tab[j] = (RefPicListTab*)frame->rpl_buf->data;
frame->rpl_tab[j] = (RefPicListTab *)frame->rpl_buf->data;

frame->frame->top_field_first = s->picture_struct == AV_PICTURE_STRUCTURE_TOP_FIELD;
frame->frame->interlaced_frame = (s->picture_struct == AV_PICTURE_STRUCTURE_TOP_FIELD) || (s->picture_struct == AV_PICTURE_STRUCTURE_BOTTOM_FIELD);
@@ -138,13 +141,13 @@ int ff_hevc_set_new_ref(HEVCContext *s, AVFrame **frame, int poc)
if (!ref)
return AVERROR(ENOMEM);

*frame = ref->frame;
s->ref = ref;
ref->poc = poc;
*frame = ref->frame;
s->ref = ref;

ref->flags = HEVC_FRAME_FLAG_OUTPUT | HEVC_FRAME_FLAG_SHORT_REF;
ref->sequence = s->seq_decode;
ref->window = s->sps->output_window;
ref->poc = poc;
ref->flags = HEVC_FRAME_FLAG_OUTPUT | HEVC_FRAME_FLAG_SHORT_REF;
ref->sequence = s->seq_decode;
ref->window = s->sps->output_window;

return 0;
}
@@ -192,7 +195,8 @@ int ff_hevc_output_frame(HEVCContext *s, AVFrame *out, int flush)
(frame->window.top_offset >> vshift) * dst->linesize[j];
dst->data[j] += off;
}
av_log(s->avctx, AV_LOG_DEBUG, "Output frame with POC %d.\n", frame->poc);
av_log(s->avctx, AV_LOG_DEBUG,
"Output frame with POC %d.\n", frame->poc);
return 1;
}

@@ -208,17 +212,17 @@ int ff_hevc_output_frame(HEVCContext *s, AVFrame *out, int flush)
static int init_slice_rpl(HEVCContext *s)
{
HEVCFrame *frame = s->ref;
int ctb_count = frame->ctb_count;
int ctb_addr_ts = s->pps->ctb_addr_rs_to_ts[s->sh.slice_segment_addr];
int ctb_count = frame->ctb_count;
int ctb_addr_ts = s->pps->ctb_addr_rs_to_ts[s->sh.slice_segment_addr];
int i;

if (s->slice_idx >= frame->rpl_buf->size / sizeof(RefPicListTab))
return AVERROR_INVALIDDATA;

for (i = ctb_addr_ts; i < ctb_count; i++)
frame->rpl_tab[i] = (RefPicListTab*)frame->rpl_buf->data + s->slice_idx;
frame->rpl_tab[i] = (RefPicListTab *)frame->rpl_buf->data + s->slice_idx;

frame->refPicList = (RefPicList*)frame->rpl_tab[ctb_addr_ts];
frame->refPicList = (RefPicList *)frame->rpl_tab[ctb_addr_ts];

return 0;
}
@@ -247,8 +251,7 @@ int ff_hevc_slice_rpl(HEVCContext *s)

/* The order of the elements is
* ST_CURR_BEF - ST_CURR_AFT - LT_CURR for the L0 and
* ST_CURR_AFT - ST_CURR_BEF - LT_CURR for the L1
*/
* ST_CURR_AFT - ST_CURR_BEF - LT_CURR for the L1 */
int cand_lists[3] = { list_idx ? ST_CURR_AFT : ST_CURR_BEF,
list_idx ? ST_CURR_BEF : ST_CURR_AFT,
LT_CURR };
@@ -258,8 +261,8 @@ int ff_hevc_slice_rpl(HEVCContext *s)
for (i = 0; i < FF_ARRAY_ELEMS(cand_lists); i++) {
RefPicList *rps = &s->rps[cand_lists[i]];
for (j = 0; j < rps->nb_refs && rpl_tmp.nb_refs < MAX_REFS; j++) {
rpl_tmp.list[rpl_tmp.nb_refs] = rps->list[j];
rpl_tmp.ref[rpl_tmp.nb_refs] = rps->ref[j];
rpl_tmp.list[rpl_tmp.nb_refs] = rps->list[j];
rpl_tmp.ref[rpl_tmp.nb_refs] = rps->ref[j];
rpl_tmp.isLongTerm[rpl_tmp.nb_refs] = (i == 2);
rpl_tmp.nb_refs++;
}
@@ -331,15 +334,14 @@ static HEVCFrame *generate_missing_ref(HEVCContext *s, int poc)
HEVCFrame *frame;
int i, x, y;


frame = alloc_frame(s);
if (!frame)
return NULL;

if (!s->sps->pixel_shift) {
for (i = 0; frame->frame->buf[i]; i++)
memset(frame->frame->buf[i]->data, 1 << (s->sps->bit_depth - 1),
frame->frame->buf[i]->size);
memset(frame->frame->buf[i]->data, 1 << (s->sps->bit_depth - 1),
frame->frame->buf[i]->size);
} else {
for (i = 0; frame->frame->data[i]; i++)
for (y = 0; y < (s->sps->height >> s->sps->vshift[i]); y++)
@@ -456,7 +458,7 @@ int ff_hevc_compute_poc(HEVCContext *s, int poc_lsb)
poc_msb = prev_poc_msb;

// For BLA picture types, POCmsb is set to 0.
if (s->nal_unit_type == NAL_BLA_W_LP ||
if (s->nal_unit_type == NAL_BLA_W_LP ||
s->nal_unit_type == NAL_BLA_W_RADL ||
s->nal_unit_type == NAL_BLA_N_LP)
poc_msb = 0;


+ 10
- 9
libavcodec/hevc_sei.c View File

@@ -25,7 +25,8 @@
#include "hevc.h"
#include "golomb.h"

static void decode_nal_sei_decoded_picture_hash(HEVCContext *s, int payload_size)
static void decode_nal_sei_decoded_picture_hash(HEVCContext *s,
int payload_size)
{
int cIdx, i;
uint8_t hash_type;
@@ -34,17 +35,16 @@ static void decode_nal_sei_decoded_picture_hash(HEVCContext *s, int payload_size
GetBitContext *gb = &s->HEVClc->gb;
hash_type = get_bits(gb, 8);


for( cIdx = 0; cIdx < 3/*((s->sps->chroma_format_idc == 0) ? 1 : 3)*/; cIdx++ ) {
if ( hash_type == 0 ) {
for (cIdx = 0; cIdx < 3/*((s->sps->chroma_format_idc == 0) ? 1 : 3)*/; cIdx++) {
if (hash_type == 0) {
s->is_md5 = 1;
for( i = 0; i < 16; i++) {
for (i = 0; i < 16; i++) {
s->md5[cIdx][i] = get_bits(gb, 8);
}
} else if( hash_type == 1 ) {
}
} else if( hash_type == 1) {
// picture_crc = get_bits(gb, 16);
skip_bits(gb, 16);
} else if( hash_type == 2 ) {
} else if (hash_type == 2) {
// picture_checksum = get_bits(gb, 32);
skip_bits(gb, 32);
}
@@ -101,7 +101,8 @@ static int decode_pic_timing(HEVCContext *s)
return 1;
}

static void active_parameter_sets(HEVCContext *s) {
static void active_parameter_sets(HEVCContext *s)
{
GetBitContext *gb = &s->HEVClc->gb;
int num_sps_ids_minus1;
int i;


+ 126
- 127
libavcodec/hevcdsp.c View File

@@ -24,80 +24,80 @@
#include "hevcdsp.h"

