def par_to_wav(params, speaker='adult', simulation_name='', pitch_var=0.0,
len_var=1.0, verbose=False, rank=1, different_folder='',
monotone=False):
"""
Creates a wave file with vocal tract lab out of the given parameters.
"""
if verbose:
print('simulating ' + speaker)
name = speaker + '_' + simulation_name
if different_folder == '':
wav_file = os.path.relpath(os.path.join('VTL_API/output/', name, '_',
str(rank), '.wav'))
else:
wav_file = os.path.relpath(different_folder)
area_file = os.path.relpath(os.path.join('VTL_API/output/', name, '_',
str(rank), '.txt'))
# gestureFile = create_gesture(name, speaker, pitch_var, len_var)
if not speaker in ('adult', 'infant'):
raise ValueError("speaker needs to be either 'adult' or 'infant'.")
if monotone:
gesture_file = os.path.relpath('VTL_API/gestures/input_%s_monotone.ges' % speaker)
else:
gesture_file = os.path.relpath('VTL_API/gestures/input_%s.ges' % speaker)
speaker_file = create_speaker(speaker, params, name, verbose=verbose, rank=rank)
# run through vocal tract lab
LIB_VTL.vtlGesToWav(speaker_file, gesture_file, wav_file, area_file)
# Repair header of wav File
with open(wav_file, 'rb') as file_:
content = file_.read()
shutil.move(wav_file, wav_file + '.bkup')
with open(wav_file, 'wb') as newfile:
newcontent = WAV_HEADER + content[68:]
newfile.write(newcontent)
os.remove(wav_file + '.bkup')
return wav_file
def par_to_wav(params, speaker='adult', simulation_name='', pitch_var=0.0,
len_var=1.0, verbose=False, rank=1, different_folder='',
monotone=False):
"""
Creates a wave file with vocal tract lab out of the given parameters.
"""
if verbose:
print('simulating ' + speaker)
name = speaker + '_' + simulation_name
if different_folder == '':
wav_file = os.path.relpath(os.path.join('VTL_API/output/', name, '_',
str(rank), '.wav'))
else:
wav_file = os.path.relpath(different_folder)
area_file = os.path.relpath(os.path.join('VTL_API/output/', name, '_',
str(rank), '.txt'))
# gestureFile = create_gesture(name, speaker, pitch_var, len_var)
if not speaker in ('adult', 'infant'):
raise ValueError("speaker needs to be either 'adult' or 'infant'.")
if monotone:
gesture_file = os.path.relpath('VTL_API/gestures/input_%s_monotone.ges' % speaker)
else:
gesture_file = os.path.relpath('VTL_API/gestures/input_%s.ges' % speaker)
speaker_file = create_speaker(speaker, params, name, verbose=verbose, rank=rank)
# run through vocal tract lab
LIB_VTL.vtlGesToWav(speaker_file, gesture_file, wav_file, area_file)
# Repair header of wav File
with open(wav_file, 'rb') as file_:
content = file_.read()
shutil.move(wav_file, wav_file + '.bkup')
with open(wav_file, 'wb') as newfile:
newcontent = WAV_HEADER + content[68:]
newfile.write(newcontent)
os.remove(wav_file + '.bkup')
return wav_file
def parToWave(params,
speaker='adult',
simulation_name='',
pitch_var=0.0,
len_var=1.0,
verbose=False,
rank=1,
different_folder='',
monotone=False):
if verbose:
print 'simulating ' + speaker
# announce simulation task
name = speaker + '_' + simulation_name # create simulation name
dllFile = pth.abspath(
'VTL_API/VocalTractLabApi.so') # C library to be used
if different_folder == '':
wavFile = pth.relpath('VTL_API/output/' + name + '_' + str(rank) +
'.wav') # output sound file
else:
wavFile = pth.relpath(different_folder)
areaFile = pth.relpath('VTL_API/output/' + name + '_' + str(rank) +
'.txt') # output area file
#########################################################################################
#
# Generate gesture score and modify speaker
#
#########################################################################################
if speaker == 'adult':
if monotone:
gestureFile = pth.relpath(
'VTL_API/gestures/input_adult_monotone.ges')
else:
gestureFile = pth.relpath('VTL_API/gestures/input_adult.ges')
if speaker == 'infant':
if monotone:
gestureFile = pth.relpath(
'VTL_API/gestures/input_infant_monotone.ges')
else:
gestureFile = pth.relpath('VTL_API/gestures/input_infant.ges')
# create gesture score and define created .ges file
if (speaker != 'adult') and (
speaker != 'infant'
): # only create speaker if neither predefined adult nor infant used
print 'Error: no valid speaker detected!'
