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 __all__ = ["open", "openfp", "Error"]
  77
  78 class Error(Exception):
  79     pass
  80
  81 WAVE_FORMAT_PCM = 0x0001
  82
  83 _array_fmts = None, 'b', 'h', None, 'l'
  84
  85 # Determine endian-ness
  86 import struct
  87 if struct.pack("h", 1) == "\000\001":
  88     big_endian = 1
  89 else:
  90     big_endian = 0
  91
  92 from chunk import Chunk
  93
  94 class Wave_read:
  95     """Variables used in this class:
  96
  97     These variables are available to the user though appropriate
  98     methods of this class:
  99     _file -- the open file with methods read(), close(), and seek()
 100               set through the __init__() method
 101     _nchannels -- the number of audio channels
 102               available through the getnchannels() method
 103     _nframes -- the number of audio frames
 104               available through the getnframes() method
 105     _sampwidth -- the number of bytes per audio sample
 106               available through the getsampwidth() method
 107     _framerate -- the sampling frequency
 108               available through the getframerate() method
 109     _comptype -- the AIFF-C compression type ('NONE' if AIFF)
 110               available through the getcomptype() method
 111     _compname -- the human-readable AIFF-C compression type
 112               available through the getcomptype() method
 113     _soundpos -- the position in the audio stream
 114               available through the tell() method, set through the
 115               setpos() method
 116
 117     These variables are used internally only:
 118     _fmt_chunk_read -- 1 iff the FMT chunk has been read
 119     _data_seek_needed -- 1 iff positioned correctly in audio
 120               file for readframes()
 121     _data_chunk -- instantiation of a chunk class for the DATA chunk
 122     _framesize -- size of one frame in the file
 123     """
 124
 125     def initfp(self, file):
 126         self._convert = None
 127         self._soundpos = 0
 128         self._file = Chunk(file, bigendian = 0)
 129         if self._file.getname() != 'RIFF':
 130             raise Error, 'file does not start with RIFF id'
 131         if self._file.read(4) != 'WAVE':
 132             raise Error, 'not a WAVE file'
 133         self._fmt_chunk_read = 0
 134         self._data_chunk = None
 135         while 1:
 136             self._data_seek_needed = 1
 137             try:
 138                 chunk = Chunk(self._file, bigendian = 0)
 139             except EOFError:
 140                 break
 141             chunkname = chunk.getname()
 142             if chunkname == 'fmt ':
 143                 self._read_fmt_chunk(chunk)
 144                 self._fmt_chunk_read = 1
 145             elif chunkname == 'data':
 146                 if not self._fmt_chunk_read:
 147                     raise Error, 'data chunk before fmt chunk'
 148                 self._data_chunk = chunk
 149                 self._nframes = chunk.chunksize // self._framesize
 150                 self._data_seek_needed = 0
 151                 break
 152             chunk.skip()
 153         if not self._fmt_chunk_read or not self._data_chunk:
 154             raise Error, 'fmt chunk and/or data chunk missing'
 155
 156     def __init__(self, f):
 157         self._i_opened_the_file = None
 158         if isinstance(f, basestring):
 159             f = __builtin__.open(f, 'rb')
 160             self._i_opened_the_file = f
 161         # else, assume it is an open file object already
 162         try:
 163             self.initfp(f)
 164         except:
 165             if self._i_opened_the_file:
 166                 f.close()
 167             raise
 168
 169     def __del__(self):
 170         self.close()
 171     #
 172     # User visible methods.
 173     #
 174     def getfp(self):
 175         return self._file
 176
 177     def rewind(self):
 178         self._data_seek_needed = 1
 179         self._soundpos = 0
 180
 181     def close(self):
 182         if self._i_opened_the_file:
 183             self._i_opened_the_file.close()
 184             self._i_opened_the_file = None
 185         self._file = None
 186
 187     def tell(self):
 188         return self._soundpos
 189
 190     def getnchannels(self):
 191         return self._nchannels
 192
 193     def getnframes(self):
 194         return self._nframes
 195
 196     def getsampwidth(self):
 197         return self._sampwidth
 198
 199     def getframerate(self):
 200         return self._framerate
 201
 202     def getcomptype(self):
 203         return self._comptype
 204
 205     def getcompname(self):
 206         return self._compname
 207
 208     def getparams(self):
 209         return self.getnchannels(), self.getsampwidth(), \
 210                self.getframerate(), self.getnframes(), \
 211                self.getcomptype(), self.getcompname()
 212
 213     def getmarkers(self):
 214         return None
 215
 216     def getmark(self, id):
 217         raise Error, 'no marks'
 218
 219     def setpos(self, pos):
 220         if pos < 0 or pos > self._nframes:
 221             raise Error, 'position not in range'
 222         self._soundpos = pos
 223         self._data_seek_needed = 1
 224
 225     def readframes(self, nframes):
 226         if self._data_seek_needed:
 227             self._data_chunk.seek(0, 0)
 228             pos = self._soundpos * self._framesize
 229             if pos:
 230                 self._data_chunk.seek(pos, 0)
 231             self._data_seek_needed = 0
 232         if nframes == 0:
 233             return ''
 234         if self._sampwidth > 1 and big_endian:
 235             # unfortunately the fromfile() method does not take
 236             # something that only looks like a file object, so
 237             # we have to reach into the innards of the chunk object
 238             import array
 239             chunk = self._data_chunk
 240             data = array.array(_array_fmts[self._sampwidth])
 241             nitems = nframes * self._nchannels
 242             if nitems * self._sampwidth > chunk.chunksize - chunk.size_read:
 243                 nitems = (chunk.chunksize - chunk.size_read) / self._sampwidth
 244             data.fromfile(chunk.file.file, nitems)
 245             # "tell" data chunk how much was read
 246             chunk.size_read = chunk.size_read + nitems * self._sampwidth
 247             # do the same for the outermost chunk
 248             chunk = chunk.file
 249             chunk.size_read = chunk.size_read + nitems * self._sampwidth
 250             data.byteswap()
 251             data = data.tostring()
 252         else:
 253             data = self._data_chunk.read(nframes * self._framesize)
 254         if self._convert and data:
 255             data = self._convert(data)
 256         self._soundpos = self._soundpos + len(data) // (self._nchannels * self._sampwidth)
 257         return data
 258
 259     #
 260     # Internal methods.
