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/base64.h"
37 #include "libavutil/avstring.h"
38 #include "libavcodec/get_bits.h"
47 #include "rtpdec_formats.h"
49 struct PayloadContext {
50 // sdp setup parameters
54 int packetization_mode;
56 int packet_types_received[32];
60 static int sdp_parse_fmtp_config_h264(AVStream *stream,
61 PayloadContext *h264_data,
62 char *attr, char *value)
64 AVCodecContext *codec = stream->codec;
65 assert(codec->codec_id == CODEC_ID_H264);
66 assert(h264_data != NULL);
68 if (!strcmp(attr, "packetization-mode")) {
69 av_log(codec, AV_LOG_DEBUG, "RTP Packetization Mode: %d\n", atoi(value));
70 h264_data->packetization_mode = atoi(value);
73 * 0 or not present: Single NAL mode (Only nals from 1-23 are allowed)
74 * 1: Non-interleaved Mode: 1-23, 24 (STAP-A), 28 (FU-A) are allowed.
75 * 2: Interleaved Mode: 25 (STAP-B), 26 (MTAP16), 27 (MTAP24), 28 (FU-A),
76 * and 29 (FU-B) are allowed.
78 if (h264_data->packetization_mode > 1)
79 av_log(codec, AV_LOG_ERROR,
80 "Interleaved RTP mode is not supported yet.");
81 } else if (!strcmp(attr, "profile-level-id")) {
82 if (strlen(value) == 6) {
84 // 6 characters=3 bytes, in hex.
92 profile_idc = strtol(buffer, NULL, 16);
95 profile_iop = strtol(buffer, NULL, 16);
98 level_idc = strtol(buffer, NULL, 16);
100 av_log(codec, AV_LOG_DEBUG,
101 "RTP Profile IDC: %x Profile IOP: %x Level: %x\n",
102 profile_idc, profile_iop, level_idc);
103 h264_data->profile_idc = profile_idc;
104 h264_data->profile_iop = profile_iop;
105 h264_data->level_idc = level_idc;
107 } else if (!strcmp(attr, "sprop-parameter-sets")) {
108 uint8_t start_sequence[] = { 0, 0, 0, 1 };
109 codec->extradata_size = 0;
110 codec->extradata = NULL;
113 char base64packet[1024];
114 uint8_t decoded_packet[1024];
116 char *dst = base64packet;
118 while (*value && *value != ','
119 && (dst - base64packet) < sizeof(base64packet) - 1) {
127 packet_size = av_base64_decode(decoded_packet, base64packet,
128 sizeof(decoded_packet));
129 if (packet_size > 0) {
130 uint8_t *dest = av_malloc(packet_size + sizeof(start_sequence) +
131 codec->extradata_size +
132 FF_INPUT_BUFFER_PADDING_SIZE);
134 if (codec->extradata_size) {
135 memcpy(dest, codec->extradata, codec->extradata_size);
136 av_free(codec->extradata);
139 memcpy(dest + codec->extradata_size, start_sequence,
140 sizeof(start_sequence));
141 memcpy(dest + codec->extradata_size + sizeof(start_sequence),
142 decoded_packet, packet_size);
143 memset(dest + codec->extradata_size + sizeof(start_sequence) +
144 packet_size, 0, FF_INPUT_BUFFER_PADDING_SIZE);
146 codec->extradata = dest;
147 codec->extradata_size += sizeof(start_sequence) + packet_size;
149 av_log(codec, AV_LOG_ERROR,
150 "Unable to allocate memory for extradata!");
151 return AVERROR(ENOMEM);
155 av_log(codec, AV_LOG_DEBUG, "Extradata set to %p (size: %d)!",
156 codec->extradata, codec->extradata_size);
161 // return 0 on packet, no more left, 1 on packet, 1 on partial packet
162 static int h264_handle_packet(AVFormatContext *ctx, PayloadContext *data,
163 AVStream *st, AVPacket *pkt, uint32_t *timestamp,
164 const uint8_t *buf, int len, int flags)
169 uint8_t start_sequence[] = { 0, 0, 0, 1 };
172 av_log(ctx, AV_LOG_ERROR, "Empty H264 RTP packet\n");
173 return AVERROR_INVALIDDATA;
181 /* Simplify the case (these are all the nal types used internally by
182 * the h264 codec). */
183 if (type >= 1 && type <= 23)
186 case 0: // undefined, but pass them through
188 av_new_packet(pkt, len + sizeof(start_sequence));
189 memcpy(pkt->data, start_sequence, sizeof(start_sequence));
190 memcpy(pkt->data + sizeof(start_sequence), buf, len);
192 data->packet_types_received[nal & 0x1f]++;
196 case 24: // STAP-A (one packet, multiple nals)
197 // consume the STAP-A NAL
200 // first we are going to figure out the total size
203 int total_length = 0;
