def recognize(wav_file):
#BASE_PATH = os.path.dirname(os.path.realpath(__file__))
#HMDIR = os.path.join(BASE_PATH, "hmm")
#LMDIR = os.path.join(BASE_PATH, "lm/en-us.lm.bin")
#DICTD = os.path.join(BASE_PATH, "dict/en_in.dic")
sound = "try1.wav"
model_path = get_model_path()
data_path = get_data_path()
config = DefaultConfig()
config.set_string('-hmm', "hmm/")
config.set_string('-lm', 'lm\en-us.lm.bin')
config.set_string('-dict', 'dict\en_in.dic')
#decoder = Decoder(config)
"""
Run speech recognition on a given file.
"""
speech_rec = Decoder(config)
print("Decoder Initialized")
wav_file = wave.open(wav_file, 'rb')
print("AudioFile Loaded")
speech_rec.decode_raw(wav_file)
print("Audio file decoded")
result = speech_rec.get_hyp()
print("Result Ready\n")
return result
def record(listen_time):
THRESHOLD = None
WAVE_OUTPUT_FILENAME = "livewav.wav"
p = pyaudio.PyAudio()
if THRESHOLD == None:
THRESHOLD = fetchThreshold()
print THRESHOLD
stream = p.open(format=FORMAT,
channels=1,
rate=RATE,
input=True,
frames_per_buffer=CHUNK)
print "* recording"
frames = []
detected = False
for i in range(0, RATE / CHUNK * listen_time):
data = stream.read(CHUNK)
frames.append(data)
score = getScore(data)
if score < THRESHOLD:
continue
else:
detected = True
if not detected:
print "nothing detected"
return("")
print "* done recording"
# stream.stop_stream()
stream.close()
p.terminate()
# write data to WAVE file
data = ''.join(frames)
wf = wave.open(WAVE_OUTPUT_FILENAME, 'wb')
wf.setnchannels(1)
wf.setsampwidth(p.get_sample_size(FORMAT))
wf.setframerate(RATE)
wf.writeframes(data)
wf.close()
sysdir = os.getcwd()
wavfile = sysdir + "/livewav.wav"
config = Decoder.default_config()
config.set_string('-hmm', hmdir)
config.set_string('-lm', lmdir)
config.set_string('-dict', dictd)
config.set_string('-logfn', '/dev/null')
speechRec = Decoder(config)
with open(wavfile, 'rb') as wavFile:
speechRec.decode_raw(wavFile)
#result = speechRec.get_hyp()
return(speechRec.hyp().hypstr)
def passiverecord(THRESHOLD=None):
FORMAT = pyaudio.paInt16
CHANNELS = 1
LISTEN_TIME = 2
WAVE_OUTPUT_FILENAME = "passive.wav"
p = pyaudio.PyAudio()
if THRESHOLD == None:
THRESHOLD = fetchThreshold()
print THRESHOLD
stream = p.open(format=FORMAT,
channels=1,
rate=RATE,
input=True,
frames_per_buffer=CHUNK)
# print "* recording"
frames = []
lastN = [THRESHOLD * 1.2 for i in range(30)]
for i in range(0, RATE / CHUNK * LISTEN_TIME):
data = stream.read(CHUNK)
frames.append(data)
score = getScore(data)
lastN.pop(0)
lastN.append(score)
average = sum(lastN) / float(len(lastN))
# print average,THRESHOLD * 0.8
if average < THRESHOLD * 0.8:
break
# print "* done recording"
# stream.stop_stream()
stream.close()
p.terminate()
# write data to WAVE file
data = ''.join(frames)
wf = wave.open(WAVE_OUTPUT_FILENAME, 'wb')
wf.setnchannels(1)
wf.setsampwidth(p.get_sample_size(FORMAT))
wf.setframerate(RATE)
wf.writeframes(data)
wf.close()
sysdir = os.getcwd()
wavfile = sysdir + "/passive.wav"
