class VoiceRssTTS(TTSInterface):
"""Uses The VoiceRss TTS service to synthesize sentences in a
specific language, e.g. en-us.
The main function synthesize returns a string which contain a RIFF
wave file audio of the synthesized text."""
def __init__(self, cfg):
"""Intitialize: just remember the configuration."""
self.cfg = cfg
super(VoiceRssTTS, self).__init__(cfg)
self.preprocessing = TTSPreprocessing(self.cfg, self.cfg['TTS']['VoiceRss']['preprocessing'])
@cache.persistent_cache(True, 'VoiceRssTTS.get_tts_mp3.')
def get_tts_mp3(self, language, text):
"""Access the VoiceRss TTS service and get synthesized audio
for a text.
Returns a string with a WAV stream."""
baseurl = "http://api.voicerss.org"
values = {'src': text.encode('utf8'),
'hl': language,
'c': 'MP3',
'f': '16khz_16bit_mono',
'key': self.cfg['TTS']['VoiceRss']['api_key']}
data = urllib.urlencode(values)
request = urllib2.Request(baseurl, data)
request.add_header("User-Agent", "Mozilla/5.0 (X11; U; Linux i686) Gecko/20071127 Firefox/2.0.0.11")
try:
mp3response = urllib2.urlopen(request)
return mp3response.read()
except (urllib2.HTTPError, urllib2.URLError):
raise TTSException("SpeechTech TTS error.")
def synthesize(self, text):
"""Synthesize the text and return it in a string
with audio in default format and sample rate."""
wav = b""
try:
if text:
text = self.preprocessing.process(text)
mp3 = self.get_tts_mp3(self.cfg['TTS']['VoiceRss']['language'], text)
wav = audio.convert_mp3_to_wav(self.cfg, mp3)
wav = audio.change_tempo(self.cfg, self.cfg['TTS']['VoiceRss']['tempo'], wav)
return wav
else:
return b""
except TTSException as e:
m = unicode(e) + " Text: %s" % text
self.cfg['Logging']['system_logger'].exception(m)
return b""
return wav
class FliteTTS(TTSInterface):
"""Uses Flite TTS to synthesize sentences in English.
The main function `synthesize' returns a string which contains a RIFF wave
file audio of the synthesized text.
"""
def __init__(self, cfg):
super(FliteTTS, self).__init__(cfg)
self.preprocessing = TTSPreprocessing(
self.cfg, self.cfg['TTS']['Flite']['preprocessing'])
@cache.persistent_cache(True, 'FliteTTS.get_tts_wav.')
def get_tts_wav(self, voice, text):
"""Runs flite from the command line and gets the synthesized audio.
Note that the returned audio is in the re-sampled PCM audio format.
"""
handle, wav_file_name = mkstemp('TmpSpeechFile.wav')
if voice not in ['awb', 'rms', 'slt', 'kal', 'awb_time', 'kal16']:
voice = 'awb'
try:
subprocess.call("flite -voice %s -t \"%s\" -o %s 2> /dev/null" %
(voice, text, wav_file_name),
shell=True)
wav = audio.load_wav(self.cfg, wav_file_name)
except:
raise TTSException("No data synthesized.")
return wav
def synthesize(self, text):
"""\
Synthesizes the text and returns it as a string with audio in default
format and sample rate.
"""
try:
text = self.preprocessing.process(text)
wav = self.get_tts_wav(self.cfg['TTS']['Flite']['voice'], text)
wav = audio.change_tempo(self.cfg,
self.cfg['TTS']['Flite']['tempo'], wav)
except TTSException:
m = e + "Text: %" % text
self.cfg['Logging']['system_logger'].exception(m)
return b""
return wav
class FliteTTS(TTSInterface):
"""Uses Flite TTS to synthesize sentences in English.
The main function `synthesize' returns a string which contains a RIFF wave
file audio of the synthesized text.
