2 * This file is part of FFmpeg.
4 * FFmpeg is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU Lesser General Public
6 * License as published by the Free Software Foundation; either
7 * version 2.1 of the License, or (at your option) any later version.
9 * FFmpeg is distributed in the hope that it will be useful,
10 * but WITHOUT ANY WARRANTY; without even the implied warranty of
11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
12 * Lesser General Public License for more details.
14 * You should have received a copy of the GNU Lesser General Public
15 * License along with FFmpeg; if not, write to the Free Software
16 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
19 #include "libavutil/common.h"
20 #include "libavutil/opt.h"
33 typedef struct H265MetadataContext {
36 CodedBitstreamContext *cbc;
37 CodedBitstreamFragment access_unit;
43 AVRational sample_aspect_ratio;
46 int video_full_range_flag;
48 int transfer_characteristics;
49 int matrix_coefficients;
51 int chroma_sample_loc_type;
54 int poc_proportional_to_timing_flag;
55 int num_ticks_poc_diff_one;
61 } H265MetadataContext;
64 static int h265_metadata_update_vps(AVBSFContext *bsf,
67 H265MetadataContext *ctx = bsf->priv_data;
69 if (ctx->tick_rate.num && ctx->tick_rate.den) {
72 av_reduce(&num, &den, ctx->tick_rate.num, ctx->tick_rate.den,
73 UINT32_MAX > INT_MAX ? UINT32_MAX : INT_MAX);
75 vps->vps_time_scale = num;
76 vps->vps_num_units_in_tick = den;
78 vps->vps_timing_info_present_flag = 1;
80 if (ctx->num_ticks_poc_diff_one > 0) {
81 vps->vps_num_ticks_poc_diff_one_minus1 =
82 ctx->num_ticks_poc_diff_one - 1;
83 vps->vps_poc_proportional_to_timing_flag = 1;
84 } else if (ctx->num_ticks_poc_diff_one == 0) {
85 vps->vps_poc_proportional_to_timing_flag = 0;
92 static int h265_metadata_update_sps(AVBSFContext *bsf,
95 H265MetadataContext *ctx = bsf->priv_data;
97 int crop_unit_x, crop_unit_y;
99 if (ctx->sample_aspect_ratio.num && ctx->sample_aspect_ratio.den) {
101 static const AVRational sar_idc[] = {
102 { 0, 0 }, // Unspecified (never written here).
103 { 1, 1 }, { 12, 11 }, { 10, 11 }, { 16, 11 },
104 { 40, 33 }, { 24, 11 }, { 20, 11 }, { 32, 11 },
105 { 80, 33 }, { 18, 11 }, { 15, 11 }, { 64, 33 },
106 { 160, 99 }, { 4, 3 }, { 3, 2 }, { 2, 1 },
110 av_reduce(&num, &den, ctx->sample_aspect_ratio.num,
111 ctx->sample_aspect_ratio.den, 65535);
113 for (i = 1; i < FF_ARRAY_ELEMS(sar_idc); i++) {
114 if (num == sar_idc[i].num &&
115 den == sar_idc[i].den)
118 if (i == FF_ARRAY_ELEMS(sar_idc)) {
119 sps->vui.aspect_ratio_idc = 255;
120 sps->vui.sar_width = num;
121 sps->vui.sar_height = den;
123 sps->vui.aspect_ratio_idc = i;
125 sps->vui.aspect_ratio_info_present_flag = 1;
129 #define SET_OR_INFER(field, value, present_flag, infer) do { \
133 } else if (!present_flag) \
137 if (ctx->video_format >= 0 ||
138 ctx->video_full_range_flag >= 0 ||
139 ctx->colour_primaries >= 0 ||
140 ctx->transfer_characteristics >= 0 ||
