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/intreadwrite.h"
39 #include "libavutil/avstring.h"
45 #include "rtpdec_formats.h"
47 struct PayloadContext {
48 // sdp setup parameters
52 int packetization_mode;
54 int packet_types_received[32];
59 #define COUNT_NAL_TYPE(data, nal) data->packet_types_received[(nal) & 0x1f]++
61 #define COUNT_NAL_TYPE(data, nal) do { } while (0)
64 static const uint8_t start_sequence[] = { 0, 0, 0, 1 };
66 static void parse_profile_level_id(AVFormatContext *s,
67 PayloadContext *h264_data,
71 // 6 characters=3 bytes, in hex.
79 profile_idc = strtol(buffer, NULL, 16);
82 profile_iop = strtol(buffer, NULL, 16);
85 level_idc = strtol(buffer, NULL, 16);
87 av_log(s, AV_LOG_DEBUG,
88 "RTP Profile IDC: %x Profile IOP: %x Level: %x\n",
89 profile_idc, profile_iop, level_idc);
90 h264_data->profile_idc = profile_idc;
91 h264_data->profile_iop = profile_iop;
92 h264_data->level_idc = level_idc;
95 static int parse_sprop_parameter_sets(AVFormatContext *s,
96 AVCodecContext *codec,
99 char base64packet[1024];
100 uint8_t decoded_packet[1024];
104 char *dst = base64packet;
106 while (*value && *value != ','
107 && (dst - base64packet) < sizeof(base64packet) - 1) {
115 packet_size = av_base64_decode(decoded_packet, base64packet,
116 sizeof(decoded_packet));
117 if (packet_size > 0) {
118 uint8_t *dest = av_realloc(codec->extradata,
119 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 codec->extradata = dest;
129 memcpy(dest + codec->extradata_size, start_sequence,
130 sizeof(start_sequence));
131 memcpy(dest + codec->extradata_size + sizeof(start_sequence),
132 decoded_packet, packet_size);
133 memset(dest + codec->extradata_size + sizeof(start_sequence) +
134 packet_size, 0, FF_INPUT_BUFFER_PADDING_SIZE);
136 codec->extradata_size += sizeof(start_sequence) + packet_size;
140 av_log(s, AV_LOG_DEBUG, "Extradata set to %p (size: %d)\n",
141 codec->extradata, codec->extradata_size);
146 static int sdp_parse_fmtp_config_h264(AVFormatContext *s,
148 PayloadContext *h264_data,
149 char *attr, char *value)
151 AVCodecContext *codec = stream->codec;
153 if (!strcmp(attr, "packetization-mode")) {
154 av_log(s, AV_LOG_DEBUG, "RTP Packetization Mode: %d\n", atoi(value));
155 h264_data->packetization_mode = atoi(value);
157 * Packetization Mode:
158 * 0 or not present: Single NAL mode (Only nals from 1-23 are allowed)
159 * 1: Non-interleaved Mode: 1-23, 24 (STAP-A), 28 (FU-A) are allowed.
160 * 2: Interleaved Mode: 25 (STAP-B), 26 (MTAP16), 27 (MTAP24), 28 (FU-A),
161 * and 29 (FU-B) are allowed.
163 if (h264_data->packetization_mode > 1)
164 av_log(s, AV_LOG_ERROR,
165 "Interleaved RTP mode is not supported yet.\n");
166 } else if (!strcmp(attr, "profile-level-id")) {
167 if (strlen(value) == 6)
168 parse_profile_level_id(s, h264_data, value);
169 } else if (!strcmp(attr, "sprop-parameter-sets")) {
170 codec->extradata_size = 0;
171 av_freep(&codec->extradata);
172 return parse_sprop_parameter_sets(s, codec, value);
177 static int h264_handle_packet_stap_a(AVFormatContext *ctx, AVPacket *pkt,
178 const uint8_t *buf, int len)
181 int total_length = 0;
185 for (pass = 0; pass < 2; pass++) {
186 const uint8_t *src = buf;
189 while (src_len > 2) {
190 uint16_t nal_size = AV_RB16(src);
192 // consume the length of the aggregate
196 if (nal_size <= src_len) {
199 total_length += sizeof(start_sequence) + nal_size;
202 memcpy(dst, start_sequence, sizeof(start_sequence));
203 dst += sizeof(start_sequence);
204 memcpy(dst, src, nal_size);
205 COUNT_NAL_TYPE(data, *src);
209 av_log(ctx, AV_LOG_ERROR,
210 "nal size exceeds length: %d %d\n", nal_size, src_len);
211 return AVERROR_INVALIDDATA;
214 // eat what we handled
219 av_log(ctx, AV_LOG_ERROR,
220 "Consumed more bytes than we got! (%d)\n", src_len);
224 /* now we know the total size of the packet (with the
225 * start sequences added) */
226 if ((ret = av_new_packet(pkt, total_length)) < 0)