static const int8_t transform[32][32] = {
{ 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64 },
{ 90, 90, 88, 85, 82, 78, 73, 67, 61, 54, 46, 38, 31, 22, 13, 4,
-4, -13, -22, -31, -38, -46, -54, -61, -67, -73, -78, -82, -85, -88, -90, -90 },
{ 90, 87, 80, 70, 57, 43, 25, 9, -9, -25, -43, -57, -70, -80, -87, -90,
-90, -87, -80, -70, -57, -43, -25, -9, 9, 25, 43, 57, 70, 80, 87, 90 },
{ 90, 82, 67, 46, 22, -4, -31, -54, -73, -85, -90, -88, -78, -61, -38, -13,
13, 38, 61, 78, 88, 90, 85, 73, 54, 31, 4, -22, -46, -67, -82, -90 },
{ 89, 75, 50, 18, -18, -50, -75, -89, -89, -75, -50, -18, 18, 50, 75, 89,
89, 75, 50, 18, -18, -50, -75, -89, -89, -75, -50, -18, 18, 50, 75, 89 },
{ 88, 67, 31, -13, -54, -82, -90, -78, -46, -4, 38, 73, 90, 85, 61, 22,
-22, -61, -85, -90, -73, -38, 4, 46, 78, 90, 82, 54, 13, -31, -67, -88 },
{ 87, 57, 9, -43, -80, -90, -70, -25, 25, 70, 90, 80, 43, -9, -57, -87,
-87, -57, -9, 43, 80, 90, 70, 25, -25, -70, -90, -80, -43, 9, 57, 87 },
{ 85, 46, -13, -67, -90, -73, -22, 38, 82, 88, 54, -4, -61, -90, -78, -31,
31, 78, 90, 61, 4, -54, -88, -82, -38, 22, 73, 90, 67, 13, -46, -85 },
{ 83, 36, -36, -83, -83, -36, 36, 83, 83, 36, -36, -83, -83, -36, 36, 83,
83, 36, -36, -83, -83, -36, 36, 83, 83, 36, -36, -83, -83, -36, 36, 83 },
{ 82, 22, -54, -90, -61, 13, 78, 85, 31, -46, -90, -67, 4, 73, 88, 38,
-38, -88, -73, -4, 67, 90, 46, -31, -85, -78, -13, 61, 90, 54, -22, -82 },
{ 80, 9, -70, -87, -25, 57, 90, 43, -43, -90, -57, 25, 87, 70, -9, -80,
-80, -9, 70, 87, 25, -57, -90, -43, 43, 90, 57, -25, -87, -70, 9, 80 },
{ 78, -4, -82, -73, 13, 85, 67, -22, -88, -61, 31, 90, 54, -38, -90, -46,
46, 90, 38, -54, -90, -31, 61, 88, 22, -67, -85, -13, 73, 82, 4, -78 },
{ 75, -18, -89, -50, 50, 89, 18, -75, -75, 18, 89, 50, -50, -89, -18, 75,
75, -18, -89, -50, 50, 89, 18, -75, -75, 18, 89, 50, -50, -89, -18, 75 },
{ 73, -31, -90, -22, 78, 67, -38, -90, -13, 82, 61, -46, -88, -4, 85, 54,
-54, -85, 4, 88, 46, -61, -82, 13, 90, 38, -67, -78, 22, 90, 31, -73 },
{ 70, -43, -87, 9, 90, 25, -80, -57, 57, 80, -25, -90, -9, 87, 43, -70,
-70, 43, 87, -9, -90, -25, 80, 57, -57, -80, 25, 90, 9, -87, -43, 70 },
{ 67, -54, -78, 38, 85, -22, -90, 4, 90, 13, -88, -31, 82, 46, -73, -61,
61, 73, -46, -82, 31, 88, -13, -90, -4, 90, 22, -85, -38, 78, 54, -67 },
{ 64, -64, -64, 64, 64, -64, -64, 64, 64, -64, -64, 64, 64, -64, -64, 64,
64, -64, -64, 64, 64, -64, -64, 64, 64, -64, -64, 64, 64, -64, -64, 64 },
{ 61, -73, -46, 82, 31, -88, -13, 90, -4, -90, 22, 85, -38, -78, 54, 67,
-67, -54, 78, 38, -85, -22, 90, 4, -90, 13, 88, -31, -82, 46, 73, -61 },
{ 57, -80, -25, 90, -9, -87, 43, 70, -70, -43, 87, 9, -90, 25, 80, -57,
-57, 80, 25, -90, 9, 87, -43, -70, 70, 43, -87, -9, 90, -25, -80, 57 },
{ 54, -85, -4, 88, -46, -61, 82, 13, -90, 38, 67, -78, -22, 90, -31, -73,
73, 31, -90, 22, 78, -67, -38, 90, -13, -82, 61, 46, -88, 4, 85, -54 },
{ 50, -89, 18, 75, -75, -18, 89, -50, -50, 89, -18, -75, 75, 18, -89, 50,
50, -89, 18, 75, -75, -18, 89, -50, -50, 89, -18, -75, 75, 18, -89, 50 },
{ 46, -90, 38, 54, -90, 31, 61, -88, 22, 67, -85, 13, 73, -82, 4, 78,
-78, -4, 82, -73, -13, 85, -67, -22, 88, -61, -31, 90, -54, -38, 90, -46 },
{ 43, -90, 57, 25, -87, 70, 9, -80, 80, -9, -70, 87, -25, -57, 90, -43,
-43, 90, -57, -25, 87, -70, -9, 80, -80, 9, 70, -87, 25, 57, -90, 43 },
{ 38, -88, 73, -4, -67, 90, -46, -31, 85, -78, 13, 61, -90, 54, 22, -82,
82, -22, -54, 90, -61, -13, 78, -85, 31, 46, -90, 67, 4, -73, 88, -38 },
{ 36, -83, 83, -36, -36, 83, -83, 36, 36, -83, 83, -36, -36, 83, -83, 36,
36, -83, 83, -36, -36, 83, -83, 36, 36, -83, 83, -36, -36, 83, -83, 36 },
{ 31, -78, 90, -61, 4, 54, -88, 82, -38, -22, 73, -90, 67, -13, -46, 85,
-85, 46, 13, -67, 90, -73, 22, 38, -82, 88, -54, -4, 61, -90, 78, -31 },
{ 25, -70, 90, -80, 43, 9, -57, 87, -87, 57, -9, -43, 80, -90, 70, -25,
-25, 70, -90, 80, -43, -9, 57, -87, 87, -57, 9, 43, -80, 90, -70, 25 },
{ 22, -61, 85, -90, 73, -38, -4, 46, -78, 90, -82, 54, -13, -31, 67, -88,
88, -67, 31, 13, -54, 82, -90, 78, -46, 4, 38, -73, 90, -85, 61, -22 },
{ 18, -50, 75, -89, 89, -75, 50, -18, -18, 50, -75, 89, -89, 75, -50, 18,
18, -50, 75, -89, 89, -75, 50, -18, -18, 50, -75, 89, -89, 75, -50, 18 },
{ 13, -38, 61, -78, 88, -90, 85, -73, 54, -31, 4, 22, -46, 67, -82, 90,
-90, 82, -67, 46, -22, -4, 31, -54, 73, -85, 90, -88, 78, -61, 38, -13 },
{ 9, -25, 43, -57, 70, -80, 87, -90, 90, -87, 80, -70, 57, -43, 25, -9,
-9, 25, -43, 57, -70, 80, -87, 90, -90, 87, -80, 70, -57, 43, -25, 9 },
{ 4, -13, 22, -31, 38, -46, 54, -61, 67, -73, 78, -82, 85, -88, 90, -90,
90, -90, 88, -85, 82, -78, 73, -67, 61, -54, 46, -38, 31, -22, 13, -4 },
{ 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64,
64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 64 },
{ 90, 90, 88, 85, 82, 78, 73, 67, 61, 54, 46, 38, 31, 22, 13, 4,
-4, -13, -22, -31, -38, -46, -54, -61, -67, -73, -78, -82, -85, -88, -90, -90 },
{ 90, 87, 80, 70, 57, 43, 25, 9, -9, -25, -43, -57, -70, -80, -87, -90,
-90, -87, -80, -70, -57, -43, -25, -9, 9, 25, 43, 57, 70, 80, 87, 90 },
{ 90, 82, 67, 46, 22, -4, -31, -54, -73, -85, -90, -88, -78, -61, -38, -13,
13, 38, 61, 78, 88, 90, 85, 73, 54, 31, 4, -22, -46, -67, -82, -90 },
{ 89, 75, 50, 18, -18, -50, -75, -89, -89, -75, -50, -18, 18, 50, 75, 89,
89, 75, 50, 18, -18, -50, -75, -89, -89, -75, -50, -18, 18, 50, 75, 89 },
{ 88, 67, 31, -13, -54, -82, -90, -78, -46, -4, 38, 73, 90, 85, 61, 22,
-22, -61, -85, -90, -73, -38, 4, 46, 78, 90, 82, 54, 13, -31, -67, -88 },
{ 87, 57, 9, -43, -80, -90, -70, -25, 25, 70, 90, 80, 43, -9, -57, -87,
-87, -57, -9, 43, 80, 90, 70, 25, -25, -70, -90, -80, -43, 9, 57, 87 },
{ 85, 46, -13, -67, -90, -73, -22, 38, 82, 88, 54, -4, -61, -90, -78, -31,
31, 78, 90, 61, 4, -54, -88, -82, -38, 22, 73, 90, 67, 13, -46, -85 },
{ 83, 36, -36, -83, -83, -36, 36, 83, 83, 36, -36, -83, -83, -36, 36, 83,
83, 36, -36, -83, -83, -36, 36, 83, 83, 36, -36, -83, -83, -36, 36, 83 },
{ 82, 22, -54, -90, -61, 13, 78, 85, 31, -46, -90, -67, 4, 73, 88, 38,
-38, -88, -73, -4, 67, 90, 46, -31, -85, -78, -13, 61, 90, 54, -22, -82 },
{ 80, 9, -70, -87, -25, 57, 90, 43, -43, -90, -57, 25, 87, 70, -9, -80,
-80, -9, 70, 87, 25, -57, -90, -43, 43, 90, 57, -25, -87, -70, 9, 80 },
{ 78, -4, -82, -73, 13, 85, 67, -22, -88, -61, 31, 90, 54, -38, -90, -46,
46, 90, 38, -54, -90, -31, 61, 88, 22, -67, -85, -13, 73, 82, 4, -78 },
{ 75, -18, -89, -50, 50, 89, 18, -75, -75, 18, 89, 50, -50, -89, -18, 75,
75, -18, -89, -50, 50, 89, 18, -75, -75, 18, 89, 50, -50, -89, -18, 75 },
{ 73, -31, -90, -22, 78, 67, -38, -90, -13, 82, 61, -46, -88, -4, 85, 54,
-54, -85, 4, 88, 46, -61, -82, 13, 90, 38, -67, -78, 22, 90, 31, -73 },
{ 70, -43, -87, 9, 90, 25, -80, -57, 57, 80, -25, -90, -9, 87, 43, -70,
-70, 43, 87, -9, -90, -25, 80, 57, -57, -80, 25, 90, 9, -87, -43, 70 },
{ 67, -54, -78, 38, 85, -22, -90, 4, 90, 13, -88, -31, 82, 46, -73, -61,
61, 73, -46, -82, 31, 88, -13, -90, -4, 90, 22, -85, -38, 78, 54, -67 },
{ 64, -64, -64, 64, 64, -64, -64, 64, 64, -64, -64, 64, 64, -64, -64, 64,
64, -64, -64, 64, 64, -64, -64, 64, 64, -64, -64, 64, 64, -64, -64, 64 },
{ 61, -73, -46, 82, 31, -88, -13, 90, -4, -90, 22, 85, -38, -78, 54, 67,
-67, -54, 78, 38, -85, -22, 90, 4, -90, 13, 88, -31, -82, 46, 73, -61 },
{ 57, -80, -25, 90, -9, -87, 43, 70, -70, -43, 87, 9, -90, 25, 80, -57,
-57, 80, 25, -90, 9, 87, -43, -70, 70, 43, -87, -9, 90, -25, -80, 57 },
{ 54, -85, -4, 88, -46, -61, 82, 13, -90, 38, 67, -78, -22, 90, -31, -73,
73, 31, -90, 22, 78, -67, -38, 90, -13, -82, 61, 46, -88, 4, 85, -54 },
{ 50, -89, 18, 75, -75, -18, 89, -50, -50, 89, -18, -75, 75, 18, -89, 50,
50, -89, 18, 75, -75, -18, 89, -50, -50, 89, -18, -75, 75, 18, -89, 50 },
{ 46, -90, 38, 54, -90, 31, 61, -88, 22, 67, -85, 13, 73, -82, 4, 78,
-78, -4, 82, -73, -13, 85, -67, -22, 88, -61, -31, 90, -54, -38, 90, -46 },
{ 43, -90, 57, 25, -87, 70, 9, -80, 80, -9, -70, 87, -25, -57, 90, -43,
-43, 90, -57, -25, 87, -70, -9, 80, -80, 9, 70, -87, 25, 57, -90, 43 },
{ 38, -88, 73, -4, -67, 90, -46, -31, 85, -78, 13, 61, -90, 54, 22, -82,
82, -22, -54, 90, -61, -13, 78, -85, 31, 46, -90, 67, 4, -73, 88, -38 },
{ 36, -83, 83, -36, -36, 83, -83, 36, 36, -83, 83, -36, -36, 83, -83, 36,
36, -83, 83, -36, -36, 83, -83, 36, 36, -83, 83, -36, -36, 83, -83, 36 },
{ 31, -78, 90, -61, 4, 54, -88, 82, -38, -22, 73, -90, 67, -13, -46, 85,
-85, 46, 13, -67, 90, -73, 22, 38, -82, 88, -54, -4, 61, -90, 78, -31 },
{ 25, -70, 90, -80, 43, 9, -57, 87, -87, 57, -9, -43, 80, -90, 70, -25,
-25, 70, -90, 80, -43, -9, 57, -87, 87, -57, 9, 43, -80, 90, -70, 25 },
{ 22, -61, 85, -90, 73, -38, -4, 46, -78, 90, -82, 54, -13, -31, 67, -88,
88, -67, 31, 13, -54, 82, -90, 78, -46, 4, 38, -73, 90, -85, 61, -22 },
{ 18, -50, 75, -89, 89, -75, 50, -18, -18, 50, -75, 89, -89, 75, -50, 18,
18, -50, 75, -89, 89, -75, 50, -18, -18, 50, -75, 89, -89, 75, -50, 18 },
{ 13, -38, 61, -78, 88, -90, 85, -73, 54, -31, 4, 22, -46, 67, -82, 90,
-90, 82, -67, 46, -22, -4, 31, -54, 73, -85, 90, -88, 78, -61, 38, -13 },
{ 9, -25, 43, -57, 70, -80, 87, -90, 90, -87, 80, -70, 57, -43, 25, -9,
-9, 25, -43, 57, -70, 80, -87, 90, -90, 87, -80, 70, -57, 43, -25, 9 },
{ 4, -13, 22, -31, 38, -46, 54, -61, 67, -73, 78, -82, 85, -88, 90, -90,
90, -90, 88, -85, 82, -78, 73, -67, 61, -54, 46, -38, 31, -22, 13, -4 },
};