 261     #
 262
 263     def _read_fmt_chunk(self, chunk):
 264         wFormatTag, self._nchannels, self._framerate, dwAvgBytesPerSec, wBlockAlign = struct.unpack('<hhllh', chunk.read(14))
 265         if wFormatTag == WAVE_FORMAT_PCM:
 266             sampwidth = struct.unpack('<h', chunk.read(2))[0]
 267             self._sampwidth = (sampwidth + 7) // 8
 268         else:
 269             raise Error, 'unknown format: %r' % (wFormatTag,)
 270         self._framesize = self._nchannels * self._sampwidth
 271         self._comptype = 'NONE'
 272         self._compname = 'not compressed'
 273
 274 class Wave_write:
 275     """Variables used in this class:
 276
 277     These variables are user settable through appropriate methods
 278     of this class:
 279     _file -- the open file with methods write(), close(), tell(), seek()
 280               set through the __init__() method
 281     _comptype -- the AIFF-C compression type ('NONE' in AIFF)
 282               set through the setcomptype() or setparams() method
 283     _compname -- the human-readable AIFF-C compression type
 284               set through the setcomptype() or setparams() method
 285     _nchannels -- the number of audio channels
 286               set through the setnchannels() or setparams() method
 287     _sampwidth -- the number of bytes per audio sample
 288               set through the setsampwidth() or setparams() method
 289     _framerate -- the sampling frequency
 290               set through the setframerate() or setparams() method
 291     _nframes -- the number of audio frames written to the header
 292               set through the setnframes() or setparams() method
 293
 294     These variables are used internally only:
 295     _datalength -- the size of the audio samples written to the header
 296     _nframeswritten -- the number of frames actually written
 297     _datawritten -- the size of the audio samples actually written
 298     """
 299
 300     def __init__(self, f):
 301         self._i_opened_the_file = None
 302         if isinstance(f, basestring):
 303             f = __builtin__.open(f, 'wb')
 304             self._i_opened_the_file = f
 305         try:
 306             self.initfp(f)
 307         except:
 308             if self._i_opened_the_file:
 309                 f.close()
 310             raise
 311
 312     def initfp(self, file):
 313         self._file = file
 314         self._convert = None
 315         self._nchannels = 0
 316         self._sampwidth = 0
 317         self._framerate = 0
 318         self._nframes = 0
 319         self._nframeswritten = 0
 320         self._datawritten = 0
 321         self._datalength = 0
 322
 323     def __del__(self):
 324         self.close()
 325
 326     #
 327     # User visible methods.
