1 | """Stuff to parse Sun and NeXT audio files. |
---|
2 | |
---|
3 | An audio file consists of a header followed by the data. The structure |
---|
4 | of the header is as follows. |
---|
5 | |
---|
6 | +---------------+ |
---|
7 | | magic word | |
---|
8 | +---------------+ |
---|
9 | | header size | |
---|
10 | +---------------+ |
---|
11 | | data size | |
---|
12 | +---------------+ |
---|
13 | | encoding | |
---|
14 | +---------------+ |
---|
15 | | sample rate | |
---|
16 | +---------------+ |
---|
17 | | # of channels | |
---|
18 | +---------------+ |
---|
19 | | info | |
---|
20 | | | |
---|
21 | +---------------+ |
---|
22 | |
---|
23 | The magic word consists of the 4 characters '.snd'. Apart from the |
---|
24 | info field, all header fields are 4 bytes in size. They are all |
---|
25 | 32-bit unsigned integers encoded in big-endian byte order. |
---|
26 | |
---|
27 | The header size really gives the start of the data. |
---|
28 | The data size is the physical size of the data. From the other |
---|
29 | parameters the number of frames can be calculated. |
---|
30 | The encoding gives the way in which audio samples are encoded. |
---|
31 | Possible values are listed below. |
---|
32 | The info field currently consists of an ASCII string giving a |
---|
33 | human-readable description of the audio file. The info field is |
---|
34 | padded with NUL bytes to the header size. |
---|
35 | |
---|
36 | Usage. |
---|
37 | |
---|
38 | Reading audio files: |
---|
39 | f = sunau.open(file, 'r') |
---|
40 | where file is either the name of a file or an open file pointer. |
---|
41 | The open file pointer must have methods read(), seek(), and close(). |
---|
42 | When the setpos() and rewind() methods are not used, the seek() |
---|
43 | method is not necessary. |
---|
44 | |
---|
45 | This returns an instance of a class with the following public methods: |
---|
46 | getnchannels() -- returns number of audio channels (1 for |
---|
47 | mono, 2 for stereo) |
---|
48 | getsampwidth() -- returns sample width in bytes |
---|
49 | getframerate() -- returns sampling frequency |
---|
50 | getnframes() -- returns number of audio frames |
---|
51 | getcomptype() -- returns compression type ('NONE' or 'ULAW') |
---|
52 | getcompname() -- returns human-readable version of |
---|
53 | compression type ('not compressed' matches 'NONE') |
---|
54 | getparams() -- returns a tuple consisting of all of the |
---|
55 | above in the above order |
---|
56 | getmarkers() -- returns None (for compatibility with the |
---|
57 | aifc module) |
---|
58 | getmark(id) -- raises an error since the mark does not |
---|
59 | exist (for compatibility with the aifc module) |
---|
60 | readframes(n) -- returns at most n frames of audio |
---|
61 | rewind() -- rewind to the beginning of the audio stream |
---|
62 | setpos(pos) -- seek to the specified position |
---|
63 | tell() -- return the current position |
---|
64 | close() -- close the instance (make it unusable) |
---|
65 | The position returned by tell() and the position given to setpos() |
---|
66 | are compatible and have nothing to do with the actual position in the |
---|
67 | file. |
---|
68 | The close() method is called automatically when the class instance |
---|
69 | is destroyed. |
---|
70 | |
---|
