avformat/mpeg: demux ivtv captions
[ffmpeg.git] / libavcodec / h264_sei.c
blob15a52322095bf287590a3b4b69fea4406acf7ace
1 /*
2 * H.26L/H.264/AVC/JVT/14496-10/... SEI decoding
3 * Copyright (c) 2003 Michael Niedermayer <michaelni@gmx.at>
5 * This file is part of FFmpeg.
7 * FFmpeg is free software; you can redistribute it and/or
8 * modify it under the terms of the GNU Lesser General Public
9 * License as published by the Free Software Foundation; either
10 * version 2.1 of the License, or (at your option) any later version.
12 * FFmpeg is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15 * Lesser General Public License for more details.
17 * You should have received a copy of the GNU Lesser General Public
18 * License along with FFmpeg; if not, write to the Free Software
19 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
22 /**
23 * @file
24 * H.264 / AVC / MPEG-4 part10 SEI decoding.
25 * @author Michael Niedermayer <michaelni@gmx.at>
28 #include <limits.h>
29 #include <stdio.h>
30 #include <string.h>
31 #include "libavutil/error.h"
32 #include "libavutil/log.h"
33 #include "libavutil/macros.h"
34 #include "libavutil/mem.h"
35 #include "bytestream.h"
36 #include "get_bits.h"
37 #include "golomb.h"
38 #include "h264_ps.h"
39 #include "h264_sei.h"
40 #include "sei.h"
42 #define AVERROR_PS_NOT_FOUND FFERRTAG(0xF8,'?','P','S')
44 static const uint8_t sei_num_clock_ts_table[9] = {
45 1, 1, 1, 2, 2, 3, 3, 2, 3
48 void ff_h264_sei_uninit(H264SEIContext *h)
50 h->recovery_point.recovery_frame_cnt = -1;
52 h->picture_timing.dpb_output_delay = 0;
53 h->picture_timing.cpb_removal_delay = -1;
55 h->picture_timing.present = 0;
56 h->buffering_period.present = 0;
57 h->common.frame_packing.present = 0;
58 h->common.display_orientation.present = 0;
59 h->common.afd.present = 0;
61 ff_h2645_sei_reset(&h->common);
64 int ff_h264_sei_process_picture_timing(H264SEIPictureTiming *h, const SPS *sps,
65 void *logctx)
67 GetBitContext gb;
68 av_unused int ret;
70 ret = init_get_bits8(&gb, h->payload, h->payload_size_bytes);
71 av_assert1(ret >= 0);
73 if (sps->nal_hrd_parameters_present_flag ||
74 sps->vcl_hrd_parameters_present_flag) {
75 h->cpb_removal_delay = get_bits_long(&gb, sps->cpb_removal_delay_length);
76 h->dpb_output_delay = get_bits_long(&gb, sps->dpb_output_delay_length);
78 if (sps->pic_struct_present_flag) {
79 unsigned int i, num_clock_ts;
81 h->pic_struct = get_bits(&gb, 4);
82 h->ct_type = 0;
84 if (h->pic_struct > H264_SEI_PIC_STRUCT_FRAME_TRIPLING)
85 return AVERROR_INVALIDDATA;
87 num_clock_ts = sei_num_clock_ts_table[h->pic_struct];
88 h->timecode_cnt = 0;
89 for (i = 0; i < num_clock_ts; i++) {
90 if (get_bits(&gb, 1)) { /* clock_timestamp_flag */
91 H264SEITimeCode *tc = &h->timecode[h->timecode_cnt++];
92 unsigned int full_timestamp_flag;
93 unsigned int counting_type, cnt_dropped_flag;
94 h->ct_type |= 1 << get_bits(&gb, 2);
95 skip_bits(&gb, 1); /* nuit_field_based_flag */
96 counting_type = get_bits(&gb, 5); /* counting_type */
97 full_timestamp_flag = get_bits(&gb, 1);
98 skip_bits(&gb, 1); /* discontinuity_flag */
99 cnt_dropped_flag = get_bits(&gb, 1); /* cnt_dropped_flag */
100 if (cnt_dropped_flag && counting_type > 1 && counting_type < 7)
101 tc->dropframe = 1;
102 tc->frame = get_bits(&gb, 8); /* n_frames */
103 if (full_timestamp_flag) {
104 tc->full = 1;
105 tc->seconds = get_bits(&gb, 6); /* seconds_value 0..59 */
106 tc->minutes = get_bits(&gb, 6); /* minutes_value 0..59 */
107 tc->hours = get_bits(&gb, 5); /* hours_value 0..23 */
108 } else {
