2 * RTP H264 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/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 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 == AV_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),
83 * and 29 (FU-B) are allowed.
85 if (h264_data->packetization_mode > 1)
86 av_log(codec, AV_LOG_ERROR,
87 "Interleaved RTP mode is not supported yet.\n");
88 } else if (!strcmp(attr, "profile-level-id")) {
89 if (strlen(value) == 6) {
91 // 6 characters=3 bytes, in hex.
99 profile_idc = strtol(buffer, NULL, 16);
100 buffer[0] = value[2];
101 buffer[1] = value[3];
102 profile_iop = strtol(buffer, NULL, 16);
103 buffer[0] = value[4];
104 buffer[1] = value[5];
105 level_idc = strtol(buffer, NULL, 16);
107 av_log(codec, AV_LOG_DEBUG,
108 "RTP Profile IDC: %x Profile IOP: %x Level: %x\n",
109 profile_idc, profile_iop, level_idc);
110 h264_data->profile_idc = profile_idc;
111 h264_data->profile_iop = profile_iop;
112 h264_data->level_idc = level_idc;
114 } else if (!strcmp(attr, "sprop-parameter-sets")) {
115 codec->extradata_size = 0;
116 av_freep(&codec->extradata);
119 char base64packet[1024];
120 uint8_t decoded_packet[1024];
122 char *dst = base64packet;
124 while (*value && *value != ','
125 && (dst - base64packet) < sizeof(base64packet) - 1) {
133 packet_size = av_base64_decode(decoded_packet, base64packet,
134 sizeof(decoded_packet));
135 if (packet_size > 0) {
136 uint8_t *dest = av_malloc(packet_size + sizeof(start_sequence) +
137 codec->extradata_size +
138 FF_INPUT_BUFFER_PADDING_SIZE);
140 av_log(codec, AV_LOG_ERROR,
141 "Unable to allocate memory for extradata!\n");
142 return AVERROR(ENOMEM);
144 if (codec->extradata_size) {
145 memcpy(dest, codec->extradata, codec->extradata_size);
146 av_free(codec->extradata);
149 memcpy(dest + codec->extradata_size, start_sequence,
150 sizeof(start_sequence));
151 memcpy(dest + codec->extradata_size + sizeof(start_sequence),
152 decoded_packet, packet_size);
153 memset(dest + codec->extradata_size + sizeof(start_sequence) +
154 packet_size, 0, FF_INPUT_BUFFER_PADDING_SIZE);
156 codec->extradata = dest;
157 codec->extradata_size += sizeof(start_sequence) + packet_size;
160 av_log(codec, AV_LOG_DEBUG, "Extradata set to %p (size: %d)!\n",
161 codec->extradata, codec->extradata_size);
166 // return 0 on packet, no more left, 1 on packet, 1 on partial packet
167 static int h264_handle_packet(AVFormatContext *ctx, PayloadContext *data,
168 AVStream *st, AVPacket *pkt, uint32_t *timestamp,
169 const uint8_t *buf, int len, uint16_t seq,
177 av_log(ctx, AV_LOG_ERROR, "Empty H264 RTP packet\n");
178 return AVERROR_INVALIDDATA;
186 /* Simplify the case (these are all the nal types used internally by
187 * the h264 codec). */
188 if (type >= 1 && type <= 23)
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);
196 COUNT_NAL_TYPE(data, nal);
199 case 24: // STAP-A (one packet, multiple nals)
200 // consume the STAP-A NAL
203 // first we are going to figure out the total size
206 int total_length = 0;
209 for (pass = 0; pass < 2; pass++) {
210 const uint8_t *src = buf;
213 while (src_len > 2) {
214 uint16_t nal_size = AV_RB16(src);
216 // consume the length of the aggregate
220 if (nal_size <= src_len) {
223 total_length += sizeof(start_sequence) + nal_size;
227 memcpy(dst, start_sequence, sizeof(start_sequence));
228 dst += sizeof(start_sequence);
229 memcpy(dst, src, nal_size);
230 COUNT_NAL_TYPE(data, *src);
234 av_log(ctx, AV_LOG_ERROR,
235 "nal size exceeds length: %d %d\n", nal_size, src_len);
238 // eat what we handled
243 av_log(ctx, AV_LOG_ERROR,
244 "Consumed more bytes than we got! (%d)\n", src_len);
248 /* now we know the total size of the packet (with the
249 * start sequences added) */
250 av_new_packet(pkt, total_length);
253 assert(dst - pkt->data == total_length);
263 av_log(ctx, AV_LOG_ERROR,
264 "Unhandled type (%d) (See RFC for implementation details\n",
266 result = AVERROR(ENOSYS);
269 case 28: // FU-A (fragmented nal)
271 len--; // skip the fu_indicator
273 // these are the same as above, we just redo them here for clarity
274 uint8_t fu_indicator = nal;
275 uint8_t fu_header = *buf;
276 uint8_t start_bit = fu_header >> 7;
277 uint8_t av_unused end_bit = (fu_header & 0x40) >> 6;
278 uint8_t nal_type = fu_header & 0x1f;
279 uint8_t reconstructed_nal;
281 // Reconstruct this packet's true nal; only the data follows.