DECLARE_ALIGNED(16, const int8_t, ff_hevc_epel_filters[7][16]) = {
{ -2, 58, 10, -2,-2, 58, 10, -2,-2, 58, 10, -2,-2, 58, 10, -2 },
{ -4, 54, 16, -2,-4, 54, 16, -2,-4, 54, 16, -2,-4, 54, 16, -2 },
{ -6, 46, 28, -4,-6, 46, 28, -4,-6, 46, 28, -4,-6, 46, 28, -4 },
{ -4, 36, 36, -4,-4, 36, 36, -4,-4, 36, 36, -4,-4, 36, 36, -4 },
{ -4, 28, 46, -6,-4, 28, 46, -6,-4, 28, 46, -6,-4, 28, 46, -6 },
{ -2, 16, 54, -4,-2, 16, 54, -4,-2, 16, 54, -4,-2, 16, 54, -4 },
{ -2, 10, 58, -2,-2, 10, 58, -2,-2, 10, 58, -2,-2, 10, 58, -2 },
{ -2, 58, 10, -2, -2, 58, 10, -2, -2, 58, 10, -2, -2, 58, 10, -2 },
{ -4, 54, 16, -2, -4, 54, 16, -2, -4, 54, 16, -2, -4, 54, 16, -2 },
{ -6, 46, 28, -4, -6, 46, 28, -4, -6, 46, 28, -4, -6, 46, 28, -4 },
{ -4, 36, 36, -4, -4, 36, 36, -4, -4, 36, 36, -4, -4, 36, 36, -4 },
{ -4, 28, 46, -6, -4, 28, 46, -6, -4, 28, 46, -6, -4, 28, 46, -6 },
{ -2, 16, 54, -4, -2, 16, 54, -4, -2, 16, 54, -4, -2, 16, 54, -4 },
{ -2, 10, 58, -2, -2, 10, 58, -2, -2, 10, 58, -2, -2, 10, 58, -2 },
};

#define BIT_DEPTH 8
@@ -117,67 +117,66 @@ void ff_hevc_dsp_init(HEVCDSPContext *hevcdsp, int bit_depth)
#undef FUNC
#define FUNC(a, depth) a ## _ ## depth

#define HEVC_DSP(depth) \
hevcdsp->put_pcm = FUNC(put_pcm, depth); \
hevcdsp->transquant_bypass[0] = FUNC(transquant_bypass4x4, depth); \
hevcdsp->transquant_bypass[1] = FUNC(transquant_bypass8x8, depth); \
hevcdsp->transquant_bypass[2] = FUNC(transquant_bypass16x16, depth); \
hevcdsp->transquant_bypass[3] = FUNC(transquant_bypass32x32, depth); \
hevcdsp->transform_skip = FUNC(transform_skip, depth); \
hevcdsp->transform_4x4_luma_add = FUNC(transform_4x4_luma_add, depth); \
hevcdsp->transform_add[0] = FUNC(transform_4x4_add, depth); \
hevcdsp->transform_add[1] = FUNC(transform_8x8_add, depth); \
hevcdsp->transform_add[2] = FUNC(transform_16x16_add, depth); \
hevcdsp->transform_add[3] = FUNC(transform_32x32_add, depth); \
#define HEVC_DSP(depth) \
hevcdsp->put_pcm = FUNC(put_pcm, depth); \
hevcdsp->transquant_bypass[0] = FUNC(transquant_bypass4x4, depth); \
hevcdsp->transquant_bypass[1] = FUNC(transquant_bypass8x8, depth); \
hevcdsp->transquant_bypass[2] = FUNC(transquant_bypass16x16, depth); \
hevcdsp->transquant_bypass[3] = FUNC(transquant_bypass32x32, depth); \
hevcdsp->transform_skip = FUNC(transform_skip, depth); \
hevcdsp->transform_4x4_luma_add = FUNC(transform_4x4_luma_add, depth); \
hevcdsp->transform_add[0] = FUNC(transform_4x4_add, depth); \
hevcdsp->transform_add[1] = FUNC(transform_8x8_add, depth); \
hevcdsp->transform_add[2] = FUNC(transform_16x16_add, depth); \
hevcdsp->transform_add[3] = FUNC(transform_32x32_add, depth); \
\
hevcdsp->sao_band_filter[0] = FUNC(sao_band_filter_0, depth); \
hevcdsp->sao_band_filter[1] = FUNC(sao_band_filter_1, depth); \
hevcdsp->sao_band_filter[2] = FUNC(sao_band_filter_2, depth); \
hevcdsp->sao_band_filter[3] = FUNC(sao_band_filter_3, depth); \
\
hevcdsp->sao_edge_filter[0] = FUNC(sao_edge_filter_0, depth); \
hevcdsp->sao_edge_filter[1] = FUNC(sao_edge_filter_1, depth); \
hevcdsp->sao_edge_filter[2] = FUNC(sao_edge_filter_2, depth); \
hevcdsp->sao_edge_filter[3] = FUNC(sao_edge_filter_3, depth); \
\
hevcdsp->put_hevc_qpel[0][0] = FUNC(put_hevc_qpel_pixels, depth); \
hevcdsp->put_hevc_qpel[0][1] = FUNC(put_hevc_qpel_h1, depth); \
hevcdsp->put_hevc_qpel[0][2] = FUNC(put_hevc_qpel_h2, depth); \
hevcdsp->put_hevc_qpel[0][3] = FUNC(put_hevc_qpel_h3, depth); \
hevcdsp->put_hevc_qpel[1][0] = FUNC(put_hevc_qpel_v1, depth); \
hevcdsp->put_hevc_qpel[1][1] = FUNC(put_hevc_qpel_h1v1, depth); \
hevcdsp->put_hevc_qpel[1][2] = FUNC(put_hevc_qpel_h2v1, depth); \
hevcdsp->put_hevc_qpel[1][3] = FUNC(put_hevc_qpel_h3v1, depth); \
hevcdsp->put_hevc_qpel[2][0] = FUNC(put_hevc_qpel_v2, depth); \
hevcdsp->put_hevc_qpel[2][1] = FUNC(put_hevc_qpel_h1v2, depth); \
hevcdsp->put_hevc_qpel[2][2] = FUNC(put_hevc_qpel_h2v2, depth); \
hevcdsp->put_hevc_qpel[2][3] = FUNC(put_hevc_qpel_h3v2, depth); \
hevcdsp->put_hevc_qpel[3][0] = FUNC(put_hevc_qpel_v3, depth); \
hevcdsp->put_hevc_qpel[3][1] = FUNC(put_hevc_qpel_h1v3, depth); \
hevcdsp->put_hevc_qpel[3][2] = FUNC(put_hevc_qpel_h2v3, depth); \
hevcdsp->put_hevc_qpel[3][3] = FUNC(put_hevc_qpel_h3v3, depth); \
\
hevcdsp->put_hevc_epel[0][0] = FUNC(put_hevc_epel_pixels, depth); \
hevcdsp->put_hevc_epel[0][1] = FUNC(put_hevc_epel_h, depth); \
hevcdsp->put_hevc_epel[1][0] = FUNC(put_hevc_epel_v, depth); \
hevcdsp->put_hevc_epel[1][1] = FUNC(put_hevc_epel_hv, depth); \
\
hevcdsp->sao_band_filter[0] = FUNC(sao_band_filter_0, depth); \
hevcdsp->sao_band_filter[1] = FUNC(sao_band_filter_1, depth); \
hevcdsp->sao_band_filter[2] = FUNC(sao_band_filter_2, depth); \
hevcdsp->sao_band_filter[3] = FUNC(sao_band_filter_3, depth); \
\
hevcdsp->sao_edge_filter[0] = FUNC(sao_edge_filter_0, depth); \
hevcdsp->sao_edge_filter[1] = FUNC(sao_edge_filter_1, depth); \
hevcdsp->sao_edge_filter[2] = FUNC(sao_edge_filter_2, depth); \
hevcdsp->sao_edge_filter[3] = FUNC(sao_edge_filter_3, depth); \
\
hevcdsp->put_hevc_qpel[0][0] = FUNC(put_hevc_qpel_pixels, depth); \
hevcdsp->put_hevc_qpel[0][1] = FUNC(put_hevc_qpel_h1, depth); \
hevcdsp->put_hevc_qpel[0][2] = FUNC(put_hevc_qpel_h2, depth); \
hevcdsp->put_hevc_qpel[0][3] = FUNC(put_hevc_qpel_h3, depth); \
hevcdsp->put_hevc_qpel[1][0] = FUNC(put_hevc_qpel_v1, depth); \
hevcdsp->put_hevc_qpel[1][1] = FUNC(put_hevc_qpel_h1v1, depth); \
hevcdsp->put_hevc_qpel[1][2] = FUNC(put_hevc_qpel_h2v1, depth); \
hevcdsp->put_hevc_qpel[1][3] = FUNC(put_hevc_qpel_h3v1, depth); \
hevcdsp->put_hevc_qpel[2][0] = FUNC(put_hevc_qpel_v2, depth); \
hevcdsp->put_hevc_qpel[2][1] = FUNC(put_hevc_qpel_h1v2, depth); \
hevcdsp->put_hevc_qpel[2][2] = FUNC(put_hevc_qpel_h2v2, depth); \
hevcdsp->put_hevc_qpel[2][3] = FUNC(put_hevc_qpel_h3v2, depth); \
hevcdsp->put_hevc_qpel[3][0] = FUNC(put_hevc_qpel_v3, depth); \
hevcdsp->put_hevc_qpel[3][1] = FUNC(put_hevc_qpel_h1v3, depth); \
hevcdsp->put_hevc_qpel[3][2] = FUNC(put_hevc_qpel_h2v3, depth); \
hevcdsp->put_hevc_qpel[3][3] = FUNC(put_hevc_qpel_h3v3, depth); \
\
hevcdsp->put_hevc_epel[0][0] = FUNC(put_hevc_epel_pixels, depth); \
hevcdsp->put_hevc_epel[0][1] = FUNC(put_hevc_epel_h, depth); \
hevcdsp->put_hevc_epel[1][0] = FUNC(put_hevc_epel_v, depth); \
hevcdsp->put_hevc_epel[1][1] = FUNC(put_hevc_epel_hv, depth); \
\
\
hevcdsp->put_unweighted_pred = FUNC(put_unweighted_pred, depth); \
hevcdsp->put_weighted_pred_avg = FUNC(put_weighted_pred_avg, depth); \
\
hevcdsp->weighted_pred = FUNC(weighted_pred, depth); \
hevcdsp->weighted_pred_avg = FUNC(weighted_pred_avg, depth); \
hevcdsp->hevc_h_loop_filter_luma = FUNC(hevc_h_loop_filter_luma, depth); \
hevcdsp->hevc_v_loop_filter_luma = FUNC(hevc_v_loop_filter_luma, depth); \
hevcdsp->hevc_h_loop_filter_chroma = FUNC(hevc_h_loop_filter_chroma, depth); \
hevcdsp->hevc_v_loop_filter_chroma = FUNC(hevc_v_loop_filter_chroma, depth); \
hevcdsp->hevc_h_loop_filter_luma_c = FUNC(hevc_h_loop_filter_luma, depth); \
hevcdsp->hevc_v_loop_filter_luma_c = FUNC(hevc_v_loop_filter_luma, depth); \
hevcdsp->put_unweighted_pred = FUNC(put_unweighted_pred, depth); \
hevcdsp->put_weighted_pred_avg = FUNC(put_weighted_pred_avg, depth); \
\
hevcdsp->weighted_pred = FUNC(weighted_pred, depth); \
hevcdsp->weighted_pred_avg = FUNC(weighted_pred_avg, depth); \
hevcdsp->hevc_h_loop_filter_luma = FUNC(hevc_h_loop_filter_luma, depth); \
hevcdsp->hevc_v_loop_filter_luma = FUNC(hevc_v_loop_filter_luma, depth); \
hevcdsp->hevc_h_loop_filter_chroma = FUNC(hevc_h_loop_filter_chroma, depth); \
hevcdsp->hevc_v_loop_filter_chroma = FUNC(hevc_v_loop_filter_chroma, depth); \
hevcdsp->hevc_h_loop_filter_luma_c = FUNC(hevc_h_loop_filter_luma, depth); \
hevcdsp->hevc_v_loop_filter_luma_c = FUNC(hevc_v_loop_filter_luma, depth); \
hevcdsp->hevc_h_loop_filter_chroma_c = FUNC(hevc_h_loop_filter_chroma, depth); \
hevcdsp->hevc_v_loop_filter_chroma_c = FUNC(hevc_v_loop_filter_chroma, depth);


switch (bit_depth) {
case 9:
HEVC_DSP(9);