print 'Choose either "adult" or "infant" speaker.'
return True
else:
speakerFile = create_speaker(speaker,
params,
name,
verbose=verbose,
rank=rank)
#########################################################################################
#
# Simulate Vocal Tract
#
#########################################################################################
lib = cdll.LoadLibrary(dllFile) # load C library
lib.vtlGesToWav(speakerFile, gestureFile, wavFile, areaFile)
# call function to generate sound file from gesture score
#########################################################################################
#
# Repair header of wav File
#
#########################################################################################
backup = wavFile + '.bkup'
system('cp ' + wavFile + ' ' + backup) # create backup of wav file
f = open(wavFile, 'r') # open wav file for reading
content = f.read() # read content of wav file
f.close() # close wav file
system('rm ' + wavFile)
system('touch ' + wavFile) # create empty file with the same name
newfile = open(wavFile, 'w') # open new file for writing
header = 'RIFF' + chr(0x8C) + chr(0x87) + chr(0x00) + chr(
0x00) + 'WAVEfmt' + chr(0x20) + chr(0x10) + chr(0x00) + chr(
0x00) + chr(0x00) + chr(0x01) + chr(0x00) + chr(0x01) + chr(
0x00) + '"V' + chr(0x00) + chr(0x00) + 'D' + chr(0xAC) + chr(
0x00) + chr(0x00) + chr(0x02) + chr(0x00) + chr(
0x10) + chr(0x00) + 'data'
# define working header
newcontent = header + content[68:] # concatenate header with sound data
newfile.write(newcontent) # write new file
newfile.close() # close new file
system('rm ' + backup)
return wavFile
def parToWave(params, speaker='adult', simulation_name='', pitch_var=0.0, len_var=1.0, verbose=False, rank=1, different_folder='', monotone=False):
if verbose:
print 'simulating '+speaker
# announce simulation task
name = speaker+'_'+simulation_name # create simulation name
dllFile = pth.abspath('VTL_API/VocalTractLabApi.so') # C library to be used
if different_folder == '':
wavFile = pth.relpath('VTL_API/output/'+name+'_'+str(rank)+'.wav') # output sound file
else:
wavFile = pth.relpath(different_folder)
areaFile = pth.relpath('VTL_API/output/'+name+'_'+str(rank)+'.txt') # output area file
#########################################################################################
#
# Generate gesture score and modify speaker
#
#########################################################################################
if speaker == 'adult':
if monotone:
gestureFile = pth.relpath('VTL_API/gestures/input_adult_monotone.ges')
else:
gestureFile = pth.relpath('VTL_API/gestures/input_adult.ges')
if speaker == 'infant':
if monotone:
gestureFile = pth.relpath('VTL_API/gestures/input_infant_monotone.ges')
else:
gestureFile = pth.relpath('VTL_API/gestures/input_infant.ges')
# create gesture score and define created .ges file
if (speaker != 'adult') and (speaker != 'infant'): # only create speaker if neither predefined adult nor infant used
print 'Error: no valid speaker detected!'
print 'Choose either "adult" or "infant" speaker.'
return True
else:
speakerFile = create_speaker(speaker, params, name, verbose=verbose, rank=rank)
#########################################################################################
#
# Simulate Vocal Tract
#
#########################################################################################
lib = cdll.LoadLibrary(dllFile) # load C library
lib.vtlGesToWav(speakerFile, gestureFile, wavFile, areaFile)
# call function to generate sound file from gesture score
#########################################################################################
#
# Repair header of wav File
#
#########################################################################################
backup = wavFile+'.bkup'
system('cp '+wavFile+' '+backup) # create backup of wav file
f = open(wavFile, 'r') # open wav file for reading
content = f.read() # read content of wav file
f.close() # close wav file
system('rm '+wavFile)
system('touch '+wavFile) # create empty file with the same name
newfile = open(wavFile, 'w') # open new file for writing
header = 'RIFF'+chr(0x8C)+chr(0x87)+chr(0x00)+chr(0x00)+'WAVEfmt'+chr(0x20)+chr(0x10)+chr(0x00)+chr(0x00)+chr(0x00)+chr(0x01)+chr(0x00)+chr(0x01)+chr(0x00)+'"V'+chr(0x00)+chr(0x00)+'D'+chr(0xAC)+chr(0x00)+chr(0x00)+chr(0x02)+chr(0x00)+chr(0x10)+chr(0x00)+'data'
# define working header
newcontent = header+content[68:] # concatenate header with sound data
newfile.write(newcontent) # write new file
newfile.close() # close new file
system('rm '+backup)
return wavFile
