Message ID | 20221205060929.2511871-3-fei.w.wang@intel.com |
---|---|
State | New |
Headers | show |
Series | [FFmpeg-devel,v2,01/13] lavc/hevc_ps: remove profile limitation of pps_range_extensions() | expand |
Context | Check | Description |
---|---|---|
yinshiyou/make_loongarch64 | success | Make finished |
yinshiyou/make_fate_loongarch64 | success | Make fate finished |
andriy/make_x86 | success | Make finished |
andriy/make_fate_x86 | success | Make fate finished |
On Ma, 2022-12-05 at 14:09 +0800, Fei Wang wrote: > From: Linjie Fu <linjie.fu@intel.com> > > 1. Add extension syntax according to 7.3.2.2.3/7.3.2.3.3 in T-REC-H.265- > 201911. > 2. Keep using parsed PPS when bitstream overread for compatibility. For > example, the clip PS_A_VIDYO_3.bit in FATE test has incomplete extension > syntax which will be overread and un-decodable if without this change. > 3. Format brace in pps_range_extensions(). > > Signed-off-by: Linjie Fu <linjie.justin.fu@gmail.com> > Signed-off-by: Haihao Xiang <haihao.xiang@intel.com> > Signed-off-by: Fei Wang <fei.w.wang@intel.com> > --- > libavcodec/hevc.h | 3 + > libavcodec/hevc_ps.c | 288 +++++++++++++++++++++++++++++++++++++++++-- > libavcodec/hevc_ps.h | 69 +++++++++++ > 3 files changed, 349 insertions(+), 11 deletions(-) > > diff --git a/libavcodec/hevc.h b/libavcodec/hevc.h > index 1804755327..6b454a75c1 100644 > --- a/libavcodec/hevc.h > +++ b/libavcodec/hevc.h > @@ -154,6 +154,9 @@ enum { > // get near that, though, so set a lower limit here with the maximum > // possible value for 4K video (at most 135 16x16 Ctb rows). > HEVC_MAX_ENTRY_POINT_OFFSETS = HEVC_MAX_TILE_COLUMNS * 135, > + > + // A.3.7: Screen content coding extensions > + HEVC_MAX_PALETTE_PREDICTOR_SIZE = 128, > }; > > > diff --git a/libavcodec/hevc_ps.c b/libavcodec/hevc_ps.c > index ad92b6bcbc..3181962918 100644 > --- a/libavcodec/hevc_ps.c > +++ b/libavcodec/hevc_ps.c > @@ -853,7 +853,7 @@ int ff_hevc_parse_sps(HEVCSPS *sps, GetBitContext *gb, > unsigned int *sps_id, > HEVCWindow *ow; > int ret = 0; > int log2_diff_max_min_transform_block_size; > - int bit_depth_chroma, start, vui_present, sublayer_ordering_info; > + int bit_depth_chroma, start, vui_present, sublayer_ordering_info, > num_comps; > int i; > > // Coded parameters > @@ -1074,8 +1074,12 @@ int ff_hevc_parse_sps(HEVCSPS *sps, GetBitContext *gb, > unsigned int *sps_id, > decode_vui(gb, avctx, apply_defdispwin, sps); > > if (get_bits1(gb)) { // sps_extension_flag > - sps->sps_range_extension_flag = get_bits1(gb); > - skip_bits(gb, 7); //sps_extension_7bits = get_bits(gb, 7); > + sps->sps_range_extension_flag = get_bits1(gb); > + sps->sps_multilayer_extension_flag = get_bits1(gb); > + sps->sps_3d_extension_flag = get_bits1(gb); > + sps->sps_scc_extension_flag = get_bits1(gb); > + skip_bits(gb, 4); // sps_extension_4bits > + > if (sps->sps_range_extension_flag) { > sps->transform_skip_rotation_enabled_flag = get_bits1(gb); > sps->transform_skip_context_enabled_flag = get_bits1(gb); > @@ -1101,6 +1105,57 @@ int ff_hevc_parse_sps(HEVCSPS *sps, GetBitContext *gb, > unsigned int *sps_id, > av_log(avctx, AV_LOG_WARNING, > "cabac_bypass_alignment_enabled_flag not yet > implemented\n"); > } > + > + if (sps->sps_multilayer_extension_flag) { > + skip_bits1(gb); // inter_view_mv_vert_constraint_flag > + av_log(avctx, AV_LOG_WARNING, > + "sps_multilayer_extension_flag not yet implemented\n"); > + } > + > + if (sps->sps_3d_extension_flag) { > + for (i = 0; i <= 1; i++) { > + skip_bits1(gb); // iv_di_mc_enabled_flag > + skip_bits1(gb); // iv_mv_scal_enabled_flag > + if (i == 0) { > + get_ue_golomb_long(gb); // log2_ivmc_sub_pb_size_minus3 > + skip_bits1(gb); // iv_res_pred_enabled_flag > + skip_bits1(gb); // depth_ref_enabled_flag > + skip_bits1(gb); // vsp_mc_enabled_flag > + skip_bits1(gb); // dbbp_enabled_flag > + } else { > + skip_bits1(gb); // tex_mc_enabled_flag > + get_ue_golomb_long(gb); // log2_ivmc_sub_pb_size_minus3 > + skip_bits1(gb); // intra_contour_enabled_flag > + skip_bits1(gb); // intra_dc_only_wedge_enabled_flag > + skip_bits1(gb); // cqt_cu_part_pred_enabled_flag > + skip_bits1(gb); // inter_dc_only_enabled_flag > + skip_bits1(gb); // skip_intra_enabled_flag > + } > + } > + av_log(avctx, AV_LOG_WARNING, > + "sps_3d_extension_flag not yet implemented\n"); > + } > + > + if (sps->sps_scc_extension_flag) { > + sps->sps_curr_pic_ref_enabled_flag = get_bits1(gb); > + sps->palette_mode_enabled_flag = get_bits1(gb); > + if (sps->palette_mode_enabled_flag) { > + sps->palette_max_size = get_ue_golomb_long(gb); > + sps->delta_palette_max_predictor_size = > get_ue_golomb_long(gb); > + sps->sps_palette_predictor_initializers_present_flag = > get_bits1(gb); > + > + if (sps->sps_palette_predictor_initializers_present_flag) { > + sps->sps_num_palette_predictor_initializers_minus1 = > get_ue_golomb_long(gb); > + num_comps = !sps->chroma_format_idc ? 