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 Libav.
7 * Libav 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 * Libav 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 Libav; if not, write to the Free Software
19 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
22 #include "libavutil/avstring.h"
23 #include "libavutil/base64.h"
27 #include "rtpdec_formats.h"
29 #define RTP_HEVC_PAYLOAD_HEADER_SIZE 2
30 #define RTP_HEVC_FU_HEADER_SIZE 1
31 #define RTP_HEVC_DONL_FIELD_SIZE 2
32 #define RTP_HEVC_DOND_FIELD_SIZE 1
33 #define HEVC_SPECIFIED_NAL_UNIT_TYPES 48
35 /* SDP out-of-band signaling data */
36 struct PayloadContext {
39 uint8_t *sps, *pps, *vps, *sei;
40 int sps_size, pps_size, vps_size, sei_size;
43 static const uint8_t start_sequence[] = { 0x00, 0x00, 0x00, 0x01 };
45 static av_cold int hevc_sdp_parse_fmtp_config(AVFormatContext *s,
47 PayloadContext *hevc_data,
48 const char *attr, const char *value)
50 /* profile-space: 0-3 */
51 /* profile-id: 0-31 */
52 if (!strcmp(attr, "profile-id")) {
53 hevc_data->profile_id = atoi(value);
54 av_log(s, AV_LOG_TRACE, "SDP: found profile-id: %d\n", hevc_data->profile_id);
59 /* interop-constraints: [base16] */
60 /* profile-compatibility-indicator: [base16] */
61 /* sprop-sub-layer-id: 0-6, defines highest possible value for TID, default: 6 */
62 /* recv-sub-layer-id: 0-6 */
63 /* max-recv-level-id: 0-255 */
64 /* tx-mode: MSM,SSM */
65 /* sprop-vps: [base64] */
66 /* sprop-sps: [base64] */
67 /* sprop-pps: [base64] */
68 /* sprop-sei: [base64] */
69 if (!strcmp(attr, "sprop-vps") || !strcmp(attr, "sprop-sps") ||
70 !strcmp(attr, "sprop-pps") || !strcmp(attr, "sprop-sei")) {
71 uint8_t **data_ptr = NULL;
73 if (!strcmp(attr, "sprop-vps")) {
74 data_ptr = &hevc_data->vps;
75 size_ptr = &hevc_data->vps_size;
76 } else if (!strcmp(attr, "sprop-sps")) {
77 data_ptr = &hevc_data->sps;
78 size_ptr = &hevc_data->sps_size;
79 } else if (!strcmp(attr, "sprop-pps")) {
80 data_ptr = &hevc_data->pps;
81 size_ptr = &hevc_data->pps_size;
82 } else if (!strcmp(attr, "sprop-sei")) {
83 data_ptr = &hevc_data->sei;
84 size_ptr = &hevc_data->sei_size;
87 ff_h264_parse_sprop_parameter_sets(s, data_ptr,
91 /* max-lsr, max-lps, max-cpb, max-dpb, max-br, max-tr, max-tc */
94 /* sprop-max-don-diff: 0-32767
96 When the RTP stream depends on one or more other RTP
97 streams (in this case tx-mode MUST be equal to "MSM" and
98 MSM is in use), this parameter MUST be present and the
99 value MUST be greater than 0.
