FFmpeg
rtpdec_h264.c
Go to the documentation of this file.
1 /*
2  * RTP H.264 Protocol (RFC3984)
3  * Copyright (c) 2006 Ryan Martell
4  *
5  * This file is part of FFmpeg.
6  *
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.
11  *
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.
16  *
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
20  */
21 
22 /**
23  * @file
24  * @brief H.264 / RTP Code (RFC3984)
25  * @author Ryan Martell <rdm4@martellventures.com>
26  *
27  * @note Notes:
28  * Notes:
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,
33  * FU-B packet types)
34  */
35 
36 #include "libavutil/attributes.h"
37 #include "libavutil/base64.h"
38 #include "libavutil/intreadwrite.h"
39 #include "libavutil/avstring.h"
40 #include "avformat.h"
41 
42 #include "rtpdec.h"
43 #include "rtpdec_formats.h"
44 
45 struct PayloadContext {
46  // sdp setup parameters
51 #ifdef DEBUG
52  int packet_types_received[32];
53 #endif
54 };
55 
56 #ifdef DEBUG
57 #define COUNT_NAL_TYPE(data, nal) data->packet_types_received[(nal) & 0x1f]++
58 #define NAL_COUNTERS data->packet_types_received
59 #else
60 #define COUNT_NAL_TYPE(data, nal) do { } while (0)
61 #define NAL_COUNTERS NULL
62 #endif
63 #define NAL_MASK 0x1f
64 
65 static const uint8_t start_sequence[] = { 0, 0, 0, 1 };
66 
68  PayloadContext *h264_data,
69  const char *value)
70 {
71  char buffer[3];
72  // 6 characters=3 bytes, in hex.
76 
77  buffer[0] = value[0];
78  buffer[1] = value[1];
79  buffer[2] = '\0';
80  profile_idc = strtol(buffer, NULL, 16);
81  buffer[0] = value[2];
82  buffer[1] = value[3];
83  profile_iop = strtol(buffer, NULL, 16);
84  buffer[0] = value[4];
85  buffer[1] = value[5];
86  level_idc = strtol(buffer, NULL, 16);
87 
89  "RTP Profile IDC: %x Profile IOP: %x Level: %x\n",
90  profile_idc, profile_iop, level_idc);
91  h264_data->profile_idc = profile_idc;
92  h264_data->profile_iop = profile_iop;
93  h264_data->level_idc = level_idc;
94 }
95 
97  uint8_t **data_ptr, int *size_ptr,
98  const char *value)
99 {
100  char base64packet[1024];
101  uint8_t decoded_packet[1024];
102  int packet_size;
103 
104  while (*value) {
105  char *dst = base64packet;
106 
107  while (*value && *value != ','
108  && (dst - base64packet) < sizeof(base64packet) - 1) {
109  *dst++ = *value++;
110  }
111  *dst++ = '\0';
112 
113  if (*value == ',')
114  value++;
115 
116  packet_size = av_base64_decode(decoded_packet, base64packet,
117  sizeof(decoded_packet));
118  if (packet_size > 0) {
119  uint8_t *dest = av_realloc(*data_ptr,
120  packet_size + sizeof(start_sequence) +
121  *size_ptr +
123  if (!dest) {
124  av_log(s, AV_LOG_ERROR,
125  "Unable to allocate memory for extradata!\n");
126  return AVERROR(ENOMEM);
127  }
128  *data_ptr = dest;
129 
130  memcpy(dest + *size_ptr, start_sequence,
131  sizeof(start_sequence));
132  memcpy(dest + *size_ptr + sizeof(start_sequence),
133  decoded_packet, packet_size);
134  memset(dest + *size_ptr + sizeof(start_sequence) +
135  packet_size, 0, AV_INPUT_BUFFER_PADDING_SIZE);
136 
137  *size_ptr += sizeof(start_sequence) + packet_size;
138  }
139  }
140 
141  return 0;
142 }
143 
145  AVStream *stream,
146  PayloadContext *h264_data,
147  const char *attr, const char *value)
148 {
149  AVCodecParameters *par = stream->codecpar;
150 
151  if (!strcmp(attr, "packetization-mode")) {
152  av_log(s, AV_LOG_DEBUG, "RTP Packetization Mode: %d\n", atoi(value));
153  h264_data->packetization_mode = atoi(value);
154  /*
155  * Packetization Mode:
156  * 0 or not present: Single NAL mode (Only nals from 1-23 are allowed)
157  * 1: Non-interleaved Mode: 1-23, 24 (STAP-A), 28 (FU-A) are allowed.