71 | Writing audio files: |
---|
72 | f = sunau.open(file, 'w') |
---|
73 | where file is either the name of a file or an open file pointer. |
---|
74 | The open file pointer must have methods write(), tell(), seek(), and |
---|
75 | close(). |
---|
76 | |
---|
77 | This returns an instance of a class with the following public methods: |
---|
78 | setnchannels(n) -- set the number of channels |
---|
79 | setsampwidth(n) -- set the sample width |
---|
80 | setframerate(n) -- set the frame rate |
---|
81 | setnframes(n) -- set the number of frames |
---|
82 | setcomptype(type, name) |
---|
83 | -- set the compression type and the |
---|
84 | human-readable compression type |
---|
85 | setparams(tuple)-- set all parameters at once |
---|
86 | tell() -- return current position in output file |
---|
87 | writeframesraw(data) |
---|
88 | -- write audio frames without pathing up the |
---|
89 | file header |
---|
90 | writeframes(data) |
---|
91 | -- write audio frames and patch up the file header |
---|
92 | close() -- patch up the file header and close the |
---|
93 | output file |
---|
94 | You should set the parameters before the first writeframesraw or |
---|
95 | writeframes. The total number of frames does not need to be set, |
---|
96 | but when it is set to the correct value, the header does not have to |
---|
97 | be patched up. |
---|
98 | It is best to first set all parameters, perhaps possibly the |
---|
99 | compression type, and then write audio frames using writeframesraw. |
---|
100 | When all frames have been written, either call writeframes('') or |
---|
101 | close() to patch up the sizes in the header. |
---|
102 | The close() method is called automatically when the class instance |
---|
103 | is destroyed. |
---|
104 | """ |
---|
105 | |
---|
106 | # from <multimedia/audio_filehdr.h> |
---|
107 | AUDIO_FILE_MAGIC = 0x2e736e64 |
---|
108 | AUDIO_FILE_ENCODING_MULAW_8 = 1 |
---|
109 | AUDIO_FILE_ENCODING_LINEAR_8 = 2 |
---|
110 | AUDIO_FILE_ENCODING_LINEAR_16 = 3 |
---|
111 | AUDIO_FILE_ENCODING_LINEAR_24 = 4 |
---|
112 | AUDIO_FILE_ENCODING_LINEAR_32 = 5 |
---|
113 | AUDIO_FILE_ENCODING_FLOAT = 6 |
---|
114 | AUDIO_FILE_ENCODING_DOUBLE = 7 |
---|
115 | AUDIO_FILE_ENCODING_ADPCM_G721 = 23 |
---|
116 | AUDIO_FILE_ENCODING_ADPCM_G722 = 24 |
---|
117 | AUDIO_FILE_ENCODING_ADPCM_G723_3 = 25 |
---|
118 | AUDIO_FILE_ENCODING_ADPCM_G723_5 = 26 |
---|
119 | AUDIO_FILE_ENCODING_ALAW_8 = 27 |
---|
120 | |
---|
121 | # from <multimedia/audio_hdr.h> |
---|
122 | AUDIO_UNKNOWN_SIZE = 0xFFFFFFFFL # ((unsigned)(~0)) |
---|
123 | |
---|
124 | _simple_encodings = [AUDIO_FILE_ENCODING_MULAW_8, |
---|
125 | AUDIO_FILE_ENCODING_LINEAR_8, |
---|
126 | AUDIO_FILE_ENCODING_LINEAR_16, |
---|
127 | AUDIO_FILE_ENCODING_LINEAR_24, |
---|
128 | AUDIO_FILE_ENCODING_LINEAR_32, |
---|
129 | AUDIO_FILE_ENCODING_ALAW_8] |
---|
130 | |
---|
131 | class Error(Exception): |
---|
132 | pass |
---|
133 | |
---|
134 | def _read_u32(file): |
---|
135 | x = 0L |
---|
136 | for i in range(4): |
---|
137 | byte = file.read(1) |
---|
138 | if byte == '': |
---|
139 | raise EOFError |
---|
140 | x = x*256 + ord(byte) |
---|
141 | return x |
---|
142 | |
---|
143 | def _write_u32(file, x): |
---|
144 | data = [] |
---|
145 | for i in range(4): |
---|
146 | d, m = divmod(x, 256) |