"""
def __init__(self, cfg):
super(FliteTTS, self).__init__(cfg)
self.preprocessing = TTSPreprocessing(self.cfg, self.cfg['TTS']['Flite']['preprocessing'])
@cache.persistent_cache(True, 'FliteTTS.get_tts_wav.')
def get_tts_wav(self, voice, text):
"""Runs flite from the command line and gets the synthesized audio.
Note that the returned audio is in the re-sampled PCM audio format.
"""
handle, wav_file_name = mkstemp('TmpSpeechFile.wav')
if voice not in ['awb', 'rms', 'slt', 'kal', 'awb_time', 'kal16']:
voice = 'awb'
try:
subprocess.call("flite -voice %s -t \"%s\" -o %s 2> /dev/null" %
(voice, text, wav_file_name), shell=True)
wav = audio.load_wav(self.cfg, wav_file_name)
except:
raise TTSException("No data synthesized.")
return wav
def synthesize(self, text):
"""\
Synthesizes the text and returns it as a string with audio in default
format and sample rate.
"""
try:
text = self.preprocessing.process(text)
wav = self.get_tts_wav(self.cfg['TTS']['Flite']['voice'], text)
wav = audio.change_tempo(self.cfg, self.cfg['TTS']['Flite']['tempo'], wav)
except TTSException:
m = e + "Text: %" % text
self.cfg['Logging']['system_logger'].exception(m)
return b""
return wav
def __init__(self, cfg):
super(SpeechtechTTS, self).__init__(cfg)
self.preprocessing = TTSPreprocessing(self.cfg, self.cfg['TTS']['SpeechTech']['preprocessing'])
class SpeechtechTTS(TTSInterface):
""" Uses SpeechTech TTS service to synthesize sentences in a specific language, e.g. en, cs.
The main function synthesize returns a string which contain a RIFF wave file audio of the synthesized text.
"""
def __init__(self, cfg):
super(SpeechtechTTS, self).__init__(cfg)
self.preprocessing = TTSPreprocessing(self.cfg, self.cfg['TTS']['SpeechTech']['preprocessing'])
@cache.lru_cache(10000)
@cache.persistent_cache(True, 'SpeechtechTTS.get_tts_mp3.')
def get_tts_mp3(self, voice, text):
""" Access SpeechTech TTS service and get synthesized audio.
Returns a string with wav in it.
"""
try:
TIMEOUT = 30
LOGIN = self.cfg['TTS']['SpeechTech']['login']
PASSWD = self.cfg['TTS']['SpeechTech']['password']
REALM = 'TTS-Server'
ROOT_URI = 'http://tts2.speechtech.cz'
URI = '%s/add_to_queue' % ROOT_URI
if hasattr(socket, 'setdefaulttimeout'):
socket.setdefaulttimeout(TIMEOUT)
# Create an OpenerDirector with support for Basic HTTP Authentication...
auth_handler = urllib2.HTTPDigestAuthHandler()
auth_handler.add_password(realm=REALM,
uri=URI,
user=LOGIN,
passwd=PASSWD)
opener = urllib2.build_opener(auth_handler)
urllib2.install_opener(opener)
params = urllib.urlencode([('text', text.encode('utf8')), ('engine', voice.encode('utf8'))])
task_id = urllib2.urlopen('%s/add_to_queue' % ROOT_URI, params).read().strip()
uri = None
i = 20
while i:
params = urllib.urlencode([('task_id', task_id)])
resp = urllib2.urlopen('%s/query_status' % ROOT_URI, params).read().splitlines()
code = int(resp[0])
#print 'Status is', code
if code == 3:
uri = resp[1]
#print 'Result available at', uri
break
time.sleep(0.2)
i -= 1
if uri:
request = urllib2.Request(uri)
request.add_header("User-Agent", "Mozilla/5.0 (X11; U; Linux i686) Gecko/20071127 Firefox/2.0.0.11")
mp3response = urllib2.urlopen(request)
return mp3response.read()
except (urllib2.HTTPError, urllib2.URLError):
raise TTSException("SpeechTech TTS error.")
raise TTSException("Time out: No data synthesized.")
