2 * RTP H264 Protocol (RFC3984)
3 * Copyright (c) 2006 Ryan Martell
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
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/avstring.h"
39 #include "libavcodec/get_bits.h"
46 #include "rtpdec_formats.h"
48 struct PayloadContext {
49 // sdp setup parameters
53 int packetization_mode;
55 int packet_types_received[32];
60 #define COUNT_NAL_TYPE(data, nal) data->packet_types_received[(nal) & 0x1f]++
62 #define COUNT_NAL_TYPE(data, nal) do { } while (0)
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 static int parse_sprop_parameter_sets(AVFormatContext *s,
97 AVCodecContext *codec,
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_malloc(packet_size + sizeof(start_sequence) +
120 codec->extradata_size +
121 FF_INPUT_BUFFER_PADDING_SIZE);
123 av_log(s, AV_LOG_ERROR,
124 "Unable to allocate memory for extradata!\n");
125 return AVERROR(ENOMEM);
127 if (codec->extradata_size) {
128 memcpy(dest, codec->extradata, codec->extradata_size);
129 av_free(codec->extradata);
132 memcpy(dest + codec->extradata_size, start_sequence,
133 sizeof(start_sequence));
134 memcpy(dest + codec->extradata_size + sizeof(start_sequence),
135 decoded_packet, packet_size);
136 memset(dest + codec->extradata_size + sizeof(start_sequence) +
137 packet_size, 0, FF_INPUT_BUFFER_PADDING_SIZE);
139 codec->extradata = dest;
140 codec->extradata_size += sizeof(start_sequence) + packet_size;
144 av_log(s, AV_LOG_DEBUG, "Extradata set to %p (size: %d)\n",
145 codec->extradata, codec->extradata_size);
150 static int sdp_parse_fmtp_config_h264(AVFormatContext *s,
152 PayloadContext *h264_data,
153 char *attr, char *value)
155 AVCodecContext *codec = stream->codec;
156 assert(codec->codec_id == AV_CODEC_ID_H264);
157 assert(h264_data != NULL);
159 if (!strcmp(attr, "packetization-mode")) {
160 av_log(s, AV_LOG_DEBUG, "RTP Packetization Mode: %d\n", atoi(value));
161 h264_data->packetization_mode = atoi(value);
163 * Packetization Mode:
164 * 0 or not present: Single NAL mode (Only nals from 1-23 are allowed)
165 * 1: Non-interleaved Mode: 1-23, 24 (STAP-A), 28 (FU-A) are allowed.
166 * 2: Interleaved Mode: 25 (STAP-B), 26 (MTAP16), 27 (MTAP24), 28 (FU-A),
167 * and 29 (FU-B) are allowed.
169 if (h264_data->packetization_mode > 1)
170 av_log(s, AV_LOG_ERROR,
171 "Interleaved RTP mode is not supported yet.\n");
172 } else if (!strcmp(attr, "profile-level-id")) {
173 if (strlen(value) == 6)
174 parse_profile_level_id(s, h264_data, value);
175 } else if (!strcmp(attr, "sprop-parameter-sets")) {
176 codec->extradata_size = 0;
177 av_freep(&codec->extradata);
178 return parse_sprop_parameter_sets(s, codec, value);
183 // return 0 on packet, no more left, 1 on packet, 1 on partial packet
184 static int h264_handle_packet(AVFormatContext *ctx, PayloadContext *data,
185 AVStream *st, AVPacket *pkt, uint32_t *timestamp,
186 const uint8_t *buf, int len, uint16_t seq,
194 av_log(ctx, AV_LOG_ERROR, "Empty H264 RTP packet\n");
195 return AVERROR_INVALIDDATA;
203 /* Simplify the case (these are all the nal types used internally by
204 * the h264 codec). */
205 if (type >= 1 && type <= 23)
208 case 0: // undefined, but pass them through
210 if ((result = av_new_packet(pkt, len + sizeof(start_sequence))) < 0)
212 memcpy(pkt->data, start_sequence, sizeof(start_sequence));
213 memcpy(pkt->data + sizeof(start_sequence), buf, len);
214 COUNT_NAL_TYPE(data, nal);
217 case 24: // STAP-A (one packet, multiple nals)
218 // consume the STAP-A NAL
221 // first we are going to figure out the total size
224 int total_length = 0;
227 for (pass = 0; pass < 2; pass++) {
228 const uint8_t *src = buf;
231 while (src_len > 2) {
232 uint16_t nal_size = AV_RB16(src);
234 // consume the length of the aggregate
238 if (nal_size <= src_len) {
241 total_length += sizeof(start_sequence) + nal_size;
245 memcpy(dst, start_sequence, sizeof(start_sequence));
246 dst += sizeof(start_sequence);
247 memcpy(dst, src, nal_size);
248 COUNT_NAL_TYPE(data, *src);
252 av_log(ctx, AV_LOG_ERROR,
253 "nal size exceeds length: %d %d\n", nal_size, src_len);
256 // eat what we handled
261 av_log(ctx, AV_LOG_ERROR,
262 "Consumed more bytes than we got! (%d)\n", src_len);
266 /* now we know the total size of the packet (with the
267 * start sequences added) */
268 if ((result = av_new_packet(pkt, total_length)) < 0)
272 assert(dst - pkt->data == total_length);
282 av_log(ctx, AV_LOG_ERROR,
283 "Unhandled type (%d) (See RFC for implementation details\n",
285 result = AVERROR(ENOSYS);
288 case 28: // FU-A (fragmented nal)
290 len--; // skip the fu_indicator
292 // these are the same as above, we just redo them here for clarity
293 uint8_t fu_indicator = nal;
294 uint8_t fu_header = *buf;
295 uint8_t start_bit = fu_header >> 7;
296 uint8_t av_unused end_bit = (fu_header & 0x40) >> 6;
297 uint8_t nal_type = fu_header & 0x1f;
298 uint8_t reconstructed_nal;
300 // Reconstruct this packet's true nal; only the data follows.
