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, FU-B packet types)
35 * 1) RTCP sender reports for udp streams are required..
39 #include "libavutil/base64.h"
40 #include "libavutil/avstring.h"
41 #include "libavcodec/get_bits.h"
50 #include "rtpdec_formats.h"
53 RTP/H264 specific private data.
55 struct PayloadContext {
56 //sdp setup parameters
57 uint8_t profile_idc; ///< from the sdp setup parameters.
58 uint8_t profile_iop; ///< from the sdp setup parameters.
59 uint8_t level_idc; ///< from the sdp setup parameters.
60 int packetization_mode; ///< from the sdp setup parameters.
62 int packet_types_received[32];
66 /* ---------------- private code */
67 static int sdp_parse_fmtp_config_h264(AVStream * stream,
68 PayloadContext * h264_data,
69 char *attr, char *value)
71 AVCodecContext *codec = stream->codec;
72 assert(codec->codec_id == CODEC_ID_H264);
73 assert(h264_data != NULL);
75 if (!strcmp(attr, "packetization-mode")) {
76 av_log(codec, AV_LOG_DEBUG, "RTP Packetization Mode: %d\n", atoi(value));
77 h264_data->packetization_mode = atoi(value);
80 0 or not present: Single NAL mode (Only nals from 1-23 are allowed)
81 1: Non-interleaved Mode: 1-23, 24 (STAP-A), 28 (FU-A) are allowed.
82 2: Interleaved Mode: 25 (STAP-B), 26 (MTAP16), 27 (MTAP24), 28 (FU-A), and 29 (FU-B) are allowed.
84 if (h264_data->packetization_mode > 1)
85 av_log(codec, AV_LOG_ERROR,
86 "Interleaved RTP mode is not supported yet.");
87 } else if (!strcmp(attr, "profile-level-id")) {
88 if (strlen(value) == 6) {
90 // 6 characters=3 bytes, in hex.
95 buffer[0] = value[0]; buffer[1] = value[1]; buffer[2] = '\0';
96 profile_idc = strtol(buffer, NULL, 16);
97 buffer[0] = value[2]; buffer[1] = value[3];
98 profile_iop = strtol(buffer, NULL, 16);
99 buffer[0] = value[4]; buffer[1] = value[5];
100 level_idc = strtol(buffer, NULL, 16);
102 // set the parameters...
103 av_log(codec, AV_LOG_DEBUG,
104 "RTP Profile IDC: %x Profile IOP: %x Level: %x\n",
105 profile_idc, profile_iop, level_idc);
106 h264_data->profile_idc = profile_idc;
107 h264_data->profile_iop = profile_iop;
108 h264_data->level_idc = level_idc;
110 } else if (!strcmp(attr, "sprop-parameter-sets")) {
111 uint8_t start_sequence[] = { 0, 0, 0, 1 };
112 codec->extradata_size= 0;
113 codec->extradata= NULL;
116 char base64packet[1024];
117 uint8_t decoded_packet[1024];
119 char *dst = base64packet;
121 while (*value && *value != ','
122 && (dst - base64packet) < sizeof(base64packet) - 1) {
130 packet_size= av_base64_decode(decoded_packet, base64packet, sizeof(decoded_packet));
131 if (packet_size > 0) {
132 uint8_t *dest = av_malloc(packet_size + sizeof(start_sequence) +
133 codec->extradata_size +
134 FF_INPUT_BUFFER_PADDING_SIZE);
137 if(codec->extradata_size)
140 memcpy(dest, codec->extradata, codec->extradata_size);
141 av_free(codec->extradata);
144 memcpy(dest+codec->extradata_size, start_sequence, sizeof(start_sequence));
145 memcpy(dest+codec->extradata_size+sizeof(start_sequence), decoded_packet, packet_size);
146 memset(dest+codec->extradata_size+sizeof(start_sequence)+
147 packet_size, 0, FF_INPUT_BUFFER_PADDING_SIZE);
149 codec->extradata= dest;
150 codec->extradata_size+= sizeof(start_sequence)+packet_size;
152 av_log(codec, AV_LOG_ERROR, "Unable to allocate memory for extradata!");
153 return AVERROR(ENOMEM);
157 av_log(codec, AV_LOG_DEBUG, "Extradata set to %p (size: %d)!", codec->extradata, codec->extradata_size);
162 // return 0 on packet, no more left, 1 on packet, 1 on partial packet...
163 static int h264_handle_packet(AVFormatContext *ctx,
164 PayloadContext *data,
167 uint32_t * timestamp,
174 uint8_t start_sequence[] = { 0, 0, 0, 1 };
177 av_log(ctx, AV_LOG_ERROR, "Empty H264 RTP packet\n");
178 return AVERROR_INVALIDDATA;
188 if (type >= 1 && type <= 23)
189 type = 1; // simplify the case. (these are all the nal types used internally by the h264 codec)
191 case 0: // undefined, but pass them through
193 av_new_packet(pkt, len+sizeof(start_sequence));
194 memcpy(pkt->data, start_sequence, sizeof(start_sequence));
195 memcpy(pkt->data+sizeof(start_sequence), buf, len);
197 data->packet_types_received[nal & 0x1f]++;
201 case 24: // STAP-A (one packet, multiple nals)
202 // consume the STAP-A NAL
205 // first we are going to figure out the total size....