# decoded=decodepassive()
speechRec = Decoder(hmm=hmdir, lm=lmdir, dict=dictd)
with open(wavfile, 'rb') as wavFile:
speechRec.decode_raw(wavFile)
result = speechRec.get_hyp()
return (result[0])
def record(listen_time):
THRESHOLD=None
WAVE_OUTPUT_FILENAME = "livewav.wav"
p = pyaudio.PyAudio()
if THRESHOLD == None:
THRESHOLD = fetchThreshold()
print THRESHOLD
stream = p.open(format=FORMAT,
channels=1,
rate=RATE,
input=True,
frames_per_buffer=CHUNK)
print "* recording"
frames = []
detected=False
for i in range(0, RATE / CHUNK * listen_time):
data = stream.read(CHUNK)
frames.append(data)
score = getScore(data)
if score < THRESHOLD:
continue
else:
detected=True
if not detected:
print "nothing detected"
return("")
print "* done recording"
#stream.stop_stream()
stream.close()
p.terminate()
# write data to WAVE file
data = ''.join(frames)
wf = wave.open(WAVE_OUTPUT_FILENAME, 'wb')
wf.setnchannels(1)
wf.setsampwidth(p.get_sample_size(FORMAT))
wf.setframerate(RATE)
wf.writeframes(data)
wf.close()
sysdir = os.getcwd()
wavfile = sysdir+"/livewav.wav"
speechRec = Decoder(hmm=hmdir, lm=lmdir, dict=dictd)
with open(wavfile, 'rb') as wavFile:
speechRec.decode_raw(wavFile)
result = speechRec.get_hyp()
return(result[0])
def best_sphinx_speech_result(pyaudio, wav_name, profile):
if not have_sphinx_dictionary:
if not profile.has_key("words"):
raise(KeyError("Pass the possible words in in profile"))
compile("sentences.txt", "dictionary.dic", "language_model.lm", profile["words"])
global have_sphinx_dictionary
have_sphinx_dictionary = True
wav_file = file(wav_name, 'rb')
speechRec = Decoder(
hmm = "/usr/local/share/pocketsphinx/model/hmm/en_US/hub4wsj_sc_8k",
lm = "language_model.lm",
dict = "dictionary.dic"
)
speechRec.decode_raw(wav_file)
results = speechRec.get_hyp()
return results[0]
class CMUSphinxRecognizer(BaseRecognizer):
def __init__(self):
config = Decoder.default_config()
config.set_string('-hmm', SPHINX_HMM)
config.set_string('-lm', SPHINX_LM)
config.set_string('-dict', SPHINX_DICT)
self.decoder = Decoder(config)
def recognize(self, raw_audio):
file_path = self.__save_file(raw_audio)
with open(file_path, 'r') as wav_fp:
self.decoder.decode_raw(wav_fp)
hypothesis = self.decoder.hyp()
return hypothesis.hypstr, hypothesis.best_score, hypothesis.prob
@staticmethod
def __save_file(data):
tmp_fp = NamedTemporaryFile(delete=False)
tmp_fp.write(data)
tmp_fp.close()
return tmp_fp.name
'''
Created on Dec 29, 2013
@author: Mindaugas Greibus
'''
import sys, os
from pocketsphinx import Decoder
MODELDIR = "../models"
# Create a decoder with certain model
config = Decoder.default_config()
config.set_string('-hmm', os.path.join(MODELDIR, 'hmm/lt.cd_cont_200/'))
config.set_string('-jsgf', os.path.join(MODELDIR, 'lm/robotas.gram'))
config.set_string('-dict', os.path.join(MODELDIR, 'dict/robotas.dict'))
decoder = Decoder(config)
decoder.decode_raw(
open(os.path.join(MODELDIR, '../test/audio/varyk_pirmyn-16k.wav'), 'rb'))