1 : 3; > + for (int comp = 0; comp < num_comps; comp++) > + for (i = 0; i <= sps- > >sps_num_palette_predictor_initializers_minus1; i++) > + sps->sps_palette_predictor_initializer[comp][i] = > + get_bits(gb, !comp ? sps->bit_depth : > sps->bit_depth_chroma); > + } > + } > + sps->motion_vector_resolution_control_idc = get_bits(gb, 2); > + sps->intra_boundary_filtering_disabled_flag = get_bits1(gb); > + } > } > if (apply_defdispwin) { > sps->output_window.left_offset += sps- > >vui.def_disp_win.left_offset; > @@ -1265,10 +1320,162 @@ static void hevc_pps_free(void *opaque, uint8_t > *data) > av_freep(&pps); > } > > -static int pps_range_extensions(GetBitContext *gb, AVCodecContext *avctx, > - HEVCPPS *pps, HEVCSPS *sps) { > - int i; > +static void colour_mapping_octants(GetBitContext *gb, HEVCPPS *pps, int > inp_depth, > + int idx_y, int idx_cb, int idx_cr, int > inp_length) > +{ > + unsigned int split_octant_flag, part_num_y, coded_res_flag, res_coeff_q, > res_coeff_r; > + int bit_depth_cm_input_y, bit_depth_cm_output_y, cm_res_bits; > + > + part_num_y = 1 << pps->cm_y_part_num_log2; > + > + split_octant_flag = inp_depth < pps->cm_octant_depth ? get_bits1(gb) : 0; > + > + if (split_octant_flag) > + for (int k = 0; k < 2; k++) > + for (int m = 0; m < 2; m++) > + for (int n = 0; n < 2; n++) > + colour_mapping_octants(gb, pps, inp_depth + 1, > + idx_y + part_num_y * k * > inp_length / 2, > + idx_cb + m * inp_length / 2, > + idx_cr + n * inp_length / 2, > + inp_length / 2); > + else > + for (int i = 0; i < part_num_y; i++) { > + for (int j = 0; j < 4; j++) { > + coded_res_flag = get_bits1(gb); > + if (coded_res_flag) > + for (int c = 0; c < 3; c++) { > + res_coeff_q = get_ue_golomb_long(gb); > + bit_depth_cm_input_y = 8 + pps- > >luma_bit_depth_cm_input_minus8; > + bit_depth_cm_output_y = 8 + pps- > >luma_bit_depth_cm_output_minus8; > + cm_res_bits = FFMAX(0, 10 + bit_depth_cm_input_y - > bit_depth_cm_output_y - > + pps->cm_res_quant_bits - (pps- > >cm_delta_flc_bits_minus1 + 1)); > + res_coeff_r = cm_res_bits ? get_bits(gb, cm_res_bits) > : 0; > + if (res_coeff_q || res_coeff_r) > + skip_bits1(gb); > + } > + } > + } > +} > > +static void colour_mapping_table(GetBitContext *gb, HEVCPPS *pps) > +{ > + pps->num_cm_ref_layers_minus1 = get_ue_golomb_long(gb); > + for (int i = 0; i <= pps->num_cm_ref_layers_minus1; i++) > + pps->cm_ref_layer_id[i] = get_bits(gb, 6); > + > + pps->cm_octant_depth = get_bits(gb, 2); > + pps->cm_y_part_num_log2 = get_bits(gb, 2); > + > + pps->luma_bit_depth_cm_input_minus8 = get_ue_golomb_long(gb); > + pps->chroma_bit_depth_cm_input_minus8 = get_ue_golomb_long(gb); > + pps->luma_bit_depth_cm_output_minus8 = get_ue_golomb_long(gb); > + pps->chroma_bit_depth_cm_output_minus8 = get_ue_golomb_long(gb); > + > + pps->cm_res_quant_bits = get_bits(gb, 2); > + pps->cm_delta_flc_bits_minus1 = get_bits(gb, 2); > + > + if (pps->cm_octant_depth == 1) { > + pps->cm_adapt_threshold_u_delta = get_se_golomb_long(gb); > + pps->cm_adapt_threshold_v_delta = get_se_golomb_long(gb); > + } > + > + colour_mapping_octants(gb, pps, 0, 0, 0, 0, 1 << pps->cm_octant_depth); > +} > + > +static int pps_multilayer_extension(GetBitContext *gb, AVCodecContext *avctx, > + HEVCPPS *pps, HEVCSPS *sps) > +{ > + pps->poc_reset_info_present_flag = get_bits1(gb); > + pps->pps_infer_scaling_list_flag = get_bits1(gb); > + if (pps->pps_infer_scaling_list_flag) > + pps->pps_scaling_list_ref_layer_id = get_bits(gb, 6); > + > + pps->num_ref_loc_offsets = get_ue_golomb_long(gb); > + for (int i = 0; i < pps->num_ref_loc_offsets; i++) { > + pps->ref_loc_offset_layer_id[i] = get_bits(gb, 6); > + pps->scaled_ref_layer_offset_present_flag[i] = get_bits1(gb); > + if (pps->scaled_ref_layer_offset_present_flag[i]) { > + pps->scaled_ref_layer_left_offset[pps- > >ref_loc_offset_layer_id[i]] = get_se_golomb_long(gb); > + pps->scaled_ref_layer_top_offset[pps- > >ref_loc_offset_layer_id[i]] = get_se_golomb_long(gb); > + pps->scaled_ref_layer_right_offset[pps- > >ref_loc_offset_layer_id[i]] = get_se_golomb_long(gb); > + pps->scaled_ref_layer_bottom_offset[pps- > >ref_loc_offset_layer_id[i]] = get_se_golomb_long(gb); > + } > + > + pps->ref_region_offset_present_flag[i] = get_bits1(gb); > + if (pps->ref_region_offset_present_flag[i]) { > + pps->ref_region_left_offset[pps->ref_loc_offset_layer_id[i]] = > get_se_golomb_long(gb); > + pps->ref_region_top_offset[pps->ref_loc_offset_layer_id[i]] = > get_se_golomb_long(gb); > + pps->ref_region_right_offset[pps->ref_loc_offset_layer_id[i]] = > get_se_golomb_long(gb); > + pps->ref_region_bottom_offset[pps->ref_loc_offset_layer_id[i]] = > get_se_golomb_long(gb); > + } > + > + pps->resample_phase_set_present_flag[i] = get_bits1(gb); > + if (pps->resample_phase_set_present_flag[i]) { > + pps->phase_hor_luma[pps->ref_loc_offset_layer_id[i]] = > get_ue_golomb_long(gb); > + pps->phase_ver_luma[pps->ref_loc_offset_layer_id[i]] = > get_ue_golomb_long(gb); > + pps->phase_hor_chroma_plus8[pps->ref_loc_offset_layer_id[i]] = > get_ue_golomb_long(gb); > + pps->phase_ver_chroma_plus8[pps->ref_loc_offset_layer_id[i]] = > get_ue_golomb_long(gb); > + } > + } > + > + pps->colour_mapping_enabled_flag = get_bits1(gb); > + if (pps->colour_mapping_enabled_flag) > + colour_mapping_table(gb, pps); > + > + return 0; > +} > + > +static void delta_dlt(GetBitContext *gb, HEVCPPS *pps) > +{ > + unsigned int num_val_delta_dlt, max_diff = 0; > + int min_diff_minus1 = -1; > + unsigned int len; > + > + num_val_delta_dlt = get_bits(gb, pps- > >pps_bit_depth_for_depth_layers_minus8 + 8); > + if (num_val_delta_dlt) { > + if (num_val_delta_dlt > 1) > + max_diff = get_bits(gb, pps- > >pps_bit_depth_for_depth_layers_minus8 + 8); > + if (num_val_delta_dlt > 2 && max_diff) { > + len = av_log2(max_diff) + 1; > + min_diff_minus1 = get_bits(gb, len); > + } > + if (max_diff > (min_diff_minus1 + 1)) > + for (int k = 1; k < num_val_delta_dlt; k++) { > + len = av_log2(max_diff - (min_diff_minus1 + 1)) + 1; > + skip_bits(gb, len); // delta_val_diff_minus_min > + } > + } > +} > + > +static int pps_3d_extension(GetBitContext *gb, AVCodecContext *avctx, > + HEVCPPS *pps, HEVCSPS *sps) > +{ > + unsigned int pps_depth_layers_minus1; > + int i, j; > + > + if (get_bits1(gb)) { // dlts_present_flag > + pps_depth_layers_minus1 = get_bits(gb, 6); > + pps->pps_bit_depth_for_depth_layers_minus8 = get_bits(gb, 4); > + for (i = 0; i <= pps_depth_layers_minus1; i++) { Please use the same coding style, for (int i = 0; ......) > + if (get_bits1(gb)) { // dlt_flag[i] > + if (!get_bits1(gb)) { // dlt_pred_flag[i] > + if (get_bits1(gb)) { // dlt_val_flags_present_flag[i] > + for (j = 0; j <= ((1 << (pps- > >pps_bit_depth_for_depth_layers_minus8 + 8)) - 1); j++) for (int j = 0; ......) Thanks Haihao > + skip_bits1(gb); // dlt_value_flag[i][j] > + } else > + delta_dlt(gb, pps); > + } > + } > + } > + } > + > + return 0; > +} > + > +static int pps_range_extensions(GetBitContext *gb, AVCodecContext *avctx, > + HEVCPPS *pps, HEVCSPS *sps) > +{ > if (pps->transform_skip_enabled_flag) { > pps->log2_max_transform_skip_block_size = get_ue_golomb_long(gb) + 2; > } > @@ -1282,7 +1489,7 @@ static int pps_range_extensions(GetBitContext *gb, > AVCodecContext *avctx, > "chroma_qp_offset_list_len_minus1 shall be in the range > [0, 5].\n"); > return AVERROR_INVALIDDATA; > } > - for (i = 0; i <= pps->chroma_qp_offset_list_len_minus1; i++) { > + for (int i = 0; i <= pps->chroma_qp_offset_list_len_minus1; i++) { > pps->cb_qp_offset_list[i] = get_se_golomb_long(gb); > if (pps->cb_qp_offset_list[i]) { > av_log(avctx, AV_LOG_WARNING, > @@ -1306,6 +1513,47 @@ static int pps_range_extensions(GetBitContext *gb, > AVCodecContext *avctx, > return(0); > } > > +static int pps_scc_extension(GetBitContext *gb, AVCodecContext *avctx, > + HEVCPPS *pps, HEVCSPS *sps) > +{ > + int num_comps, ret; > + > + pps->pps_curr_pic_ref_enabled_flag = get_bits1(gb); > + if (pps->residual_adaptive_colour_transform_enabled_flag = get_bits1(gb)) > { > + pps->pps_slice_act_qp_offsets_present_flag = get_bits1(gb); > + pps->pps_act_y_qp_offset = get_se_golomb_long(gb) - 5; > + pps->pps_act_cb_qp_offset = get_se_golomb_long(gb) - 5; > + pps->pps_act_cr_qp_offset = get_se_golomb_long(gb) - 3; > + > +#define CHECK_QP_OFFSET(name) (pps->pps_act_ ## name ## _qp_offset <= -12 || > \ > + pps->pps_act_ ## name ## _qp_offset >= 12) > + ret = CHECK_QP_OFFSET(y) || CHECK_QP_OFFSET(cb) || > CHECK_QP_OFFSET(cr); > +#undef CHECK_QP_OFFSET > + if (ret) { > + av_log(avctx, AV_LOG_ERROR, > + "PpsActQpOffsetY/Cb/Cr shall be in the range of [-12, > 12].\n"); > + return AVERROR_INVALIDDATA; > + } > + } > + > + if (pps->pps_palette_predictor_initializers_present_flag = get_bits1(gb)) > { > + if ((pps->pps_num_palette_predictor_initializers = > get_ue_golomb_long(gb)) > 0) { > + pps->monochrome_palette_flag = get_bits1(gb); > + pps->luma_bit_depth_entry_minus8 = get_ue_golomb_long(gb); > + if (!pps->monochrome_palette_flag) > + pps->chroma_bit_depth_entry_minus8 = get_ue_golomb_long(gb); > + num_comps = pps->monochrome_palette_flag ? 