"""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, IOError 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

def what(filename):

[34] Fix | Delete

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

[35] Fix | Delete

res = whathdr(filename)

[36] Fix | Delete

return res

[37] Fix | Delete

[38] Fix | Delete

[39] Fix | Delete

def whathdr(filename):

[40] Fix | Delete

"""Recognize sound headers"""

[41] Fix | Delete

f = open(filename, 'rb')

[42] Fix | Delete

h = f.read(512)

[43] Fix | Delete

for tf in tests:

[44] Fix | Delete

res = tf(h, f)

[45] Fix | Delete

if res:

[46] Fix | Delete

return res

[47] Fix | Delete

return None

[48] Fix | Delete

[49] Fix | Delete

[50] Fix | Delete

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

[51] Fix | Delete

# Subroutines per sound header type #

[52] Fix | Delete

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

[53] Fix | Delete

[54] Fix | Delete

tests = []

[55] Fix | Delete

[56] Fix | Delete

def test_aifc(h, f):

[57] Fix | Delete

import aifc

[58] Fix | Delete

if h[:4] != 'FORM':

[59] Fix | Delete

return None

[60] Fix | Delete

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

[61] Fix | Delete

fmt = 'aifc'

[62] Fix | Delete

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

[63] Fix | Delete

fmt = 'aiff'

[64] Fix | Delete

else:

[65] Fix | Delete

return None

[66] Fix | Delete

f.seek(0)

[67] Fix | Delete

try:

[68] Fix | Delete

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

[69] Fix | Delete

except (EOFError, aifc.Error):

[70] Fix | Delete

return None

[71] Fix | Delete

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

[72] Fix | Delete

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

[73] Fix | Delete

[74] Fix | Delete

tests.append(test_aifc)

[75] Fix | Delete

[76] Fix | Delete

[77] Fix | Delete

def test_au(h, f):

[78] Fix | Delete

if h[:4] == '.snd':

[79] Fix | Delete

f = get_long_be

[80] Fix | Delete

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

[81] Fix | Delete

f = get_long_le

[82] Fix | Delete

else:

[83] Fix | Delete

return None

[84] Fix | Delete

type = 'au'

[85] Fix | Delete

hdr_size = f(h[4:8])

[86] Fix | Delete

data_size = f(h[8:12])

[87] Fix | Delete

encoding = f(h[12:16])

[88] Fix | Delete

rate = f(h[16:20])

[89] Fix | Delete

nchannels = f(h[20:24])

[90] Fix | Delete

sample_size = 1 # default

[91] Fix | Delete

if encoding == 1:

[92] Fix | Delete

sample_bits = 'U'

[93] Fix | Delete

elif encoding == 2:

[94] Fix | Delete

sample_bits = 8

[95] Fix | Delete

elif encoding == 3:

[96] Fix | Delete

sample_bits = 16

[97] Fix | Delete

sample_size = 2

[98] Fix | Delete

else:

[99] Fix | Delete

sample_bits = '?'

[100] Fix | Delete

frame_size = sample_size * nchannels

[101] Fix | Delete

return type, rate, nchannels, data_size//frame_size, sample_bits

[102] Fix | Delete

[103] Fix | Delete

tests.append(test_au)

[104] Fix | Delete

[105] Fix | Delete

[106] Fix | Delete

def test_hcom(h, f):

[107] Fix | Delete

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

[108] Fix | Delete

return None

[109] Fix | Delete

divisor = get_long_be(h[128+16:128+20])

[110] Fix | Delete

return 'hcom', 22050//divisor, 1, -1, 8

[111] Fix | Delete

[112] Fix | Delete

tests.append(test_hcom)

[113] Fix | Delete

[114] Fix | Delete

[115] Fix | Delete

def test_voc(h, f):

[116] Fix | Delete

if h[:20] != 'Creative Voice File\032':

[117] Fix | Delete

return None

[118] Fix | Delete

sbseek = get_short_le(h[20:22])

[119] Fix | Delete

rate = 0

[120] Fix | Delete

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

[121] Fix | Delete

ratecode = ord(h[sbseek+4])

[122] Fix | Delete

rate = int(1000000.0 / (256 - ratecode))

[123] Fix | Delete

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

[124] Fix | Delete

[125] Fix | Delete

tests.append(test_voc)

[126] Fix | Delete

[127] Fix | Delete

[128] Fix | Delete

def test_wav(h, f):

[129] Fix | Delete

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

[130] Fix | Delete

if h[:4] != 'RIFF' or h[8:12] != 'WAVE' or h[12:16] != 'fmt ':