211 for(pass= 0; pass<2; pass++) {
212 const uint8_t *src= buf;
215 while (src_len > 2) {
216 uint16_t nal_size = AV_RB16(src); // this going to be a problem if unaligned (can it be?)
218 // consume the length of the aggregate...
222 if (nal_size <= src_len) {
225 total_length+= sizeof(start_sequence)+nal_size;
229 memcpy(dst, start_sequence, sizeof(start_sequence));
230 dst+= sizeof(start_sequence);
231 memcpy(dst, src, nal_size);
233 data->packet_types_received[*src & 0x1f]++;
238 av_log(ctx, AV_LOG_ERROR,
239 "nal size exceeds length: %d %d\n", nal_size, src_len);
242 // eat what we handled...
247 av_log(ctx, AV_LOG_ERROR,
248 "Consumed more bytes than we got! (%d)\n", src_len);
252 // now we know the total size of the packet (with the start sequences added)
253 av_new_packet(pkt, total_length);
256 assert(dst-pkt->data==total_length);
266 av_log(ctx, AV_LOG_ERROR,
267 "Unhandled type (%d) (See RFC for implementation details\n",
269 result = AVERROR(ENOSYS);
272 case 28: // FU-A (fragmented nal)
274 len--; // skip the fu_indicator
276 // these are the same as above, we just redo them here for clarity...
277 uint8_t fu_indicator = nal;
278 uint8_t fu_header = *buf; // read the fu_header.
279 uint8_t start_bit = fu_header >> 7;
280 // uint8_t end_bit = (fu_header & 0x40) >> 6;
281 uint8_t nal_type = (fu_header & 0x1f);
282 uint8_t reconstructed_nal;
284 // reconstruct this packet's true nal; only the data follows..
285 reconstructed_nal = fu_indicator & (0xe0); // the original nal forbidden bit and NRI are stored in this packet's nal;
286 reconstructed_nal |= nal_type;
288 // skip the fu_header...
294 data->packet_types_received[nal_type]++;
297 // copy in the start sequence, and the reconstructed nal....
298 av_new_packet(pkt, sizeof(start_sequence)+sizeof(nal)+len);
299 memcpy(pkt->data, start_sequence, sizeof(start_sequence));
300 pkt->data[sizeof(start_sequence)]= reconstructed_nal;
301 memcpy(pkt->data+sizeof(start_sequence)+sizeof(nal), buf, len);
303 av_new_packet(pkt, len);
304 memcpy(pkt->data, buf, len);
307 av_log(ctx, AV_LOG_ERROR, "Too short data for FU-A H264 RTP packet\n");
308 result = AVERROR_INVALIDDATA;
312 case 30: // undefined
313 case 31: // undefined
315 av_log(ctx, AV_LOG_ERROR, "Undefined type (%d)", type);
316 result = AVERROR_INVALIDDATA;
320 pkt->stream_index = st->index;
325 /* ---------------- public code */
326 static PayloadContext *h264_new_context(void)
328 return av_mallocz(sizeof(PayloadContext) + FF_INPUT_BUFFER_PADDING_SIZE);
331 static void h264_free_context(PayloadContext *data)
336 for (ii = 0; ii < 32; ii++) {
337 if (data->packet_types_received[ii])
338 av_log(NULL, AV_LOG_DEBUG, "Received %d packets of type %d\n",
339 data->packet_types_received[ii], ii);
346 static int parse_h264_sdp_line(AVFormatContext *s, int st_index,
347 PayloadContext *h264_data, const char *line)
350 AVCodecContext *codec;
351 const char *p = line;
356 stream = s->streams[st_index];
357 codec = stream->codec;
359 if (av_strstart(p, "framesize:", &p)) {
363 // remove the protocol identifier..
364 while (*p && *p == ' ') p++; // strip spaces.
365 while (*p && *p != ' ') p++; // eat protocol identifier
366 while (*p && *p == ' ') p++; // strip trailing spaces.
367 while (*p && *p != '-' && (dst - buf1) < sizeof(buf1) - 1) {
372 // a='framesize:96 320-240'
373 // set our parameters..
374 codec->width = atoi(buf1);
375 codec->height = atoi(p + 1); // skip the -
376 codec->pix_fmt = PIX_FMT_YUV420P;
377 } else if (av_strstart(p, "fmtp:", &p)) {
378 return ff_parse_fmtp(stream, h264_data, p, sdp_parse_fmtp_config_h264);
379 } else if (av_strstart(p, "cliprect:", &p)) {
380 // could use this if we wanted.
383 return 0; // keep processing it the normal way...
387 This is the structure for expanding on the dynamic rtp protocols (makes everything static. yay!)
389 RTPDynamicProtocolHandler ff_h264_dynamic_handler = {
391 .codec_type = AVMEDIA_TYPE_VIDEO,
392 .codec_id = CODEC_ID_H264,
393 .parse_sdp_a_line = parse_h264_sdp_line,
394 .alloc = h264_new_context,
395 .free = h264_free_context,
396 .parse_packet = h264_handle_packet