"""Routines to help recognizing sound files.

[0] Fix | Delete

[1] Fix | Delete

Function whathdr() recognizes various types of sound file headers.

[2] Fix | Delete

It understands almost all headers that SOX can decode.

[3] Fix | Delete

[4] Fix | Delete

The return tuple contains the following items, in this order:

[5] Fix | Delete

- file type (as SOX understands it)

[6] Fix | Delete

- sampling rate (0 if unknown or hard to decode)

[7] Fix | Delete

- number of channels (0 if unknown or hard to decode)

[8] Fix | Delete

- number of frames in the file (-1 if unknown or hard to decode)

[9] Fix | Delete

- number of bits/sample, or 'U' for U-LAW, or 'A' for A-LAW

[10] Fix | Delete

[11] Fix | Delete

If the file doesn't have a recognizable type, it returns None.

[12] Fix | Delete

If the file can't be opened, OSError is raised.

[13] Fix | Delete

[14] Fix | Delete

To compute the total time, divide the number of frames by the

[15] Fix | Delete

sampling rate (a frame contains a sample for each channel).

[16] Fix | Delete

[17] Fix | Delete

Function what() calls whathdr(). (It used to also use some

[18] Fix | Delete

heuristics for raw data, but this doesn't work very well.)

[19] Fix | Delete

[20] Fix | Delete

Finally, the function test() is a simple main program that calls

[21] Fix | Delete

what() for all files mentioned on the argument list. For directory

[22] Fix | Delete

arguments it calls what() for all files in that directory. Default

[23] Fix | Delete

argument is "." (testing all files in the current directory). The

[24] Fix | Delete

option -r tells it to recurse down directories found inside

[25] Fix | Delete

explicitly given directories.

[26] Fix | Delete

"""

[27] Fix | Delete

[28] Fix | Delete

# The file structure is top-down except that the test program and its

[29] Fix | Delete

# subroutine come last.

[30] Fix | Delete

[31] Fix | Delete

__all__ = ['what', 'whathdr']

[32] Fix | Delete

[33] Fix | Delete

from collections import namedtuple

[34] Fix | Delete

[35] Fix | Delete

SndHeaders = namedtuple('SndHeaders',

[36] Fix | Delete

'filetype framerate nchannels nframes sampwidth')

[37] Fix | Delete

[38] Fix | Delete

SndHeaders.filetype.__doc__ = ("""The value for type indicates the data type

[39] Fix | Delete

and will be one of the strings 'aifc', 'aiff', 'au','hcom',

[40] Fix | Delete

'sndr', 'sndt', 'voc', 'wav', '8svx', 'sb', 'ub', or 'ul'.""")

[41] Fix | Delete

SndHeaders.framerate.__doc__ = ("""The sampling_rate will be either the actual

[42] Fix | Delete

value or 0 if unknown or difficult to decode.""")

[43] Fix | Delete

SndHeaders.nchannels.__doc__ = ("""The number of channels or 0 if it cannot be

[44] Fix | Delete

determined or if the value is difficult to decode.""")

[45] Fix | Delete

SndHeaders.nframes.__doc__ = ("""The value for frames will be either the number

[46] Fix | Delete

of frames or -1.""")

[47] Fix | Delete

SndHeaders.sampwidth.__doc__ = ("""Either the sample size in bits or

[48] Fix | Delete

'A' for A-LAW or 'U' for u-LAW.""")

[49] Fix | Delete

[50] Fix | Delete

def what(filename):

[51] Fix | Delete

"""Guess the type of a sound file."""

[52] Fix | Delete

res = whathdr(filename)

[53] Fix | Delete

return res

[54] Fix | Delete

[55] Fix | Delete

[56] Fix | Delete

def whathdr(filename):

[57] Fix | Delete

"""Recognize sound headers."""

[58] Fix | Delete

with open(filename, 'rb') as f:

[59] Fix | Delete

h = f.read(512)

[60] Fix | Delete

for tf in tests:

[61] Fix | Delete

res = tf(h, f)

[62] Fix | Delete

if res:

[63] Fix | Delete

return SndHeaders(*res)

[64] Fix | Delete

return None

[65] Fix | Delete

[66] Fix | Delete

[67] Fix | Delete

#-----------------------------------#

[68] Fix | Delete

# Subroutines per sound header type #

[69] Fix | Delete

#-----------------------------------#

[70] Fix | Delete

[71] Fix | Delete

tests = []

[72] Fix | Delete

[73] Fix | Delete

def test_aifc(h, f):

[74] Fix | Delete

import aifc

[75] Fix | Delete

if not h.startswith(b'FORM'):

[76] Fix | Delete

return None

[77] Fix | Delete

if h[8:12] == b'AIFC':

