2 * RTP parser for HEVC/H.265 payload format (draft version 6)
3 * Copyright (c) 2014 Thomas Volkert <thomas@homer-conferencing.com>
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
23 #include "libavutil/avassert.h"
24 #include "libavutil/avstring.h"
25 #include "libavutil/base64.h"
26 #include "libavcodec/get_bits.h"
30 #include "rtpdec_formats.h"
32 #define RTP_HEVC_PAYLOAD_HEADER_SIZE 2
33 #define RTP_HEVC_FU_HEADER_SIZE 1
34 #define RTP_HEVC_DONL_FIELD_SIZE 2
35 #define RTP_HEVC_DOND_FIELD_SIZE 1
36 #define RTP_HEVC_AP_NALU_LENGTH_FIELD_SIZE 2
37 #define HEVC_SPECIFIED_NAL_UNIT_TYPES 48
39 /* SDP out-of-band signaling data */
40 struct PayloadContext {
43 uint8_t *sps, *pps, *vps, *sei;
44 int sps_size, pps_size, vps_size, sei_size;
47 static const uint8_t start_sequence[] = { 0x00, 0x00, 0x00, 0x01 };
49 static av_cold int hevc_sdp_parse_fmtp_config(AVFormatContext *s,
51 PayloadContext *hevc_data,
52 const char *attr, const char *value)
54 /* profile-space: 0-3 */
55 /* profile-id: 0-31 */
56 if (!strcmp(attr, "profile-id")) {
57 hevc_data->profile_id = atoi(value);
58 av_dlog(s, "SDP: found profile-id: %d\n", hevc_data->profile_id);
63 /* interop-constraints: [base16] */
64 /* profile-compatibility-indicator: [base16] */
65 /* sprop-sub-layer-id: 0-6, defines highest possible value for TID, default: 6 */
66 /* recv-sub-layer-id: 0-6 */
67 /* max-recv-level-id: 0-255 */
68 /* tx-mode: MSM,SSM */
69 /* sprop-vps: [base64] */
70 /* sprop-sps: [base64] */
71 /* sprop-pps: [base64] */
72 /* sprop-sei: [base64] */
73 if (!strcmp(attr, "sprop-vps") || !strcmp(attr, "sprop-sps") ||
74 !strcmp(attr, "sprop-pps") || !strcmp(attr, "sprop-sei")) {
75 uint8_t **data_ptr = NULL;
77 if (!strcmp(attr, "sprop-vps")) {
78 data_ptr = &hevc_data->vps;
79 size_ptr = &hevc_data->vps_size;
80 } else if (!strcmp(attr, "sprop-sps")) {
81 data_ptr = &hevc_data->sps;
82 size_ptr = &hevc_data->sps_size;
83 } else if (!strcmp(attr, "sprop-pps")) {
84 data_ptr = &hevc_data->pps;
85 size_ptr = &hevc_data->pps_size;
86 } else if (!strcmp(attr, "sprop-sei")) {
87 data_ptr = &hevc_data->sei;
88 size_ptr = &hevc_data->sei_size;
92 ff_h264_parse_sprop_parameter_sets(s, data_ptr,
96 /* max-lsr, max-lps, max-cpb, max-dpb, max-br, max-tr, max-tc */
99 /* sprop-max-don-diff: 0-32767
101 When the RTP stream depends on one or more other RTP
102 streams (in this case tx-mode MUST be equal to "MSM" and
103 MSM is in use), this parameter MUST be present and the
104 value MUST be greater than 0.