---|
147 | data.insert(0, m) |
---|
148 | x = d |
---|
149 | for i in range(4): |
---|
150 | file.write(chr(int(data[i]))) |
---|
151 | |
---|
152 | class Au_read: |
---|
153 | |
---|
154 | def __init__(self, f): |
---|
155 | if type(f) == type(''): |
---|
156 | import __builtin__ |
---|
157 | f = __builtin__.open(f, 'rb') |
---|
158 | self.initfp(f) |
---|
159 | |
---|
160 | def __del__(self): |
---|
161 | if self._file: |
---|
162 | self.close() |
---|
163 | |
---|
164 | def initfp(self, file): |
---|
165 | self._file = file |
---|
166 | self._soundpos = 0 |
---|
167 | magic = int(_read_u32(file)) |
---|
168 | if magic != AUDIO_FILE_MAGIC: |
---|
169 | raise Error, 'bad magic number' |
---|
170 | self._hdr_size = int(_read_u32(file)) |
---|
171 | if self._hdr_size < 24: |
---|
172 | raise Error, 'header size too small' |
---|
173 | if self._hdr_size > 100: |
---|
174 | raise Error, 'header size ridiculously large' |
---|
175 | self._data_size = _read_u32(file) |
---|
176 | if self._data_size != AUDIO_UNKNOWN_SIZE: |
---|
177 | self._data_size = int(self._data_size) |
---|
178 | self._encoding = int(_read_u32(file)) |
---|
179 | if self._encoding not in _simple_encodings: |
---|
180 | raise Error, 'encoding not (yet) supported' |
---|
181 | if self._encoding in (AUDIO_FILE_ENCODING_MULAW_8, |
---|
182 | AUDIO_FILE_ENCODING_ALAW_8): |
---|
183 | self._sampwidth = 2 |
---|
184 | self._framesize = 1 |
---|
185 | elif self._encoding == AUDIO_FILE_ENCODING_LINEAR_8: |
---|
186 | self._framesize = self._sampwidth = 1 |
---|
187 | elif self._encoding == AUDIO_FILE_ENCODING_LINEAR_16: |
---|
188 | self._framesize = self._sampwidth = 2 |
---|
189 | elif self._encoding == AUDIO_FILE_ENCODING_LINEAR_24: |
---|
190 | self._framesize = self._sampwidth = 3 |
---|
191 | elif self._encoding == AUDIO_FILE_ENCODING_LINEAR_32: |
---|
192 | self._framesize = self._sampwidth = 4 |
---|
193 | else: |
---|
194 | raise Error, 'unknown encoding' |
---|
195 | self._framerate = int(_read_u32(file)) |
---|
196 | self._nchannels = int(_read_u32(file)) |
---|
197 | self._framesize = self._framesize * self._nchannels |
---|
198 | if self._hdr_size > 24: |
---|
199 | self._info = file.read(self._hdr_size - 24) |
---|
200 | for i in range(len(self._info)): |
---|
201 | if self._info[i] == '\0': |
---|
202 | self._info = self._info[:i] |
---|
203 | break |
---|
204 | else: |
---|
205 | self._info = '' |
---|
206 | try: |
---|
207 | self._data_pos = file.tell() |
---|
208 | except (AttributeError, IOError): |
---|
209 | self._data_pos = None |
---|
210 | |
---|
211 | def getfp(self): |
---|
212 | return self._file |
---|
213 | |
---|
214 | def getnchannels(self): |
---|
215 | return self._nchannels |
---|
216 | |
---|
217 | def getsampwidth(self): |
---|
218 | return self._sampwidth |
---|
219 | |
---|
220 | def getframerate(self): |
---|
221 | return self._framerate |
---|
222 | |
---|
223 | def getnframes(self): |
---|
224 | if self._data_size == AUDIO_UNKNOWN_SIZE: |
---|
225 | return AUDIO_UNKNOWN_SIZE |
---|
226 | if self._encoding in _simple_encodings: |
---|
227 | return self._data_size // self._framesize |
---|
228 | return 0 # XXX--must do some arithmetic here |
---|
229 | |
---|
230 | def getcomptype(self): |
---|
231 | if self._encoding == AUDIO_FILE_ENCODING_MULAW_8: |