141 ctx->matrix_coefficients >= 0) {
143 SET_OR_INFER(sps->vui.video_format, ctx->video_format,
144 sps->vui.video_signal_type_present_flag, 5);
146 SET_OR_INFER(sps->vui.video_full_range_flag,
147 ctx->video_full_range_flag,
148 sps->vui.video_signal_type_present_flag, 0);
150 if (ctx->colour_primaries >= 0 ||
151 ctx->transfer_characteristics >= 0 ||
152 ctx->matrix_coefficients >= 0) {
154 SET_OR_INFER(sps->vui.colour_primaries,
155 ctx->colour_primaries,
156 sps->vui.colour_description_present_flag, 2);
158 SET_OR_INFER(sps->vui.transfer_characteristics,
159 ctx->transfer_characteristics,
160 sps->vui.colour_description_present_flag, 2);
162 SET_OR_INFER(sps->vui.matrix_coefficients,
163 ctx->matrix_coefficients,
164 sps->vui.colour_description_present_flag, 2);
166 sps->vui.colour_description_present_flag = 1;
168 sps->vui.video_signal_type_present_flag = 1;
172 if (ctx->chroma_sample_loc_type >= 0) {
173 sps->vui.chroma_sample_loc_type_top_field =
174 ctx->chroma_sample_loc_type;
175 sps->vui.chroma_sample_loc_type_bottom_field =
176 ctx->chroma_sample_loc_type;
177 sps->vui.chroma_loc_info_present_flag = 1;
181 if (ctx->tick_rate.num && ctx->tick_rate.den) {
184 av_reduce(&num, &den, ctx->tick_rate.num, ctx->tick_rate.den,
185 UINT32_MAX > INT_MAX ? UINT32_MAX : INT_MAX);
187 sps->vui.vui_time_scale = num;
188 sps->vui.vui_num_units_in_tick = den;
190 sps->vui.vui_timing_info_present_flag = 1;
193 if (ctx->num_ticks_poc_diff_one > 0) {
194 sps->vui.vui_num_ticks_poc_diff_one_minus1 =
195 ctx->num_ticks_poc_diff_one - 1;
196 sps->vui.vui_poc_proportional_to_timing_flag = 1;
197 } else if (ctx->num_ticks_poc_diff_one == 0) {
198 sps->vui.vui_poc_proportional_to_timing_flag = 0;
202 if (sps->separate_colour_plane_flag || sps->chroma_format_idc == 0) {
206 crop_unit_x = 1 + (sps->chroma_format_idc < 3);
207 crop_unit_y = 1 + (sps->chroma_format_idc < 2);
209 #define CROP(border, unit) do { \
210 if (ctx->crop_ ## border >= 0) { \
211 if (ctx->crop_ ## border % unit != 0) { \
212 av_log(bsf, AV_LOG_ERROR, "Invalid value for crop_%s: " \
213 "must be a multiple of %d.\n", #border, unit); \
214 return AVERROR(EINVAL); \
216 sps->conf_win_ ## border ## _offset = \
217 ctx->crop_ ## border / unit; \
218 sps->conformance_window_flag = 1; \
221 CROP(left, crop_unit_x);
222 CROP(right, crop_unit_x);
223 CROP(top, crop_unit_y);
224 CROP(bottom, crop_unit_y);
228 sps->vui_parameters_present_flag = 1;
233 static int h265_metadata_filter(AVBSFContext *bsf, AVPacket *out)
235 H265MetadataContext *ctx = bsf->priv_data;
237 CodedBitstreamFragment *au = &ctx->access_unit;
240 err = ff_bsf_get_packet(bsf, &in);
244 err = ff_cbs_read_packet(ctx->cbc, au, in);
246 av_log(bsf, AV_LOG_ERROR, "Failed to read packet.\n");
250 if (au->nb_units == 0) {
251 av_log(bsf, AV_LOG_ERROR, "No NAL units in packet.\n");
252 err = AVERROR_INVALIDDATA;
256 // If an AUD is present, it must be the first NAL unit.