106 if (!strcmp(attr, "sprop-max-don-diff")) {
108 hevc_data->using_donl_field = 1;
109 av_dlog(s, "Found sprop-max-don-diff in SDP, DON field usage is: %d\n",
110 hevc_data->using_donl_field);
113 /* sprop-depack-buf-nalus: 0-32767 */
114 if (!strcmp(attr, "sprop-depack-buf-nalus")) {
116 hevc_data->using_donl_field = 1;
117 av_dlog(s, "Found sprop-depack-buf-nalus in SDP, DON field usage is: %d\n",
118 hevc_data->using_donl_field);
121 /* sprop-depack-buf-bytes: 0-4294967295 */
123 /* sprop-segmentation-id: 0-3 */
124 /* sprop-spatial-segmentation-idc: [base16] */
125 /* dec-parallel-ca: */
131 static av_cold int hevc_parse_sdp_line(AVFormatContext *ctx, int st_index,
132 PayloadContext *hevc_data, const char *line)
134 AVStream *current_stream;
135 AVCodecContext *codec;
136 const char *sdp_line_ptr = line;
141 current_stream = ctx->streams[st_index];
142 codec = current_stream->codec;
144 if (av_strstart(sdp_line_ptr, "framesize:", &sdp_line_ptr)) {
145 ff_h264_parse_framesize(codec, sdp_line_ptr);
146 } else if (av_strstart(sdp_line_ptr, "fmtp:", &sdp_line_ptr)) {
147 int ret = ff_parse_fmtp(ctx, current_stream, hevc_data, sdp_line_ptr,
148 hevc_sdp_parse_fmtp_config);
149 if (hevc_data->vps_size || hevc_data->sps_size ||
150 hevc_data->pps_size || hevc_data->sei_size) {
151 av_freep(&codec->extradata);
152 codec->extradata_size = hevc_data->vps_size + hevc_data->sps_size +
153 hevc_data->pps_size + hevc_data->sei_size;
154 codec->extradata = av_malloc(codec->extradata_size +
155 FF_INPUT_BUFFER_PADDING_SIZE);
156 if (!codec->extradata) {
157 ret = AVERROR(ENOMEM);
158 codec->extradata_size = 0;
161 memcpy(codec->extradata + pos, hevc_data->vps, hevc_data->vps_size);
162 pos += hevc_data->vps_size;
163 memcpy(codec->extradata + pos, hevc_data->sps, hevc_data->sps_size);
164 pos += hevc_data->sps_size;
165 memcpy(codec->extradata + pos, hevc_data->pps, hevc_data->pps_size);
166 pos += hevc_data->pps_size;
167 memcpy(codec->extradata + pos, hevc_data->sei, hevc_data->sei_size);
168 pos += hevc_data->sei_size;
169 memset(codec->extradata + pos, 0, FF_INPUT_BUFFER_PADDING_SIZE);
172 av_freep(&hevc_data->vps);
173 av_freep(&hevc_data->sps);
174 av_freep(&hevc_data->pps);
175 av_freep(&hevc_data->sei);
176 hevc_data->vps_size = 0;
177 hevc_data->sps_size = 0;
178 hevc_data->pps_size = 0;
179 hevc_data->sei_size = 0;
187 static int hevc_handle_packet(AVFormatContext *ctx, PayloadContext *rtp_hevc_ctx,
188 AVStream *st, AVPacket *pkt, uint32_t *timestamp,
189 const uint8_t *buf, int len, uint16_t seq,
192 const uint8_t *rtp_pl = buf;
193 int tid, lid, nal_type;
194 int first_fragment, last_fragment, fu_type;
195 uint8_t new_nal_header[2];
198 /* sanity check for size of input packet: 1 byte payload at least */
199 if (len < RTP_HEVC_PAYLOAD_HEADER_SIZE + 1) {
200 av_log(ctx, AV_LOG_ERROR, "Too short RTP/HEVC packet, got %d bytes\n", len);
201 return AVERROR_INVALIDDATA;
205 * decode the HEVC payload header according to section 4 of draft version 6:
208 * 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5
209 * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
210 * |F| Type | LayerId | TID |
211 * +-------------+-----------------+
213 * Forbidden zero (F): 1 bit
214 * NAL unit type (Type): 6 bits
215 * NUH layer ID (LayerId): 6 bits
216 * NUH temporal ID plus 1 (TID): 3 bits