def synthesize(self, text):
""" Synthesize the text and returns it in a string with audio in default format and sample rate. """
wav = b""
try:
if text:
text = self.preprocessing.process(text)
mp3 = self.get_tts_mp3(self.cfg['TTS']['SpeechTech']['voice'], text)
wav = audio.convert_mp3_to_wav(self.cfg, mp3)
wav = audio.change_tempo(self.cfg, self.cfg['TTS']['SpeechTech']['tempo'], wav)
# if self.cfg['TTS']['debug']:
# m = "TTS cache hits %d and misses %d " % (self.get_tts_mp3.hits, self.get_tts_mp3.misses)
# self.cfg['Logging']['system_logger'].debug(m)
return wav
else:
return b""
except TTSException as e:
m = unicode(e) + " Text: %s" % text
self.cfg['Logging']['system_logger'].exception(m)
return b""
return wav
def __init__(self, cfg):
super(FliteTTS, self).__init__(cfg)
self.preprocessing = TTSPreprocessing(
self.cfg, self.cfg['TTS']['Flite']['preprocessing'])
class GoogleTTS(TTSInterface):
"""
Uses Google TTS service to synthesize sentences in a specific language,
e.g. en, cs.
The main function synthesize returns a string which contain a RIFF wave
file audio of the synthesized text.
"""
def __init__(self, cfg):
super(GoogleTTS, self).__init__(cfg)
self.preprocessing = TTSPreprocessing(
self.cfg, self.cfg['TTS']['Google']['preprocessing'])
@cache.persistent_cache(True, 'GoogleTTS.get_tts_mp3.')
def get_tts_mp3(self, language, text, rate=1.0):
"""
Access Google TTS service and get synthesized audio.
Note that the returned audio is in the MP3 format.
Returns a string with MP3 in it.
"""
baseurl = "http://translate.google.com/translate_tts"
values = {'q': text.encode('utf8'), 'tl': language, 'rate': rate}
if self.cfg['TTS']['Google']['debug']:
print values
data = urllib.urlencode(values)
request = urllib2.Request(baseurl, data)
request.add_header(
"User-Agent",
"Mozilla/5.0 (Macintosh; Intel Mac OS X 10_7_3) AppleWebKit/535.19 (KHTML, like Gecko) Chrome/18.0.1025.163 Safari/535.19"
)
request.add_header(
'Referer', 'http://www.gstatic.com/translate/sound_player2.swf')
mp3response = urllib2.urlopen(request)
return mp3response.read()
def synthesize(self, text):
"""
Synthesize the text and returns it in a string with audio in default
format and sample rate.
"""
try:
if text:
text = self.preprocessing.process(text)
mp3 = self.get_tts_mp3(
self.cfg['TTS']['Google']['language'], text,
self.cfg['TTS']['Google'].get('tempo', 1.0))
wav = audio.convert_mp3_to_wav(self.cfg, mp3)
wav = audio.change_tempo(self.cfg,
self.cfg['TTS']['Google']['tempo'],
wav)
return wav
else:
return b""
except TTSException as e:
m = unicode(e) + " Text: %s" % text
self.cfg['Logging']['system_logger'].exception(m)
return b""
def __init__(self, cfg):
super(GoogleTTS, self).__init__(cfg)
self.preprocessing = TTSPreprocessing(
self.cfg, self.cfg['TTS']['Google']['preprocessing'])
class GoogleTTS(TTSInterface):
"""
Uses Google TTS service to synthesize sentences in a specific language,
e.g. en, cs.
The main function synthesize returns a string which contain a RIFF wave
file audio of the synthesized text.
"""
def __init__(self, cfg):
super(GoogleTTS, self).__init__(cfg)
self.preprocessing = TTSPreprocessing(
self.cfg, self.cfg['TTS']['Google']['preprocessing'])
@cache.persistent_cache(True, 'GoogleTTS.get_tts_mp3.')
def get_tts_mp3(self, language, text, rate=1.0):
"""
Access Google TTS service and get synthesized audio.