257 if (au->units[0].type == HEVC_NAL_AUD) {
258 if (ctx->aud == REMOVE)
259 ff_cbs_delete_unit(ctx->cbc, au, 0);
261 if (ctx->aud == INSERT) {
262 H265RawAUD *aud = &ctx->aud_nal;
263 int pic_type = 0, temporal_id = 8, layer_id = 0;
265 for (i = 0; i < au->nb_units; i++) {
266 const H265RawNALUnitHeader *nal = au->units[i].content;
269 if (nal->nuh_temporal_id_plus1 < temporal_id + 1)
270 temporal_id = nal->nuh_temporal_id_plus1 - 1;
272 if (au->units[i].type <= HEVC_NAL_RSV_VCL31) {
273 const H265RawSlice *slice = au->units[i].content;
274 layer_id = nal->nuh_layer_id;
275 if (slice->header.slice_type == HEVC_SLICE_B &&
278 if (slice->header.slice_type == HEVC_SLICE_P &&
284 aud->nal_unit_header = (H265RawNALUnitHeader) {
285 .nal_unit_type = HEVC_NAL_AUD,
286 .nuh_layer_id = layer_id,
287 .nuh_temporal_id_plus1 = temporal_id + 1,
289 aud->pic_type = pic_type;
291 err = ff_cbs_insert_unit_content(ctx->cbc, au,
292 0, HEVC_NAL_AUD, aud, NULL);
294 av_log(bsf, AV_LOG_ERROR, "Failed to insert AUD.\n");
300 for (i = 0; i < au->nb_units; i++) {
301 if (au->units[i].type == HEVC_NAL_VPS) {
302 err = h265_metadata_update_vps(bsf, au->units[i].content);
306 if (au->units[i].type == HEVC_NAL_SPS) {
307 err = h265_metadata_update_sps(bsf, au->units[i].content);
313 err = ff_cbs_write_packet(ctx->cbc, out, au);
315 av_log(bsf, AV_LOG_ERROR, "Failed to write packet.\n");
319 err = av_packet_copy_props(out, in);
325 ff_cbs_fragment_uninit(ctx->cbc, au);
328 av_packet_unref(out);
334 static int h265_metadata_init(AVBSFContext *bsf)
336 H265MetadataContext *ctx = bsf->priv_data;
337 CodedBitstreamFragment *au = &ctx->access_unit;
340 err = ff_cbs_init(&ctx->cbc, AV_CODEC_ID_HEVC, bsf);
344 if (bsf->par_in->extradata) {
345 err = ff_cbs_read_extradata(ctx->cbc, au, bsf->par_in);
347 av_log(bsf, AV_LOG_ERROR, "Failed to read extradata.\n");
351 for (i = 0; i < au->nb_units; i++) {
352 if (au->units[i].type == HEVC_NAL_VPS) {
353 err = h265_metadata_update_vps(bsf, au->units[i].content);
357 if (au->units[i].type == HEVC_NAL_SPS) {
358 err = h265_metadata_update_sps(bsf, au->units[i].content);
364 err = ff_cbs_write_extradata(ctx->cbc, bsf->par_out, au);
366 av_log(bsf, AV_LOG_ERROR, "Failed to write extradata.\n");
373 ff_cbs_fragment_uninit(ctx->cbc, au);
377 static void h265_metadata_close(AVBSFContext *bsf)
379 H265MetadataContext *ctx = bsf->priv_data;
380 ff_cbs_close(&ctx->cbc);
383 #define OFFSET(x) offsetof(H265MetadataContext, x)
384 #define FLAGS (AV_OPT_FLAG_VIDEO_PARAM|AV_OPT_FLAG_BSF_PARAM)
385 static const AVOption h265_metadata_options[] = {
386 { "aud", "Access Unit Delimiter NAL units",
387 OFFSET(aud), AV_OPT_TYPE_INT,
388 { .i64 = PASS }, PASS, REMOVE, FLAGS, "aud" },
389 { "pass", NULL, 0, AV_OPT_TYPE_CONST,
390 { .i64 = PASS }, .flags = FLAGS, .unit = "aud" },