1 : 3; > + for (int comp = 0; comp < num_comps; comp++) > + for (int i = 0; i < pps- > >pps_num_palette_predictor_initializers; i++) > + pps->pps_palette_predictor_initializer[comp][i] = > + get_bits(gb, 8 + (!comp ? pps- > >luma_bit_depth_entry_minus8 : > + pps- > >chroma_bit_depth_entry_minus8)); > + } > + } > + > + return 0; > +} > + > static inline int setup_pps(AVCodecContext *avctx, GetBitContext *gb, > HEVCPPS *pps, HEVCSPS *sps) > { > @@ -1658,12 +1906,31 @@ int ff_hevc_decode_nal_pps(GetBitContext *gb, > AVCodecContext *avctx, > pps->slice_header_extension_present_flag = get_bits1(gb); > > if (get_bits1(gb)) { // pps_extension_present_flag > - pps->pps_range_extensions_flag = get_bits1(gb); > - skip_bits(gb, 7); // pps_extension_7bits > + pps->pps_range_extensions_flag = get_bits1(gb); > + pps->pps_multilayer_extension_flag = get_bits1(gb); > + pps->pps_3d_extension_flag = get_bits1(gb); > + pps->pps_scc_extension_flag = get_bits1(gb); > + skip_bits(gb, 4); // pps_extension_4bits > + > if (pps->pps_range_extensions_flag) { > if ((ret = pps_range_extensions(gb, avctx, pps, sps)) < 0) > goto err; > } > + > + if (pps->pps_multilayer_extension_flag) { > + if ((ret = pps_multilayer_extension(gb, avctx, pps, sps)) < 0) > + goto err; > + } > + > + if (pps->pps_3d_extension_flag) { > + if ((ret = pps_3d_extension(gb, avctx, pps, sps)) < 0) > + goto err; > + } > + > + if (pps->pps_scc_extension_flag) { > + if ((ret = pps_scc_extension(gb, avctx, pps, sps)) < 0) > + goto err; > + } > } > > ret = setup_pps(avctx, gb, pps, sps); > @@ -1671,9 +1938,8 @@ int ff_hevc_decode_nal_pps(GetBitContext *gb, > AVCodecContext *avctx, > goto err; > > if (get_bits_left(gb) < 0) { > - av_log(avctx, AV_LOG_ERROR, > + av_log(avctx, AV_LOG_WARNING, > "Overread PPS by %d bits\n", -get_bits_left(gb)); > - goto err; > } > > remove_pps(ps, pps_id); > diff --git a/libavcodec/hevc_ps.h b/libavcodec/hevc_ps.h > index 18894cfed1..1f704108e3 100644 > --- a/libavcodec/hevc_ps.h > +++ b/libavcodec/hevc_ps.h > @@ -210,6 +210,20 @@ typedef struct HEVCSPS { > int persistent_rice_adaptation_enabled_flag; > int cabac_bypass_alignment_enabled_flag; > > + int sps_multilayer_extension_flag; > + int sps_3d_extension_flag; > + > + int sps_scc_extension_flag; > + int sps_curr_pic_ref_enabled_flag; > + int palette_mode_enabled_flag; > + int palette_max_size; > + int delta_palette_max_predictor_size; > + int sps_palette_predictor_initializers_present_flag; > + int sps_num_palette_predictor_initializers_minus1; > + int > sps_palette_predictor_initializer[3][HEVC_MAX_PALETTE_PREDICTOR_SIZE]; > + int motion_vector_resolution_control_idc; > + int intra_boundary_filtering_disabled_flag; > + > ///< coded frame dimension in various units > int width; > int height; > @@ -284,6 +298,9 @@ typedef struct HEVCPPS { > uint8_t slice_header_extension_present_flag; > uint8_t log2_max_transform_skip_block_size; > uint8_t pps_range_extensions_flag; > + uint8_t pps_multilayer_extension_flag; > + uint8_t pps_3d_extension_flag; > + uint8_t pps_scc_extension_flag; > uint8_t cross_component_prediction_enabled_flag; > uint8_t chroma_qp_offset_list_enabled_flag; > uint8_t diff_cu_chroma_qp_offset_depth; > @@ -293,6 +310,58 @@ typedef struct HEVCPPS { > uint8_t log2_sao_offset_scale_luma; > uint8_t log2_sao_offset_scale_chroma; > > + // Multilayer extension parameters > + uint8_t poc_reset_info_present_flag; > + uint8_t pps_infer_scaling_list_flag; > + uint8_t pps_scaling_list_ref_layer_id; > + uint16_t num_ref_loc_offsets; > + uint8_t ref_loc_offset_layer_id[64]; > + uint8_t scaled_ref_layer_offset_present_flag[64]; > + int8_t scaled_ref_layer_left_offset[64]; > + int8_t scaled_ref_layer_top_offset[64]; > + int8_t scaled_ref_layer_right_offset[64]; > + int8_t scaled_ref_layer_bottom_offset[64]; > + uint8_t ref_region_offset_present_flag[64]; > + int8_t ref_region_left_offset[64]; > + int8_t ref_region_top_offset[64]; > + int8_t ref_region_right_offset[64]; > + int8_t ref_region_bottom_offset[64]; > + uint8_t resample_phase_set_present_flag[64]; > + uint16_t phase_hor_luma[64]; > + uint16_t phase_ver_luma[64]; > + uint16_t phase_hor_chroma_plus8[64]; > + uint16_t phase_ver_chroma_plus8[64]; > + uint8_t colour_mapping_enabled_flag; > + uint16_t num_cm_ref_layers_minus1; > + uint8_t cm_ref_layer_id[63]; > + uint8_t