235 static int h264_handle_packet_fu_a(AVFormatContext *ctx, AVPacket *pkt,
236 const uint8_t *buf, int len)
238 uint8_t fu_indicator, fu_header, start_bit, nal_type, nal;
242 av_log(ctx, AV_LOG_ERROR, "Too short data for FU-A H264 RTP packet\n");
243 return AVERROR_INVALIDDATA;
246 fu_indicator = buf[0];
248 start_bit = fu_header >> 7;
249 nal_type = fu_header & 0x1f;
250 nal = fu_indicator & 0xe0 | nal_type;
252 // skip the fu_indicator and fu_header
257 COUNT_NAL_TYPE(data, nal_type);
258 /* copy in the start sequence, and the reconstructed nal */
259 if ((ret = av_new_packet(pkt, sizeof(start_sequence) + sizeof(nal) + len)) < 0)
261 memcpy(pkt->data, start_sequence, sizeof(start_sequence));
262 pkt->data[sizeof(start_sequence)] = nal;
263 memcpy(pkt->data + sizeof(start_sequence) + sizeof(nal), buf, len);
265 if ((ret = av_new_packet(pkt, len)) < 0)
267 memcpy(pkt->data, buf, len);
273 // return 0 on packet, no more left, 1 on packet, 1 on partial packet
274 static int h264_handle_packet(AVFormatContext *ctx, PayloadContext *data,
275 AVStream *st, AVPacket *pkt, uint32_t *timestamp,
276 const uint8_t *buf, int len, uint16_t seq,
284 av_log(ctx, AV_LOG_ERROR, "Empty H264 RTP packet\n");
285 return AVERROR_INVALIDDATA;
290 /* Simplify the case (these are all the nal types used internally by
291 * the h264 codec). */
292 if (type >= 1 && type <= 23)
295 case 0: // undefined, but pass them through
297 if ((result = av_new_packet(pkt, len + sizeof(start_sequence))) < 0)
299 memcpy(pkt->data, start_sequence, sizeof(start_sequence));
300 memcpy(pkt->data + sizeof(start_sequence), buf, len);
301 COUNT_NAL_TYPE(data, nal);
304 case 24: // STAP-A (one packet, multiple nals)
305 // consume the STAP-A NAL
308 // first we are going to figure out the total size
309 result = h264_handle_packet_stap_a(ctx, pkt, buf, len);
316 av_log(ctx, AV_LOG_ERROR,
317 "Unhandled type (%d) (See RFC for implementation details\n",
319 result = AVERROR(ENOSYS);
322 case 28: // FU-A (fragmented nal)
323 result = h264_handle_packet_fu_a(ctx, pkt, buf, len);
326 case 30: // undefined
327 case 31: // undefined
329 av_log(ctx, AV_LOG_ERROR, "Undefined type (%d)\n", type);
330 result = AVERROR_INVALIDDATA;
334 pkt->stream_index = st->index;
339 static PayloadContext *h264_new_context(void)
341 return av_mallocz(sizeof(PayloadContext) + FF_INPUT_BUFFER_PADDING_SIZE);
344 static void h264_free_context(PayloadContext *data)
349 for (ii = 0; ii < 32; ii++) {
350 if (data->packet_types_received[ii])
351 av_log(NULL, AV_LOG_DEBUG, "Received %d packets of type %d\n",
352 data->packet_types_received[ii], ii);
359 static av_cold int h264_init(AVFormatContext *s, int st_index,
360 PayloadContext *data)
364 s->streams[st_index]->need_parsing = AVSTREAM_PARSE_FULL;
368 static int parse_h264_sdp_line(AVFormatContext *s, int st_index,
369 PayloadContext *h264_data, const char *line)
372 AVCodecContext *codec;
373 const char *p = line;
378 stream = s->streams[st_index];
379 codec = stream->codec;
381 if (av_strstart(p, "framesize:", &p)) {
385 // remove the protocol identifier
386 while (*p && *p == ' ')
387 p++; // strip spaces.
388 while (*p && *p != ' ')
389 p++; // eat protocol identifier
390 while (*p && *p == ' ')
391 p++; // strip trailing spaces.
392 while (*p && *p != '-' && (dst - buf1) < sizeof(buf1) - 1)
396 // a='framesize:96 320-240'
397 // set our parameters
398 codec->width = atoi(buf1);
399 codec->height = atoi(p + 1); // skip the -
400 } else if (av_strstart(p, "fmtp:", &p)) {
401 return ff_parse_fmtp(s, stream, h264_data, p, sdp_parse_fmtp_config_h264);
402 } else if (av_strstart(p, "cliprect:", &p)) {
403 // could use this if we wanted.
409 RTPDynamicProtocolHandler ff_h264_dynamic_handler = {
411 .codec_type = AVMEDIA_TYPE_VIDEO,
412 .codec_id = AV_CODEC_ID_H264,
414 .parse_sdp_a_line = parse_h264_sdp_line,
415 .alloc = h264_new_context,
416 .free = h264_free_context,
417 .parse_packet = h264_handle_packet