Lib/wave.py

   1 """Stuff to parse WAVE files.
   2
   3 Usage.
   4
   5 Reading WAVE files:
   6       f = wave.open(file, 'r')
   7 where file is either the name of a file or an open file pointer.
   8 The open file pointer must have methods read(), seek(), and close().
   9 When the setpos() and rewind() methods are not used, the seek()
  10 method is not  necessary.
  11
  12 This returns an instance of a class with the following public methods:
  13       getnchannels()  -- returns number of audio channels (1 for
  14                          mono, 2 for stereo)
  15       getsampwidth()  -- returns sample width in bytes
  16       getframerate()  -- returns sampling frequency
  17       getnframes()    -- returns number of audio frames
  18       getcomptype()   -- returns compression type ('NONE' for linear samples)
  19       getcompname()   -- returns human-readable version of
  20                          compression type ('not compressed' linear samples)
  21       getparams()     -- returns a tuple consisting of all of the
  22                          above in the above order
  23       getmarkers()    -- returns None (for compatibility with the
  24                          aifc module)
  25       getmark(id)     -- raises an error since the mark does not
  26                          exist (for compatibility with the aifc module)
  27       readframes(n)   -- returns at most n frames of audio
  28       rewind()        -- rewind to the beginning of the audio stream
  29       setpos(pos)     -- seek to the specified position
  30       tell()          -- return the current position
  31       close()         -- close the instance (make it unusable)
  32 The position returned by tell() and the position given to setpos()
  33 are compatible and have nothing to do with the actual position in the
  34 file.
  35 The close() method is called automatically when the class instance
  36 is destroyed.
  37
  38 Writing WAVE files:
  39       f = wave.open(file, 'w')
  40 where file is either the name of a file or an open file pointer.
  41 The open file pointer must have methods write(), tell(), seek(), and
  42 close().
  43
  44 This returns an instance of a class with the following public methods:
  45       setnchannels(n) -- set the number of channels
  46       setsampwidth(n) -- set the sample width
  47       setframerate(n) -- set the frame rate
  48       setnframes(n)   -- set the number of frames
  49       setcomptype(type, name)
  50                       -- set the compression type and the
  51                          human-readable compression type
  52       setparams(tuple)
  53                       -- set all parameters at once
  54       tell()          -- return current position in output file
  55       writeframesraw(data)
  56                       -- write audio frames without pathing up the
  57                          file header
  58       writeframes(data)
  59                       -- write audio frames and patch up the file header
  60       close()         -- patch up the file header and close the
  61                          output file
  62 You should set the parameters before the first writeframesraw or
  63 writeframes.  The total number of frames does not need to be set,
  64 but when it is set to the correct value, the header does not have to
  65 be patched up.
  66 It is best to first set all parameters, perhaps possibly the
  67 compression type, and then write audio frames using writeframesraw.
  68 When all frames have been written, either call writeframes('') or
  69 close() to patch up the sizes in the header.
  70 The close() method is called automatically when the class instance
  71 is destroyed.
  72 """
  73
  74 import __builtin__
  75
  76 class Error(Exception):
  77     pass
  78
  79 WAVE_FORMAT_PCM = 0x0001
  80
  81 _array_fmts = None, 'b', 'h', None, 'l'
  82
  83 # Determine endian-ness
  84 import struct
  85 if struct.pack("h", 1) == "\000\001":
  86     big_endian = 1
  87 else:
  88     big_endian = 0
  89
  90 from chunk import Chunk
  91
  92 class Wave_read:
  93     """Variables used in this class:
  94
  95     These variables are available to the user though appropriate
  96     methods of this class:
  97     _file -- the open file with methods read(), close(), and seek()
  98               set through the __init__() method
  99     _nchannels -- the number of audio channels
 100               available through the getnchannels() method
 101     _nframes -- the number of audio frames
 102               available through the getnframes() method
 103     _sampwidth -- the number of bytes per audio sample
 104               available through the getsampwidth() method
 105     _framerate -- the sampling frequency
 106               available through the getframerate() method
 107     _comptype -- the AIFF-C compression type ('NONE' if AIFF)
 108               available through the getcomptype() method
 109     _compname -- the human-readable AIFF-C compression type
 110               available through the getcomptype() method
 111     _soundpos -- the position in the audio stream
 112               available through the tell() method, set through the
 113               setpos() method
 114