Note that the returned audio is in the MP3 format.
Returns a string with MP3 in it.
"""
baseurl = "http://translate.google.com/translate_tts"
values = {'q': text.encode('utf8'), 'tl': language, 'rate': rate}
if self.cfg['ASR']['Google']['debug']:
print values
data = urllib.urlencode(values)
request = urllib2.Request(baseurl, data)
request.add_header("User-Agent", "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_7_3) AppleWebKit/535.19 (KHTML, like Gecko) Chrome/18.0.1025.163 Safari/535.19")
request.add_header('Referer', 'http://www.gstatic.com/translate/sound_player2.swf')
mp3response = urllib2.urlopen(request)
return mp3response.read()
def synthesize(self, text):
"""
Synthesize the text and returns it in a string with audio in default
format and sample rate.
"""
wav = b""
try:
if text:
text = self.preprocessing.process(text)
mp3 = self.get_tts_mp3(self.cfg['TTS']['Google']['language'],
text,
self.cfg['TTS']['Google'].get('tempo', 1.0))
wav = audio.convert_mp3_to_wav(self.cfg, mp3)
wav = audio.change_tempo(self.cfg, self.cfg['TTS']['Google']['tempo'], wav)
return wav
else:
return b""
except TTSException as e:
m = unicode(e) + " Text: %s" % text
self.cfg['Logging']['system_logger'].exception(m)
return b""
return wav
def __init__(self, cfg):
super(SpeechtechTTS, self).__init__(cfg)
self.preprocessing = TTSPreprocessing(
self.cfg, self.cfg['TTS']['SpeechTech']['preprocessing'])
class SpeechtechTTS(TTSInterface):
""" Uses SpeechTech TTS service to synthesize sentences in a specific language, e.g. en, cs.
The main function synthesize returns a string which contain a RIFF wave file audio of the synthesized text.
"""
def __init__(self, cfg):
super(SpeechtechTTS, self).__init__(cfg)
self.preprocessing = TTSPreprocessing(
self.cfg, self.cfg['TTS']['SpeechTech']['preprocessing'])
@cache.lru_cache(10000)
@cache.persistent_cache(True, 'SpeechtechTTS.get_tts_mp3.')
def get_tts_mp3(self, voice, text):
""" Access SpeechTech TTS service and get synthesized audio.
Returns a string with wav in it.
"""
try:
TIMEOUT = 30
LOGIN = self.cfg['TTS']['SpeechTech']['login']
PASSWD = self.cfg['TTS']['SpeechTech']['password']
REALM = 'TTS-Server'
ROOT_URI = 'http://tts2.speechtech.cz'
URI = '%s/add_to_queue' % ROOT_URI
if hasattr(socket, 'setdefaulttimeout'):
socket.setdefaulttimeout(TIMEOUT)
# Create an OpenerDirector with support for Basic HTTP Authentication...
auth_handler = urllib2.HTTPDigestAuthHandler()
auth_handler.add_password(realm=REALM,
uri=URI,
user=LOGIN,
passwd=PASSWD)
opener = urllib2.build_opener(auth_handler)
urllib2.install_opener(opener)
params = urllib.urlencode([('text', text.encode('utf8')),
('engine', voice.encode('utf8'))])
task_id = urllib2.urlopen('%s/add_to_queue' % ROOT_URI,
params).read().strip()
uri = None
i = 20
while i:
params = urllib.urlencode([('task_id', task_id)])
resp = urllib2.urlopen('%s/query_status' % ROOT_URI,
params).read().splitlines()
code = int(resp[0])
#print 'Status is', code
if code == 3:
uri = resp[1]
#print 'Result available at', uri
break
time.sleep(0.2)
i -= 1
if uri:
request = urllib2.Request(uri)
request.add_header(
"User-Agent",
"Mozilla/5.0 (X11; U; Linux i686) Gecko/20071127 Firefox/2.0.0.11"
)
mp3response = urllib2.urlopen(request)
return mp3response.read()
except (urllib2.HTTPError, urllib2.URLError):
raise TTSException("SpeechTech TTS error.")