158  * 2: Interleaved Mode: 25 (STAP-B), 26 (MTAP16), 27 (MTAP24), 28 (FU-A),
159  * and 29 (FU-B) are allowed.
160  */
161  if (h264_data->packetization_mode > 1)
162  av_log(s, AV_LOG_ERROR,
163  "Interleaved RTP mode is not supported yet.\n");
164  } else if (!strcmp(attr, "profile-level-id")) {
165  if (strlen(value) == 6)
166  parse_profile_level_id(s, h264_data, value);
167  } else if (!strcmp(attr, "sprop-parameter-sets")) {
168  int ret;
169  if (*value == 0 || value[strlen(value) - 1] == ',') {
170  av_log(s, AV_LOG_WARNING, "Missing PPS in sprop-parameter-sets, ignoring\n");
171  return 0;
172  }
173  par->extradata_size = 0;
174  av_freep(&par->extradata);
176  &par->extradata_size, value);
177  av_log(s, AV_LOG_DEBUG, "Extradata set to %p (size: %d)\n",
178  par->extradata, par->extradata_size);
179  return ret;
180  }
181  return 0;
182 }
183 
185 {
186  char buf1[50];
187  char *dst = buf1;
188 
189  // remove the protocol identifier
190  while (*p && *p == ' ')
191  p++; // strip spaces.
192  while (*p && *p != ' ')
193  p++; // eat protocol identifier
194  while (*p && *p == ' ')
195  p++; // strip trailing spaces.
196  while (*p && *p != '-' && (dst - buf1) < sizeof(buf1) - 1)
197  *dst++ = *p++;
198  *dst = '\0';
199 
200  // a='framesize:96 320-240'
201  // set our parameters
202  par->width = atoi(buf1);
203  par->height = atoi(p + 1); // skip the -
204 }
205 
207  const uint8_t *buf, int len,
208  int skip_between, int *nal_counters,
209  int nal_mask)
210 {
211  int pass = 0;
212  int total_length = 0;
213  uint8_t *dst = NULL;
214  int ret;
215 
216  // first we are going to figure out the total size
217  for (pass = 0; pass < 2; pass++) {
218  const uint8_t *src = buf;
219  int src_len = len;
220 
221  while (src_len > 2) {
222  uint16_t nal_size = AV_RB16(src);
223 
224  // consume the length of the aggregate
225  src += 2;
226  src_len -= 2;
227 
228  if (nal_size <= src_len) {
229  if (pass == 0) {
230  // counting
231  total_length += sizeof(start_sequence) + nal_size;
232  } else {
233  // copying
234  memcpy(dst, start_sequence, sizeof(start_sequence));
235  dst += sizeof(start_sequence);
236  memcpy(dst, src, nal_size);
237  if (nal_counters)
238  nal_counters[(*src) & nal_mask]++;
239  dst += nal_size;
240  }
241  } else {
242  av_log(ctx, AV_LOG_ERROR,
243  "nal size exceeds length: %d %d\n", nal_size, src_len);
244  return AVERROR_INVALIDDATA;
245  }
246 
247  // eat what we handled
248  src += nal_size + skip_between;
249  src_len -= nal_size + skip_between;
250  }
251 
252  if (pass == 0) {
253  /* now we know the total size of the packet (with the
254  * start sequences added) */
255  if ((ret = av_new_packet(pkt, total_length)) < 0)
256  return ret;
257  dst = pkt->data;
258  }
259  }
260 
261  return 0;
262 }
263 
265  int start_bit, const uint8_t *nal_header,
266  int nal_header_len)
267 {
268  int ret;
269  int tot_len = len;
270  int pos = 0;
271  if (start_bit)
272  tot_len += sizeof(start_sequence) + nal_header_len;
273  if ((ret = av_new_packet(pkt, tot_len)) < 0)
274  return ret;
275  if (start_bit) {
276  memcpy(pkt->data + pos, start_sequence, sizeof(start_sequence));
277  pos += sizeof(start_sequence);
278  memcpy(pkt->data + pos, nal_header, nal_header_len);
279  pos += nal_header_len;
280  }