[78] Fix | Delete

fmt = 'aifc'

[79] Fix | Delete

elif h[8:12] == b'AIFF':

[80] Fix | Delete

fmt = 'aiff'

[81] Fix | Delete

else:

[82] Fix | Delete

return None

[83] Fix | Delete

f.seek(0)

[84] Fix | Delete

try:

[85] Fix | Delete

a = aifc.open(f, 'r')

[86] Fix | Delete

except (EOFError, aifc.Error):

[87] Fix | Delete

return None

[88] Fix | Delete

return (fmt, a.getframerate(), a.getnchannels(),

[89] Fix | Delete

a.getnframes(), 8 * a.getsampwidth())

[90] Fix | Delete

[91] Fix | Delete

tests.append(test_aifc)

[92] Fix | Delete

[93] Fix | Delete

[94] Fix | Delete

def test_au(h, f):

[95] Fix | Delete

if h.startswith(b'.snd'):

[96] Fix | Delete

func = get_long_be

[97] Fix | Delete

elif h[:4] in (b'\0ds.', b'dns.'):

[98] Fix | Delete

func = get_long_le

[99] Fix | Delete

else:

[100] Fix | Delete

return None

[101] Fix | Delete

filetype = 'au'

[102] Fix | Delete

hdr_size = func(h[4:8])

[103] Fix | Delete

data_size = func(h[8:12])

[104] Fix | Delete

encoding = func(h[12:16])

[105] Fix | Delete

rate = func(h[16:20])

[106] Fix | Delete

nchannels = func(h[20:24])

[107] Fix | Delete

sample_size = 1 # default

[108] Fix | Delete

if encoding == 1:

[109] Fix | Delete

sample_bits = 'U'

[110] Fix | Delete

elif encoding == 2:

[111] Fix | Delete

sample_bits = 8

[112] Fix | Delete

elif encoding == 3:

[113] Fix | Delete

sample_bits = 16

[114] Fix | Delete

sample_size = 2

[115] Fix | Delete

else:

[116] Fix | Delete

sample_bits = '?'

[117] Fix | Delete

frame_size = sample_size * nchannels

[118] Fix | Delete

if frame_size:

[119] Fix | Delete

nframe = data_size / frame_size

[120] Fix | Delete

else:

[121] Fix | Delete

nframe = -1

[122] Fix | Delete

return filetype, rate, nchannels, nframe, sample_bits

[123] Fix | Delete

[124] Fix | Delete

tests.append(test_au)

[125] Fix | Delete

[126] Fix | Delete

[127] Fix | Delete

def test_hcom(h, f):

[128] Fix | Delete

if h[65:69] != b'FSSD' or h[128:132] != b'HCOM':

[129] Fix | Delete

return None

[130] Fix | Delete

divisor = get_long_be(h[144:148])

[131] Fix | Delete

if divisor:

[132] Fix | Delete

rate = 22050 / divisor

[133] Fix | Delete

else:

[134] Fix | Delete

rate = 0

[135] Fix | Delete

return 'hcom', rate, 1, -1, 8

[136] Fix | Delete

[137] Fix | Delete

tests.append(test_hcom)

[138] Fix | Delete

[139] Fix | Delete

[140] Fix | Delete

def test_voc(h, f):

[141] Fix | Delete

if not h.startswith(b'Creative Voice File\032'):

[142] Fix | Delete

return None

[143] Fix | Delete

sbseek = get_short_le(h[20:22])

[144] Fix | Delete

rate = 0

[145] Fix | Delete

if 0 <= sbseek < 500 and h[sbseek] == 1:

[146] Fix | Delete

ratecode = 256 - h[sbseek+4]

[147] Fix | Delete

if ratecode:

[148] Fix | Delete

rate = int(1000000.0 / ratecode)

[149] Fix | Delete

return 'voc', rate, 1, -1, 8

[150] Fix | Delete

[151] Fix | Delete

tests.append(test_voc)

[152] Fix | Delete

[153] Fix | Delete

[154] Fix | Delete

def test_wav(h, f):

[155] Fix | Delete

import wave

[156] Fix | Delete

# 'RIFF' <len> 'WAVE' 'fmt ' <len>

[157] Fix | Delete

if not h.startswith(b'RIFF') or h[8:12] != b'WAVE' or h[12:16] != b'fmt ':

[158] Fix | Delete

return None

[159] Fix | Delete

f.seek(0)

[160] Fix | Delete

try:

[161] Fix | Delete

w = wave.open(f, 'r')

[162] Fix | Delete

except (EOFError, wave.Error):

[163] Fix | Delete

return None

[164] Fix | Delete

return ('wav', w.getframerate(), w.getnchannels(),

[165] Fix | Delete

w.getnframes(), 8*w.getsampwidth())

