9294
|
1 MPlayer container format draft 0.01
|
|
2
|
|
3
|
|
4
|
|
5 Intro:
|
|
6
|
|
7 Features / goals:
|
|
8 (supported by the format, not necessary by a specific implementation)
|
|
9
|
|
10 Simple
|
|
11 use the same encoding for nearly all fields
|
|
12 Extendible
|
|
13 no limit for the possible values for all fields (using universal vlc)
|
|
14 allow adding of new headers in the future
|
|
15 allow adding more fields at the end of headers
|
|
16 Compact
|
|
17 ~0.2% overhead, for normal bitrates
|
|
18 index is <10kb per hour (1 keyframe every 3sec)
|
|
19 Error resistant
|
|
20 seeking / playback without an index
|
|
21 headers & index can be repeated
|
|
22 audio packet reshuffle
|
|
23 checksums to allow quick redownloading of damaged parts
|
|
24 gurantees that no mpeg startcode prefixes occur in the headers with the
|
|
25 exception of codec specific headers (really?)
|
|
26
|
|
27
|
|
28
|
|
29 Definitions:
|
|
30
|
|
31 MUST the specific part must be done to conform to this standard
|
|
32 SHOULD its recommanded to be done that way but its not strictly required
|
|
33
|
|
34
|
|
35
|
|
36 Syntax:
|
|
37
|
|
38 packet header
|
|
39 forward ptr v
|
|
40 backward ptr v
|
|
41
|
|
42 align_byte
|
|
43 while(not byte aligned)
|
|
44 one f(1)
|
|
45
|
|
46 reserved_bytes
|
|
47 for(i=0; i<forward_ptr - length_of_non_reserved; i++)
|
|
48 reserved u(8)
|
|
49
|
|
50 main header:
|
|
51 packet header
|
|
52 main_startcode f(64)
|
|
53 version v
|
|
54 stream_count v
|
|
55 file_size v
|
|
56 length_in msec v
|
|
57 reserved_bytes
|
|
58 checksum u(32)
|
|
59
|
|
60 stream_header:
|
|
61 packet_header
|
|
62 stream_startcode f(64)
|
|
63 stream_id v
|
|
64 stream_class v
|
|
65 fourcc v
|
|
66 average_bitrate v
|
|
67 language_code v
|
|
68 time_base v
|
|
69 lsb_timestamp_length v
|
|
70 fixed_fps u(1)
|
|
71 codec_specific_header_flag u(1)
|
|
72 reserved u(6)
|
|
73
|
|
74 video_stream_header:
|
|
75 stream_header
|
|
76 width v
|
|
77 height v
|
|
78 sample_width v
|
|
79 sample_height v
|
|
80 colorspace_type v
|
|
81 depth v
|
|
82 reserved_bytes
|
|
83 checksum u(32)
|
|
84
|
|
85 audio_stream_header:
|
|
86 stream_header
|
|
87 samplerate v
|
|
88 channel_count v
|
|
89 sub_packet_size v
|
|
90 shuffle_type v
|
|
91 reserved_bytes
|
|
92 checksum u(32)
|
|
93
|
|
94 codec_specific_header:
|
|
95 packet_header
|
|
96 codec_specific_startcode f(64)
|
|
97 stream_id v
|
|
98 codec_specific data
|
|
99 checksum
|
|
100
|
|
101 frame
|
|
102 packet header
|
|
103 if(keyframe){
|
|
104 keyframe_startcode f(64)
|
|
105 }
|
|
106 lsb_timestamp u(lsb_timestamp_length)
|
|
107 stream_id u(log2_stream_count)
|
|
108 priority u(2)
|
|
109 checksum_flag u(1)
|
|
110 msb_timestamp_flag u(1)
|
|
111 align_byte
|
|
112 if(msb_timestamp_flag)
|
|
113 msb_timestamp v
|
|
114 bitstream
|
|
115 if(checksum_flag)
|
|
116 checksum u(32)
|
|
117
|
|
118 Index:
|
|
119 packet header
|
|
120 index_startcode f(64)
|
|
121 stream_id v
|
|
122 index_length v
|
|
123 for(i=0; i<index_length; i++){
|
|
124 index_timestamp v
|
|
125 index_position v
|
|
126 }
|
|
127 checksum u(32)
|
|
128
|
|
129 info_header: (optional)
|
|
130 packet header
|
|
131 info_startcode f(64)
|
|
132 title sz
|
|
133 author sz
|
|
134 copyright sz
|
|
135 description sz
|
|
136 checksum u(32)
|
|
137
|
|
138 v
|
|
139 value=0
|
|
140 do{
|
|
141 more_data u(1)
|
|
142 data u(7)
|
|
143 value= 128*value + data
|
|
144 }while(more_data)
|
|
145 value-=4
|
|
146
|
|
147 sz (zero terminated string)
|
|
148 for(i=0; next_byte != 0; i++){
|
|
149 string[i] u(8)
|
|
150 }
|
|
151 zero_byte f(8)
|
|
152 Note: the string MUST not be 0 bytes long (only a zero byte coded), instead
|
|
153 {'?', 0} should be used instead
|
|
154
|
|
155 f(x) n fixed bits
|
|
156 u(x) unsigned number encoded in x bits in MSB first order
|
|
157
|
|
158
|
|
159 forward_ptr
|
|
160 backward_ptr
|
|
161 pointer to the next / previous packet
|
|
162 Note: a frame with 0 bytes means that its skiped
|
|
163
|
|
164 version
|
|
165 0 for now
|
|
166
|
|
167 file_size
|
|
168 size in bytes, can be 0 if not meaningfull (realtime streams, ...)