+ 222
- 210
libavcodec/hevcdsp_template.c View File

@@ -23,9 +23,10 @@
#include "get_bits.h"
#include "bit_depth_template.c"
#include "hevcdsp.h"

#include "hevc.h"

#define SET(dst, x) (dst) = (x)
#define SET(dst, x) (dst) = (x)
#define SCALE(dst, x) (dst) = av_clip_int16(((x) + add) >> shift)
#define ADD_AND_SCALE(dst, x) (dst) = av_clip_pixel((dst) + av_clip_int16(((x) + add) >> shift))

@@ -46,7 +47,7 @@ static void FUNC(put_pcm)(uint8_t *_dst, ptrdiff_t _stride, int size,
static void FUNC(transquant_bypass4x4)(uint8_t *_dst, int16_t *coeffs, ptrdiff_t _stride)
{
int x, y;
pixel *dst = (pixel*)_dst;
pixel *dst = (pixel *)_dst;
ptrdiff_t stride = _stride / sizeof(pixel);

for (y = 0; y < 4; y++) {
@@ -163,8 +164,8 @@ static void FUNC(transform_4x4_luma_add)(uint8_t *_dst, int16_t *coeffs, ptrdiff
#undef TR_4x4_LUMA
}

#define TR_4(dst, src, dstep, sstep, assign) \
do { \
#define TR_4(dst, src, dstep, sstep, assign) \
do { \
const int e0 = transform[8*0][0] * src[0*sstep] + \
transform[8*2][0] * src[2*sstep]; \
const int e1 = transform[8*0][1] * src[0*sstep] + \
@@ -207,48 +208,50 @@ static void FUNC(transform_4x4_add)(uint8_t *_dst, int16_t *coeffs, ptrdiff_t _s

#define TR_8(dst, src, dstep, sstep, assign) \
do { \
int i, j; \
int e_8[4]; \
int o_8[4] = { 0 }; \
for (i = 0; i < 4; i++) \
for (j = 1; j < 8; j += 2) \
o_8[i] += transform[4*j][i] * src[j*sstep]; \
TR_4(e_8, src, 1, 2*sstep, SET); \
\
for (i = 0; i < 4; i++) { \
assign(dst[i*dstep], e_8[i] + o_8[i]); \
assign(dst[(7-i)*dstep], e_8[i] - o_8[i]); \
} \
int i, j; \
int e_8[4]; \
int o_8[4] = { 0 }; \
for (i = 0; i < 4; i++) \
for (j = 1; j < 8; j += 2) \
o_8[i] += transform[4 * j][i] * src[j * sstep]; \
TR_4(e_8, src, 1, 2 * sstep, SET); \
\
for (i = 0; i < 4; i++) { \
assign(dst[i * dstep], e_8[i] + o_8[i]); \
assign(dst[(7 - i) * dstep], e_8[i] - o_8[i]); \
} \
} while (0)
#define TR_16(dst, src, dstep, sstep, assign) \
do { \
int i, j; \
int e_16[8]; \
int o_16[8] = { 0 }; \
for (i = 0; i < 8; i++) \
for (j = 1; j < 16; j += 2) \
o_16[i] += transform[2*j][i] * src[j*sstep]; \
TR_8(e_16, src, 1, 2*sstep, SET); \
\
for (i = 0; i < 8; i++) { \
assign(dst[i*dstep], e_16[i] + o_16[i]); \
assign(dst[(15-i)*dstep], e_16[i] - o_16[i]); \
} \

#define TR_16(dst, src, dstep, sstep, assign) \
do { \
int i, j; \
int e_16[8]; \
int o_16[8] = { 0 }; \
for (i = 0; i < 8; i++) \
for (j = 1; j < 16; j += 2) \
o_16[i] += transform[2 * j][i] * src[j * sstep]; \
TR_8(e_16, src, 1, 2 * sstep, SET); \
\
for (i = 0; i < 8; i++) { \
assign(dst[i * dstep], e_16[i] + o_16[i]); \
assign(dst[(15 - i) * dstep], e_16[i] - o_16[i]); \
} \
} while (0)
#define TR_32(dst, src, dstep, sstep, assign) \
do { \
int i, j; \
int e_32[16]; \
int o_32[16] = { 0 }; \
for (i = 0; i < 16; i++) \
for (j = 1; j < 32; j += 2) \
o_32[i] += transform[j][i] * src[j*sstep]; \
TR_16(e_32, src, 1, 2*sstep, SET); \
\
for (i = 0; i < 16; i++) { \
assign(dst[i*dstep], e_32[i] + o_32[i]); \
assign(dst[(31-i)*dstep], e_32[i] - o_32[i]); \
} \

#define TR_32(dst, src, dstep, sstep, assign) \
do { \
int i, j; \
int e_32[16]; \
int o_32[16] = { 0 }; \
for (i = 0; i < 16; i++) \
for (j = 1; j < 32; j += 2) \
o_32[i] += transform[j][i] * src[j * sstep]; \
TR_16(e_32, src, 1, 2 * sstep, SET); \
\
for (i = 0; i < 16; i++) { \
assign(dst[i * dstep], e_32[i] + o_32[i]); \
assign(dst[(31 - i) * dstep], e_32[i] - o_32[i]); \
} \
} while (0)

#define TR_8_1(dst, src) TR_8(dst, src, 8, 8, SCALE)
@@ -262,7 +265,7 @@ static void FUNC(transform_4x4_add)(uint8_t *_dst, int16_t *coeffs, ptrdiff_t _s
static void FUNC(transform_8x8_add)(uint8_t *_dst, int16_t *coeffs, ptrdiff_t _stride)
{
int i;
pixel *dst = (pixel*)_dst;
pixel *dst = (pixel *)_dst;
ptrdiff_t stride = _stride / sizeof(pixel);
int shift = 7;
int add = 1 << (shift - 1);
@@ -344,7 +347,7 @@ static void FUNC(transform_32x32_add)(uint8_t *_dst, int16_t *coeffs, ptrdiff_t

static void FUNC(sao_band_filter)(uint8_t *_dst, uint8_t *_src,
ptrdiff_t _stride, SAOParams *sao,
int *borders, int width, int height,
int *borders, int width, int height,
int c_idx, int class)
{
pixel *dst = (pixel*)_dst;
@@ -402,7 +405,8 @@ static void FUNC(sao_band_filter_0)(uint8_t *dst, uint8_t *src,
int *borders, int width, int height,
int c_idx)
{
FUNC(sao_band_filter)(dst, src, stride, sao, borders, width, height, c_idx, 0);
FUNC(sao_band_filter)(dst, src, stride, sao, borders,
width, height, c_idx, 0);
}

static void FUNC(sao_band_filter_1)(uint8_t *dst, uint8_t *src,
@@ -410,7 +414,8 @@ static void FUNC(sao_band_filter_1)(uint8_t *dst, uint8_t *src,
int *borders, int width, int height,
int c_idx)
{
FUNC(sao_band_filter)(dst, src, stride, sao, borders, width, height, c_idx, 1);
FUNC(sao_band_filter)(dst, src, stride, sao, borders,
width, height, c_idx, 1);
}

static void FUNC(sao_band_filter_2)(uint8_t *dst, uint8_t *src,
@@ -444,10 +449,10 @@ static void FUNC(sao_edge_filter_0)(uint8_t *_dst, uint8_t *_src,
int sao_eo_class = sao->eo_class[c_idx];

static const int8_t pos[4][2][2] = {
{{ -1, 0}, { 1, 0}}, // horizontal
{{ 0, -1}, { 0, 1}}, // vertical
{{ -1, -1}, { 1, 1}}, // 45 degree
{{ 1, -1}, {-1, 1}}, // 135 degree
{ { -1, 0 }, { 1, 0 } }, // horizontal
{ { 0, -1 }, { 0, 1 } }, // vertical
{ { -1, -1 }, { 1, 1 } }, // 45 degree
{ { 1, -1 }, { -1, 1 } }, // 135 degree
};
static const uint8_t edge_idx[] = { 1, 2, 0, 3, 4 };

@@ -469,7 +474,7 @@ static void FUNC(sao_edge_filter_0)(uint8_t *_dst, uint8_t *_src,
int y_stride = 0;
for (y = 0; y < height; y++) {
dst[y_stride] = av_clip_pixel(src[y_stride] + offset_val);
y_stride += stride;
y_stride += stride;
}
init_x = 1;
}
@@ -478,21 +483,20 @@ static void FUNC(sao_edge_filter_0)(uint8_t *_dst, uint8_t *_src,
int x_stride = width - 1;
for (x = 0; x < height; x++) {
dst[x_stride] = av_clip_pixel(src[x_stride] + offset_val);
x_stride += stride;
x_stride += stride;
}
width --;
width--;
}

}
if (sao_eo_class != SAO_EO_HORIZ ) {
if (borders[1]){
if (sao_eo_class != SAO_EO_HORIZ) {
if (borders[1]) {
int offset_val = sao_offset_val[0];
for (x = init_x; x < width; x++) {
dst[x] = av_clip_pixel(src[x] + offset_val);
}
init_y = 1;
}
if (borders[3]){
if (borders[3]) {
int offset_val = sao_offset_val[0];
int y_stride = stride * (height - 1);
for (x = init_x; x < width; x++) {
@@ -502,11 +506,11 @@ static void FUNC(sao_edge_filter_0)(uint8_t *_dst, uint8_t *_src,
}
}
{
int y_stride = init_y * stride;
int pos_0_0 = pos[sao_eo_class][0][0];
int pos_0_1 = pos[sao_eo_class][0][1];
int pos_1_0 = pos[sao_eo_class][1][0];
int pos_1_1 = pos[sao_eo_class][1][1];
int y_stride = init_y * stride;
int pos_0_0 = pos[sao_eo_class][0][0];
int pos_0_1 = pos[sao_eo_class][0][1];
int pos_1_0 = pos[sao_eo_class][1][0];
int pos_1_1 = pos[sao_eo_class][1][1];

int y_stride_0_1 = (init_y + pos_0_1) * stride;
int y_stride_1_1 = (init_y + pos_1_1) * stride;
@@ -542,22 +546,22 @@ static void FUNC(sao_edge_filter_0)(uint8_t *_dst, uint8_t *_src,
static void FUNC(sao_edge_filter_1)(uint8_t *_dst, uint8_t *_src,
ptrdiff_t _stride, SAOParams *sao,
int *borders, int _width, int _height,
int c_idx,
uint8_t vert_edge, uint8_t horiz_edge, uint8_t diag_edge)
int c_idx, uint8_t vert_edge,
uint8_t horiz_edge, uint8_t diag_edge)
{
int x, y;
pixel *dst = (pixel*)_dst;
pixel *src = (pixel*)_src;
pixel *dst = (pixel *)_dst;
pixel *src = (pixel *)_src;
ptrdiff_t stride = _stride / sizeof(pixel);
int chroma = !!c_idx;
int *sao_offset_val = sao->offset_val[c_idx];
int sao_eo_class = sao->eo_class[c_idx];
int sao_eo_class = sao->eo_class[c_idx];

static const int8_t pos[4][2][2] = {
{{ -1, 0}, { 1, 0 }}, // horizontal
{{ 0, -1}, { 0, 1 }}, // vertical
{{ -1, -1}, { 1, 1 }}, // 45 degree
{{ 1, -1}, {-1, 1 }}, // 135 degree
{ { -1, 0 }, { 1, 0 } }, // horizontal
{ { 0, -1 }, { 0, 1 } }, // vertical
{ { -1, -1 }, { 1, 1 } }, // 45 degree
{ { 1, -1 }, { -1, 1 } }, // 135 degree
};
static const uint8_t edge_idx[] = { 1, 2, 0, 3, 4 };