101 if (!strcmp(attr, "sprop-max-don-diff")) {
103 hevc_data->using_donl_field = 1;
104 av_log(s, AV_LOG_TRACE, "Found sprop-max-don-diff in SDP, DON field usage is: %d\n",
105 hevc_data->using_donl_field);
108 /* sprop-depack-buf-nalus: 0-32767 */
109 if (!strcmp(attr, "sprop-depack-buf-nalus")) {
111 hevc_data->using_donl_field = 1;
112 av_log(s, AV_LOG_TRACE, "Found sprop-depack-buf-nalus in SDP, DON field usage is: %d\n",
113 hevc_data->using_donl_field);
116 /* sprop-depack-buf-bytes: 0-4294967295 */
118 /* sprop-segmentation-id: 0-3 */
119 /* sprop-spatial-segmentation-idc: [base16] */
120 /* dec-parallel-ca: */
126 static av_cold int hevc_parse_sdp_line(AVFormatContext *ctx, int st_index,
127 PayloadContext *hevc_data, const char *line)
129 AVStream *current_stream;
130 AVCodecParameters *par;
131 const char *sdp_line_ptr = line;
136 current_stream = ctx->streams[st_index];
137 par = current_stream->codecpar;
139 if (av_strstart(sdp_line_ptr, "framesize:", &sdp_line_ptr)) {
140 ff_h264_parse_framesize(par, sdp_line_ptr);
141 } else if (av_strstart(sdp_line_ptr, "fmtp:", &sdp_line_ptr)) {
142 int ret = ff_parse_fmtp(ctx, current_stream, hevc_data, sdp_line_ptr,
143 hevc_sdp_parse_fmtp_config);
144 if (hevc_data->vps_size || hevc_data->sps_size ||
145 hevc_data->pps_size || hevc_data->sei_size) {
146 av_freep(&par->extradata);
147 par->extradata_size = hevc_data->vps_size + hevc_data->sps_size +
148 hevc_data->pps_size + hevc_data->sei_size;
149 par->extradata = av_malloc(par->extradata_size +
150 AV_INPUT_BUFFER_PADDING_SIZE);
151 if (!par->extradata) {
152 ret = AVERROR(ENOMEM);
153 par->extradata_size = 0;
156 memcpy(par->extradata + pos, hevc_data->vps, hevc_data->vps_size);
157 pos += hevc_data->vps_size;
158 memcpy(par->extradata + pos, hevc_data->sps, hevc_data->sps_size);
159 pos += hevc_data->sps_size;
160 memcpy(par->extradata + pos, hevc_data->pps, hevc_data->pps_size);
161 pos += hevc_data->pps_size;
162 memcpy(par->extradata + pos, hevc_data->sei, hevc_data->sei_size);
163 pos += hevc_data->sei_size;
164 memset(par->extradata + pos, 0, AV_INPUT_BUFFER_PADDING_SIZE);
167 av_freep(&hevc_data->vps);
168 av_freep(&hevc_data->sps);
169 av_freep(&hevc_data->pps);
170 av_freep(&hevc_data->sei);
171 hevc_data->vps_size = 0;
172 hevc_data->sps_size = 0;
173 hevc_data->pps_size = 0;
174 hevc_data->sei_size = 0;
182 static int hevc_handle_packet(AVFormatContext *ctx, PayloadContext *rtp_hevc_ctx,
183 AVStream *st, AVPacket *pkt, uint32_t *timestamp,
184 const uint8_t *buf, int len, uint16_t seq,
187 const uint8_t *rtp_pl = buf;
188 int tid, lid, nal_type;
189 int first_fragment, last_fragment, fu_type;
190 uint8_t new_nal_header[2];
193 /* sanity check for size of input packet: 1 byte payload at least */
194 if (len < RTP_HEVC_PAYLOAD_HEADER_SIZE + 1) {
195 av_log(ctx, AV_LOG_ERROR, "Too short RTP/HEVC packet, got %d bytes\n", len);
196 return AVERROR_INVALIDDATA;
200 * decode the HEVC payload header according to section 4 of draft version 6:
203 * 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5
204 * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
205 * |F| Type | LayerId | TID |
206 * +-------------+-----------------+
208 * Forbidden zero (F): 1 bit
209 * NAL unit type (Type): 6 bits
210 * NUH layer ID (LayerId): 6 bits
211 * NUH temporal ID plus 1 (TID): 3 bits