cm_octant_depth; > + uint8_t cm_y_part_num_log2; > + uint16_t luma_bit_depth_cm_input_minus8; > + uint16_t chroma_bit_depth_cm_input_minus8; > + uint16_t luma_bit_depth_cm_output_minus8; > + uint16_t chroma_bit_depth_cm_output_minus8; > + uint8_t cm_res_quant_bits; > + uint8_t cm_delta_flc_bits_minus1; > + int8_t cm_adapt_threshold_u_delta; > + int8_t cm_adapt_threshold_v_delta; > + > + // 3D extension parameters > + uint8_t pps_bit_depth_for_depth_layers_minus8; > + > + // SCC extension parameters > + uint8_t pps_curr_pic_ref_enabled_flag; > + uint8_t residual_adaptive_colour_transform_enabled_flag; > + uint8_t pps_slice_act_qp_offsets_present_flag; > + int8_t pps_act_y_qp_offset; // _plus5 > + int8_t pps_act_cb_qp_offset; // _plus5 > + int8_t pps_act_cr_qp_offset; // _plus3 > + uint8_t pps_palette_predictor_initializers_present_flag; > + uint8_t pps_num_palette_predictor_initializers; > + uint8_t monochrome_palette_flag; > + uint8_t luma_bit_depth_entry_minus8; > + uint8_t chroma_bit_depth_entry_minus8; > + uint8_t > pps_palette_predictor_initializer[3][HEVC_MAX_PALETTE_PREDICTOR_SIZE]; > + > // Inferred parameters > unsigned int *column_width; ///< ColumnWidth > unsigned int *row_height; ///< RowHeight
diff --git a/libavcodec/hevc.h b/libavcodec/hevc.h index 1804755327..6b454a75c1 100644 --- a/libavcodec/hevc.h +++ b/libavcodec/hevc.h @@ -154,6 +154,9 @@ enum { // get near that, though, so set a lower limit here with the maximum // possible value for 4K video (at most 135 16x16 Ctb rows). HEVC_MAX_ENTRY_POINT_OFFSETS = HEVC_MAX_TILE_COLUMNS * 135, + + // A.3.7: Screen content coding extensions + HEVC_MAX_PALETTE_PREDICTOR_SIZE = 128, }; diff --git a/libavcodec/hevc_ps.c b/libavcodec/hevc_ps.c index ad92b6bcbc..3181962918 100644 --- a/libavcodec/hevc_ps.c +++ b/libavcodec/hevc_ps.c @@ -853,7 +853,7 @@ int ff_hevc_parse_sps(HEVCSPS *sps, GetBitContext *gb, unsigned int *sps_id, HEVCWindow *ow; int ret = 0; int log2_diff_max_min_transform_block_size; - int bit_depth_chroma, start, vui_present, sublayer_ordering_info; + int bit_depth_chroma, start, vui_present, sublayer_ordering_info, num_comps; int i; // Coded parameters @@ -1074,8 +1074,12 @@ int ff_hevc_parse_sps(HEVCSPS *sps, GetBitContext *gb, unsigned int *sps_id, decode_vui(gb, avctx, apply_defdispwin, sps); if (get_bits1(gb)) { // sps_extension_flag - sps->sps_range_extension_flag = get_bits1(gb); - skip_bits(gb, 7); //sps_extension_7bits = get_bits(gb, 7); + sps->sps_range_extension_flag = get_bits1(gb); + sps->sps_multilayer_extension_flag = get_bits1(gb); + sps->sps_3d_extension_flag = get_bits1(gb); + sps->sps_scc_extension_flag = get_bits1(gb); + skip_bits(gb, 4); // sps_extension_4bits + if (sps->sps_range_extension_flag) { sps->transform_skip_rotation_enabled_flag = get_bits1(gb); sps->transform_skip_context_enabled_flag = get_bits1(gb); @@ -1101,6 +1105,57 @@ int ff_hevc_parse_sps(HEVCSPS *sps, GetBitContext *gb, unsigned int *sps_id, av_log(avctx, AV_LOG_WARNING, "cabac_bypass_alignment_enabled_flag not yet implemented\n"); } + + if (sps->sps_multilayer_extension_flag) { + skip_bits1(gb); // inter_view_mv_vert_constraint_flag + av_log(avctx, AV_LOG_WARNING, + "sps_multilayer_extension_flag not yet implemented\n"); + } + + if (sps->sps_3d_extension_flag) { + for (i = 0; i <= 1; i++) { + skip_bits1(gb); // iv_di_mc_enabled_flag + skip_bits1(gb); // iv_mv_scal_enabled_flag + if (i == 0) { + get_ue_golomb_long(gb); // log2_ivmc_sub_pb_size_minus3 + skip_bits1(gb); // iv_res_pred_enabled_flag + skip_bits1(gb); // depth_ref_enabled_flag + skip_bits1(gb); // vsp_mc_enabled_flag + skip_bits1(gb); // dbbp_enabled_flag + } else { + skip_bits1(gb); // tex_mc_enabled_flag + get_ue_golomb_long(gb); // log2_ivmc_sub_pb_size_minus3 + skip_bits1(gb); // intra_contour_enabled_flag + skip_bits1(gb); // intra_dc_only_wedge_enabled_flag + skip_bits1(gb); // cqt_cu_part_pred_enabled_flag + skip_bits1(gb); // inter_dc_only_enabled_flag + skip_bits1(gb); // skip_intra_enabled_flag + } + } + av_log(avctx, AV_LOG_WARNING, + "sps_3d_extension_flag not yet implemented\n"); + } + + if (sps->sps_scc_extension_flag) { + sps->sps_curr_pic_ref_enabled_flag = get_bits1(gb); + sps->palette_mode_enabled_flag = get_bits1(gb); + if (sps->palette_mode_enabled_flag) { + sps->palette_max_size = get_ue_golomb_long(gb); + sps->delta_palette_max_predictor_size = get_ue_golomb_long(gb); + sps->sps_palette_predictor_initializers_present_flag = get_bits1(gb); + + if (sps->sps_palette_predictor_initializers_present_flag) { + sps->sps_num_palette_predictor_initializers_minus1 = get_ue_golomb_long(gb); + num_comps = !sps->chroma_format_idc ? 