---|
232 | return 'ULAW' |
---|
233 | elif self._encoding == AUDIO_FILE_ENCODING_ALAW_8: |
---|
234 | return 'ALAW' |
---|
235 | else: |
---|
236 | return 'NONE' |
---|
237 | |
---|
238 | def getcompname(self): |
---|
239 | if self._encoding == AUDIO_FILE_ENCODING_MULAW_8: |
---|
240 | return 'CCITT G.711 u-law' |
---|
241 | elif self._encoding == AUDIO_FILE_ENCODING_ALAW_8: |
---|
242 | return 'CCITT G.711 A-law' |
---|
243 | else: |
---|
244 | return 'not compressed' |
---|
245 | |
---|
246 | def getparams(self): |
---|
247 | return self.getnchannels(), self.getsampwidth(), \ |
---|
248 | self.getframerate(), self.getnframes(), \ |
---|
249 | self.getcomptype(), self.getcompname() |
---|
250 | |
---|
251 | def getmarkers(self): |
---|
252 | return None |
---|
253 | |
---|
254 | def getmark(self, id): |
---|
255 | raise Error, 'no marks' |
---|
256 | |
---|
257 | def readframes(self, nframes): |
---|
258 | if self._encoding in _simple_encodings: |
---|
259 | if nframes == AUDIO_UNKNOWN_SIZE: |
---|
260 | data = self._file.read() |
---|
261 | else: |
---|
262 | data = self._file.read(nframes * self._framesize) |
---|
263 | self._soundpos += len(data) // self._framesize |
---|
264 | if self._encoding == AUDIO_FILE_ENCODING_MULAW_8: |
---|
265 | import audioop |
---|
266 | data = audioop.ulaw2lin(data, self._sampwidth) |
---|
267 | return data |
---|
268 | return None # XXX--not implemented yet |
---|
269 | |
---|
270 | def rewind(self): |
---|
271 | if self._data_pos is None: |
---|
272 | raise IOError('cannot seek') |
---|
273 | self._file.seek(self._data_pos) |
---|
274 | self._soundpos = 0 |
---|
275 | |
---|
276 | def tell(self): |
---|
277 | return self._soundpos |
---|
278 | |
---|
279 | def setpos(self, pos): |
---|
280 | if pos < 0 or pos > self.getnframes(): |
---|
281 | raise Error, 'position not in range' |
---|
282 | if self._data_pos is None: |
---|
283 | raise IOError('cannot seek') |
---|
284 | self._file.seek(self._data_pos + pos * self._framesize) |
---|
285 | self._soundpos = pos |
---|
286 | |
---|
287 | def close(self): |
---|
288 | self._file = None |
---|
289 | |
---|
290 | class Au_write: |
---|
291 | |
---|
292 | def __init__(self, f): |
---|
293 | if type(f) == type(''): |
---|
294 | import __builtin__ |
---|
295 | f = __builtin__.open(f, 'wb') |
---|
296 | self.initfp(f) |
---|
297 | |
---|
298 | def __del__(self): |
---|
299 | if self._file: |
---|
300 | self.close() |
---|
301 | |
---|
302 | def initfp(self, file): |
---|
303 | self._file = file |
---|
304 | self._framerate = 0 |
---|
305 | self._nchannels = 0 |
---|
306 | self._sampwidth = 0 |
---|
307 | self._framesize = 0 |
---|
308 | self._nframes = AUDIO_UNKNOWN_SIZE |
---|
309 | self._nframeswritten = 0 |
---|
310 | self._datawritten = 0 |
---|
311 | self._datalength = 0 |
---|
312 | self._info = '' |
---|
313 | self._comptype = 'ULAW' # default is U-law |
---|
314 | |
---|
315 | def setnchannels(self, nchannels): |
---|
316 | if self._nframeswritten: |
---|
317 | raise Error, 'cannot change parameters after starting to write' |
---|
318 | if nchannels not in (1, 2, 4): |
---|
319 | raise Error, 'only 1, 2, or 4 channels supported' |
---|
320 | self._nchannels = nchannels |
---|
321 | |
---|
322 | def getnchannels(self): |
---|