raise TTSException("Time out: No data synthesized.")
def synthesize(self, text):
""" Synthesize the text and returns it in a string with audio in default format and sample rate. """
wav = b""
try:
if text:
text = self.preprocessing.process(text)
mp3 = self.get_tts_mp3(self.cfg['TTS']['SpeechTech']['voice'],
text)
wav = audio.convert_mp3_to_wav(self.cfg, mp3)
wav = audio.change_tempo(
self.cfg, self.cfg['TTS']['SpeechTech']['tempo'], wav)
# if self.cfg['TTS']['debug']:
# m = "TTS cache hits %d and misses %d " % (self.get_tts_mp3.hits, self.get_tts_mp3.misses)
# self.cfg['Logging']['system_logger'].debug(m)
return wav
else:
return b""
except TTSException as e:
m = unicode(e) + " Text: %s" % text
self.cfg['Logging']['system_logger'].exception(m)
return b""
return wav
class GoogleTTS(TTSInterface):
"""
Uses Google TTS service to synthesize sentences in a specific language,
e.g. en, cs.
The main function synthesize returns a string which contain a RIFF wave
file audio of the synthesized text.
"""
def __init__(self, cfg):
super(GoogleTTS, self).__init__(cfg)
self.preprocessing = TTSPreprocessing(
self.cfg, self.cfg['TTS']['Google']['preprocessing'])
@cache.persistent_cache(True, 'GoogleTTS.get_tts_mp3.')
def get_tts_mp3(self, language, text, rate=1.0):
"""
Access Google TTS service and get synthesized audio.
Note that the returned audio is in the MP3 format.
Returns a string with MP3 in it.
"""
baseurl = "http://translate.google.com/translate_tts"
values = {
'q': text.encode('utf8'),
'tl': language,
'rate': rate,
'ie': 'UTF-8',
'total': 1,
'idx': 0,
'client': 't'
}
if self.cfg['TTS']['Google']['debug']:
print values
data = urllib.urlencode(values)
# not using the "data" parameter of urllib2.Request => GET method will be used
request = urllib2.Request(baseurl + '?' + str(data))
request.add_header(
"User-Agent",
"Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:41.0) Gecko/20100101 Firefox/41.0"
)
request.add_header("Accept", "*/*")
mp3response = urllib2.urlopen(request)
return mp3response.read()
def synthesize(self, text):
"""
Synthesize the text and returns it in a string with audio in default
format and sample rate.
"""
try:
if text:
text = self.preprocessing.process(text)
mp3 = self.get_tts_mp3(
self.cfg['TTS']['Google']['language'], text,
self.cfg['TTS']['Google'].get('tempo', 1.0))
wav = audio.convert_mp3_to_wav(self.cfg, mp3)
wav = audio.change_tempo(self.cfg,
self.cfg['TTS']['Google']['tempo'],
wav)
return wav
else:
return b""
except TTSException as e:
m = unicode(e) + " Text: %s" % text
self.cfg['Logging']['system_logger'].exception(m)
return b""
class SpeechtechTTS(TTSInterface):
""" Uses SpeechTech TTS service to synthesize sentences in a specific language, e.g. en, cs.
The main function synthesize returns a string which contain a RIFF wave file audio of the synthesized text.
"""
def __init__(self, cfg):
super(SpeechtechTTS, self).__init__(cfg)
self.preprocessing = TTSPreprocessing(self.cfg, self.cfg['TTS']['SpeechTech']['preprocessing'])
@cache.lru_cache(10000)
@cache.persistent_cache(True, 'SpeechtechTTS.get_tts_mp3.')
def get_tts_mp3(self, voice, text):
""" Access SpeechTech TTS service and get synthesized audio.
Returns a string with wav in it.