282 /* The original nal forbidden bit and NRI are stored in this
284 reconstructed_nal = fu_indicator & 0xe0;
285 reconstructed_nal |= nal_type;
287 // skip the fu_header
292 COUNT_NAL_TYPE(data, nal_type);
294 /* copy in the start sequence, and the reconstructed nal */
295 av_new_packet(pkt, sizeof(start_sequence) + sizeof(nal) + len);
296 memcpy(pkt->data, start_sequence, sizeof(start_sequence));
297 pkt->data[sizeof(start_sequence)] = reconstructed_nal;
298 memcpy(pkt->data + sizeof(start_sequence) + sizeof(nal), buf, len);
300 av_new_packet(pkt, len);
301 memcpy(pkt->data, buf, len);
304 av_log(ctx, AV_LOG_ERROR, "Too short data for FU-A H264 RTP packet\n");
305 result = AVERROR_INVALIDDATA;
309 case 30: // undefined
310 case 31: // undefined
312 av_log(ctx, AV_LOG_ERROR, "Undefined type (%d)\n", type);
313 result = AVERROR_INVALIDDATA;
317 pkt->stream_index = st->index;
322 static PayloadContext *h264_new_context(void)
324 return av_mallocz(sizeof(PayloadContext) + FF_INPUT_BUFFER_PADDING_SIZE);
327 static void h264_free_context(PayloadContext *data)
332 for (ii = 0; ii < 32; ii++) {
333 if (data->packet_types_received[ii])
334 av_log(NULL, AV_LOG_DEBUG, "Received %d packets of type %d\n",
335 data->packet_types_received[ii], ii);
342 static av_cold int h264_init(AVFormatContext *s, int st_index,
343 PayloadContext *data)
347 s->streams[st_index]->need_parsing = AVSTREAM_PARSE_FULL;
351 static int parse_h264_sdp_line(AVFormatContext *s, int st_index,
352 PayloadContext *h264_data, const char *line)
355 AVCodecContext *codec;
356 const char *p = line;
361 stream = s->streams[st_index];
362 codec = stream->codec;
364 if (av_strstart(p, "framesize:", &p)) {
368 // remove the protocol identifier
369 while (*p && *p == ' ')
370 p++; // strip spaces.
371 while (*p && *p != ' ')
372 p++; // eat protocol identifier
373 while (*p && *p == ' ')
374 p++; // strip trailing spaces.
375 while (*p && *p != '-' && (dst - buf1) < sizeof(buf1) - 1)
379 // a='framesize:96 320-240'
380 // set our parameters
381 codec->width = atoi(buf1);
382 codec->height = atoi(p + 1); // skip the -
383 } else if (av_strstart(p, "fmtp:", &p)) {
384 return ff_parse_fmtp(stream, h264_data, p, sdp_parse_fmtp_config_h264);
385 } else if (av_strstart(p, "cliprect:", &p)) {
386 // could use this if we wanted.
392 RTPDynamicProtocolHandler ff_h264_dynamic_handler = {
394 .codec_type = AVMEDIA_TYPE_VIDEO,
395 .codec_id = AV_CODEC_ID_H264,
397 .parse_sdp_a_line = parse_h264_sdp_line,
398 .alloc = h264_new_context,
399 .free = h264_free_context,
400 .parse_packet = h264_handle_packet