213 nal_type = (buf[0] >> 1) & 0x3f;
214 lid = ((buf[0] << 5) & 0x20) | ((buf[1] >> 3) & 0x1f);
217 /* sanity check for correct layer ID */
219 /* future scalable or 3D video coding extensions */
220 avpriv_report_missing_feature(ctx, "Multi-layer HEVC coding");
221 return AVERROR_PATCHWELCOME;
224 /* sanity check for correct temporal ID */
226 av_log(ctx, AV_LOG_ERROR, "Illegal temporal ID in RTP/HEVC packet\n");
227 return AVERROR_INVALIDDATA;
230 /* sanity check for correct NAL unit type */
232 av_log(ctx, AV_LOG_ERROR, "Unsupported (HEVC) NAL type (%d)\n", nal_type);
233 return AVERROR_INVALIDDATA;
237 /* video parameter set (VPS) */
239 /* sequence parameter set (SPS) */
241 /* picture parameter set (PPS) */
243 /* supplemental enhancement information (SEI) */
245 /* single NAL unit packet */
247 /* create A/V packet */
248 if ((res = av_new_packet(pkt, sizeof(start_sequence) + len)) < 0)
250 /* A/V packet: copy start sequence */
251 memcpy(pkt->data, start_sequence, sizeof(start_sequence));
252 /* A/V packet: copy NAL unit data */
253 memcpy(pkt->data + sizeof(start_sequence), buf, len);
256 /* aggregated packet (AP) - with two or more NAL units */
258 /* pass the HEVC payload header */
259 buf += RTP_HEVC_PAYLOAD_HEADER_SIZE;
260 len -= RTP_HEVC_PAYLOAD_HEADER_SIZE;
262 /* pass the HEVC DONL field */
263 if (rtp_hevc_ctx->using_donl_field) {
264 buf += RTP_HEVC_DONL_FIELD_SIZE;
265 len -= RTP_HEVC_DONL_FIELD_SIZE;
268 res = ff_h264_handle_aggregated_packet(ctx, pkt, buf, len,
269 rtp_hevc_ctx->using_donl_field ?
270 RTP_HEVC_DOND_FIELD_SIZE : 0,
275 /* fragmentation unit (FU) */
277 /* pass the HEVC payload header */
278 buf += RTP_HEVC_PAYLOAD_HEADER_SIZE;
279 len -= RTP_HEVC_PAYLOAD_HEADER_SIZE;
282 * decode the FU header
289 * Start fragment (S): 1 bit
290 * End fragment (E): 1 bit
293 first_fragment = buf[0] & 0x80;
294 last_fragment = buf[0] & 0x40;
295 fu_type = buf[0] & 0x3f;
297 /* pass the HEVC FU header */
298 buf += RTP_HEVC_FU_HEADER_SIZE;
299 len -= RTP_HEVC_FU_HEADER_SIZE;
301 /* pass the HEVC DONL field */
302 if (rtp_hevc_ctx->using_donl_field) {
303 buf += RTP_HEVC_DONL_FIELD_SIZE;
304 len -= RTP_HEVC_DONL_FIELD_SIZE;
307 av_log(ctx, AV_LOG_TRACE, " FU type %d with %d bytes\n", fu_type, len);
310 /* sanity check for size of input packet: 1 byte payload at least */
311 av_log(ctx, AV_LOG_ERROR,
312 "Too short RTP/HEVC packet, got %d bytes of NAL unit type %d\n",
314 return AVERROR_INVALIDDATA;
317 if (first_fragment && last_fragment) {
318 av_log(ctx, AV_LOG_ERROR, "Illegal combination of S and E bit in RTP/HEVC packet\n");
319 return AVERROR_INVALIDDATA;
322 new_nal_header[0] = (rtp_pl[0] & 0x81) | (fu_type << 1);
323 new_nal_header[1] = rtp_pl[1];
325 res = ff_h264_handle_frag_packet(pkt, buf, len, first_fragment,
326 new_nal_header, sizeof(new_nal_header));
331 /* Temporal scalability control information (TSCI) */
332 avpriv_report_missing_feature(ctx, "PACI packets for RTP/HEVC");
333 res = AVERROR_PATCHWELCOME;
337 pkt->stream_index = st->index;
342 RTPDynamicProtocolHandler ff_hevc_dynamic_handler = {
344 .codec_type = AVMEDIA_TYPE_VIDEO,
345 .codec_id = AV_CODEC_ID_HEVC,
346 .need_parsing = AVSTREAM_PARSE_FULL,
347 .priv_data_size = sizeof(PayloadContext),
348 .parse_sdp_a_line = hevc_parse_sdp_line,
349 .parse_packet = hevc_handle_packet,