281  memcpy(pkt->data + pos, buf, len);
282  return 0;
283 }
284 
286  const uint8_t *buf, int len,
287  int *nal_counters, int nal_mask)
288 {
289  uint8_t fu_indicator, fu_header, start_bit, nal_type, nal;
290 
291  if (len < 3) {
292  av_log(ctx, AV_LOG_ERROR, "Too short data for FU-A H.264 RTP packet\n");
293  return AVERROR_INVALIDDATA;
294  }
295 
296  fu_indicator = buf[0];
297  fu_header = buf[1];
298  start_bit = fu_header >> 7;
299  nal_type = fu_header & 0x1f;
300  nal = fu_indicator & 0xe0 | nal_type;
301 
302  // skip the fu_indicator and fu_header
303  buf += 2;
304  len -= 2;
305 
306  if (start_bit && nal_counters)
307  nal_counters[nal_type & nal_mask]++;
308  return ff_h264_handle_frag_packet(pkt, buf, len, start_bit, &nal, 1);
309 }
310 
311 // return 0 on packet, no more left, 1 on packet, 1 on partial packet
313  AVStream *st, AVPacket *pkt, uint32_t *timestamp,
314  const uint8_t *buf, int len, uint16_t seq,
315  int flags)
316 {
317  uint8_t nal;
318  uint8_t type;
319  int result = 0;
320 
321  if (!len) {
322  av_log(ctx, AV_LOG_ERROR, "Empty H.264 RTP packet\n");
323  return AVERROR_INVALIDDATA;
324  }
325  nal = buf[0];
326  type = nal & 0x1f;
327 
328  /* Simplify the case (these are all the NAL types used internally by
329  * the H.264 codec). */
330  if (type >= 1 && type <= 23)
331  type = 1;
332  switch (type) {
333  case 0: // undefined, but pass them through
334  case 1:
335  if ((result = av_new_packet(pkt, len + sizeof(start_sequence))) < 0)
336  return result;
337  memcpy(pkt->data, start_sequence, sizeof(start_sequence));
338  memcpy(pkt->data + sizeof(start_sequence), buf, len);
339  COUNT_NAL_TYPE(data, nal);
340  break;
341 
342  case 24: // STAP-A (one packet, multiple nals)
343  // consume the STAP-A NAL
344  buf++;
345  len--;
346  result = ff_h264_handle_aggregated_packet(ctx, data, pkt, buf, len, 0,
348  break;
349 
350  case 25: // STAP-B
351  case 26: // MTAP-16
352  case 27: // MTAP-24
353  case 29: // FU-B
354  avpriv_report_missing_feature(ctx, "RTP H.264 NAL unit type %d", type);
355  result = AVERROR_PATCHWELCOME;
356  break;
357 
358  case 28: // FU-A (fragmented nal)
359  result = h264_handle_packet_fu_a(ctx, data, pkt, buf, len,
361  break;
362 
363  case 30: // undefined
364  case 31: // undefined
365  default:
366  av_log(ctx, AV_LOG_ERROR, "Undefined type (%d)\n", type);
367  result = AVERROR_INVALIDDATA;
368  break;
369  }
370 
371  pkt->stream_index = st->index;
372 
373  return result;
374 }
375 
377 {
378 #ifdef DEBUG
379  int ii;
380 
381  for (ii = 0; ii < 32; ii++) {
382  if (data->packet_types_received[ii])
383  av_log(NULL, AV_LOG_DEBUG, "Received %d packets of type %d\n",
384  data->packet_types_received[ii], ii);
385  }
386 #endif
387 }
388 
389 static int parse_h264_sdp_line(AVFormatContext *s, int st_index,
390  PayloadContext *h264_data, const char *line)
391 {
392  AVStream *stream;
393  const char *p = line;
394 
395  if (st_index < 0)
396  return 0;
397 
398  stream = s->streams[st_index];
399 
400  if (av_strstart(p, "framesize:", &p)) {
401  ff_h264_parse_framesize(stream->codecpar, p);
402  } else if (av_strstart(p, "fmtp:", &p)) {
403  return ff_parse_fmtp(s, stream, h264_data, p, sdp_parse_fmtp_config_h264);
404  } else if (av_strstart(p, "cliprect:", &p)) {
405  // could use this if we wanted.