|
|
169
|
|
170 length_in_msec
|
|
171 length of the file in milli seconds (can be 0 if realtime or such)
|
|
172
|
|
173 stream_id
|
|
174 Note: streams with a lower relative class MUST have a lower relative id
|
|
175 so a stream with class 0 MUST allways have a id which is lower then any
|
|
176 stream with class > 0
|
|
177 if there is a stream with id n>0 then there MUST be a stream with id n-1
|
|
178
|
|
179 stream_class
|
|
180 0 video
|
|
181 32 audio
|
|
182 64 subtiles
|
|
183 Note the remaining values are reserved and MUST NOT be used
|
|
184
|
|
185 fourcc
|
|
186 identification for the codec
|
|
187 example: 'h'<<24 + '2'<<16 + '6'<<8 + '4'
|
|
188
|
|
189 language_code
|
|
190 something like 'u'<<24 + 's'<<16 + 'e'<<8 + 'n' (US english), can be 0
|
|
191 if unknown
|
|
192
|
|
193 time_base
|
|
194 the number of timer ticks per second, this MUST be equal to the fps
|
|
195 if the fixed_fps is 1
|
|
196 MUST be < 2^15
|
|
197
|
|
198 lsb_timestamp_length
|
|
199 length in bits of the lsb_timestamp
|
|
200 MUST be <16
|
|
201
|
|
202 fixed_fps
|
|
203 1 indicates that the fps is fixed
|
|
204
|
|
205 codec_specific_header_flag
|
|
206 1 indicates that this stream has a codec specific header
|
|
207
|
|
208 msb_timestamp_flag
|
|
209 indicates that the msb_timestamp is coded
|
|
210 MUST be 1 for keyframes
|
|
211
|
|
212 msb_timestamp
|
|
213 most significant bits of the timestamp, SHOULD be 0 for the first frame
|
|
214
|
|
215 lsb_timestamp
|
|
216 most significant bits of the timestamp in time_base precission, with
|
|
217 lsb_timestamp_length bits
|
|
218 Example: IBBP display order
|
|
219 keyframe msb_timestamp=0 lsb_timestamp=0 -> timestamp=0
|
|
220 frame lsb_timestamp=3 -> timestamp=3
|
|
221 frame lsb_timestamp=1 -> timestamp=1
|
|
222 frame lsb_timestamp=2 -> timestamp=2
|
|
223 ...