[166] Fix | Delete

[167] Fix | Delete

tests.append(test_wav)

[168] Fix | Delete

[169] Fix | Delete

[170] Fix | Delete

def test_8svx(h, f):

[171] Fix | Delete

if not h.startswith(b'FORM') or h[8:12] != b'8SVX':

[172] Fix | Delete

return None

[173] Fix | Delete

# Should decode it to get #channels -- assume always 1

[174] Fix | Delete

return '8svx', 0, 1, 0, 8

[175] Fix | Delete

[176] Fix | Delete

tests.append(test_8svx)

[177] Fix | Delete

[178] Fix | Delete

[179] Fix | Delete

def test_sndt(h, f):

[180] Fix | Delete

if h.startswith(b'SOUND'):

[181] Fix | Delete

nsamples = get_long_le(h[8:12])

[182] Fix | Delete

rate = get_short_le(h[20:22])

[183] Fix | Delete

return 'sndt', rate, 1, nsamples, 8

[184] Fix | Delete

[185] Fix | Delete

tests.append(test_sndt)

[186] Fix | Delete

[187] Fix | Delete

[188] Fix | Delete

def test_sndr(h, f):

[189] Fix | Delete

if h.startswith(b'\0\0'):

[190] Fix | Delete

rate = get_short_le(h[2:4])

[191] Fix | Delete

if 4000 <= rate <= 25000:

[192] Fix | Delete

return 'sndr', rate, 1, -1, 8

[193] Fix | Delete

[194] Fix | Delete

tests.append(test_sndr)

[195] Fix | Delete

[196] Fix | Delete

[197] Fix | Delete

#-------------------------------------------#

[198] Fix | Delete

# Subroutines to extract numbers from bytes #

[199] Fix | Delete

#-------------------------------------------#

[200] Fix | Delete

[201] Fix | Delete

def get_long_be(b):

[202] Fix | Delete

return (b[0] << 24) | (b[1] << 16) | (b[2] << 8) | b[3]

[203] Fix | Delete

[204] Fix | Delete

def get_long_le(b):

[205] Fix | Delete

return (b[3] << 24) | (b[2] << 16) | (b[1] << 8) | b[0]

[206] Fix | Delete

[207] Fix | Delete

def get_short_be(b):

[208] Fix | Delete

return (b[0] << 8) | b[1]

[209] Fix | Delete

[210] Fix | Delete

def get_short_le(b):

[211] Fix | Delete

return (b[1] << 8) | b[0]

[212] Fix | Delete

[213] Fix | Delete

[214] Fix | Delete

#--------------------#

[215] Fix | Delete

# Small test program #

[216] Fix | Delete

#--------------------#

[217] Fix | Delete

[218] Fix | Delete

def test():

[219] Fix | Delete

import sys

[220] Fix | Delete

recursive = 0

[221] Fix | Delete

if sys.argv[1:] and sys.argv[1] == '-r':

[222] Fix | Delete

del sys.argv[1:2]

[223] Fix | Delete

recursive = 1

[224] Fix | Delete

try:

[225] Fix | Delete

if sys.argv[1:]:

[226] Fix | Delete

testall(sys.argv[1:], recursive, 1)

[227] Fix | Delete

else:

[228] Fix | Delete

testall(['.'], recursive, 1)

[229] Fix | Delete

except KeyboardInterrupt:

[230] Fix | Delete

sys.stderr.write('\n[Interrupted]\n')

[231] Fix | Delete

sys.exit(1)

[232] Fix | Delete

[233] Fix | Delete

def testall(list, recursive, toplevel):

[234] Fix | Delete

import sys

[235] Fix | Delete

import os

[236] Fix | Delete

for filename in list:

[237] Fix | Delete

if os.path.isdir(filename):

[238] Fix | Delete

print(filename + '/:', end=' ')

[239] Fix | Delete

if recursive or toplevel:

[240] Fix | Delete

print('recursing down:')

[241] Fix | Delete

import glob

[242] Fix | Delete

names = glob.glob(os.path.join(glob.escape(filename), '*'))

[243] Fix | Delete

testall(names, recursive, 0)

[244] Fix | Delete

else:

[245] Fix | Delete

print('*** directory (use -r) ***')

[246] Fix | Delete

else:

[247] Fix | Delete

print(filename + ':', end=' ')

[248] Fix | Delete

sys.stdout.flush()

[249] Fix | Delete

try:

[250] Fix | Delete

print(what(filename))

[251] Fix | Delete

except OSError:

[252] Fix | Delete

print('*** not found ***')

[253] Fix | Delete

[254] Fix | Delete

if __name__ == '__main__':

[255] Fix | Delete

test()

[256] Fix | Delete

[257] Fix | Delete