406  }
407 
408  return 0;
409 }
410 
412  .enc_name = "H264",
413  .codec_type = AVMEDIA_TYPE_VIDEO,
414  .codec_id = AV_CODEC_ID_H264,
415  .need_parsing = AVSTREAM_PARSE_FULL,
416  .priv_data_size = sizeof(PayloadContext),
417  .parse_sdp_a_line = parse_h264_sdp_line,
418  .close = h264_close_context,
420 };
uint8_t profile_idc
Definition: rtpdec_h264.c:47
const RTPDynamicProtocolHandler ff_h264_dynamic_handler
Definition: rtpdec_h264.c:411
AVPacket pkt
Definition: rtpdec_qt.c:37
#define NULL
Definition: coverity.c:32
#define AVERROR_INVALIDDATA
Invalid data found when processing input.
Definition: error.h:59
void * av_realloc(void *ptr, size_t size)
Allocate, reallocate, or free a block of memory.
Definition: mem.c:135
ptrdiff_t const GLvoid * data
Definition: opengl_enc.c:100
#define AV_LOG_WARNING
Something somehow does not look correct.
Definition: log.h:182
RTP/JPEG specific private data.
Definition: rdt.c:83
int index
stream index in AVFormatContext
Definition: avformat.h:882
static int h264_handle_packet_fu_a(AVFormatContext *ctx, PayloadContext *data, AVPacket *pkt, const uint8_t *buf, int len, int *nal_counters, int nal_mask)
Definition: rtpdec_h264.c:285
int packetization_mode
Definition: rtpdec_h264.c:50
GLint GLenum type
Definition: opengl_enc.c:104
uint64_t_TMPL AV_WL64 unsigned int_TMPL AV_WL32 unsigned int_TMPL AV_WL24 unsigned int_TMPL AV_WL16 uint64_t_TMPL AV_WB64 unsigned int_TMPL AV_WB32 unsigned int_TMPL AV_WB24 unsigned int_TMPL AV_RB16
Definition: bytestream.h:87
#define src
Definition: vp8dsp.c:254
This struct describes the properties of an encoded stream.
Definition: avcodec.h:3949
Macro definitions for various function/variable attributes.
Format I/O context.
Definition: avformat.h:1358
uint8_t
int width
Video only.
Definition: avcodec.h:4023
GLsizei GLboolean const GLfloat * value
Definition: opengl_enc.c:108
AVStream ** streams
A list of all streams in the file.
Definition: avformat.h:1426
uint8_t * data
Definition: avcodec.h:1477
int ff_h264_handle_frag_packet(AVPacket *pkt, const uint8_t *buf, int len, int start_bit, const uint8_t *nal_header, int nal_header_len)
Definition: rtpdec_h264.c:264
#define av_log(a,...)
int av_new_packet(AVPacket *pkt, int size)
Allocate the payload of a packet and initialize its fields with default values.
Definition: avpacket.c:86
uint32_t timestamp
current frame timestamp
Definition: rtpdec_ac3.c:31
#define AV_LOG_ERROR
Something went wrong and cannot losslessly be recovered.
Definition: log.h:176
static void h264_close_context(PayloadContext *data)
Definition: rtpdec_h264.c:376
static int sdp_parse_fmtp_config_h264(AVFormatContext *s, AVStream *stream, PayloadContext *h264_data, const char *attr, const char *value)
Definition: rtpdec_h264.c:144
#define AV_LOG_DEBUG
Stuff which is only useful for libav* developers.
Definition: log.h:197
Definition: graph2dot.c:48
static int parse_h264_sdp_line(AVFormatContext *s, int st_index, PayloadContext *h264_data, const char *line)
Definition: rtpdec_h264.c:389
int extradata_size
Size of the extradata content in bytes.
Definition: avcodec.h:3975
#define pass
Definition: fft_template.c:619
void ff_h264_parse_framesize(AVCodecParameters *par, const char *p)
Definition: rtpdec_h264.c:184
AVFormatContext * ctx
Definition: movenc.c:48
#define s(width, name)
Definition: cbs_vp9.c:257
uint8_t profile_iop
Definition: rtpdec_h264.c:48
#define COUNT_NAL_TYPE(data, nal)
Definition: rtpdec_h264.c:60
Stream structure.
