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 #include "libavutil/base64.h"
36 #include "libavutil/avstring.h"
37 #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];
59 static int sdp_parse_fmtp_config_h264(AVStream * stream,
60 PayloadContext * h264_data,
61 char *attr, char *value)
63 AVCodecContext *codec = stream->codec;
64 assert(codec->codec_id == CODEC_ID_H264);
65 assert(h264_data != NULL);
67 if (!strcmp(attr, "packetization-mode")) {
68 av_log(codec, AV_LOG_DEBUG, "RTP Packetization Mode: %d\n", atoi(value));
69 h264_data->packetization_mode = atoi(value);
72 0 or not present: Single NAL mode (Only nals from 1-23 are allowed)
73 1: Non-interleaved Mode: 1-23, 24 (STAP-A), 28 (FU-A) are allowed.
74 2: Interleaved Mode: 25 (STAP-B), 26 (MTAP16), 27 (MTAP24), 28 (FU-A), and 29 (FU-B) are allowed.
76 if (h264_data->packetization_mode > 1)
77 av_log(codec, AV_LOG_ERROR,
78 "Interleaved RTP mode is not supported yet.");
79 } else if (!strcmp(attr, "profile-level-id")) {
80 if (strlen(value) == 6) {
82 // 6 characters=3 bytes, in hex.
87 buffer[0] = value[0]; buffer[1] = value[1]; buffer[2] = '\0';
88 profile_idc = strtol(buffer, NULL, 16);
89 buffer[0] = value[2]; buffer[1] = value[3];
90 profile_iop = strtol(buffer, NULL, 16);
91 buffer[0] = value[4]; buffer[1] = value[5];
92 level_idc = strtol(buffer, NULL, 16);
94 av_log(codec, AV_LOG_DEBUG,
95 "RTP Profile IDC: %x Profile IOP: %x Level: %x\n",
96 profile_idc, profile_iop, level_idc);
97 h264_data->profile_idc = profile_idc;
98 h264_data->profile_iop = profile_iop;
99 h264_data->level_idc = level_idc;
101 } else if (!strcmp(attr, "sprop-parameter-sets")) {
102 uint8_t start_sequence[] = { 0, 0, 0, 1 };
103 codec->extradata_size= 0;
104 codec->extradata= NULL;
107 char base64packet[1024];
108 uint8_t decoded_packet[1024];
110 char *dst = base64packet;
112 while (*value && *value != ','
113 && (dst - base64packet) < sizeof(base64packet) - 1) {
121 packet_size= av_base64_decode(decoded_packet, base64packet, sizeof(decoded_packet));
122 if (packet_size > 0) {
123 uint8_t *dest = av_malloc(packet_size + sizeof(start_sequence) +
124 codec->extradata_size +
125 FF_INPUT_BUFFER_PADDING_SIZE);
128 if(codec->extradata_size)
130 memcpy(dest, codec->extradata, codec->extradata_size);
131 av_free(codec->extradata);
134 memcpy(dest+codec->extradata_size, start_sequence, sizeof(start_sequence));
135 memcpy(dest+codec->extradata_size+sizeof(start_sequence), 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;
142 av_log(codec, AV_LOG_ERROR, "Unable to allocate memory for extradata!");
143 return AVERROR(ENOMEM);
147 av_log(codec, AV_LOG_DEBUG, "Extradata set to %p (size: %d)!", codec->extradata, codec->extradata_size);
152 // return 0 on packet, no more left, 1 on packet, 1 on partial packet...
153 static int h264_handle_packet(AVFormatContext *ctx,
154 PayloadContext *data,
157 uint32_t * timestamp,
164 uint8_t start_sequence[] = { 0, 0, 0, 1 };
167 av_log(ctx, AV_LOG_ERROR, "Empty H264 RTP packet\n");
168 return AVERROR_INVALIDDATA;
176 if (type >= 1 && type <= 23)
177 type = 1; // simplify the case. (these are all the nal types used internally by the h264 codec)
179 case 0: // undefined, but pass them through
181 av_new_packet(pkt, len+sizeof(start_sequence));
182 memcpy(pkt->data, start_sequence, sizeof(start_sequence));
183 memcpy(pkt->data+sizeof(start_sequence), buf, len);
185 data->packet_types_received[nal & 0x1f]++;
189 case 24: // STAP-A (one packet, multiple nals)
190 // consume the STAP-A NAL
193 // first we are going to figure out the total size....
199 for(pass= 0; pass<2; pass++) {
200 const uint8_t *src= buf;
203 while (src_len > 2) {
204 uint16_t nal_size = AV_RB16(src);
206 // consume the length of the aggregate...