391 { "insert", NULL, 0, AV_OPT_TYPE_CONST,
392 { .i64 = INSERT }, .flags = FLAGS, .unit = "aud" },
393 { "remove", NULL, 0, AV_OPT_TYPE_CONST,
394 { .i64 = REMOVE }, .flags = FLAGS, .unit = "aud" },
396 { "sample_aspect_ratio", "Set sample aspect ratio (table E-1)",
397 OFFSET(sample_aspect_ratio), AV_OPT_TYPE_RATIONAL,
398 { .dbl = 0.0 }, 0, 65535, FLAGS },
400 { "video_format", "Set video format (table E-2)",
401 OFFSET(video_format), AV_OPT_TYPE_INT,
402 { .i64 = -1 }, -1, 7, FLAGS },
403 { "video_full_range_flag", "Set video full range flag",
404 OFFSET(video_full_range_flag), AV_OPT_TYPE_INT,
405 { .i64 = -1 }, -1, 1, FLAGS },
406 { "colour_primaries", "Set colour primaries (table E-3)",
407 OFFSET(colour_primaries), AV_OPT_TYPE_INT,
408 { .i64 = -1 }, -1, 255, FLAGS },
409 { "transfer_characteristics", "Set transfer characteristics (table E-4)",
410 OFFSET(transfer_characteristics), AV_OPT_TYPE_INT,
411 { .i64 = -1 }, -1, 255, FLAGS },
412 { "matrix_coefficients", "Set matrix coefficients (table E-5)",
413 OFFSET(matrix_coefficients), AV_OPT_TYPE_INT,
414 { .i64 = -1 }, -1, 255, FLAGS },
416 { "chroma_sample_loc_type", "Set chroma sample location type (figure E-1)",
417 OFFSET(chroma_sample_loc_type), AV_OPT_TYPE_INT,
418 { .i64 = -1 }, -1, 6, FLAGS },
421 "Set VPS and VUI tick rate (num_units_in_tick / time_scale)",
422 OFFSET(tick_rate), AV_OPT_TYPE_RATIONAL,
423 { .dbl = 0.0 }, 0, UINT_MAX, FLAGS },
424 { "num_ticks_poc_diff_one",
425 "Set VPS and VUI number of ticks per POC increment",
426 OFFSET(num_ticks_poc_diff_one), AV_OPT_TYPE_INT,
427 { .i64 = -1 }, -1, INT_MAX, FLAGS },
429 { "crop_left", "Set left border crop offset",
430 OFFSET(crop_left), AV_OPT_TYPE_INT,
431 { .i64 = -1 }, -1, HEVC_MAX_WIDTH, FLAGS },
432 { "crop_right", "Set right border crop offset",
433 OFFSET(crop_right), AV_OPT_TYPE_INT,
434 { .i64 = -1 }, -1, HEVC_MAX_WIDTH, FLAGS },
435 { "crop_top", "Set top border crop offset",
436 OFFSET(crop_top), AV_OPT_TYPE_INT,
437 { .i64 = -1 }, -1, HEVC_MAX_HEIGHT, FLAGS },
438 { "crop_bottom", "Set bottom border crop offset",
439 OFFSET(crop_bottom), AV_OPT_TYPE_INT,
440 { .i64 = -1 }, -1, HEVC_MAX_HEIGHT, FLAGS },
445 static const AVClass h265_metadata_class = {
446 .class_name = "h265_metadata_bsf",
447 .item_name = av_default_item_name,
448 .option = h265_metadata_options,
449 .version = LIBAVUTIL_VERSION_INT,
452 static const enum AVCodecID h265_metadata_codec_ids[] = {
453 AV_CODEC_ID_HEVC, AV_CODEC_ID_NONE,
456 const AVBitStreamFilter ff_hevc_metadata_bsf = {
457 .name = "hevc_metadata",
458 .priv_data_size = sizeof(H265MetadataContext),
459 .priv_class = &h265_metadata_class,
460 .init = &h265_metadata_init,
461 .close = &h265_metadata_close,
462 .filter = &h265_metadata_filter,
463 .codec_ids = h265_metadata_codec_ids,