"""
try:
TIMEOUT = 30
LOGIN = self.cfg['TTS']['SpeechTech']['login']
PASSWD = self.cfg['TTS']['SpeechTech']['password']
ROOT_URI = 'http://services.speechtech.cz/tts/v3'
if hasattr(socket, 'setdefaulttimeout'):
socket.setdefaulttimeout(TIMEOUT)
mgr = urllib2.HTTPPasswordMgrWithDefaultRealm()
mgr.add_password(None, ROOT_URI, LOGIN, PASSWD)
auth = urllib2.HTTPDigestAuthHandler(mgr)
opener = urllib2.build_opener(auth)
mp3data = opener.open(ROOT_URI+'/synth', data=json.dumps({'text': text, 'engine': voice})).read()
return mp3data
except (urllib2.HTTPError, urllib2.URLError) as e:
raise TTSException("SpeechTech TTS error: " + unicode(e))
raise TTSException("Time out: No data synthesized.")
def synthesize(self, text):
""" Synthesize the text and returns it in a string with audio in default format and sample rate. """
wav = b""
try:
if text:
text = self.preprocessing.process(text)
mp3 = self.get_tts_mp3(self.cfg['TTS']['SpeechTech']['voice'], text)
wav = audio.convert_mp3_to_wav(self.cfg, mp3)
wav = audio.change_tempo(self.cfg, self.cfg['TTS']['SpeechTech']['tempo'], wav)
# if self.cfg['TTS']['debug']:
# m = "TTS cache hits %d and misses %d " % (self.get_tts_mp3.hits, self.get_tts_mp3.misses)
# self.cfg['Logging']['system_logger'].debug(m)
return wav
else:
return b""
except TTSException as e:
m = unicode(e) + " Text: %s" % text
self.cfg['Logging']['system_logger'].exception(m)
return b""
return wav
class GoogleTTS(TTSInterface):
"""
Uses Google TTS service to synthesize sentences in a specific language,
e.g. en, cs.
The main function synthesize returns a string which contain a RIFF wave
file audio of the synthesized text.
"""
def __init__(self, cfg):
super(GoogleTTS, self).__init__(cfg)
self.preprocessing = TTSPreprocessing(
self.cfg, self.cfg['TTS']['Google']['preprocessing'])
@cache.persistent_cache(True, 'GoogleTTS.get_tts_mp3.')
def get_tts_mp3(self, language, text, rate=1.0):
"""
Access Google TTS service and get synthesized audio.
Note that the returned audio is in the MP3 format.
Returns a string with MP3 in it.
"""
baseurl = "http://translate.google.com/translate_tts"
values = {'q': text.encode('utf8'),
'tl': language,
'rate': rate,
'ie': 'UTF-8',
'total': 1,
'idx': 0,
'client': 't'}
if self.cfg['TTS']['Google']['debug']:
print values
data = urllib.urlencode(values)
# not using the "data" parameter of urllib2.Request => GET method will be used
request = urllib2.Request(baseurl + '?' + str(data))
request.add_header("User-Agent", "Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:41.0) Gecko/20100101 Firefox/41.0")
request.add_header("Accept", "*/*")
mp3response = urllib2.urlopen(request)
return mp3response.read()
def synthesize(self, text):
"""
Synthesize the text and returns it in a string with audio in default
format and sample rate.
"""
try:
if text:
text = self.preprocessing.process(text)
mp3 = self.get_tts_mp3(self.cfg['TTS']['Google']['language'],
text,
self.cfg['TTS']['Google'].get('tempo', 1.0))
wav = audio.convert_mp3_to_wav(self.cfg, mp3)
wav = audio.change_tempo(self.cfg, self.cfg['TTS']['Google']['tempo'], wav)
return wav
else:
return b""
except TTSException as e:
m = unicode(e) + " Text: %s" % text
self.cfg['Logging']['system_logger'].exception(m)
return b""
def __init__(self, cfg):
"""Intitialize: just remember the configuration."""
self.cfg = cfg
super(VoiceRssTTS, self).__init__(cfg)
self.preprocessing = TTSPreprocessing(self.cfg, self.cfg['TTS']['VoiceRss']['preprocessing'])