@@ -579,7 +583,7 @@ static void FUNC(sao_edge_filter_1)(uint8_t *_dst, uint8_t *_src,
int y_stride = 0;
for (y = 0; y < height; y++) {
dst[y_stride] = av_clip_pixel(src[y_stride] + offset_val);
y_stride += stride;
y_stride += stride;
}
init_x = 1;
}
@@ -588,18 +592,17 @@ static void FUNC(sao_edge_filter_1)(uint8_t *_dst, uint8_t *_src,
int x_stride = width - 1;
for (x = 0; x < height; x++) {
dst[x_stride] = av_clip_pixel(src[x_stride] + offset_val);
x_stride += stride;
x_stride += stride;
}
width--;
}

}
{
int y_stride = init_y * stride;
int pos_0_0 = pos[sao_eo_class][0][0];
int pos_0_1 = pos[sao_eo_class][0][1];
int pos_1_0 = pos[sao_eo_class][1][0];
int pos_1_1 = pos[sao_eo_class][1][1];
int y_stride = init_y * stride;
int pos_0_0 = pos[sao_eo_class][0][0];
int pos_0_1 = pos[sao_eo_class][0][1];
int pos_1_0 = pos[sao_eo_class][1][0];
int pos_1_1 = pos[sao_eo_class][1][1];

int y_stride_0_1 = (init_y + pos_0_1) * stride;
int y_stride_1_1 = (init_y + pos_1_1) * stride;
@@ -635,22 +638,22 @@ static void FUNC(sao_edge_filter_1)(uint8_t *_dst, uint8_t *_src,
static void FUNC(sao_edge_filter_2)(uint8_t *_dst, uint8_t *_src,
ptrdiff_t _stride, SAOParams *sao,
int *borders, int _width, int _height,
int c_idx,
uint8_t vert_edge, uint8_t horiz_edge, uint8_t diag_edge)
int c_idx, uint8_t vert_edge,
uint8_t horiz_edge, uint8_t diag_edge)
{
int x, y;
pixel *dst = (pixel*)_dst;
pixel *src = (pixel*)_src;
pixel *dst = (pixel *)_dst;
pixel *src = (pixel *)_src;
ptrdiff_t stride = _stride / sizeof(pixel);
int chroma = !!c_idx;
int *sao_offset_val = sao->offset_val[c_idx];
int sao_eo_class = sao->eo_class[c_idx];
int sao_eo_class = sao->eo_class[c_idx];

static const int8_t pos[4][2][2] = {
{{ -1, 0}, { 1, 0}}, // horizontal
{{ 0, -1}, { 0, 1}}, // vertical
{{ -1, -1}, { 1, 1}}, // 45 degree
{{ 1, -1}, {-1, 1}}, // 135 degree
{ { -1, 0 }, { 1, 0 } }, // horizontal
{ { 0, -1 }, { 0, 1 } }, // vertical
{ { -1, -1 }, { 1, 1 } }, // 45 degree
{ { 1, -1 }, { -1, 1 } }, // 135 degree
};
static const uint8_t edge_idx[] = { 1, 2, 0, 3, 4 };

@@ -667,14 +670,14 @@ static void FUNC(sao_edge_filter_2)(uint8_t *_dst, uint8_t *_src,
src = src + (init_y * stride + init_x);
init_y = init_x = 0;
if (sao_eo_class != SAO_EO_HORIZ) {
if (borders[1]){
if (borders[1]) {
int offset_val = sao_offset_val[0];
for (x = init_x; x < width; x++) {
dst[x] = av_clip_pixel(src[x] + offset_val);
}
init_y = 1;
}
if (borders[3]){
if (borders[3]) {
int offset_val = sao_offset_val[0];
int y_stride = stride * (height - 1);
for (x = init_x; x < width; x++) {
@@ -684,11 +687,11 @@ static void FUNC(sao_edge_filter_2)(uint8_t *_dst, uint8_t *_src,
}
}
{
int y_stride = init_y * stride;
int pos_0_0 = pos[sao_eo_class][0][0];
int pos_0_1 = pos[sao_eo_class][0][1];
int pos_1_0 = pos[sao_eo_class][1][0];
int pos_1_1 = pos[sao_eo_class][1][1];
int y_stride = init_y * stride;
int pos_0_0 = pos[sao_eo_class][0][0];
int pos_0_1 = pos[sao_eo_class][0][1];
int pos_1_0 = pos[sao_eo_class][1][0];
int pos_1_1 = pos[sao_eo_class][1][1];

int y_stride_0_1 = (init_y + pos_0_1) * stride;
int y_stride_1_1 = (init_y + pos_1_1) * stride;
@@ -723,22 +726,22 @@ static void FUNC(sao_edge_filter_2)(uint8_t *_dst, uint8_t *_src,
static void FUNC(sao_edge_filter_3)(uint8_t *_dst, uint8_t *_src,
ptrdiff_t _stride, SAOParams *sao,
int *borders, int _width, int _height,
int c_idx,
uint8_t vert_edge, uint8_t horiz_edge, uint8_t diag_edge)
int c_idx, uint8_t vert_edge,
uint8_t horiz_edge, uint8_t diag_edge)
{
int x, y;
pixel *dst = (pixel*)_dst;
pixel *src = (pixel*)_src;
pixel *dst = (pixel *)_dst;
pixel *src = (pixel *)_src;
ptrdiff_t stride = _stride / sizeof(pixel);
int chroma = !!c_idx;
int *sao_offset_val = sao->offset_val[c_idx];
int sao_eo_class = sao->eo_class[c_idx];

static const int8_t pos[4][2][2] = {
{{ -1, 0}, { 1, 0}}, // horizontal
{{ 0, -1}, { 0, 1}}, // vertical
{{ -1, -1}, { 1, 1}}, // 45 degree
{{ 1, -1}, {-1, 1}}, // 135 degree
{ { -1, 0 }, { 1, 0 } }, // horizontal
{ { 0, -1 }, { 0, 1 } }, // vertical
{ { -1, -1 }, { 1, 1 } }, // 45 degree
{ { 1, -1 }, { -1, 1 } }, // 135 degree
};
static const uint8_t edge_idx[] = { 1, 2, 0, 3, 4 };

@@ -752,16 +755,16 @@ static void FUNC(sao_edge_filter_3)(uint8_t *_dst, uint8_t *_src,
height = (4 >> chroma) + 2;


dst = dst + (init_y * stride + init_x);
src = src + (init_y * stride + init_x);
dst = dst + (init_y * stride + init_x);
src = src + (init_y * stride + init_x);
init_y = init_x = 0;

{
int y_stride = init_y * stride;
int pos_0_0 = pos[sao_eo_class][0][0];
int pos_0_1 = pos[sao_eo_class][0][1];
int pos_1_0 = pos[sao_eo_class][1][0];
int pos_1_1 = pos[sao_eo_class][1][1];
int y_stride = init_y * stride;
int pos_0_0 = pos[sao_eo_class][0][0];
int pos_0_1 = pos[sao_eo_class][0][1];
int pos_1_0 = pos[sao_eo_class][1][0];
int pos_1_1 = pos[sao_eo_class][1][1];

int y_stride_0_1 = (init_y + pos_0_1) * stride;
int y_stride_1_1 = (init_y + pos_1_1) * stride;
@@ -815,7 +818,7 @@ static void FUNC(put_hevc_qpel_pixels)(int16_t *dst, ptrdiff_t dststride,
int width, int height, int16_t* mcbuffer)
{
int x, y;
pixel *src = (pixel*)_src;
pixel *src = (pixel *)_src;
ptrdiff_t srcstride = _srcstride / sizeof(pixel);

for (y = 0; y < height; y++) {
@@ -826,15 +829,17 @@ static void FUNC(put_hevc_qpel_pixels)(int16_t *dst, ptrdiff_t dststride,
}
}

#define QPEL_FILTER_1(src, stride) \
(-src[x-3*stride] + 4*src[x-2*stride] - 10*src[x-stride] + 58*src[x] + \
17*src[x+stride] - 5*src[x+2*stride] + 1*src[x+3*stride])
#define QPEL_FILTER_2(src, stride) \
(-src[x-3*stride] + 4*src[x-2*stride] - 11*src[x-stride] + 40*src[x] + \
40*src[x+stride] - 11*src[x+2*stride] + 4*src[x+3*stride] - src[x+4*stride])
#define QPEL_FILTER_3(src, stride) \
(src[x-2*stride] - 5*src[x-stride] + 17*src[x] + 58*src[x+stride] \
- 10*src[x+2*stride] + 4*src[x+3*stride] - src[x+4*stride])
#define QPEL_FILTER_1(src, stride) \
(-src[x - 3 * stride] + 4 * src[x - 2 * stride] - 10 * src[x - stride] + 58 * src[x] + \
17 * src[x + stride] - 5 * src[x + 2 * stride] + 1 * src[x + 3 * stride])

#define QPEL_FILTER_2(src, stride) \
(-src[x - 3 * stride] + 4 * src[x - 2 * stride] - 11 * src[x - stride] + 40 * src[x] + \
40 * src[x + stride] - 11 * src[x + 2 * stride] + 4 * src[x + 3 * stride] - src[x + 4 * stride])

#define QPEL_FILTER_3(src, stride) \
(src[x - 2 * stride] - 5 * src[x - stride] + 17 * src[x] + 58 * src[x + stride] \
- 10 * src[x + 2 * stride] + 4 * src[x + 3 * stride] - src[x + 4 * stride])


#define PUT_HEVC_QPEL_H(H) \
@@ -849,7 +854,7 @@ static void FUNC(put_hevc_qpel_h ## H)(int16_t *dst, ptrdiff_t dststride, \
\
for (y = 0; y < height; y++) { \
for (x = 0; x < width; x++) \
dst[x] = QPEL_FILTER_ ## H (src, 1) >> (BIT_DEPTH - 8); \
dst[x] = QPEL_FILTER_ ## H(src, 1) >> (BIT_DEPTH - 8); \
src += srcstride; \
dst += dststride; \
} \
@@ -867,42 +872,42 @@ static void FUNC(put_hevc_qpel_v ## V)(int16_t *dst, ptrdiff_t dststride, \
\
for (y = 0; y < height; y++) { \
for (x = 0; x < width; x++) \
dst[x] = QPEL_FILTER_ ## V (src, srcstride) >> (BIT_DEPTH - 8); \
dst[x] = QPEL_FILTER_ ## V(src, srcstride) >> (BIT_DEPTH - 8); \
src += srcstride; \
dst += dststride; \
} \
}