[131] Fix | Delete

return None

[132] Fix | Delete

style = get_short_le(h[20:22])

[133] Fix | Delete

nchannels = get_short_le(h[22:24])

[134] Fix | Delete

rate = get_long_le(h[24:28])

[135] Fix | Delete

sample_bits = get_short_le(h[34:36])

[136] Fix | Delete

return 'wav', rate, nchannels, -1, sample_bits

[137] Fix | Delete

[138] Fix | Delete

tests.append(test_wav)

[139] Fix | Delete

[140] Fix | Delete

[141] Fix | Delete

def test_8svx(h, f):

[142] Fix | Delete

if h[:4] != 'FORM' or h[8:12] != '8SVX':

[143] Fix | Delete

return None

[144] Fix | Delete

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

[145] Fix | Delete

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

[146] Fix | Delete

[147] Fix | Delete

tests.append(test_8svx)

[148] Fix | Delete

[149] Fix | Delete

[150] Fix | Delete

def test_sndt(h, f):

[151] Fix | Delete

if h[:5] == 'SOUND':

[152] Fix | Delete

nsamples = get_long_le(h[8:12])

[153] Fix | Delete

rate = get_short_le(h[20:22])

[154] Fix | Delete

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

[155] Fix | Delete

[156] Fix | Delete

tests.append(test_sndt)

[157] Fix | Delete

[158] Fix | Delete

[159] Fix | Delete

def test_sndr(h, f):

[160] Fix | Delete

if h[:2] == '\0\0':

[161] Fix | Delete

rate = get_short_le(h[2:4])

[162] Fix | Delete

if 4000 <= rate <= 25000:

[163] Fix | Delete

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

[164] Fix | Delete

[165] Fix | Delete

tests.append(test_sndr)

[166] Fix | Delete

[167] Fix | Delete

[168] Fix | Delete

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

[169] Fix | Delete

# Subroutines to extract numbers from strings #

[170] Fix | Delete

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

[171] Fix | Delete

[172] Fix | Delete

def get_long_be(s):

[173] Fix | Delete

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

[174] Fix | Delete

[175] Fix | Delete

def get_long_le(s):

[176] Fix | Delete

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

[177] Fix | Delete

[178] Fix | Delete

def get_short_be(s):

[179] Fix | Delete

return (ord(s[0])<<8) | ord(s[1])

[180] Fix | Delete

[181] Fix | Delete

def get_short_le(s):

[182] Fix | Delete

return (ord(s[1])<<8) | ord(s[0])

[183] Fix | Delete

[184] Fix | Delete

[185] Fix | Delete

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

[186] Fix | Delete

# Small test program #

[187] Fix | Delete

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

[188] Fix | Delete

[189] Fix | Delete

def test():

[190] Fix | Delete

import sys

[191] Fix | Delete

recursive = 0

[192] Fix | Delete

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

[193] Fix | Delete

del sys.argv[1:2]

[194] Fix | Delete

recursive = 1

[195] Fix | Delete

try:

[196] Fix | Delete

if sys.argv[1:]:

[197] Fix | Delete

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

[198] Fix | Delete

else:

[199] Fix | Delete

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

[200] Fix | Delete

except KeyboardInterrupt:

[201] Fix | Delete

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

[202] Fix | Delete

sys.exit(1)

[203] Fix | Delete

[204] Fix | Delete

def testall(list, recursive, toplevel):

[205] Fix | Delete

import sys

[206] Fix | Delete

import os

[207] Fix | Delete

for filename in list:

[208] Fix | Delete

if os.path.isdir(filename):

[209] Fix | Delete

print filename + '/:',

[210] Fix | Delete

if recursive or toplevel:

[211] Fix | Delete

print 'recursing down:'

[212] Fix | Delete

import glob

[213] Fix | Delete

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

[214] Fix | Delete

testall(names, recursive, 0)

[215] Fix | Delete

else:

[216] Fix | Delete

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

[217] Fix | Delete

else:

[218] Fix | Delete

print filename + ':',

[219] Fix | Delete

sys.stdout.flush()

[220] Fix | Delete

try:

[221] Fix | Delete

print what(filename)

[222] Fix | Delete

except IOError:

[223] Fix | Delete

print '*** not found ***'

[224] Fix | Delete

[225] Fix | Delete

if __name__ == '__main__':

[226] Fix | Delete

test()

[227] Fix | Delete

[228] Fix | Delete