 115     These variables are used internally only:
 116     _fmt_chunk_read -- 1 iff the FMT chunk has been read
 117     _data_seek_needed -- 1 iff positioned correctly in audio
 118               file for readframes()
 119     _data_chunk -- instantiation of a chunk class for the DATA chunk
 120     _framesize -- size of one frame in the file
 121     """
 122
 123     def initfp(self, file):
 124         self._convert = None
 125         self._soundpos = 0
 126         self._file = Chunk(file, bigendian = 0)
 127         if self._file.getname() != 'RIFF':
 128             raise Error, 'file does not start with RIFF id'
 129         if self._file.read(4) != 'WAVE':
 130             raise Error, 'not a WAVE file'
 131         self._fmt_chunk_read = 0
 132         self._data_chunk = None
 133         while 1:
 134             self._data_seek_needed = 1
 135             try:
 136                 chunk = Chunk(self._file, bigendian = 0)
 137             except EOFError:
 138                 break
 139             chunkname = chunk.getname()
 140             if chunkname == 'fmt ':
 141                 self._read_fmt_chunk(chunk)
 142                 self._fmt_chunk_read = 1
 143             elif chunkname == 'data':
 144                 if not self._fmt_chunk_read:
 145                     raise Error, 'data chunk before fmt chunk'
 146                 self._data_chunk = chunk
 147                 self._nframes = chunk.chunksize / self._framesize
 148                 self._data_seek_needed = 0
 149                 break
 150             chunk.skip()
 151         if not self._fmt_chunk_read or not self._data_chunk:
 152             raise Error, 'fmt chunk and/or data chunk missing'
 153
 154     def __init__(self, f):
 155         self._i_opened_the_file = None
 156         if type(f) == type(''):
 157             f = __builtin__.open(f, 'rb')
 158             self._i_opened_the_file = f
 159         # else, assume it is an open file object already
 160         self.initfp(f)
 161
 162     def __del__(self):
 163         self.close()
 164     #
 165     # User visible methods.
 166     #
 167     def getfp(self):
 168         return self._file
 169
 170     def rewind(self):
 171         self._data_seek_needed = 1
 172         self._soundpos = 0
 173
 174     def close(self):
 175         if self._i_opened_the_file:
 176             self._i_opened_the_file.close()
 177             self._i_opened_the_file = None
 178         self._file = None
 179
 180     def tell(self):
 181         return self._soundpos
 182
 183     def getnchannels(self):
 184         return self._nchannels
 185
 186     def getnframes(self):
 187         return self._nframes
 188
 189     def getsampwidth(self):
 190         return self._sampwidth
 191
 192     def getframerate(self):
 193         return self._framerate
 194
 195     def getcomptype(self):
 196         return self._comptype
 197
 198     def getcompname(self):
 199         return self._compname
 200
 201     def getparams(self):
 202         return self.getnchannels(), self.getsampwidth(), \
 203                self.getframerate(), self.getnframes(), \
 204                self.getcomptype(), self.getcompname()
 205
 206     def getmarkers(self):
 207         return None
 208
 209     def getmark(self, id):
 210         raise Error, 'no marks'
 211
 212     def setpos(self, pos):
 213         if pos < 0 or pos > self._nframes:
 214             raise Error, 'position not in range'
 215         self._soundpos = pos
 216         self._data_seek_needed = 1
 217
 218     def readframes(self, nframes):
 219         if self._data_seek_needed:
 220             self._data_chunk.seek(0, 0)
 221             pos = self._soundpos * self._framesize
 222             if pos:
 223                 self._data_chunk.seek(pos, 0)
 224             self._data_seek_needed = 0
 225         if nframes == 0:
 226             return ''
 227         if self._sampwidth > 1 and big_endian:
 228             # unfortunately the fromfile() method does not take
 229             # something that only looks like a file object, so
 230             # we have to reach into the innards of the chunk object
 231             import array
 232             chunk = self._data_chunk
 233             data = array.array(_array_fmts[self._sampwidth])
 234             nitems = nframes * self._nchannels
 235             if nitems * self._sampwidth > chunk.chunksize - chunk.size_read:
 236                 nitems = (chunk.chunksize - chunk.size_read) / self._sampwidth
 237             data.fromfile(chunk.file.file, nitems)
 238             # "tell" data chunk how much was read
 239             chunk.size_read = chunk.size_read + nitems * self._sampwidth
 240             # do the same for the outermost chunk
 241             chunk = chunk.file
 242             chunk.size_read = chunk.size_read + nitems * self._sampwidth
 243             data.byteswap()
 244             data = data.tostring()
 245         else:
 246             data = self._data_chunk.read(nframes * self._framesize)
 247         if self._convert and data:
 248             data = self._convert(data)
 249         self._soundpos = self._soundpos + len(data) / (self._nchannels * self._sampwidth)
 250         return data
 251
 252     #
 253     # Internal methods.