#define PUT_HEVC_QPEL_HV(H, V) \
#define PUT_HEVC_QPEL_HV(H, V) \
static void FUNC(put_hevc_qpel_h ## H ## v ## V)(int16_t *dst, ptrdiff_t dststride, \
uint8_t *_src, ptrdiff_t _srcstride,\
int width, int height, \
int16_t* mcbuffer) \
{ \
int x, y; \
pixel *src = (pixel*)_src; \
ptrdiff_t srcstride = _srcstride / sizeof(pixel); \
\
int16_t tmp_array[(MAX_PB_SIZE + 7)*MAX_PB_SIZE]; \
int16_t *tmp = tmp_array; \
\
src -= ff_hevc_qpel_extra_before[V] * srcstride; \
\
for (y = 0; y < height + ff_hevc_qpel_extra[V]; y++) { \
for (x = 0; x < width; x++) \
tmp[x] = QPEL_FILTER_ ## H (src, 1) >> (BIT_DEPTH - 8); \
src += srcstride; \
tmp += MAX_PB_SIZE; \
} \
\
tmp = tmp_array + ff_hevc_qpel_extra_before[V] * MAX_PB_SIZE; \
\
for (y = 0; y < height; y++) { \
for (x = 0; x < width; x++) \
dst[x] = QPEL_FILTER_ ## V (tmp, MAX_PB_SIZE) >> 6; \
tmp += MAX_PB_SIZE; \
dst += dststride; \
} \
int width, int height, \
int16_t* mcbuffer) \
{ \
int x, y; \
pixel *src = (pixel*)_src; \
ptrdiff_t srcstride = _srcstride / sizeof(pixel); \
\
int16_t tmp_array[(MAX_PB_SIZE + 7) * MAX_PB_SIZE]; \
int16_t *tmp = tmp_array; \
\
src -= ff_hevc_qpel_extra_before[V] * srcstride; \
\
for (y = 0; y < height + ff_hevc_qpel_extra[V]; y++) { \
for (x = 0; x < width; x++) \
tmp[x] = QPEL_FILTER_ ## H(src, 1) >> (BIT_DEPTH - 8); \
src += srcstride; \
tmp += MAX_PB_SIZE; \
} \
\
tmp = tmp_array + ff_hevc_qpel_extra_before[V] * MAX_PB_SIZE; \
\
for (y = 0; y < height; y++) { \
for (x = 0; x < width; x++) \
dst[x] = QPEL_FILTER_ ## V(tmp, MAX_PB_SIZE) >> 6; \
tmp += MAX_PB_SIZE; \
dst += dststride; \
} \
}

PUT_HEVC_QPEL_H(1)
@@ -927,7 +932,7 @@ static void FUNC(put_hevc_epel_pixels)(int16_t *dst, ptrdiff_t dststride,
int16_t* mcbuffer)
{
int x, y;
pixel *src = (pixel*)_src;
pixel *src = (pixel *)_src;
ptrdiff_t srcstride = _srcstride / sizeof(pixel);

for (y = 0; y < height; y++) {
@@ -939,8 +944,8 @@ static void FUNC(put_hevc_epel_pixels)(int16_t *dst, ptrdiff_t dststride,
}
}

#define EPEL_FILTER(src, stride) \
(filter_0*src[x-stride] + filter_1*src[x] + filter_2*src[x+stride] + filter_3*src[x+2*stride])
#define EPEL_FILTER(src, stride) \
(filter_0 * src[x - stride] + filter_1 * src[x] + filter_2 * src[x + stride] + filter_3 * src[x + 2 * stride])

static void FUNC(put_hevc_epel_h)(int16_t *dst, ptrdiff_t dststride,
uint8_t *_src, ptrdiff_t _srcstride,
@@ -948,8 +953,8 @@ static void FUNC(put_hevc_epel_h)(int16_t *dst, ptrdiff_t dststride,
int16_t* mcbuffer)
{
int x, y;
pixel *src = (pixel*)_src;
ptrdiff_t srcstride = _srcstride / sizeof(pixel);
pixel *src = (pixel *)_src;
ptrdiff_t srcstride = _srcstride / sizeof(pixel);
const int8_t *filter = ff_hevc_epel_filters[mx - 1];
int8_t filter_0 = filter[0];
int8_t filter_1 = filter[1];
@@ -970,9 +975,9 @@ static void FUNC(put_hevc_epel_v)(int16_t *dst, ptrdiff_t dststride,
int16_t* mcbuffer)
{
int x, y;
pixel *src = (pixel*)_src;
pixel *src = (pixel *)_src;
ptrdiff_t srcstride = _srcstride / sizeof(pixel);
const int8_t *filter = ff_hevc_epel_filters[my-1];
const int8_t *filter = ff_hevc_epel_filters[my - 1];
int8_t filter_0 = filter[0];
int8_t filter_1 = filter[1];
int8_t filter_2 = filter[2];
@@ -993,15 +998,15 @@ static void FUNC(put_hevc_epel_hv)(int16_t *dst, ptrdiff_t dststride,
int16_t* mcbuffer)
{
int x, y;
pixel *src = (pixel*)_src;
pixel *src = (pixel *)_src;
ptrdiff_t srcstride = _srcstride / sizeof(pixel);
const int8_t *filter_h = ff_hevc_epel_filters[mx-1];
const int8_t *filter_v = ff_hevc_epel_filters[my-1];
const int8_t *filter_h = ff_hevc_epel_filters[mx - 1];
const int8_t *filter_v = ff_hevc_epel_filters[my - 1];
int8_t filter_0 = filter_h[0];
int8_t filter_1 = filter_h[1];
int8_t filter_2 = filter_h[2];
int8_t filter_3 = filter_h[3];
int16_t tmp_array[(MAX_PB_SIZE + 3)*MAX_PB_SIZE];
int16_t tmp_array[(MAX_PB_SIZE + 3) * MAX_PB_SIZE];
int16_t *tmp = tmp_array;

src -= EPEL_EXTRA_BEFORE * srcstride;
@@ -1014,7 +1019,7 @@ static void FUNC(put_hevc_epel_hv)(int16_t *dst, ptrdiff_t dststride,
tmp += MAX_PB_SIZE;
}

tmp = tmp_array + EPEL_EXTRA_BEFORE * MAX_PB_SIZE;
tmp = tmp_array + EPEL_EXTRA_BEFORE * MAX_PB_SIZE;
filter_0 = filter_v[0];
filter_1 = filter_v[1];
filter_2 = filter_v[2];
@@ -1033,7 +1038,7 @@ static void FUNC(put_unweighted_pred)(uint8_t *_dst, ptrdiff_t _dststride,
int width, int height)
{
int x, y;
pixel *dst = (pixel*)_dst;
pixel *dst = (pixel *)_dst;
ptrdiff_t dststride = _dststride / sizeof(pixel);

int shift = 14 - BIT_DEPTH;
@@ -1057,7 +1062,7 @@ static void FUNC(put_weighted_pred_avg)(uint8_t *_dst, ptrdiff_t _dststride,
int width, int height)
{
int x, y;
pixel *dst = (pixel*)_dst;
pixel *dst = (pixel *)_dst;
ptrdiff_t dststride = _dststride / sizeof(pixel);

int shift = 14 + 1 - BIT_DEPTH;
@@ -1088,14 +1093,14 @@ static void FUNC(weighted_pred)(uint8_t denom, int16_t wlxFlag, int16_t olxFlag,
int ox;
int x , y;
int offset;
pixel *dst = (pixel*)_dst;
pixel *dst = (pixel *)_dst;
ptrdiff_t dststride = _dststride / sizeof(pixel);

shift = 14 - BIT_DEPTH;
log2Wd = denom + shift;
offset = 1 << (log2Wd - 1);
wx = wlxFlag;
ox = olxFlag * (1 << (BIT_DEPTH - 8));
wx = wlxFlag;
ox = olxFlag * (1 << (BIT_DEPTH - 8));

for (y = 0; y < height; y++) {
for (x = 0; x < width; x++) {
@@ -1105,15 +1110,17 @@ static void FUNC(weighted_pred)(uint8_t denom, int16_t wlxFlag, int16_t olxFlag,
dst[x] = av_clip_pixel(src[x] * wx + ox);
}
}
dst += dststride;
src += srcstride;
dst += dststride;
src += srcstride;
}
}

static void FUNC(weighted_pred_avg)(uint8_t denom, int16_t wl0Flag, int16_t wl1Flag,
static void FUNC(weighted_pred_avg)(uint8_t denom,
int16_t wl0Flag, int16_t wl1Flag,
int16_t ol0Flag, int16_t ol1Flag,
uint8_t *_dst, ptrdiff_t _dststride,
int16_t *src1, int16_t *src2, ptrdiff_t srcstride,
int16_t *src1, int16_t *src2,
ptrdiff_t srcstride,
int width, int height)
{
int shift;
@@ -1123,13 +1130,13 @@ static void FUNC(weighted_pred_avg)(uint8_t denom, int16_t wl0Flag, int16_t wl1F
int o0;
int o1;
int x , y;
pixel *dst = (pixel*)_dst;
pixel *dst = (pixel *)_dst;
ptrdiff_t dststride = _dststride / sizeof(pixel);

shift = 14 - BIT_DEPTH;
log2Wd = denom + shift;
w0 = wl0Flag;
w1 = wl1Flag;
w0 = wl0Flag;
w1 = wl1Flag;
o0 = (ol0Flag) * (1 << (BIT_DEPTH - 8));
o1 = (ol1Flag) * (1 << (BIT_DEPTH - 8));

@@ -1145,43 +1152,44 @@ static void FUNC(weighted_pred_avg)(uint8_t denom, int16_t wl0Flag, int16_t wl1F
}

// line zero
#define P3 pix[-4*xstride]
#define P2 pix[-3*xstride]
#define P1 pix[-2*xstride]
#define P3 pix[-4 * xstride]
#define P2 pix[-3 * xstride]
#define P1 pix[-2 * xstride]
#define P0 pix[-xstride]
#define Q0 pix[0]
#define Q1 pix[xstride]
#define Q2 pix[2*xstride]
#define Q3 pix[3*xstride]
#define Q2 pix[2 * xstride]
#define Q3 pix[3 * xstride]

// line three. used only for deblocking decision
#define TP3 pix[-4*xstride+3*ystride]
#define TP2 pix[-3*xstride+3*ystride]
#define TP1 pix[-2*xstride+3*ystride]
#define TP0 pix[-xstride+3*ystride]
#define TQ0 pix[3*ystride]
#define TQ1 pix[xstride+3*ystride]
#define TQ2 pix[2*xstride+3*ystride]
#define TQ3 pix[3*xstride+3*ystride]

static void FUNC(hevc_loop_filter_luma)(uint8_t *_pix, ptrdiff_t _xstride,
ptrdiff_t _ystride, int *_beta, int *_tc,
#define TP3 pix[-4 * xstride + 3 * ystride]
#define TP2 pix[-3 * xstride + 3 * ystride]
#define TP1 pix[-2 * xstride + 3 * ystride]
#define TP0 pix[-xstride+3 * ystride]
#define TQ0 pix[3 * ystride]
#define TQ1 pix[xstride+3 * ystride]
#define TQ2 pix[2 * xstride + 3 * ystride]
#define TQ3 pix[3 * xstride + 3 * ystride]

static void FUNC(hevc_loop_filter_luma)(uint8_t *_pix,
ptrdiff_t _xstride, ptrdiff_t _ystride,
int *_beta, int *_tc,
uint8_t *_no_p, uint8_t *_no_q)
{
int d, j;
pixel *pix = (pixel*)_pix;
pixel *pix = (pixel *)_pix;
ptrdiff_t xstride = _xstride / sizeof(pixel);
ptrdiff_t ystride = _ystride / sizeof(pixel);

for (j = 0; j < 2; j++) {
const int dp0 = abs(P2 - 2 * P1 + P0);
const int dq0 = abs(Q2 - 2 * Q1 + Q0);
const int dp0 = abs(P2 - 2 * P1 + P0);
const int dq0 = abs(Q2 - 2 * Q1 + Q0);
const int dp3 = abs(TP2 - 2 * TP1 + TP0);
const int dq3 = abs(TQ2 - 2 * TQ1 + TQ0);
const int d0 = dp0 + dq0;
const int d3 = dp3 + dq3;
int beta = _beta[j] << (BIT_DEPTH - 8);
const int tc = _tc[j] << (BIT_DEPTH - 8);
const int tc = _tc[j] << (BIT_DEPTH - 8);
const int no_p = _no_p[j];
const int no_q = _no_q[j];