|
|
224 keyframe msb_timestamp=1 lsb_timestamp=1 -> timestamp=257
|
|
225 frame msb_timestamp=0 lsb_timestamp=255->timestamp=255
|
|
226 frame msb_timestamp=1 lsb_timestamp=0 -> timestamp=256
|
|
227 frame lsb_timestamp=4 -> timestamp=260
|
|
228 frame lsb_timestamp=2 -> timestamp=258
|
|
229 frame lsb_timestamp=3 -> timestamp=259
|
|
230
|
|
231 width/height
|
|
232 MUST be set to the coded width/height
|
|
233
|
|
234 sample_width/sample_height (aspect ratio)
|
|
235 sample_width is the horizontal distance between samples
|
|
236 sample_width and sample_height MUST be relative prime if not zero
|
|
237 MUST be 0 if unknown
|
|
238
|
|
239 depth
|
|
240 for compatibility with some win32 codecs
|
|
241
|
|
242 priority
|
|
243 if 0 then the frame isnt used as reference (b frame) and can be droped
|
|
244 MUST be > 0 for keyframes
|
|
245
|
|
246 sub_packet_size
|
|
247 size of an audio packet
|
|
248 Note a subpacket MUST be in exactly one packet, it cannot be split
|
|
249
|
|
250 shuffle_type
|
|
251 audio is often encoded in small fixed size packets, and to increase the
|
|
252 error robustness these can be shuffled
|
|
253 0 -> no shuffle
|
|
254 1-16 -> interleave packets by 2^n
|
|
255
|
|
256 checksum
|
|
257 crc32 checksum using the generator polynomial=0x04c11db7 (same as ogg)
|
|
258 bit 23 is forced to 1 to avoid MPEG startcode emulation
|
|
259
|
|
260 checksum_flag
|
|
261 indicates that the frame_checksum is coded
|
|
262 must be 1 for the last non keyframe before a keyframe
|
|
263
|
|
264 frame_checksum
|
|
265 identical to checksum, but instead of covering just the current
|
|
266 packet, it covers all frames of the same stream id since the last
|
|
267 checksum
|
|
268 this field is only coded if checksum_flag=1
|
|
269
|
|
270 index_timestamp
|
|
271 value in time_base precission, relative to the last index_timestamp
|
|
272
|
|
273 index_position
|
|
274 position in bytes of the first byte of the keyframe header, relative
|
|
275 to the last index_position
|
|
276
|
|
277
|
|
278
|
|
279 Structure:
|
|
280
|
|
281 the headers MUST be in exactly the following order (to simplify demuxer design)
|
|
282 main header
|
|
283 stream_header (id=0)
|
|
284 codec_specific_header (id=0)
|
|
285 stream_header (id=1)
|
|
286 codec_specific_header (id=1)
|
|
287 ...
|
|
288 stream_header (id=n)
|
|
289 codec_specific_header (id=n)
|
|
290
|
|
291 headers may be repated, but if they are then they MUST all be repeated together
|
|
292 and repeated headers MUST be identical
|
|
293
|
|
294 headers MUST be repeated every 10sec at least ? FIXME
|
|
295
|
|
296
|
|
297 Sample code (GPL, & untested)
|
|
298
|
|
299 typedef BufferContext{
|
|
300 uint8_t *buf;
|
|
301 uint8_t *buf_ptr;
|
|
302 }BufferContext;
|
|
303
|
|
304 static inline uint64_t get_bytes(BufferContext *bc, int count){
|
|
305 uint64_t val=0;
|
|
306
|
|
307 assert(count>0 && count<9)
|
|
308
|
|
309 for(i=0; i<count; i++){
|
|
310 val <<=8;
|
|
311 val += *(bc->buf_ptr++);
|
|
312 }
|
|
313
|
|
314 return val;
|
|
315 }
|
|
316
|
|
317 static inline void put_bytes(BufferContext *bc, int count, uint64_t val){
|
|
318 uint64_t val=0;
|
|
319
|
|
320 assert(count>0 && count<9)
|
|
321
|
|
322 for(i=count-1; i>=0; i--){
|
|
323 *(bc->buf_ptr++)= val >> (8*i);
|
|
324 }
|
|
325
|
|
326 return val;
|
|
327 }
|
|
328
|
|
329 static inline uint64_t get_v(BufferContext *bc){
|
|
330 uint64_t val= 0;
|
|
331
|
|
332 for(;;){
|
|
333 int tmp= *(bc->buf_ptr++);
|
|
334 if(tmp&0x80)
|
|
335 val= (val<<7) + tmp - 0x80;
|
|
336 else