 328     #
 329     def setnchannels(self, nchannels):
 330         if self._datawritten:
 331             raise Error, 'cannot change parameters after starting to write'
 332         if nchannels < 1:
 333             raise Error, 'bad # of channels'
 334         self._nchannels = nchannels
 335
 336     def getnchannels(self):
 337         if not self._nchannels:
 338             raise Error, 'number of channels not set'
 339         return self._nchannels
 340
 341     def setsampwidth(self, sampwidth):
 342         if self._datawritten:
 343             raise Error, 'cannot change parameters after starting to write'
 344         if sampwidth < 1 or sampwidth > 4:
 345             raise Error, 'bad sample width'
 346         self._sampwidth = sampwidth
 347
 348     def getsampwidth(self):
 349         if not self._sampwidth:
 350             raise Error, 'sample width not set'
 351         return self._sampwidth
 352
 353     def setframerate(self, framerate):
 354         if self._datawritten:
 355             raise Error, 'cannot change parameters after starting to write'
 356         if framerate <= 0:
 357             raise Error, 'bad frame rate'
 358         self._framerate = framerate
 359
 360     def getframerate(self):
 361         if not self._framerate:
 362             raise Error, 'frame rate not set'
 363         return self._framerate
 364
 365     def setnframes(self, nframes):
 366         if self._datawritten:
 367             raise Error, 'cannot change parameters after starting to write'
 368         self._nframes = nframes
 369
 370     def getnframes(self):
 371         return self._nframeswritten
 372
 373     def setcomptype(self, comptype, compname):
 374         if self._datawritten:
 375             raise Error, 'cannot change parameters after starting to write'
 376         if comptype not in ('NONE',):
 377             raise Error, 'unsupported compression type'
 378         self._comptype = comptype
 379         self._compname = compname
 380
 381     def getcomptype(self):
 382         return self._comptype
 383
 384     def getcompname(self):
 385         return self._compname
 386
 387     def setparams(self, (nchannels, sampwidth, framerate, nframes, comptype, compname)):
 388         if self._datawritten:
 389             raise Error, 'cannot change parameters after starting to write'
 390         self.setnchannels(nchannels)
 391         self.setsampwidth(sampwidth)
 392         self.setframerate(framerate)
 393         self.setnframes(nframes)
 394         self.setcomptype(comptype, compname)
 395
 396     def getparams(self):
 397         if not self._nchannels or not self._sampwidth or not self._framerate:
 398             raise Error, 'not all parameters set'
 399         return self._nchannels, self._sampwidth, self._framerate, \
 400               self._nframes, self._comptype, self._compname
 401
 402     def setmark(self, id, pos, name):
 403         raise Error, 'setmark() not supported'
 404
 405     def getmark(self, id):
 406         raise Error, 'no marks'
 407
 408     def getmarkers(self):
 409         return None
 410
 411     def tell(self):
 412         return self._nframeswritten
 413
 414     def writeframesraw(self, data):
 415         self._ensure_header_written(len(data))
 416         nframes = len(data) // (self._sampwidth * self._nchannels)
 417         if self._convert:
 418             data = self._convert(data)
 419         if self._sampwidth > 1 and big_endian:
 420             import array
 421             data = array.array(_array_fmts[self._sampwidth], data)
 422             data.byteswap()
 423             data.tofile(self._file)
 424             self._datawritten = self._datawritten + len(data) * self._sampwidth
 425         else:
 426             self._file.write(data)
 427             self._datawritten = self._datawritten + len(data)
 428         self._nframeswritten = self._nframeswritten + nframes
 429
 430     def writeframes(self, data):
 431         self.writeframesraw(data)
 432         if self._datalength != self._datawritten:
 433             self._patchheader()
 434
 435     def close(self):
 436         if self._file:
 437             self._ensure_header_written(0)
 438             if self._datalength != self._datawritten:
 439                 self._patchheader()
 440             self._file.flush()
 441             self._file = None
 442         if self._i_opened_the_file:
 443             self._i_opened_the_file.close()
 444             self._i_opened_the_file = None
 445
 446     #
 447     # Internal methods.
 448     #
 449
 450     def _ensure_header_written(self, datasize):
 451         if not self._datawritten:
 452             if not self._nchannels:
 453                 raise Error, '# channels not specified'
 454             if not self._sampwidth:
 455                 raise Error, 'sample width not specified'
 456             if not self._framerate:
 457                 raise Error, 'sampling rate not specified'
 458             self._write_header(datasize)
 459
 460     def _write_header(self, initlength):
 461         self._file.write('RIFF')
 462         if not self._nframes:
 463             self._nframes = initlength / (self._nchannels * self._sampwidth)
 464         self._datalength = self._nframes * self._nchannels * self._sampwidth
 465         self._form_length_pos = self._file.tell()
 466         self._file.write(struct.pack('<l4s4slhhllhh4s',
 467             36 + self._datalength, 'WAVE', 'fmt ', 16,
 468             WAVE_FORMAT_PCM, self._nchannels, self._framerate,
 469             self._nchannels * self._framerate * self._sampwidth,
 470             self._nchannels * self._sampwidth,
 471             self._sampwidth * 8, 'data'))
 472         self._data_length_pos = self._file.tell()
 473         self._file.write(struct.pack('<l', self._datalength))
 474
 475     def _patchheader(self):
 476         if self._datawritten == self._datalength:
 477             return
 478         curpos = self._file.tell()
 479         self._file.seek(self._form_length_pos, 0)
 480         self._file.write(struct.pack('<l', 36 + self._datawritten))
 481         self._file.seek(self._data_length_pos, 0)
 482         self._file.write(struct.pack('<l', self._datawritten))
 483         self._file.seek(curpos, 0)
 484         self._datalength = self._datawritten
 485
 486 def open(f, mode=None):
 487     if mode is None:
 488         if hasattr(f, 'mode'):
 489             mode = f.mode
 490         else:
 491             mode = 'rb'
 492     if mode in ('r', 'rb'):
 493         return Wave_read(f)
 494     elif mode in ('w', 'wb'):
 495         return Wave_write(f)
 496     else:
 497         raise Error, "mode must be 'r', 'rb', 'w', or 'wb'"
 498
 499 openfp = open # B/W compatibility