2 * RTP H.264 Protocol (RFC3984)
3 * Copyright (c) 2006 Ryan Martell
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
24 * @brief H.264 / RTP Code (RFC3984)
25 * @author Ryan Martell <rdm4@martellventures.com>
29 * This currently supports packetization mode:
30 * Single Nal Unit Mode (0), or
31 * Non-Interleaved Mode (1). It currently does not support
32 * Interleaved Mode (2). (This requires implementing STAP-B, MTAP16, MTAP24,
36 #include "libavutil/attributes.h"
37 #include "libavutil/base64.h"
38 #include "libavutil/intreadwrite.h"
39 #include "libavutil/avstring.h"
43 #include "rtpdec_formats.h"
45 struct PayloadContext {
46 // sdp setup parameters
50 int packetization_mode;
52 int packet_types_received[32];
57 #define COUNT_NAL_TYPE(data, nal) data->packet_types_received[(nal) & 0x1f]++
58 #define NAL_COUNTERS data->packet_types_received
60 #define COUNT_NAL_TYPE(data, nal) do { } while (0)
61 #define NAL_COUNTERS NULL
65 static const uint8_t start_sequence[] = { 0, 0, 0, 1 };
67 static void parse_profile_level_id(AVFormatContext *s,
68 PayloadContext *h264_data,
72 // 6 characters=3 bytes, in hex.
80 profile_idc = strtol(buffer, NULL, 16);
83 profile_iop = strtol(buffer, NULL, 16);
86 level_idc = strtol(buffer, NULL, 16);
88 av_log(s, AV_LOG_DEBUG,
89 "RTP Profile IDC: %x Profile IOP: %x Level: %x\n",
90 profile_idc, profile_iop, level_idc);
91 h264_data->profile_idc = profile_idc;
92 h264_data->profile_iop = profile_iop;
93 h264_data->level_idc = level_idc;
96 int ff_h264_parse_sprop_parameter_sets(AVFormatContext *s,
97 uint8_t **data_ptr, int *size_ptr,
100 char base64packet[1024];
101 uint8_t decoded_packet[1024];
105 char *dst = base64packet;
107 while (*value && *value != ','
108 && (dst - base64packet) < sizeof(base64packet) - 1) {
116 packet_size = av_base64_decode(decoded_packet, base64packet,
117 sizeof(decoded_packet));
118 if (packet_size > 0) {
119 uint8_t *dest = av_realloc(*data_ptr,
120 packet_size + sizeof(start_sequence) +
122 AV_INPUT_BUFFER_PADDING_SIZE);
124 av_log(s, AV_LOG_ERROR,
125 "Unable to allocate memory for extradata!\n");
126 return AVERROR(ENOMEM);
130 memcpy(dest + *size_ptr, start_sequence,
131 sizeof(start_sequence));
132 memcpy(dest + *size_ptr + sizeof(start_sequence),
133 decoded_packet, packet_size);
134 memset(dest + *size_ptr + sizeof(start_sequence) +
135 packet_size, 0, AV_INPUT_BUFFER_PADDING_SIZE);
137 *size_ptr += sizeof(start_sequence) + packet_size;
144 static int sdp_parse_fmtp_config_h264(AVFormatContext *s,
146 PayloadContext *h264_data,
147 const char *attr, const char *value)
149 AVCodecParameters *par = stream->codecpar;
151 if (!strcmp(attr, "packetization-mode")) {
152 av_log(s, AV_LOG_DEBUG, "RTP Packetization Mode: %d\n", atoi(value));
153 h264_data->packetization_mode = atoi(value);
155 * Packetization Mode:
156 * 0 or not present: Single NAL mode (Only nals from 1-23 are allowed)
157 * 1: Non-interleaved Mode: 1-23, 24 (STAP-A), 28 (FU-A) are allowed.
158 * 2: Interleaved Mode: 25 (STAP-B), 26 (MTAP16), 27 (MTAP24), 28 (FU-A),
159 * and 29 (FU-B) are allowed.
161 if (h264_data->packetization_mode > 1)
162 av_log(s, AV_LOG_ERROR,
163 "Interleaved RTP mode is not supported yet.\n");
164 } else if (!strcmp(attr, "profile-level-id")) {
165 if (strlen(value) == 6)
166 parse_profile_level_id(s, h264_data, value);
167 } else if (!strcmp(attr, "sprop-parameter-sets")) {
169 if (value[strlen(value) - 1] == ',') {
170 av_log(s, AV_LOG_WARNING, "Missing PPS in sprop-parameter-sets, ignoring\n");
173 par->extradata_size = 0;
174 av_freep(&par->extradata);
175 ret = ff_h264_parse_sprop_parameter_sets(s, &par->extradata,
176 &par->extradata_size, value);
177 av_log(s, AV_LOG_DEBUG, "Extradata set to %p (size: %d)\n",
178 par->extradata, par->extradata_size);
184 void ff_h264_parse_framesize(AVCodecParameters *par, const char *p)
189 // remove the protocol identifier
190 while (*p && *p == ' ')
191 p++; // strip spaces.
192 while (*p && *p != ' ')
193 p++; // eat protocol identifier
194 while (*p && *p == ' ')
195 p++; // strip trailing spaces.