 254     #
 255
 256     def _read_fmt_chunk(self, chunk):
 257         wFormatTag, self._nchannels, self._framerate, dwAvgBytesPerSec, wBlockAlign = struct.unpack('<hhllh', chunk.read(14))
 258         if wFormatTag == WAVE_FORMAT_PCM:
 259             sampwidth = struct.unpack('<h', chunk.read(2))[0]
 260             self._sampwidth = (sampwidth + 7) / 8
 261         else:
 262             raise Error, 'unknown format: ' + `wFormatTag`
 263         self._framesize = self._nchannels * self._sampwidth
 264         self._comptype = 'NONE'
 265         self._compname = 'not compressed'
 266
 267 class Wave_write:
 268     """Variables used in this class:
 269
 270     These variables are user settable through appropriate methods
 271     of this class:
 272     _file -- the open file with methods write(), close(), tell(), seek()
 273               set through the __init__() method
 274     _comptype -- the AIFF-C compression type ('NONE' in AIFF)
 275               set through the setcomptype() or setparams() method
 276     _compname -- the human-readable AIFF-C compression type
 277               set through the setcomptype() or setparams() method
 278     _nchannels -- the number of audio channels
 279               set through the setnchannels() or setparams() method
 280     _sampwidth -- the number of bytes per audio sample
 281               set through the setsampwidth() or setparams() method
 282     _framerate -- the sampling frequency
 283               set through the setframerate() or setparams() method
 284     _nframes -- the number of audio frames written to the header
 285               set through the setnframes() or setparams() method
 286
 287     These variables are used internally only:
 288     _datalength -- the size of the audio samples written to the header
 289     _nframeswritten -- the number of frames actually written
 290     _datawritten -- the size of the audio samples actually written
 291     """
 292
 293     def __init__(self, f):
 294         self._i_opened_the_file = None
 295         if type(f) == type(''):
 296             f = __builtin__.open(f, 'wb')
 297             self._i_opened_the_file = f
 298         self.initfp(f)
 299
 300     def initfp(self, file):
 301         self._file = file
 302         self._convert = None
 303         self._nchannels = 0
 304         self._sampwidth = 0
 305         self._framerate = 0
 306         self._nframes = 0
 307         self._nframeswritten = 0
 308         self._datawritten = 0
 309         self._datalength = 0
 310
 311     def __del__(self):
 312         self.close()
 313
 314     #
 315     # User visible methods.
 316     #
 317     def setnchannels(self, nchannels):
 318         if self._datawritten:
 319             raise Error, 'cannot change parameters after starting to write'
 320         if nchannels < 1:
 321             raise Error, 'bad # of channels'
 322         self._nchannels = nchannels
 323
 324     def getnchannels(self):
 325         if not self._nchannels:
 326             raise Error, 'number of channels not set'
 327         return self._nchannels
 328
 329     def setsampwidth(self, sampwidth):
 330         if self._datawritten:
 331             raise Error, 'cannot change parameters after starting to write'
 332         if sampwidth < 1 or sampwidth > 4:
 333             raise Error, 'bad sample width'
 334         self._sampwidth = sampwidth
 335
 336     def getsampwidth(self):
 337         if not self._sampwidth:
 338             raise Error, 'sample width not set'
 339         return self._sampwidth
 340
 341     def setframerate(self, framerate):
 342         if self._datawritten:
 343             raise Error, 'cannot change parameters after starting to write'
 344         if framerate <= 0:
 345             raise Error, 'bad frame rate'
 346         self._framerate = framerate
 347
 348     def getframerate(self):
 349         if not self._framerate:
 350             raise Error, 'frame rate not set'
 351         return self._framerate
 352
 353     def setnframes(self, nframes):
 354         if self._datawritten:
 355             raise Error, 'cannot change parameters after starting to write'
 356         self._nframes = nframes
 357
 358     def getnframes(self):
 359         return self._nframeswritten
 360
 361     def setcomptype(self, comptype, compname):
 362         if self._datawritten:
 363             raise Error, 'cannot change parameters after starting to write'
 364         if comptype not in ('NONE',):
 365             raise Error, 'unsupported compression type'
 366         self._comptype = comptype