109 tc->seconds = tc->minutes = tc->hours = tc->full = 0;
110 if (get_bits(&gb, 1)) { /* seconds_flag */
111 tc->seconds = get_bits(&gb, 6);
112 if (get_bits(&gb, 1)) { /* minutes_flag */
113 tc->minutes = get_bits(&gb, 6);
114 if (get_bits(&gb, 1)) /* hours_flag */
115 tc->hours = get_bits(&gb, 5);
120 if (sps->time_offset_length > 0)
121 skip_bits(&gb,
122 sps->time_offset_length); /* time_offset */
126 av_log(logctx, AV_LOG_DEBUG, "ct_type:%X pic_struct:%d\n",
127 h->ct_type, h->pic_struct);
130 return 0;
133 static int decode_picture_timing(H264SEIPictureTiming *h, GetByteContext *gb,
134 void *logctx)
136 int size = bytestream2_get_bytes_left(gb);
138 if (size > sizeof(h->payload)) {
139 av_log(logctx, AV_LOG_ERROR, "Picture timing SEI payload too large\n");
140 return AVERROR_INVALIDDATA;
142 bytestream2_get_bufferu(gb, h->payload, size);
144 h->payload_size_bytes = size;
146 h->present = 1;
147 return 0;
150 static int decode_recovery_point(H264SEIRecoveryPoint *h, GetBitContext *gb, void *logctx)
152 unsigned recovery_frame_cnt = get_ue_golomb_long(gb);
154 if (recovery_frame_cnt >= (1<<MAX_LOG2_MAX_FRAME_NUM)) {
155 av_log(logctx, AV_LOG_ERROR, "recovery_frame_cnt %u is out of range\n", recovery_frame_cnt);
156 return AVERROR_INVALIDDATA;
159 h->recovery_frame_cnt = recovery_frame_cnt;
160 /* 1b exact_match_flag,
161 * 1b broken_link_flag,
162 * 2b changing_slice_group_idc */
163 skip_bits(gb, 4);
165 return 0;
168 static int decode_buffering_period(H264SEIBufferingPeriod *h, GetBitContext *gb,
169 const H264ParamSets *ps, void *logctx)
171 unsigned int sps_id;
172 int sched_sel_idx;
173 const SPS *sps;
175 sps_id = get_ue_golomb_31(gb);
176 if (sps_id > 31 || !ps->sps_list[sps_id]) {
177 av_log(logctx, AV_LOG_ERROR,
178 "non-existing SPS %d referenced in buffering period\n", sps_id);
179 return sps_id > 31 ? AVERROR_INVALIDDATA : AVERROR_PS_NOT_FOUND;
181 sps = ps->sps_list[sps_id];
183 // NOTE: This is really so duplicated in the standard... See H.264, D.1.1
184 if (sps->nal_hrd_parameters_present_flag) {
185 for (sched_sel_idx = 0; sched_sel_idx < sps->cpb_cnt; sched_sel_idx++) {
186 h->initial_cpb_removal_delay[sched_sel_idx] =
187 get_bits_long(gb, sps->initial_cpb_removal_delay_length);
188 // initial_cpb_removal_delay_offset
189 skip_bits(gb, sps->initial_cpb_removal_delay_length);
192 if (sps->vcl_hrd_parameters_present_flag) {
193 for (sched_sel_idx = 0; sched_sel_idx < sps->cpb_cnt; sched_sel_idx++) {
194 h->initial_cpb_removal_delay[sched_sel_idx] =
195 get_bits_long(gb, sps->initial_cpb_removal_delay_length);
196 // initial_cpb_removal_delay_offset
197 skip_bits(gb, sps->initial_cpb_removal_delay_length);
201 h->present = 1;
202 return 0;
205 static int decode_green_metadata(H264SEIGreenMetaData *h, GetByteContext *gb)
207 h->green_metadata_type = bytestream2_get_byte(gb);
209 if (h->green_metadata_type == 0) {
210 h->period_type = bytestream2_get_byte(gb);
212 if (h->period_type == 2)
213 h->num_seconds = bytestream2_get_be16(gb);
214 else if (h->period_type == 3)
215 h->num_pictures = bytestream2_get_be16(gb);
217 h->percent_non_zero_macroblocks = bytestream2_get_byte(gb);
218 h->percent_intra_coded_macroblocks = bytestream2_get_byte(gb);
219 h->percent_six_tap_filtering = bytestream2_get_byte(gb);
220 h->percent_alpha_point_deblocking_instance = bytestream2_get_byte(gb);
222 } else if (h->green_metadata_type == 1) {
223 h->xsd_metric_type = bytestream2_get_byte(gb);
224 h->xsd_metric_value = bytestream2_get_be16(gb);
227 return 0;