206 for (pass = 0; pass < 2; pass++) {
207 const uint8_t *src = buf;
210 while (src_len > 2) {
211 uint16_t nal_size = AV_RB16(src);
213 // consume the length of the aggregate
217 if (nal_size <= src_len) {
220 total_length += sizeof(start_sequence) + nal_size;
224 memcpy(dst, start_sequence, sizeof(start_sequence));
225 dst += sizeof(start_sequence);
226 memcpy(dst, src, nal_size);
228 data->packet_types_received[*src & 0x1f]++;
233 av_log(ctx, AV_LOG_ERROR,
234 "nal size exceeds length: %d %d\n", nal_size, src_len);
237 // eat what we handled
242 av_log(ctx, AV_LOG_ERROR,
243 "Consumed more bytes than we got! (%d)\n", src_len);
247 /* now we know the total size of the packet (with the
248 * start sequences added) */
249 av_new_packet(pkt, total_length);
252 assert(dst - pkt->data == total_length);
262 av_log(ctx, AV_LOG_ERROR,
263 "Unhandled type (%d) (See RFC for implementation details\n",
265 result = AVERROR(ENOSYS);
268 case 28: // FU-A (fragmented nal)
270 len--; // skip the fu_indicator
272 // these are the same as above, we just redo them here for clarity
273 uint8_t fu_indicator = nal;
274 uint8_t fu_header = *buf;
275 uint8_t start_bit = fu_header >> 7;
276 uint8_t av_unused end_bit = (fu_header & 0x40) >> 6;
277 uint8_t nal_type = fu_header & 0x1f;
278 uint8_t reconstructed_nal;
280 // Reconstruct this packet's true nal; only the data follows.
281 /* The original nal forbidden bit and NRI are stored in this
283 reconstructed_nal = fu_indicator & 0xe0;
284 reconstructed_nal |= nal_type;
286 // skip the fu_header
292 data->packet_types_received[nal_type]++;
295 /* copy in the start sequence, and the reconstructed nal */
296 av_new_packet(pkt, sizeof(start_sequence) + sizeof(nal) + len);
297 memcpy(pkt->data, start_sequence, sizeof(start_sequence));
298 pkt->data[sizeof(start_sequence)] = reconstructed_nal;
299 memcpy(pkt->data + sizeof(start_sequence) + sizeof(nal), buf, len);
301 av_new_packet(pkt, len);
302 memcpy(pkt->data, buf, len);
305 av_log(ctx, AV_LOG_ERROR, "Too short data for FU-A H264 RTP packet\n");
306 result = AVERROR_INVALIDDATA;
310 case 30: // undefined
311 case 31: // undefined
313 av_log(ctx, AV_LOG_ERROR, "Undefined type (%d)", type);
314 result = AVERROR_INVALIDDATA;
318 pkt->stream_index = st->index;
323 static PayloadContext *h264_new_context(void)
325 return av_mallocz(sizeof(PayloadContext) + FF_INPUT_BUFFER_PADDING_SIZE);
328 static void h264_free_context(PayloadContext *data)
333 for (ii = 0; ii < 32; ii++) {
334 if (data->packet_types_received[ii])
335 av_log(NULL, AV_LOG_DEBUG, "Received %d packets of type %d\n",
336 data->packet_types_received[ii], ii);
343 static int parse_h264_sdp_line(AVFormatContext *s, int st_index,
344 PayloadContext *h264_data, const char *line)
347 AVCodecContext *codec;
348 const char *p = line;
353 stream = s->streams[st_index];
354 codec = stream->codec;
356 if (av_strstart(p, "framesize:", &p)) {
360 // remove the protocol identifier
361 while (*p && *p == ' ')
362 p++; // strip spaces.
363 while (*p && *p != ' ')
364 p++; // eat protocol identifier
365 while (*p && *p == ' ')
366 p++; // strip trailing spaces.
367 while (*p && *p != '-' && (dst - buf1) < sizeof(buf1) - 1)
371 // a='framesize:96 320-240'
372 // set our parameters
373 codec->width = atoi(buf1);
374 codec->height = atoi(p + 1); // skip the -
375 codec->pix_fmt = PIX_FMT_YUV420P;
376 } else if (av_strstart(p, "fmtp:", &p)) {
377 return ff_parse_fmtp(stream, h264_data, p, sdp_parse_fmtp_config_h264);
378 } else if (av_strstart(p, "cliprect:", &p)) {
379 // could use this if we wanted.
385 RTPDynamicProtocolHandler ff_h264_dynamic_handler = {
387 .codec_type = AVMEDIA_TYPE_VIDEO,
388 .codec_id = CODEC_ID_H264,
389 .parse_sdp_a_line = parse_h264_sdp_line,
390 .alloc = h264_new_context,
391 .free = h264_free_context,
392 .parse_packet = h264_handle_packet