323 | if not self._nchannels: |
---|
324 | raise Error, 'number of channels not set' |
---|
325 | return self._nchannels |
---|
326 | |
---|
327 | def setsampwidth(self, sampwidth): |
---|
328 | if self._nframeswritten: |
---|
329 | raise Error, 'cannot change parameters after starting to write' |
---|
330 | if sampwidth not in (1, 2, 4): |
---|
331 | raise Error, 'bad sample width' |
---|
332 | self._sampwidth = sampwidth |
---|
333 | |
---|
334 | def getsampwidth(self): |
---|
335 | if not self._framerate: |
---|
336 | raise Error, 'sample width not specified' |
---|
337 | return self._sampwidth |
---|
338 | |
---|
339 | def setframerate(self, framerate): |
---|
340 | if self._nframeswritten: |
---|
341 | raise Error, 'cannot change parameters after starting to write' |
---|
342 | self._framerate = framerate |
---|
343 | |
---|
344 | def getframerate(self): |
---|
345 | if not self._framerate: |
---|
346 | raise Error, 'frame rate not set' |
---|
347 | return self._framerate |
---|
348 | |
---|
349 | def setnframes(self, nframes): |
---|
350 | if self._nframeswritten: |
---|
351 | raise Error, 'cannot change parameters after starting to write' |
---|
352 | if nframes < 0: |
---|
353 | raise Error, '# of frames cannot be negative' |
---|
354 | self._nframes = nframes |
---|
355 | |
---|
356 | def getnframes(self): |
---|
357 | return self._nframeswritten |
---|
358 | |
---|
359 | def setcomptype(self, type, name): |
---|
360 | if type in ('NONE', 'ULAW'): |
---|
361 | self._comptype = type |
---|
362 | else: |
---|
363 | raise Error, 'unknown compression type' |
---|
364 | |
---|
365 | def getcomptype(self): |
---|
366 | return self._comptype |
---|
367 | |
---|
368 | def getcompname(self): |
---|
369 | if self._comptype == 'ULAW': |
---|
370 | return 'CCITT G.711 u-law' |
---|
371 | elif self._comptype == 'ALAW': |
---|
372 | return 'CCITT G.711 A-law' |
---|
373 | else: |
---|
374 | return 'not compressed' |
---|
375 | |
---|
376 | def setparams(self, params): |
---|
377 | nchannels, sampwidth, framerate, nframes, comptype, compname = params |
---|
378 | self.setnchannels(nchannels) |
---|
379 | self.setsampwidth(sampwidth) |
---|
380 | self.setframerate(framerate) |
---|
381 | self.setnframes(nframes) |
---|
382 | self.setcomptype(comptype, compname) |
---|
383 | |
---|
384 | def getparams(self): |
---|
385 | return self.getnchannels(), self.getsampwidth(), \ |
---|
386 | self.getframerate(), self.getnframes(), \ |
---|
387 | self.getcomptype(), self.getcompname() |
---|
388 | |
---|
389 | def tell(self): |
---|
390 | return self._nframeswritten |
---|
391 | |
---|
392 | def writeframesraw(self, data): |
---|
393 | self._ensure_header_written() |
---|
394 | if self._comptype == 'ULAW': |
---|
395 | import audioop |
---|
396 | data = audioop.lin2ulaw(data, self._sampwidth) |
---|
397 | nframes = len(data) // self._framesize |
---|
398 | self._file.write(data) |
---|
399 | self._nframeswritten = self._nframeswritten + nframes |
---|
400 | self._datawritten = self._datawritten + len(data) |
---|
401 | |
---|
402 | def writeframes(self, data): |
---|
403 | self.writeframesraw(data) |
---|
404 | if self._nframeswritten != self._nframes or \ |
---|
405 | self._datalength != self._datawritten: |
---|
406 | self._patchheader() |
---|