1 : 3; + for (int comp = 0; comp < num_comps; comp++) + for (i = 0; i <= sps->sps_num_palette_predictor_initializers_minus1; i++) + sps->sps_palette_predictor_initializer[comp][i] = + get_bits(gb, !comp ? sps->bit_depth : sps->bit_depth_chroma); + } + } + sps->motion_vector_resolution_control_idc = get_bits(gb, 2); + sps->intra_boundary_filtering_disabled_flag = get_bits1(gb); + } } if (apply_defdispwin) { sps->output_window.left_offset += sps->vui.def_disp_win.left_offset; @@ -1265,10 +1320,162 @@ static void hevc_pps_free(void *opaque, uint8_t *data) av_freep(&pps); } -static int pps_range_extensions(GetBitContext *gb, AVCodecContext *avctx, - HEVCPPS *pps, HEVCSPS *sps) { - int i; +static void colour_mapping_octants(GetBitContext *gb, HEVCPPS *pps, int inp_depth, + int idx_y, int idx_cb, int idx_cr, int inp_length) +{ + unsigned int split_octant_flag, part_num_y, coded_res_flag, res_coeff_q, res_coeff_r; + int bit_depth_cm_input_y, bit_depth_cm_output_y, cm_res_bits; + + part_num_y = 1 << pps->cm_y_part_num_log2; + + split_octant_flag = inp_depth < pps->cm_octant_depth ? get_bits1(gb) : 0; + + if (split_octant_flag) + for (int k = 0; k < 2; k++) + for (int m = 0; m < 2; m++) + for (int n = 0; n < 2; n++) + colour_mapping_octants(gb, pps, inp_depth + 1, + idx_y + part_num_y * k * inp_length / 2, + idx_cb + m * inp_length / 2, + idx_cr + n * inp_length / 2, + inp_length / 2); + else + for (int i = 0; i < part_num_y; i++) { + for (int j = 0; j < 4; j++) { + coded_res_flag = get_bits1(gb); + if (coded_res_flag) + for (int c = 0; c < 3; c++) { + res_coeff_q = get_ue_golomb_long(gb); + bit_depth_cm_input_y = 8 + pps->luma_bit_depth_cm_input_minus8; + bit_depth_cm_output_y = 8 + pps->luma_bit_depth_cm_output_minus8; + cm_res_bits = FFMAX(0, 10 + bit_depth_cm_input_y - bit_depth_cm_output_y - + pps->cm_res_quant_bits - (pps->cm_delta_flc_bits_minus1 + 1)); + res_coeff_r = cm_res_bits ? get_bits(gb, cm_res_bits) : 0; + if (res_coeff_q || res_coeff_r) + skip_bits1(gb); + } + } + } +} +static void colour_mapping_table(GetBitContext *gb, HEVCPPS *pps) +{ + pps->num_cm_ref_layers_minus1 = get_ue_golomb_long(gb); + for (int i = 0; i <= pps->num_cm_ref_layers_minus1; i++) + pps->cm_ref_layer_id[i] = get_bits(gb, 6); + + pps->cm_octant_depth = get_bits(gb, 2); + pps->cm_y_part_num_log2 = get_bits(gb, 2); + + pps->luma_bit_depth_cm_input_minus8 = get_ue_golomb_long(gb); + pps->chroma_bit_depth_cm_input_minus8 = get_ue_golomb_long(gb); + pps->luma_bit_depth_cm_output_minus8 = get_ue_golomb_long(gb); + pps->chroma_bit_depth_cm_output_minus8 = get_ue_golomb_long(gb); + + pps->cm_res_quant_bits = get_bits(gb, 2); + pps->cm_delta_flc_bits_minus1 = get_bits(gb, 2); + + if (pps->cm_octant_depth == 1) { + pps->cm_adapt_threshold_u_delta = get_se_golomb_long(gb); + pps->cm_adapt_threshold_v_delta = get_se_golomb_long(gb); + } + + colour_mapping_octants(gb, pps, 0, 0, 0, 0, 1 << pps->cm_octant_depth); +} + +static int pps_multilayer_extension(GetBitContext *gb, AVCodecContext *avctx, + HEVCPPS *pps, HEVCSPS *sps) +{ + pps->poc_reset_info_present_flag = get_bits1(gb); + pps->pps_infer_scaling_list_flag = get_bits1(gb); + if (pps->pps_infer_scaling_list_flag) + pps->pps_scaling_list_ref_layer_id = get_bits(gb, 6); + + pps->num_ref_loc_offsets = get_ue_golomb_long(gb); + for (int i = 0; i < pps->num_ref_loc_offsets; i++) { + pps->ref_loc_offset_layer_id[i] = get_bits(gb, 6); + pps->scaled_ref_layer_offset_present_flag[i] = get_bits1(gb); + if (pps->scaled_ref_layer_offset_present_flag[i]) { + pps->scaled_ref_layer_left_offset[pps->ref_loc_offset_layer_id[i]] = get_se_golomb_long(gb); + pps->scaled_ref_layer_top_offset[pps->ref_loc_offset_layer_id[i]] = get_se_golomb_long(gb); + pps->scaled_ref_layer_right_offset[pps->ref_loc_offset_layer_id[i]] = get_se_golomb_long(gb); + pps->scaled_ref_layer_bottom_offset[pps->ref_loc_offset_layer_id[i]] = get_se_golomb_long(gb); + } + + pps->ref_region_offset_present_flag[i] = get_bits1(gb); + if (pps->ref_region_offset_present_flag[i]) { + pps->ref_region_left_offset[pps->ref_loc_offset_layer_id[i]] = get_se_golomb_long(gb); + pps->ref_region_top_offset[pps->ref_loc_offset_layer_id[i]] = get_se_golomb_long(gb); + pps->ref_region_right_offset[pps->ref_loc_offset_layer_id[i]] = get_se_golomb_long(gb); + pps->ref_region_bottom_offset[pps->ref_loc_offset_layer_id[i]] = get_se_golomb_long(gb); + } + + pps->resample_phase_set_present_flag[i] = get_bits1(gb); + if (pps->resample_phase_set_present_flag[i]) { + pps->phase_hor_luma[pps->ref_loc_offset_layer_id[i]] = get_ue_golomb_long(gb); + pps->phase_ver_luma[pps->ref_loc_offset_layer_id[i]] = get_ue_golomb_long(gb); + pps->phase_hor_chroma_plus8[pps->ref_loc_offset_layer_id[i]] = get_ue_golomb_long(gb); + pps->phase_ver_chroma_plus8[pps->ref_loc_offset_layer_id[i]] = get_ue_golomb_long(gb); + } + } + + pps->colour_mapping_enabled_flag = get_bits1(gb); + if (pps->colour_mapping_enabled_flag) + colour_mapping_table(gb, pps); + + return 0; +} + +static void delta_dlt(GetBitContext *gb, HEVCPPS *pps) +{ + unsigned int num_val_delta_dlt, max_diff = 0; + int min_diff_minus1 = -1; + unsigned int len; + + num_val_delta_dlt = get_bits(gb, pps->pps_bit_depth_for_depth_layers_minus8 + 8); + if (num_val_delta_dlt) { + if (num_val_delta_dlt > 1) + max_diff = get_bits(gb, pps->pps_bit_depth_for_depth_layers_minus8 + 8); + if (num_val_delta_dlt > 2 && max_diff) { + len = av_log2(max_diff) + 1; + min_diff_minus1 = get_bits(gb, len); + } + if (max_diff > (min_diff_minus1 + 1)) + for (int k = 1; k < num_val_delta_dlt; k++) { + len = av_log2(max_diff - (min_diff_minus1 + 1)) + 1; + skip_bits(gb, len); // delta_val_diff_minus_min + } + } +} + +static int pps_3d_extension(GetBitContext *gb, AVCodecContext *avctx, + HEVCPPS *pps, HEVCSPS *sps) +{ + unsigned int pps_depth_layers_minus1; + int i, j; + + if (get_bits1(gb)) { // dlts_present_flag + pps_depth_layers_minus1 = get_bits(gb, 6); + pps->pps_bit_depth_for_depth_layers_minus8 = get_bits(gb, 4); + for (i = 0; i <= pps_depth_layers_minus1; i++) { + if (get_bits1(gb)) { // dlt_flag[i] + if (!get_bits1(gb)) { // dlt_pred_flag[i] + if (get_bits1(gb)) { // dlt_val_flags_present_flag[i] + for (j = 0; j <= ((1 << (pps->pps_bit_depth_for_depth_layers_minus8 + 8)) - 1); j++) + skip_bits1(gb); // dlt_value_flag[i][j] + } else + delta_dlt(gb, pps); + } + } + } + } + + return 0; +} + +static int pps_range_extensions(GetBitContext *gb, AVCodecContext *avctx, + HEVCPPS *pps, HEVCSPS *sps) +{ if (pps->transform_skip_enabled_flag) { pps->log2_max_transform_skip_block_size = get_ue_golomb_long(gb) + 2; } @@ -1282,7 +1489,7 @@ static int pps_range_extensions(GetBitContext *gb, AVCodecContext *avctx, "chroma_qp_offset_list_len_minus1 shall be in the range [0, 5].\n"); return AVERROR_INVALIDDATA; } - for (i = 0; i <= pps->chroma_qp_offset_list_len_minus1; i++) { + for (int i = 0; i <= pps->chroma_qp_offset_list_len_minus1; i++) { pps->cb_qp_offset_list[i] = get_se_golomb_long(gb); if (pps->cb_qp_offset_list[i]) { av_log(avctx, AV_LOG_WARNING, @@ -1306,6 +1513,47 @@ static int pps_range_extensions(GetBitContext *gb, AVCodecContext *avctx, return(0); } +static int pps_scc_extension(GetBitContext *gb, AVCodecContext *avctx, + HEVCPPS *pps, HEVCSPS *sps) +{ + int num_comps, ret; + + pps->pps_curr_pic_ref_enabled_flag = get_bits1(gb); + if (pps->residual_adaptive_colour_transform_enabled_flag = get_bits1(gb)) { + pps->pps_slice_act_qp_offsets_present_flag = get_bits1(gb); + pps->pps_act_y_qp_offset = get_se_golomb_long(gb) - 5; + pps->pps_act_cb_qp_offset = get_se_golomb_long(gb) - 5; + pps->pps_act_cr_qp_offset = get_se_golomb_long(gb) - 3; + +#define CHECK_QP_OFFSET(name) (pps->pps_act_ ## name ## _qp_offset <= -12 || \ + pps->pps_act_ ## name ## _qp_offset >= 12) + ret = CHECK_QP_OFFSET(y) || CHECK_QP_OFFSET(cb) || CHECK_QP_OFFSET(cr); +#undef CHECK_QP_OFFSET + if (ret) { + av_log(avctx, AV_LOG_ERROR, + "PpsActQpOffsetY/Cb/Cr shall be in the range of [-12, 12].\n"); + return AVERROR_INVALIDDATA; + } + } + + if (pps->pps_palette_predictor_initializers_present_flag = get_bits1(gb)) { + if ((pps->pps_num_palette_predictor_initializers = get_ue_golomb_long(gb)) > 0) { + pps->monochrome_palette_flag = get_bits1(gb); + pps->luma_bit_depth_entry_minus8 = get_ue_golomb_long(gb); + if (!pps->monochrome_palette_flag) + pps->chroma_bit_depth_entry_minus8 = get_ue_golomb_long(gb); + num_comps = pps->monochrome_palette_flag ? 