|
|
337 return (val<<7) + tmp - 4;
|
|
338 }
|
|
339 }
|
|
340
|
|
341 static inline void put_v(BufferContext *bc, uint64_t val){
|
|
342 int i;
|
|
343 val+=4;
|
|
344
|
|
345 assert(val);
|
|
346
|
|
347 for(i=56;; i-=8){
|
|
348 if(val>>i) break;
|
|
349 }
|
|
350
|
|
351 for(;i>0; i-=8){
|
|
352 *(bc->buf_ptr++)= 0x80 | (val>>i);
|
|
353 }
|
|
354 *(bc->buf_ptr++)= val&0x7F;
|
|
355 }
|
|
356
|
|
357
|
|
358 Example stream
|
|
359
|
|
360 main header
|
|
361 video_stream_header (stream 0, video jpjp, timebase 30, lsb_timestamp_length=8)
|
|
362 codec_specific_header (stream 0)
|
|
363 video_stream_header (stream 1 subtitle usen, timebase 30, lsb_timestamp_length=8)
|
|
364 video_stream_header (stream 2 subtitle atde, timebase 30, lsb_timestamp_length=8)
|
|
365 audio_stream_header (stream 3, audio jpjp, timebase 1 , lsb_timestamp_length=8)
|
|
366 audio_stream_header (stream 4, audio usen, timebase 1 , lsb_timestamp_length=8)
|
|
367 index (stream 0)
|
|
368 keyframe (stream 0, msb_timestamp=0, lsb_timestamp=0)
|
|
369 keyframe (stream 1, msb_timestamp=0, lsb_timestamp=0)
|
|
370 keyframe (stream 2, msb_timestamp=0, lsb_timestamp=0)
|
|
371 keyframe (stream 3, msb_timestamp=0, lsb_timestamp=0)
|
|
372 keyframe (stream 4, msb_timestamp=0, lsb_timestamp=0)
|
|
373 frame (stream 0, lsb_timestamp=1)
|
|
374 frame (stream 0, lsb_timestamp=2)
|
|
375 ...
|
|
376 frame (stream 0, lsb_timestamp=30)
|
|
377 keyframe (stream 3, msb_timestamp=0, lsb_timestamp=1)
|
|
378 keyframe (stream 4, msb_timestamp=0, lsb_timestamp=1)
|
|
379 frame (stream 0, lsb_timestamp=31)
|
|
380 frame (stream 0, lsb_timestamp=32)
|
|
381 ...
|
|
382 frame (stream 0, lsb_timestamp=60)
|
|
383 frame (stream 1, lsb_timestamp=60)
|
|
384 frame (stream 2, lsb_timestamp=60)
|
|
385 keyframe (stream 3, msb_timestamp=0, lsb_timestamp=2)
|
|
386 keyframe (stream 4, msb_timestamp=0, lsb_timestamp=2)
|
|
387 frame (stream 0, lsb_timestamp=61)
|
|
388 frame (stream 0, lsb_timestamp=62)
|
|
389 ...
|
|
390 main header
|
|
391 video_stream_header (stream 0, video jpjp, timebase 30, lsb_timestamp_length=8)
|
|
392 codec_specific_header (stream 0)
|
|
393 video_stream_header (stream 1 subtitle usen, timebase 30, lsb_timestamp_length=8)
|
|
394 video_stream_header (stream 2 subtitle atde, timebase 30, lsb_timestamp_length=8)
|
|
395 audio_stream_header (stream 3, audio jpjp, timebase 1 , lsb_timestamp_length=8)
|
|
396 audio_stream_header (stream 4, audio usen, timebase 1 , lsb_timestamp_length=8)
|
|
397 frame (stream 0, lsb_timestamp=255)
|
|
398 frame (stream 0, msb_timestamp=1 lsb_timestamp=0)
|
|
399 frame (stream 0, lsb_timestamp=1)
|
|
400 frame (stream 0, lsb_timestamp=2)
|
|
401 frame (stream 1, msb_timestamp=1 lsb_timestamp=2)
|
|
402 frame (stream 2, msb_timestamp=1 lsb_timestamp=2)
|
|
403 frame (stream 0, lsb_timestamp=3)
|
|
404 frame (stream 0, lsb_timestamp=4)
|
|
405 ...
|
|
406 keyframe (stream 3, msb_timestamp=0, lsb_timestamp=9)
|
|
407 keyframe (stream 4, msb_timestamp=0, lsb_timestamp=9)
|
|
408 main header
|
|
409 video_stream_header (stream 0, video jpjp, timebase 30, lsb_timestamp_length=8)
|
|
410 codec_specific_header (stream 0)
|
|
411 video_stream_header (stream 1 subtitle usen, timebase 30, lsb_timestamp_length=8)
|
|
412 video_stream_header (stream 2 subtitle atde, timebase 30, lsb_timestamp_length=8)
|
|
413 audio_stream_header (stream 3, audio jpjp, timebase 1 , lsb_timestamp_length=8)
|
|
414 audio_stream_header (stream 4, audio usen, timebase 1 , lsb_timestamp_length=8)
|
|
415 index (stream 0)
|