301 /* The original nal forbidden bit and NRI are stored in this
303 reconstructed_nal = fu_indicator & 0xe0;
304 reconstructed_nal |= nal_type;
306 // skip the fu_header
311 COUNT_NAL_TYPE(data, nal_type);
313 /* copy in the start sequence, and the reconstructed nal */
314 if ((result = av_new_packet(pkt, sizeof(start_sequence) + sizeof(nal) + len)) < 0)
316 memcpy(pkt->data, start_sequence, sizeof(start_sequence));
317 pkt->data[sizeof(start_sequence)] = reconstructed_nal;
318 memcpy(pkt->data + sizeof(start_sequence) + sizeof(nal), buf, len);
320 if ((result = av_new_packet(pkt, len)) < 0)
322 memcpy(pkt->data, buf, len);
325 av_log(ctx, AV_LOG_ERROR, "Too short data for FU-A H264 RTP packet\n");
326 result = AVERROR_INVALIDDATA;
330 case 30: // undefined
331 case 31: // undefined
333 av_log(ctx, AV_LOG_ERROR, "Undefined type (%d)\n", type);
334 result = AVERROR_INVALIDDATA;
338 pkt->stream_index = st->index;
343 static PayloadContext *h264_new_context(void)
345 return av_mallocz(sizeof(PayloadContext) + FF_INPUT_BUFFER_PADDING_SIZE);
348 static void h264_free_context(PayloadContext *data)
353 for (ii = 0; ii < 32; ii++) {
354 if (data->packet_types_received[ii])
355 av_log(NULL, AV_LOG_DEBUG, "Received %d packets of type %d\n",
356 data->packet_types_received[ii], ii);
363 static av_cold int h264_init(AVFormatContext *s, int st_index,
364 PayloadContext *data)
368 s->streams[st_index]->need_parsing = AVSTREAM_PARSE_FULL;
372 static int parse_h264_sdp_line(AVFormatContext *s, int st_index,
373 PayloadContext *h264_data, const char *line)
376 AVCodecContext *codec;
377 const char *p = line;
382 stream = s->streams[st_index];
383 codec = stream->codec;
385 if (av_strstart(p, "framesize:", &p)) {
389 // remove the protocol identifier
390 while (*p && *p == ' ')
391 p++; // strip spaces.
392 while (*p && *p != ' ')
393 p++; // eat protocol identifier
394 while (*p && *p == ' ')
395 p++; // strip trailing spaces.
396 while (*p && *p != '-' && (dst - buf1) < sizeof(buf1) - 1)
400 // a='framesize:96 320-240'
401 // set our parameters
402 codec->width = atoi(buf1);
403 codec->height = atoi(p + 1); // skip the -
404 } else if (av_strstart(p, "fmtp:", &p)) {
405 return ff_parse_fmtp(s, stream, h264_data, p, sdp_parse_fmtp_config_h264);
406 } else if (av_strstart(p, "cliprect:", &p)) {
407 // could use this if we wanted.
413 RTPDynamicProtocolHandler ff_h264_dynamic_handler = {
415 .codec_type = AVMEDIA_TYPE_VIDEO,
416 .codec_id = AV_CODEC_ID_H264,
418 .parse_sdp_a_line = parse_h264_sdp_line,
419 .alloc = h264_new_context,
420 .free = h264_free_context,
421 .parse_packet = h264_handle_packet