230 int ff_h264_sei_decode(H264SEIContext *h, GetBitContext *gb,
231 const H264ParamSets *ps, void *logctx)
233 GetByteContext gbyte;
234 int master_ret = 0;
236 av_assert1((get_bits_count(gb) % 8) == 0);
237 bytestream2_init(&gbyte, gb->buffer + get_bits_count(gb) / 8,
238 get_bits_left(gb) / 8);
240 while (bytestream2_get_bytes_left(&gbyte) > 2 && bytestream2_peek_ne16(&gbyte)) {
241 GetByteContext gbyte_payload;
242 GetBitContext gb_payload;
243 int type = 0;
244 unsigned size = 0;
245 int ret = 0;
247 do {
248 if (bytestream2_get_bytes_left(&gbyte) <= 0)
249 return AVERROR_INVALIDDATA;
250 type += bytestream2_peek_byteu(&gbyte);
251 } while (bytestream2_get_byteu(&gbyte) == 255);
253 do {
254 if (bytestream2_get_bytes_left(&gbyte) <= 0)
255 return AVERROR_INVALIDDATA;
256 size += bytestream2_peek_byteu(&gbyte);
257 } while (bytestream2_get_byteu(&gbyte) == 255);
259 if (size > bytestream2_get_bytes_left(&gbyte)) {
260 av_log(logctx, AV_LOG_ERROR, "SEI type %d size %d truncated at %d\n",
261 type, size, bytestream2_get_bytes_left(&gbyte));
262 return AVERROR_INVALIDDATA;
265 bytestream2_init (&gbyte_payload, gbyte.buffer, size);
266 ret = init_get_bits8(&gb_payload, gbyte.buffer, size);
267 if (ret < 0)
268 return ret;
270 switch (type) {
271 case SEI_TYPE_PIC_TIMING: // Picture timing SEI
272 ret = decode_picture_timing(&h->picture_timing, &gbyte_payload, logctx);
273 break;
274 case SEI_TYPE_RECOVERY_POINT:
275 ret = decode_recovery_point(&h->recovery_point, &gb_payload, logctx);
276 break;
277 case SEI_TYPE_BUFFERING_PERIOD:
278 ret = decode_buffering_period(&h->buffering_period, &gb_payload, ps, logctx);
279 break;
280 case SEI_TYPE_GREEN_METADATA:
281 ret = decode_green_metadata(&h->green_metadata, &gbyte_payload);
282 break;
283 default:
284 ret = ff_h2645_sei_message_decode(&h->common, type, AV_CODEC_ID_H264,
285 &gb_payload, &gbyte_payload, logctx);
286 if (ret == FF_H2645_SEI_MESSAGE_UNHANDLED)
287 av_log(logctx, AV_LOG_DEBUG, "unknown SEI type %d\n", type);
289 if (ret < 0 && ret != AVERROR_PS_NOT_FOUND)
290 return ret;
291 if (ret < 0)
292 master_ret = ret;
294 if (get_bits_left(&gb_payload) < 0) {
295 av_log(logctx, AV_LOG_WARNING, "SEI type %d overread by %d bits\n",
296 type, -get_bits_left(&gb_payload));
299 bytestream2_skipu(&gbyte, size);
302 return master_ret;
305 const char *ff_h264_sei_stereo_mode(const H2645SEIFramePacking *h)
307 if (h->arrangement_cancel_flag == 0) {
308 switch (h->arrangement_type) {
309 case SEI_FPA_H264_TYPE_CHECKERBOARD:
310 if (h->content_interpretation_type == 2)
311 return "checkerboard_rl";
312 else
313 return "checkerboard_lr";
314 case SEI_FPA_H264_TYPE_INTERLEAVE_COLUMN:
315 if (h->content_interpretation_type == 2)
316 return "col_interleaved_rl";
317 else
318 return "col_interleaved_lr";
319 case SEI_FPA_H264_TYPE_INTERLEAVE_ROW:
320 if (h->content_interpretation_type == 2)
321 return "row_interleaved_rl";
322 else
323 return "row_interleaved_lr";
324 case SEI_FPA_TYPE_SIDE_BY_SIDE:
325 if (h->content_interpretation_type == 2)
326 return "right_left";
327 else
328 return "left_right";
329 case SEI_FPA_TYPE_TOP_BOTTOM:
330 if (h->content_interpretation_type == 2)
331 return "bottom_top";
332 else
333 return "top_bottom";
334 case SEI_FPA_TYPE_INTERLEAVE_TEMPORAL:
335 if (h->content_interpretation_type == 2)
336 return "block_rl";
337 else
338 return "block_lr";
339 case SEI_FPA_H264_TYPE_2D:
340 default:
341 return "mono";
343 } else if (h->arrangement_cancel_flag == 1) {
344 return "mono";
345 } else {
346 return NULL;