407 | |
---|
408 | def close(self): |
---|
409 | if self._file: |
---|
410 | try: |
---|
411 | self._ensure_header_written() |
---|
412 | if self._nframeswritten != self._nframes or \ |
---|
413 | self._datalength != self._datawritten: |
---|
414 | self._patchheader() |
---|
415 | self._file.flush() |
---|
416 | finally: |
---|
417 | self._file = None |
---|
418 | |
---|
419 | # |
---|
420 | # private methods |
---|
421 | # |
---|
422 | |
---|
423 | def _ensure_header_written(self): |
---|
424 | if not self._nframeswritten: |
---|
425 | if not self._nchannels: |
---|
426 | raise Error, '# of channels not specified' |
---|
427 | if not self._sampwidth: |
---|
428 | raise Error, 'sample width not specified' |
---|
429 | if not self._framerate: |
---|
430 | raise Error, 'frame rate not specified' |
---|
431 | self._write_header() |
---|
432 | |
---|
433 | def _write_header(self): |
---|
434 | if self._comptype == 'NONE': |
---|
435 | if self._sampwidth == 1: |
---|
436 | encoding = AUDIO_FILE_ENCODING_LINEAR_8 |
---|
437 | self._framesize = 1 |
---|
438 | elif self._sampwidth == 2: |
---|
439 | encoding = AUDIO_FILE_ENCODING_LINEAR_16 |
---|
440 | self._framesize = 2 |
---|
441 | elif self._sampwidth == 4: |
---|
442 | encoding = AUDIO_FILE_ENCODING_LINEAR_32 |
---|
443 | self._framesize = 4 |
---|
444 | else: |
---|
445 | raise Error, 'internal error' |
---|
446 | elif self._comptype == 'ULAW': |
---|
447 | encoding = AUDIO_FILE_ENCODING_MULAW_8 |
---|
448 | self._framesize = 1 |
---|
449 | else: |
---|
450 | raise Error, 'internal error' |
---|
451 | self._framesize = self._framesize * self._nchannels |
---|
452 | _write_u32(self._file, AUDIO_FILE_MAGIC) |
---|
453 | header_size = 25 + len(self._info) |
---|
454 | header_size = (header_size + 7) & ~7 |
---|
455 | _write_u32(self._file, header_size) |
---|
456 | if self._nframes == AUDIO_UNKNOWN_SIZE: |
---|
457 | length = AUDIO_UNKNOWN_SIZE |
---|
458 | else: |
---|
459 | length = self._nframes * self._framesize |
---|
460 | try: |
---|
461 | self._form_length_pos = self._file.tell() |
---|
462 | except (AttributeError, IOError): |
---|
463 | self._form_length_pos = None |
---|
464 | _write_u32(self._file, length) |
---|
465 | self._datalength = length |
---|
466 | _write_u32(self._file, encoding) |
---|
467 | _write_u32(self._file, self._framerate) |
---|
468 | _write_u32(self._file, self._nchannels) |
---|
469 | self._file.write(self._info) |
---|
470 | self._file.write('\0'*(header_size - len(self._info) - 24)) |
---|
471 | |
---|
472 | def _patchheader(self): |
---|
473 | if self._form_length_pos is None: |
---|
474 | raise IOError('cannot seek') |
---|
475 | self._file.seek(self._form_length_pos) |
---|
476 | _write_u32(self._file, self._datawritten) |
---|
477 | self._datalength = self._datawritten |
---|
478 | self._file.seek(0, 2) |
---|
479 | |
---|
480 | def open(f, mode=None): |
---|
481 | if mode is None: |
---|
482 | if hasattr(f, 'mode'): |
---|
483 | mode = f.mode |
---|
484 | else: |
---|
485 | mode = 'rb' |
---|
486 | if mode in ('r', 'rb'): |
---|
487 | return Au_read(f) |
---|
488 | elif mode in ('w', 'wb'): |
---|
489 | return Au_write(f) |
---|
490 | else: |
---|
491 | raise Error, "mode must be 'r', 'rb', 'w', or 'wb'" |
---|
492 | |
---|
493 | openfp = open |
---|