 367         self._compname = compname
 368
 369     def getcomptype(self):
 370         return self._comptype
 371
 372     def getcompname(self):
 373         return self._compname
 374
 375     def setparams(self, (nchannels, sampwidth, framerate, nframes, comptype, compname)):
 376         if self._datawritten:
 377             raise Error, 'cannot change parameters after starting to write'
 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         if not self._nchannels or not self._sampwidth or not self._framerate:
 386             raise Error, 'not all parameters set'
 387         return self._nchannels, self._sampwidth, self._framerate, \
 388               self._nframes, self._comptype, self._compname
 389
 390     def setmark(self, id, pos, name):
 391         raise Error, 'setmark() not supported'
 392
 393     def getmark(self, id):
 394         raise Error, 'no marks'
 395
 396     def getmarkers(self):
 397         return None
 398
 399     def tell(self):
 400         return self._nframeswritten
 401
 402     def writeframesraw(self, data):
 403         self._ensure_header_written(len(data))
 404         nframes = len(data) / (self._sampwidth * self._nchannels)
 405         if self._convert:
 406             data = self._convert(data)
 407         if self._sampwidth > 1 and big_endian:
 408             import array
 409             data = array.array(_array_fmts[self._sampwidth], data)
 410             data.byteswap()
 411             data.tofile(self._file)
 412             self._datawritten = self._datawritten + len(data) * self._sampwidth
 413         else:
 414             self._file.write(data)
 415             self._datawritten = self._datawritten + len(data)
 416         self._nframeswritten = self._nframeswritten + nframes
 417
 418     def writeframes(self, data):
 419         self.writeframesraw(data)
 420         if self._datalength != self._datawritten:
 421             self._patchheader()
 422
 423     def close(self):
 424         if self._file:
 425             self._ensure_header_written(0)
 426             if self._datalength != self._datawritten:
 427                 self._patchheader()
 428             self._file.flush()
 429             self._file = None
 430         if self._i_opened_the_file:
 431             self._i_opened_the_file.close()
 432             self._i_opened_the_file = None
 433
 434     #
 435     # Internal methods.
 436     #
 437
 438     def _ensure_header_written(self, datasize):
 439         if not self._datawritten:
 440             if not self._nchannels:
 441                 raise Error, '# channels not specified'
 442             if not self._sampwidth:
 443                 raise Error, 'sample width not specified'
 444             if not self._framerate:
 445                 raise Error, 'sampling rate not specified'
 446             self._write_header(datasize)
 447
 448     def _write_header(self, initlength):
 449         self._file.write('RIFF')
 450         if not self._nframes:
 451             self._nframes = initlength / (self._nchannels * self._sampwidth)
 452         self._datalength = self._nframes * self._nchannels * self._sampwidth
 453         self._form_length_pos = self._file.tell()
 454         self._file.write(struct.pack('<l4s4slhhllhh4s',
 455             36 + self._datalength, 'WAVE', 'fmt ', 16,
 456             WAVE_FORMAT_PCM, self._nchannels, self._framerate,
 457             self._nchannels * self._framerate * self._sampwidth,
 458             self._nchannels * self._sampwidth,
 459             self._sampwidth * 8, 'data'))
 460         self._data_length_pos = self._file.tell()
 461         self._file.write(struct.pack('<l', self._datalength))
 462
 463     def _patchheader(self):
 464         if self._datawritten == self._datalength:
 465             return
 466         curpos = self._file.tell()
 467         self._file.seek(self._form_length_pos, 0)
 468         self._file.write(struct.pack('<l', 36 + self._datawritten))
 469         self._file.seek(self._data_length_pos, 0)
 470         self._file.write(struct.pack('<l', self._datawritten))
 471         self._file.seek(curpos, 0)
 472         self._datalength = self._datawritten
 473
 474 def open(f, mode=None):
 475     if mode is None:
 476         if hasattr(f, 'mode'):
 477             mode = f.mode
 478         else:
 479             mode = 'rb'
 480     if mode in ('r', 'rb'):
 481         return Wave_read(f)
 482     elif mode in ('w', 'wb'):
 483         return Wave_write(f)
 484     else:
 485         raise Error, "mode must be 'r', 'rb', 'w', or 'wb'"
 486
 487 openfp = open # B/W compatibility