@@ -1191,9 +1199,9 @@ static void FUNC(hevc_loop_filter_luma)(uint8_t *_pix, ptrdiff_t _xstride,
} else {
const int beta_3 = beta >> 3;
const int beta_2 = beta >> 2;
const int tc25 = ((tc * 5 + 1) >> 1);
const int tc25 = ((tc * 5 + 1) >> 1);

if (abs( P3 - P0) + abs( Q3 - Q0) < beta_3 && abs( P0 - Q0) < tc25 &&
if (abs(P3 - P0) + abs(Q3 - Q0) < beta_3 && abs(P0 - Q0) < tc25 &&
abs(TP3 - TP0) + abs(TQ3 - TQ0) < beta_3 && abs(TP0 - TQ0) < tc25 &&
(d0 << 1) < beta_2 && (d3 << 1) < beta_2) {
// strong filtering
@@ -1208,14 +1216,14 @@ static void FUNC(hevc_loop_filter_luma)(uint8_t *_pix, ptrdiff_t _xstride,
const int q2 = Q2;
const int q3 = Q3;
if (!no_p) {
P0 = p0 + av_clip((( p2 + 2*p1 + 2*p0 + 2*q0 + q1 + 4 ) >> 3) - p0, -tc2, tc2);
P1 = p1 + av_clip((( p2 + p1 + p0 + q0 + 2 ) >> 2) - p1, -tc2, tc2);
P2 = p2 + av_clip((( 2*p3 + 3*p2 + p1 + p0 + q0 + 4 ) >> 3) - p2, -tc2, tc2);
P0 = p0 + av_clip(((p2 + 2 * p1 + 2 * p0 + 2 * q0 + q1 + 4) >> 3) - p0, -tc2, tc2);
P1 = p1 + av_clip(((p2 + p1 + p0 + q0 + 2) >> 2) - p1, -tc2, tc2);
P2 = p2 + av_clip(((2 * p3 + 3 * p2 + p1 + p0 + q0 + 4) >> 3) - p2, -tc2, tc2);
}
if (!no_q) {
Q0 = q0 + av_clip((( p1 + 2*p0 + 2*q0 + 2*q1 + q2 + 4 ) >> 3) - q0, -tc2, tc2);
Q1 = q1 + av_clip((( p0 + q0 + q1 + q2 + 2 ) >> 2) - q1, -tc2, tc2);
Q2 = q2 + av_clip((( 2*q3 + 3*q2 + q1 + q0 + p0 + 4 ) >> 3) - q2, -tc2, tc2);
Q0 = q0 + av_clip(((p1 + 2 * p0 + 2 * q0 + 2 * q1 + q2 + 4) >> 3) - q0, -tc2, tc2);
Q1 = q1 + av_clip(((p0 + q0 + q1 + q2 + 2) >> 2) - q1, -tc2, tc2);
Q2 = q2 + av_clip(((2 * q3 + 3 * q2 + q1 + q0 + p0 + 4) >> 3) - q2, -tc2, tc2);
}
pix += ystride;
}
@@ -1235,7 +1243,7 @@ static void FUNC(hevc_loop_filter_luma)(uint8_t *_pix, ptrdiff_t _xstride,
const int q0 = Q0;
const int q1 = Q1;
const int q2 = Q2;
int delta0 = (9 * (q0 - p0) - 3 * (q1 - p1) + 8) >> 4;
int delta0 = (9 * (q0 - p0) - 3 * (q1 - p1) + 8) >> 4;
if (abs(delta0) < 10 * tc) {
delta0 = av_clip(delta0, -tc, tc);
if (!no_p)
@@ -1264,7 +1272,7 @@ static void FUNC(hevc_loop_filter_chroma)(uint8_t *_pix, ptrdiff_t _xstride,
{
int d, j;
int no_p, no_q;
pixel *pix = (pixel*)_pix;
pixel *pix = (pixel *)_pix;
ptrdiff_t xstride = _xstride / sizeof(pixel);
ptrdiff_t ystride = _ystride / sizeof(pixel);

@@ -1294,13 +1302,15 @@ static void FUNC(hevc_loop_filter_chroma)(uint8_t *_pix, ptrdiff_t _xstride,
}

static void FUNC(hevc_h_loop_filter_chroma)(uint8_t *pix, ptrdiff_t stride,
int *tc, uint8_t *no_p, uint8_t *no_q)
int *tc, uint8_t *no_p,
uint8_t *no_q)
{
FUNC(hevc_loop_filter_chroma)(pix, stride, sizeof(pixel), tc, no_p, no_q);
}

static void FUNC(hevc_v_loop_filter_chroma)(uint8_t *pix, ptrdiff_t stride,
int *tc, uint8_t *no_p, uint8_t *no_q)
int *tc, uint8_t *no_p,
uint8_t *no_q)
{
FUNC(hevc_loop_filter_chroma)(pix, sizeof(pixel), stride, tc, no_p, no_q);
}
@@ -1309,14 +1319,16 @@ static void FUNC(hevc_h_loop_filter_luma)(uint8_t *pix, ptrdiff_t stride,
int *beta, int *tc, uint8_t *no_p,
uint8_t *no_q)
{
FUNC(hevc_loop_filter_luma)(pix, stride, sizeof(pixel), beta, tc, no_p, no_q);
FUNC(hevc_loop_filter_luma)(pix, stride, sizeof(pixel),
beta, tc, no_p, no_q);
}

static void FUNC(hevc_v_loop_filter_luma)(uint8_t *pix, ptrdiff_t stride,
int *beta, int *tc, uint8_t *no_p,
uint8_t *no_q)
{
FUNC(hevc_loop_filter_luma)(pix, sizeof(pixel), stride, beta, tc, no_p, no_q);
FUNC(hevc_loop_filter_luma)(pix, sizeof(pixel), stride,
beta, tc, no_p, no_q);
}

#undef P3


+ 6
- 6
libavcodec/hevcpred.c View File

@@ -40,16 +40,16 @@ void ff_hevc_pred_init(HEVCPredContext *hpc, int bit_depth)
#undef FUNC
#define FUNC(a, depth) a ## _ ## depth

#define HEVC_PRED(depth) \
hpc->intra_pred = FUNC(intra_pred, depth); \
#define HEVC_PRED(depth) \
hpc->intra_pred = FUNC(intra_pred, depth); \
hpc->pred_planar[0] = FUNC(pred_planar_0, depth); \
hpc->pred_planar[1] = FUNC(pred_planar_1, depth); \
hpc->pred_planar[2] = FUNC(pred_planar_2, depth); \
hpc->pred_planar[3] = FUNC(pred_planar_3, depth); \
hpc->pred_dc = FUNC(pred_dc, depth); \
hpc->pred_angular[0] = FUNC(pred_angular_0, depth);\
hpc->pred_angular[1] = FUNC(pred_angular_1, depth);\
hpc->pred_angular[2] = FUNC(pred_angular_2, depth);\
hpc->pred_dc = FUNC(pred_dc, depth); \
hpc->pred_angular[0] = FUNC(pred_angular_0, depth); \
hpc->pred_angular[1] = FUNC(pred_angular_1, depth); \
hpc->pred_angular[2] = FUNC(pred_angular_2, depth); \
hpc->pred_angular[3] = FUNC(pred_angular_3, depth);

switch (bit_depth) {


+ 72
- 67
libavcodec/hevcpred_template.c View File

@@ -92,9 +92,9 @@ static void FUNC(intra_pred)(HEVCContext *s, int x0, int y0, int log2_size, int
pixel filtered_top_array[2 * MAX_TB_SIZE + 1];

pixel *left = left_array + 1;
pixel *top = top_array + 1;
pixel *top = top_array + 1;
pixel *filtered_left = filtered_left_array + 1;
pixel *filtered_top = filtered_top_array + 1;
pixel *filtered_top = filtered_top_array + 1;

int cand_bottom_left = lc->na.cand_bottom_left && cur_tb_addr > MIN_TB_ADDR_ZS(x_tb - 1, y_tb + size_in_tbs);
int cand_left = lc->na.cand_left;
@@ -109,9 +109,9 @@ static void FUNC(intra_pred)(HEVCContext *s, int x0, int y0, int log2_size, int