196 while (*p && *p != '-' && (dst - buf1) < sizeof(buf1) - 1)
200 // a='framesize:96 320-240'
201 // set our parameters
202 par->width = atoi(buf1);
203 par->height = atoi(p + 1); // skip the -
206 int ff_h264_handle_aggregated_packet(AVFormatContext *ctx, PayloadContext *data, AVPacket *pkt,
207 const uint8_t *buf, int len,
208 int skip_between, int *nal_counters,
212 int total_length = 0;
216 // first we are going to figure out the total size
217 for (pass = 0; pass < 2; pass++) {
218 const uint8_t *src = buf;
221 while (src_len > 2) {
222 uint16_t nal_size = AV_RB16(src);
224 // consume the length of the aggregate
228 if (nal_size <= src_len) {
231 total_length += sizeof(start_sequence) + nal_size;
234 memcpy(dst, start_sequence, sizeof(start_sequence));
235 dst += sizeof(start_sequence);
236 memcpy(dst, src, nal_size);
238 nal_counters[(*src) & nal_mask]++;
242 av_log(ctx, AV_LOG_ERROR,
243 "nal size exceeds length: %d %d\n", nal_size, src_len);
244 return AVERROR_INVALIDDATA;
247 // eat what we handled
248 src += nal_size + skip_between;
249 src_len -= nal_size + skip_between;
253 /* now we know the total size of the packet (with the
254 * start sequences added) */
255 if ((ret = av_new_packet(pkt, total_length)) < 0)
264 int ff_h264_handle_frag_packet(AVPacket *pkt, const uint8_t *buf, int len,
265 int start_bit, const uint8_t *nal_header,
272 tot_len += sizeof(start_sequence) + nal_header_len;
273 if ((ret = av_new_packet(pkt, tot_len)) < 0)
276 memcpy(pkt->data + pos, start_sequence, sizeof(start_sequence));
277 pos += sizeof(start_sequence);
278 memcpy(pkt->data + pos, nal_header, nal_header_len);
279 pos += nal_header_len;
281 memcpy(pkt->data + pos, buf, len);
285 static int h264_handle_packet_fu_a(AVFormatContext *ctx, PayloadContext *data, AVPacket *pkt,
286 const uint8_t *buf, int len,
287 int *nal_counters, int nal_mask)
289 uint8_t fu_indicator, fu_header, start_bit, nal_type, nal;
292 av_log(ctx, AV_LOG_ERROR, "Too short data for FU-A H.264 RTP packet\n");
293 return AVERROR_INVALIDDATA;
296 fu_indicator = buf[0];
298 start_bit = fu_header >> 7;
299 nal_type = fu_header & 0x1f;
300 nal = fu_indicator & 0xe0 | nal_type;
302 // skip the fu_indicator and fu_header
306 if (start_bit && nal_counters)
307 nal_counters[nal_type & nal_mask]++;
308 return ff_h264_handle_frag_packet(pkt, buf, len, start_bit, &nal, 1);
311 // return 0 on packet, no more left, 1 on packet, 1 on partial packet
312 static int h264_handle_packet(AVFormatContext *ctx, PayloadContext *data,
313 AVStream *st, AVPacket *pkt, uint32_t *timestamp,
314 const uint8_t *buf, int len, uint16_t seq,
322 av_log(ctx, AV_LOG_ERROR, "Empty H.264 RTP packet\n");
323 return AVERROR_INVALIDDATA;
328 /* Simplify the case (these are all the NAL types used internally by
329 * the H.264 codec). */
330 if (type >= 1 && type <= 23)
333 case 0: // undefined, but pass them through
335 if ((result = av_new_packet(pkt, len + sizeof(start_sequence))) < 0)
337 memcpy(pkt->data, start_sequence, sizeof(start_sequence));
338 memcpy(pkt->data + sizeof(start_sequence), buf, len);
339 COUNT_NAL_TYPE(data, nal);
342 case 24: // STAP-A (one packet, multiple nals)
343 // consume the STAP-A NAL
346 result = ff_h264_handle_aggregated_packet(ctx, data, pkt, buf, len, 0,
347 NAL_COUNTERS, NAL_MASK);
354 avpriv_report_missing_feature(ctx, "RTP H.264 NAL unit type %d", type);
355 result = AVERROR_PATCHWELCOME;
358 case 28: // FU-A (fragmented nal)
359 result = h264_handle_packet_fu_a(ctx, data, pkt, buf, len,
360 NAL_COUNTERS, NAL_MASK);
363 case 30: // undefined
364 case 31: // undefined
366 av_log(ctx, AV_LOG_ERROR, "Undefined type (%d)\n", type);
367 result = AVERROR_INVALIDDATA;
371 pkt->stream_index = st->index;
376 static void h264_close_context(PayloadContext *data)
381 for (ii = 0; ii < 32; ii++) {
382 if (data->packet_types_received[ii])
383 av_log(NULL, AV_LOG_DEBUG, "Received %d packets of type %d\n",
384 data->packet_types_received[ii], ii);
389 static int parse_h264_sdp_line(AVFormatContext *s, int st_index,
390 PayloadContext *h264_data, const char *line)
393 const char *p = line;
398 stream = s->streams[st_index];
400 if (av_strstart(p, "framesize:", &p)) {
401 ff_h264_parse_framesize(stream->codecpar, p);
402 } else if (av_strstart(p, "fmtp:", &p)) {
403 return ff_parse_fmtp(s, stream, h264_data, p, sdp_parse_fmtp_config_h264);
404 } else if (av_strstart(p, "cliprect:", &p)) {
405 // could use this if we wanted.
411 RTPDynamicProtocolHandler ff_h264_dynamic_handler = {
413 .codec_type = AVMEDIA_TYPE_VIDEO,
414 .codec_id = AV_CODEC_ID_H264,
415 .need_parsing = AVSTREAM_PARSE_FULL,
416 .priv_data_size = sizeof(PayloadContext),
417 .parse_sdp_a_line = parse_h264_sdp_line,
418 .close = h264_close_context,
419 .parse_packet = h264_handle_packet,