1 : 3; + for (int comp = 0; comp < num_comps; comp++) + for (int i = 0; i < pps->pps_num_palette_predictor_initializers; i++) + pps->pps_palette_predictor_initializer[comp][i] = + get_bits(gb, 8 + (!comp ? pps->luma_bit_depth_entry_minus8 : + pps->chroma_bit_depth_entry_minus8)); + } + } + + return 0; +} + static inline int setup_pps(AVCodecContext *avctx, GetBitContext *gb, HEVCPPS *pps, HEVCSPS *sps) { @@ -1658,12 +1906,31 @@ int ff_hevc_decode_nal_pps(GetBitContext *gb, AVCodecContext *avctx, pps->slice_header_extension_present_flag = get_bits1(gb); if (get_bits1(gb)) { // pps_extension_present_flag - pps->pps_range_extensions_flag = get_bits1(gb); - skip_bits(gb, 7); // pps_extension_7bits + pps->pps_range_extensions_flag = get_bits1(gb); + pps->pps_multilayer_extension_flag = get_bits1(gb); + pps->pps_3d_extension_flag = get_bits1(gb); + pps->pps_scc_extension_flag = get_bits1(gb); + skip_bits(gb, 4); // pps_extension_4bits + if (pps->pps_range_extensions_flag) { if ((ret = pps_range_extensions(gb, avctx, pps, sps)) < 0) goto err; } + + if (pps->pps_multilayer_extension_flag) { + if ((ret = pps_multilayer_extension(gb, avctx, pps, sps)) < 0) + goto err; + } + + if (pps->pps_3d_extension_flag) { + if ((ret = pps_3d_extension(gb, avctx, pps, sps)) < 0) + goto err; + } + + if (pps->pps_scc_extension_flag) { + if ((ret = pps_scc_extension(gb, avctx, pps, sps)) < 0) + goto err; + } } ret = setup_pps(avctx, gb, pps, sps); @@ -1671,9 +1938,8 @@ int ff_hevc_decode_nal_pps(GetBitContext *gb, AVCodecContext *avctx, goto err; if (get_bits_left(gb) < 0) { - av_log(avctx, AV_LOG_ERROR, + av_log(avctx, AV_LOG_WARNING, "Overread PPS by %d bits\n", -get_bits_left(gb)); - goto err; } remove_pps(ps, pps_id); diff --git a/libavcodec/hevc_ps.h b/libavcodec/hevc_ps.h index 18894cfed1..1f704108e3 100644 --- a/libavcodec/hevc_ps.h +++ b/libavcodec/hevc_ps.h @@ -210,6 +210,20 @@ typedef struct HEVCSPS { int persistent_rice_adaptation_enabled_flag; int cabac_bypass_alignment_enabled_flag; + int sps_multilayer_extension_flag; + int sps_3d_extension_flag; + + int sps_scc_extension_flag; + int sps_curr_pic_ref_enabled_flag; + int palette_mode_enabled_flag; + int palette_max_size; + int delta_palette_max_predictor_size; + int sps_palette_predictor_initializers_present_flag; + int sps_num_palette_predictor_initializers_minus1; + int sps_palette_predictor_initializer[3][HEVC_MAX_PALETTE_PREDICTOR_SIZE]; + int motion_vector_resolution_control_idc; + int intra_boundary_filtering_disabled_flag; + ///< coded frame dimension in various units int width; int height; @@ -284,6 +298,9 @@ typedef struct HEVCPPS { uint8_t slice_header_extension_present_flag; uint8_t log2_max_transform_skip_block_size; uint8_t pps_range_extensions_flag; + uint8_t pps_multilayer_extension_flag; + uint8_t pps_3d_extension_flag; + uint8_t pps_scc_extension_flag; uint8_t cross_component_prediction_enabled_flag; uint8_t chroma_qp_offset_list_enabled_flag; uint8_t diff_cu_chroma_qp_offset_depth; @@ -293,6 +310,58 @@ typedef struct HEVCPPS { uint8_t log2_sao_offset_scale_luma; uint8_t log2_sao_offset_scale_chroma; + // Multilayer extension parameters + uint8_t poc_reset_info_present_flag; + uint8_t pps_infer_scaling_list_flag; + uint8_t pps_scaling_list_ref_layer_id; + uint16_t num_ref_loc_offsets; + uint8_t ref_loc_offset_layer_id[64]; + uint8_t scaled_ref_layer_offset_present_flag[64]; + int8_t scaled_ref_layer_left_offset[64]; + int8_t scaled_ref_layer_top_offset[64]; + int8_t scaled_ref_layer_right_offset[64]; + int8_t scaled_ref_layer_bottom_offset[64]; + uint8_t ref_region_offset_present_flag[64]; + int8_t ref_region_left_offset[64]; + int8_t ref_region_top_offset[64]; + int8_t ref_region_right_offset[64]; + int8_t ref_region_bottom_offset[64]; + uint8_t resample_phase_set_present_flag[64]; + uint16_t phase_hor_luma[64]; + uint16_t phase_ver_luma[64]; + uint16_t phase_hor_chroma_plus8[64]; + uint16_t phase_ver_chroma_plus8[64]; + uint8_t colour_mapping_enabled_flag; + uint16_t num_cm_ref_layers_minus1; + uint8_t cm_ref_layer_id[63]; + uint8_t cm_octant_depth; + uint8_t cm_y_part_num_log2; + uint16_t luma_bit_depth_cm_input_minus8; + uint16_t chroma_bit_depth_cm_input_minus8; + uint16_t luma_bit_depth_cm_output_minus8; + uint16_t chroma_bit_depth_cm_output_minus8; + uint8_t cm_res_quant_bits; + uint8_t cm_delta_flc_bits_minus1; + int8_t cm_adapt_threshold_u_delta; + int8_t cm_adapt_threshold_v_delta; + + // 3D extension parameters + uint8_t pps_bit_depth_for_depth_layers_minus8; + + // SCC extension parameters + uint8_t pps_curr_pic_ref_enabled_flag; + uint8_t residual_adaptive_colour_transform_enabled_flag; + uint8_t pps_slice_act_qp_offsets_present_flag; + int8_t pps_act_y_qp_offset; // _plus5 + int8_t pps_act_cb_qp_offset; // _plus5 + int8_t pps_act_cr_qp_offset; // _plus3 + uint8_t pps_palette_predictor_initializers_present_flag; + uint8_t pps_num_palette_predictor_initializers; + uint8_t monochrome_palette_flag; + uint8_t luma_bit_depth_entry_minus8; + uint8_t chroma_bit_depth_entry_minus8; + uint8_t pps_palette_predictor_initializer[3][HEVC_MAX_PALETTE_PREDICTOR_SIZE]; + // Inferred parameters unsigned int *column_width; ///< ColumnWidth unsigned int *row_height; ///< RowHeight