210 if (nal_size <= src_len) {
213 total_length+= sizeof(start_sequence)+nal_size;
217 memcpy(dst, start_sequence, sizeof(start_sequence));
218 dst+= sizeof(start_sequence);
219 memcpy(dst, src, nal_size);
221 data->packet_types_received[*src & 0x1f]++;
226 av_log(ctx, AV_LOG_ERROR,
227 "nal size exceeds length: %d %d\n", nal_size, src_len);
230 // eat what we handled...
235 av_log(ctx, AV_LOG_ERROR,
236 "Consumed more bytes than we got! (%d)\n", src_len);
240 // now we know the total size of the packet (with the start sequences added)
241 av_new_packet(pkt, total_length);
244 assert(dst-pkt->data==total_length);
254 av_log(ctx, AV_LOG_ERROR,
255 "Unhandled type (%d) (See RFC for implementation details\n",
257 result = AVERROR(ENOSYS);
260 case 28: // FU-A (fragmented nal)
262 len--; // skip the fu_indicator
264 // these are the same as above, we just redo them here for clarity...
265 uint8_t fu_indicator = nal;
266 uint8_t fu_header = *buf;
267 uint8_t start_bit = fu_header >> 7;
268 uint8_t av_unused end_bit = (fu_header & 0x40) >> 6;
269 uint8_t nal_type = (fu_header & 0x1f);
270 uint8_t reconstructed_nal;
272 // reconstruct this packet's true nal; only the data follows..
273 reconstructed_nal = fu_indicator & (0xe0); // the original nal forbidden bit and NRI are stored in this packet's nal;
274 reconstructed_nal |= nal_type;
276 // skip the fu_header...
282 data->packet_types_received[nal_type]++;
285 // copy in the start sequence, and the reconstructed nal....
286 av_new_packet(pkt, sizeof(start_sequence)+sizeof(nal)+len);
287 memcpy(pkt->data, start_sequence, sizeof(start_sequence));
288 pkt->data[sizeof(start_sequence)]= reconstructed_nal;
289 memcpy(pkt->data+sizeof(start_sequence)+sizeof(nal), buf, len);
291 av_new_packet(pkt, len);
292 memcpy(pkt->data, buf, len);
295 av_log(ctx, AV_LOG_ERROR, "Too short data for FU-A H264 RTP packet\n");
296 result = AVERROR_INVALIDDATA;
300 case 30: // undefined
301 case 31: // undefined
303 av_log(ctx, AV_LOG_ERROR, "Undefined type (%d)", type);
304 result = AVERROR_INVALIDDATA;
308 pkt->stream_index = st->index;
313 static PayloadContext *h264_new_context(void)
315 return av_mallocz(sizeof(PayloadContext) + FF_INPUT_BUFFER_PADDING_SIZE);
318 static void h264_free_context(PayloadContext *data)
323 for (ii = 0; ii < 32; ii++) {
324 if (data->packet_types_received[ii])
325 av_log(NULL, AV_LOG_DEBUG, "Received %d packets of type %d\n",
326 data->packet_types_received[ii], ii);
333 static int parse_h264_sdp_line(AVFormatContext *s, int st_index,
334 PayloadContext *h264_data, const char *line)
337 AVCodecContext *codec;
338 const char *p = line;
343 stream = s->streams[st_index];
344 codec = stream->codec;
346 if (av_strstart(p, "framesize:", &p)) {
350 // remove the protocol identifier..
351 while (*p && *p == ' ') p++; // strip spaces.
352 while (*p && *p != ' ') p++; // eat protocol identifier
353 while (*p && *p == ' ') p++; // strip trailing spaces.
354 while (*p && *p != '-' && (dst - buf1) < sizeof(buf1) - 1) {
359 // a='framesize:96 320-240'
360 // set our parameters..
361 codec->width = atoi(buf1);
362 codec->height = atoi(p + 1); // skip the -
363 codec->pix_fmt = PIX_FMT_YUV420P;
364 } else if (av_strstart(p, "fmtp:", &p)) {
365 return ff_parse_fmtp(stream, h264_data, p, sdp_parse_fmtp_config_h264);
366 } else if (av_strstart(p, "cliprect:", &p)) {
367 // could use this if we wanted.
373 RTPDynamicProtocolHandler ff_h264_dynamic_handler = {
375 .codec_type = AVMEDIA_TYPE_VIDEO,
376 .codec_id = CODEC_ID_H264,
377 .parse_sdp_a_line = parse_h264_sdp_line,
378 .alloc = h264_new_context,
379 .free = h264_free_context,
380 .parse_packet = h264_handle_packet