Definition: avformat.h:881
#define AVERROR_PATCHWELCOME
Not yet implemented in FFmpeg, patches welcome.
Definition: error.h:62
static void parse_profile_level_id(AVFormatContext *s, PayloadContext *h264_data, const char *value)
Definition: rtpdec_h264.c:67
uint8_t level_idc
Definition: rtpdec_h264.c:49
uint8_t * buf
the temporary storage buffer
Definition: rtpdec_asf.c:183
int ff_parse_fmtp(AVFormatContext *s, AVStream *stream, PayloadContext *data, const char *p, int(*parse_fmtp)(AVFormatContext *s, AVStream *stream, PayloadContext *data, const char *attr, const char *value))
Definition: rtpdec.c:889
const char * enc_name
Definition: rtpdec.h:116
void avpriv_report_missing_feature(void *avc, const char *msg,...) av_printf_format(2
Log a generic warning message about a missing feature.
#define flags(name, subs,...)
Definition: cbs_av1.c:561
#define NAL_MASK
Definition: rtpdec_h264.c:63
int av_strstart(const char *str, const char *pfx, const char **ptr)
Return non-zero if pfx is a prefix of str.
Definition: avstring.c:34
full parsing and repack
Definition: avformat.h:800
Main libavformat public API header.
The official guide to swscale for confused that consecutive non overlapping rectangles of slice_bottom special converter These generally are unscaled converters of common like for each output line the vertical scaler pulls lines from a ring buffer When the ring buffer does not contain the wanted line
Definition: swscale.txt:33
#define NAL_COUNTERS
Definition: rtpdec_h264.c:61
static int h264_handle_packet(AVFormatContext *ctx, PayloadContext *data, AVStream *st, AVPacket *pkt, uint32_t *timestamp, const uint8_t *buf, int len, uint16_t seq, int flags)
Definition: rtpdec_h264.c:312
int ff_h264_parse_sprop_parameter_sets(AVFormatContext *s, uint8_t **data_ptr, int *size_ptr, const char *value)
Definition: rtpdec_h264.c:96
#define AV_INPUT_BUFFER_PADDING_SIZE
Required number of additionally allocated bytes at the end of the input bitstream for decoding...
Definition: avcodec.h:790
static int parse_packet(AVFormatContext *s, AVPacket *pkt, int stream_index)
Parse a packet, add all split parts to parse_queue.
Definition: utils.c:1438
int ff_h264_handle_aggregated_packet(AVFormatContext *ctx, PayloadContext *data, AVPacket *pkt, const uint8_t *buf, int len, int skip_between, int *nal_counters, int nal_mask)
Definition: rtpdec_h264.c:206
static const uint8_t start_sequence[]
Definition: rtpdec_h264.c:65
int av_base64_decode(uint8_t *out, const char *in_str, int out_size)
Decode a base64-encoded string.
Definition: base64.c:79
uint8_t * extradata
Extra binary data needed for initializing the decoder, codec-dependent.
Definition: avcodec.h:3971
and forward the result(frame or status change) to the corresponding input.If nothing is possible
#define av_freep(p)
AVCodecParameters * codecpar
Codec parameters associated with this stream.
Definition: avformat.h:1028
int stream_index
Definition: avcodec.h:1479
Filter the word “frame” indicates either a video frame or a group of audio as stored in an AVFrame structure Format for each input and each output the list of supported formats For video that means pixel format For audio that means channel sample they are references to shared objects When the negotiation mechanism computes the intersection of the formats supported at each end of a all references to both lists are replaced with a reference to the intersection And when a single format is eventually chosen for a link amongst the remaining all references to the list are updated That means that if a filter requires that its input and output have the same format amongst a supported all it has to do is use a reference to the same list of formats query_formats can leave some formats unset and return AVERROR(EAGAIN) to cause the negotiation mechanism toagain later.That can be used by filters with complex requirements to use the format negotiated on one link to set the formats supported on another.Frame references ownership and permissions
This structure stores compressed data.
Definition: avcodec.h:1454
GLuint buffer
Definition: opengl_enc.c:101