218 nal_type = (buf[0] >> 1) & 0x3f;
219 lid = ((buf[0] << 5) & 0x20) | ((buf[1] >> 3) & 0x1f);
222 /* sanity check for correct layer ID */
224 /* future scalable or 3D video coding extensions */
225 avpriv_report_missing_feature(ctx, "Multi-layer HEVC coding\n");
226 return AVERROR_PATCHWELCOME;
229 /* sanity check for correct temporal ID */
231 av_log(ctx, AV_LOG_ERROR, "Illegal temporal ID in RTP/HEVC packet\n");
232 return AVERROR_INVALIDDATA;
235 /* sanity check for correct NAL unit type */
237 av_log(ctx, AV_LOG_ERROR, "Unsupported (HEVC) NAL type (%d)\n", nal_type);
238 return AVERROR_INVALIDDATA;
242 /* video parameter set (VPS) */
244 /* sequence parameter set (SPS) */
246 /* picture parameter set (PPS) */
248 /* supplemental enhancement information (SEI) */
250 /* single NAL unit packet */
252 /* create A/V packet */
253 if ((res = av_new_packet(pkt, sizeof(start_sequence) + len)) < 0)
255 /* A/V packet: copy start sequence */
256 memcpy(pkt->data, start_sequence, sizeof(start_sequence));
257 /* A/V packet: copy NAL unit data */
258 memcpy(pkt->data + sizeof(start_sequence), buf, len);
261 /* aggregated packet (AP) - with two or more NAL units */
263 /* pass the HEVC payload header */
264 buf += RTP_HEVC_PAYLOAD_HEADER_SIZE;
265 len -= RTP_HEVC_PAYLOAD_HEADER_SIZE;
267 /* pass the HEVC DONL field */
268 if (rtp_hevc_ctx->using_donl_field) {
269 buf += RTP_HEVC_DONL_FIELD_SIZE;
270 len -= RTP_HEVC_DONL_FIELD_SIZE;
273 res = ff_h264_handle_aggregated_packet(ctx, rtp_hevc_ctx, pkt, buf, len,
274 rtp_hevc_ctx->using_donl_field ?
275 RTP_HEVC_DOND_FIELD_SIZE : 0,
280 /* fragmentation unit (FU) */
282 /* pass the HEVC payload header */
283 buf += RTP_HEVC_PAYLOAD_HEADER_SIZE;
284 len -= RTP_HEVC_PAYLOAD_HEADER_SIZE;
287 * decode the FU header
294 * Start fragment (S): 1 bit
295 * End fragment (E): 1 bit
298 first_fragment = buf[0] & 0x80;
299 last_fragment = buf[0] & 0x40;
300 fu_type = buf[0] & 0x3f;
302 /* pass the HEVC FU header */
303 buf += RTP_HEVC_FU_HEADER_SIZE;
304 len -= RTP_HEVC_FU_HEADER_SIZE;
306 /* pass the HEVC DONL field */
307 if (rtp_hevc_ctx->using_donl_field) {
308 buf += RTP_HEVC_DONL_FIELD_SIZE;
309 len -= RTP_HEVC_DONL_FIELD_SIZE;
312 av_dlog(ctx, " FU type %d with %d bytes\n", fu_type, len);
314 /* sanity check for size of input packet: 1 byte payload at least */
317 av_log(ctx, AV_LOG_ERROR,
318 "Too short RTP/HEVC packet, got %d bytes of NAL unit type %d\n",
320 return AVERROR_INVALIDDATA;
322 return AVERROR(EAGAIN);
326 if (first_fragment && last_fragment) {
327 av_log(ctx, AV_LOG_ERROR, "Illegal combination of S and E bit in RTP/HEVC packet\n");
328 return AVERROR_INVALIDDATA;
331 new_nal_header[0] = (rtp_pl[0] & 0x81) | (fu_type << 1);
332 new_nal_header[1] = rtp_pl[1];
334 res = ff_h264_handle_frag_packet(pkt, buf, len, first_fragment,
335 new_nal_header, sizeof(new_nal_header));
340 /* Temporal scalability control information (TSCI) */
341 avpriv_report_missing_feature(ctx, "PACI packets for RTP/HEVC\n");
342 res = AVERROR_PATCHWELCOME;
346 pkt->stream_index = st->index;
351 RTPDynamicProtocolHandler ff_hevc_dynamic_handler = {
353 .codec_type = AVMEDIA_TYPE_VIDEO,
354 .codec_id = AV_CODEC_ID_HEVC,
355 .need_parsing = AVSTREAM_PARSE_FULL,
356 .priv_data_size = sizeof(PayloadContext),
357 .parse_sdp_a_line = hevc_parse_sdp_line,
358 .parse_packet = hevc_handle_packet,