if (s->pps->constrained_intra_pred_flag == 1) {
int size_in_luma_pu = PU(size_in_luma);
int on_pu_edge_x = !(x0 & ((1 << s->sps->log2_min_pu_size) - 1));
int on_pu_edge_y = !(y0 & ((1 << s->sps->log2_min_pu_size) - 1));
if(!size_in_luma_pu)
int on_pu_edge_x = !(x0 & ((1 << s->sps->log2_min_pu_size) - 1));
int on_pu_edge_y = !(y0 & ((1 << s->sps->log2_min_pu_size) - 1));
if (!size_in_luma_pu)
size_in_luma_pu++;
if (cand_bottom_left == 1 && on_pu_edge_x) {
int x_left_pu = PU(x0 - 1);
@@ -157,7 +157,8 @@ static void FUNC(intra_pred)(HEVCContext *s, int x0, int y0, int log2_size, int
}
if (cand_bottom_left) {
for (i = size + bottom_left_size; i < (size << 1); i++)
if (IS_INTRA(-1, size + bottom_left_size - 1) || !s->pps->constrained_intra_pred_flag)
if (IS_INTRA(-1, size + bottom_left_size - 1) ||
!s->pps->constrained_intra_pred_flag)
left[i] = POS(-1, size + bottom_left_size - 1);
for (i = size + bottom_left_size - 1; i >= size; i--)
if (IS_INTRA(-1, i) || !s->pps->constrained_intra_pred_flag)
@@ -178,7 +179,8 @@ static void FUNC(intra_pred)(HEVCContext *s, int x0, int y0, int log2_size, int
top[i] = POS(i, -1);
if (cand_up_right) {
for (i = size + top_right_size; i < (size << 1); i++)
if (IS_INTRA(size + top_right_size - 1, -1) || !s->pps->constrained_intra_pred_flag)
if (IS_INTRA(size + top_right_size - 1, -1) ||
!s->pps->constrained_intra_pred_flag)
top[i] = POS(size + top_right_size - 1, -1);
for (i = size + top_right_size - 1; i >= size; i--)
if (IS_INTRA(i, -1) || !s->pps->constrained_intra_pred_flag)
@@ -201,29 +203,32 @@ static void FUNC(intra_pred)(HEVCContext *s, int x0, int y0, int log2_size, int
size : (s->sps->height - y0) >> vshift;
}
if (cand_bottom_left || cand_left || cand_up_left) {
while (j>-1 && !IS_INTRA(-1, j)) j--;
while (j > -1 && !IS_INTRA(-1, j))
j--;
if (!IS_INTRA(-1, j)) {
j = 0;
while(j < size_max_x && !IS_INTRA(j, -1)) j++;
EXTEND_LEFT_CIP(top, j, j+1);
while (j < size_max_x && !IS_INTRA(j, -1))
j++;
EXTEND_LEFT_CIP(top, j, j + 1);
left[-1] = top[-1];
j = 0;
j = 0;
}
} else {
j = 0;
while (j < size_max_x && !IS_INTRA(j, -1)) j++;
while (j < size_max_x && !IS_INTRA(j, -1))
j++;
if (j > 0)
if (x0 > 0) {
EXTEND_LEFT_CIP(top, j, j+1);
EXTEND_LEFT_CIP(top, j, j + 1);
} else {
EXTEND_LEFT_CIP(top, j, j);
top[-1] = top[0];
}
left[-1] = top[-1];
j = 0;
j = 0;
}
if (cand_bottom_left || cand_left) {
EXTEND_DOWN_CIP(left, j, size_max_y-j);
EXTEND_DOWN_CIP(left, j, size_max_y - j);
}
if (!cand_left) {
EXTEND_DOWN(left, 0, size);
@@ -233,10 +238,10 @@ static void FUNC(intra_pred)(HEVCContext *s, int x0, int y0, int log2_size, int
}
if (x0 != 0 && y0 != 0) {
EXTEND_UP_CIP(left, size_max_y - 1, size_max_y);
} else if( x0 == 0) {
} else if (x0 == 0) {
EXTEND_UP_CIP_0(left, size_max_y - 1, size_max_y);
} else{
EXTEND_UP_CIP(left, size_max_y - 1, size_max_y-1);
} else {
EXTEND_UP_CIP(left, size_max_y - 1, size_max_y - 1);
}
top[-1] = left[-1];
if (y0 != 0) {
@@ -255,14 +260,14 @@ static void FUNC(intra_pred)(HEVCContext *s, int x0, int y0, int log2_size, int
left[-1] = top[0];
EXTEND_DOWN(left, 0, 2 * size);
cand_up_left = 1;
cand_left = 1;
cand_left = 1;
} else if (cand_up_right) {
EXTEND_LEFT(top, size, size);
left[-1] = top[0];
EXTEND_DOWN(left ,0 , 2 * size);
cand_up = 1;
EXTEND_DOWN(left, 0, 2 * size);
cand_up = 1;
cand_up_left = 1;
cand_left = 1;
cand_left = 1;
} else { // No samples available
top[0] = left[-1] = (1 << (BIT_DEPTH - 1));
EXTEND_RIGHT(top, 1, 2 * size - 1);
@@ -278,7 +283,7 @@ static void FUNC(intra_pred)(HEVCContext *s, int x0, int y0, int log2_size, int
}
if (!cand_up) {
top[0] = left[-1];
EXTEND_RIGHT(top, 1, size-1);
EXTEND_RIGHT(top, 1, size - 1);
}
if (!cand_up_right) {
EXTEND_RIGHT(top, size, size);
@@ -303,13 +308,13 @@ static void FUNC(intra_pred)(HEVCContext *s, int x0, int y0, int log2_size, int
// Filtering process
if (c_idx == 0 && mode != INTRA_DC && size != 4) {
int intra_hor_ver_dist_thresh[] = { 7, 1, 0 };
int min_dist_vert_hor = FFMIN(FFABS((int)(mode - 26U)),
int min_dist_vert_hor = FFMIN(FFABS((int)(mode - 26U)),
FFABS((int)(mode - 10U)));
if (min_dist_vert_hor > intra_hor_ver_dist_thresh[log2_size - 3]) {
int threshold = 1 << (BIT_DEPTH - 5);
if (s->sps->sps_strong_intra_smoothing_enable_flag &&
log2_size == 5 &&
FFABS(top[-1] + top[63] - 2 * top[31]) < threshold &&
FFABS(top[-1] + top[63] - 2 * top[31]) < threshold &&
FFABS(left[-1] + left[63] - 2 * left[31]) < threshold) {
// We can't just overwrite values in top because it could be
// a pointer into src
@@ -328,9 +333,8 @@ static void FUNC(intra_pred)(HEVCContext *s, int x0, int y0, int log2_size, int
for (i = 2 * size - 2; i >= 0; i--)
filtered_left[i] = (left[i + 1] + 2 * left[i] +
left[i - 1] + 2) >> 2;
filtered_top[-1] =
filtered_left[-1] = (left[0] + 2 * left[-1] +
top[0] + 2) >> 2;
filtered_top[-1] =
filtered_left[-1] = (left[0] + 2 * left[-1] + top[0] + 2) >> 2;
for (i = 2 * size - 2; i >= 0; i--)
filtered_top[i] = (top[i + 1] + 2 * top[i] +
top[i - 1] + 2) >> 2;
@@ -342,16 +346,17 @@ static void FUNC(intra_pred)(HEVCContext *s, int x0, int y0, int log2_size, int

switch (mode) {
case INTRA_PLANAR:
s->hpc.pred_planar[log2_size - 2]((uint8_t*)src, (uint8_t*)top,
(uint8_t*)left, stride);
s->hpc.pred_planar[log2_size - 2]((uint8_t *)src, (uint8_t *)top,
(uint8_t *)left, stride);
break;
case INTRA_DC:
s->hpc.pred_dc((uint8_t*)src, (uint8_t*)top,
(uint8_t*)left, stride, log2_size, c_idx);
s->hpc.pred_dc((uint8_t *)src, (uint8_t *)top,
(uint8_t *)left, stride, log2_size, c_idx);
break;
default:
s->hpc.pred_angular[log2_size - 2]((uint8_t*)src, (uint8_t*)top,
(uint8_t*)left, stride, c_idx, mode);
s->hpc.pred_angular[log2_size - 2]((uint8_t *)src, (uint8_t *)top,
(uint8_t *)left, stride, c_idx,
mode);
break;
}
}
@@ -361,52 +366,52 @@ static void FUNC(pred_planar_0)(uint8_t *_src, const uint8_t *_top,
ptrdiff_t stride)
{
int x, y;
pixel *src = (pixel*)_src;
const pixel *top = (const pixel*)_top;
const pixel *left = (const pixel*)_left;
pixel *src = (pixel *)_src;
const pixel *top = (const pixel *)_top;
const pixel *left = (const pixel *)_left;
for (y = 0; y < 4; y++)
for (x = 0; x < 4; x++)
POS(x, y) = ((3 - x) * left[y] + (x + 1) * top[4] +
(3 - y) * top[x] + (y + 1) * left[4] + 4) >> 3;
POS(x, y) = ((3 - x) * left[y] + (x + 1) * top[4] +
(3 - y) * top[x] + (y + 1) * left[4] + 4) >> 3;
}

static void FUNC(pred_planar_1)(uint8_t *_src, const uint8_t *_top,
const uint8_t *_left, ptrdiff_t stride)
{
int x, y;
pixel *src = (pixel*)_src;
const pixel *top = (const pixel*)_top;
const pixel *left = (const pixel*)_left;
pixel *src = (pixel *)_src;
const pixel *top = (const pixel *)_top;
const pixel *left = (const pixel *)_left;
for (y = 0; y < 8; y++)
for (x = 0; x < 8; x++)
POS(x, y) = ((7 - x) * left[y] + (x + 1) * top[8] +
(7 - y) * top[x] + (y + 1) * left[8] + 8) >> 4;
POS(x, y) = ((7 - x) * left[y] + (x + 1) * top[8] +
(7 - y) * top[x] + (y + 1) * left[8] + 8) >> 4;
}

static void FUNC(pred_planar_2)(uint8_t *_src, const uint8_t *_top,
const uint8_t *_left, ptrdiff_t stride)
{
int x, y;
pixel *src = (pixel*)_src;
const pixel *top = (const pixel*)_top;
const pixel *left = (const pixel*)_left;
pixel *src = (pixel *)_src;
const pixel *top = (const pixel *)_top;
const pixel *left = (const pixel *)_left;
for (y = 0; y < 16; y++)
for (x = 0; x < 16; x++)
POS(x, y) = ((15 - x) * left[y] + (x + 1) * top[16] +
(15 - y) * top[x] + (y + 1) * left[16] + 16) >> 5;
POS(x, y) = ((15 - x) * left[y] + (x + 1) * top[16] +
(15 - y) * top[x] + (y + 1) * left[16] + 16) >> 5;
}

static void FUNC(pred_planar_3)(uint8_t *_src, const uint8_t *_top,
const uint8_t *_left, ptrdiff_t stride)
{
int x, y;
pixel *src = (pixel*)_src;
const pixel *top = (const pixel*)_top;
const pixel *left = (const pixel*)_left;
pixel *src = (pixel *)_src;
const pixel *top = (const pixel *)_top;
const pixel *left = (const pixel *)_left;
for (y = 0; y < 32; y++)
for (x = 0; x < 32; x++)
POS(x, y) = ((31 - x) * left[y] + (x + 1) * top[32] +
(31 - y) * top[x] + (y + 1) * left[32] + 32) >> 6;
POS(x, y) = ((31 - x) * left[y] + (x + 1) * top[32] +
(31 - y) * top[x] + (y + 1) * left[32] + 32) >> 6;
}

static void FUNC(pred_dc)(uint8_t *_src, const uint8_t *_top,
@@ -414,11 +419,11 @@ static void FUNC(pred_dc)(uint8_t *_src, const uint8_t *_top,
ptrdiff_t stride, int log2_size, int c_idx)
{
int i, j, x, y;
int size = (1 << log2_size);
pixel *src = (pixel*)_src;
const pixel *top = (const pixel*)_top;
const pixel *left = (const pixel*)_left;
int dc = size;
int size = (1 << log2_size);
pixel *src = (pixel *)_src;
const pixel *top = (const pixel *)_top;
const pixel *left = (const pixel *)_left;
int dc = size;
pixel4 a;
for (i = 0; i < size; i++)
dc += left[i] + top[i];
@@ -432,7 +437,7 @@ static void FUNC(pred_dc)(uint8_t *_src, const uint8_t *_top,
AV_WN4PA(&POS(j * 4, i), a);

if (c_idx == 0 && size < 32) {
POS(0, 0) = (left[0] + 2 * dc + top[0] + 2) >> 2;
POS(0, 0) = (left[0] + 2 * dc + top[0] + 2) >> 2;
for (x = 1; x < size; x++)
POS(x, 0) = (top[x] + 3 * dc + 2) >> 2;
for (y = 1; y < size; y++)
@@ -447,13 +452,13 @@ static av_always_inline void FUNC(pred_angular)(uint8_t *_src,
int mode, int size)
{
int x, y;
pixel *src = (pixel*)_src;
const pixel *top = (const pixel*)_top;
const pixel *left = (const pixel*)_left;
pixel *src = (pixel *)_src;
const pixel *top = (const pixel *)_top;
const pixel *left = (const pixel *)_left;

static const int intra_pred_angle[] = {
32, 26, 21, 17, 13, 9, 5, 2, 0, -2, -5, -9, -13, -17, -21, -26, -32,
-26, -21, -17, -13, -9, -5, -2, 0, 2, 5, 9, 13, 17, 21, 26, 32
32, 26, 21, 17, 13, 9, 5, 2, 0, -2, -5, -9, -13, -17, -21, -26, -32,
-26, -21, -17, -13, -9, -5, -2, 0, 2, 5, 9, 13, 17, 21, 26, 32
};
static const int inv_angle[] = {
-4096, -1638, -910, -630, -482, -390, -315, -256, -315, -390, -482,
@@ -472,7 +477,7 @@ static av_always_inline void FUNC(pred_angular)(uint8_t *_src,
for (x = 0; x <= size; x++)
ref_tmp[x] = top[x - 1];
for (x = last; x <= -1; x++)
ref_tmp[x] = left[-1 + ((x * inv_angle[mode-11] + 128) >> 8)];
ref_tmp[x] = left[-1 + ((x * inv_angle[mode - 11] + 128) >> 8)];
ref = ref_tmp;
}

@@ -500,7 +505,7 @@ static av_always_inline void FUNC(pred_angular)(uint8_t *_src,
for (x = 0; x <= size; x++)
ref_tmp[x] = left[x - 1];
for (x = last; x <= -1; x++)
ref_tmp[x] = top[-1 + ((x * inv_angle[mode-11] + 128) >> 8)];
ref_tmp[x] = top[-1 + ((x * inv_angle[mode - 11] + 128) >> 8)];
ref = ref_tmp;
}



Write Preview
Loading…
Cancel
Save