# Retrieve hypothesis.
hypothesis = decoder.hyp()
print('Best hypothesis: ', hypothesis.best_score, hypothesis.hypstr)
print('Best hypothesis segments: ', [seg.word for seg in decoder.seg()])
'''
Created on Dec 29, 2013
@author: Mindaugas Greibus
'''
import sys, os
from pocketsphinx import Decoder
MODELDIR = "../models"
# Create a decoder with certain model
config = Decoder.default_config()
config.set_string('-hmm', os.path.join(MODELDIR, 'hmm/lt.cd_cont_200/'))
config.set_string('-jsgf', os.path.join(MODELDIR, 'lm/robotas.gram'))
config.set_string('-dict', os.path.join(MODELDIR, 'dict/robotas.dict'))
decoder = Decoder(config)
decoder.decode_raw(open(os.path.join(MODELDIR, '../test/audio/varyk_pirmyn-16k.wav'), 'rb'))
# Retrieve hypothesis.
hypothesis = decoder.hyp()
print ('Best hypothesis: ', hypothesis.best_score, hypothesis.hypstr)
print ('Best hypothesis segments: ', [seg.word for seg in decoder.seg()])
def record(THRESHOLD=None):
FORMAT = pyaudio.paInt16
CHANNELS = 1
LISTEN_TIME = 4
WAVE_OUTPUT_FILENAME = "livewav.wav"
p = pyaudio.PyAudio()
if THRESHOLD == None:
THRESHOLD = fetchThreshold()
print THRESHOLD
stream = p.open(format=FORMAT,
channels=1,
rate=RATE,
input=True,
frames_per_buffer=CHUNK)
print "* recording"
frames = []
lastN = [THRESHOLD * 1.2 for i in range(30)]
for i in range(0, RATE / CHUNK * LISTEN_TIME):
data = stream.read(CHUNK)
frames.append(data)
score = getScore(data)
lastN.pop(0)
lastN.append(score)
average = sum(lastN) / float(len(lastN))
#print average,THRESHOLD * 0.8
if average < THRESHOLD * 0.8:
break
print "* done recording"
#stream.stop_stream()
stream.close()
p.terminate()
# write data to WAVE file
data = ''.join(frames)
wf = wave.open(WAVE_OUTPUT_FILENAME, 'wb')
wf.setnchannels(1)
wf.setsampwidth(p.get_sample_size(FORMAT))
wf.setframerate(RATE)
wf.writeframes(data)
wf.close()
sysdir = os.getcwd()
wavfile = sysdir+"/livewav.wav"
#decoded=decodepassive()
speechRec = Decoder(hmm=hmdir, lm=lmdir, dict=dictd)
with open(wavfile, 'rb') as wavFile:
speechRec.decode_raw(wavFile)
result = speechRec.get_hyp()
return(result[0])
def decodepassive():
speechRec = Decoder(hmm = hmdir, lm = lmdir, dict = dictd)
with open(passivewav, 'rb') as passivewav:
speechRec.decode_raw(passivewav)
result = speechRec.get_hyp()
return(result[0])
p.terminate()
# write data to WAVE file
data = ''.join(all)
wf = wave.open(WAVE_OUTPUT_FILENAME, 'wb')
wf.setnchannels(CHANNELS)
wf.setsampwidth(p.get_sample_size(FORMAT))
wf.setframerate(RATE)
wf.writeframes(data)
wf.close()
if __name__ == "__main__":
hmdir = "/usr/share/pocketsphinx/model/hmm/en_US/hub4wsj_sc_8k"
lmdir = "/usr/share/pocketsphinx/model/lm/en_US/hub4.5000.DMP"
dictd = "/usr/share/pocketsphinx/model/lm/en_US/cmu07a.dic"
record()
wavfile = "/home/shridhar/pocketsphinxtest/livewav.wav"
speechRec = Decoder(hmm=hmdir, lm=lmdir, dict=dictd)
wavFile = file(wavfile, 'rb')
speechRec.decode_raw(wavFile)
result = speechRec.get_hyp()
print "Recognised text from the converted video file"
print "%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%"
print result[0]
print "%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%"
from os import environ, path
from pocketsphinx import Decoder
from sphinxbase import *
print('a')
model_dir = '/home/wmlab/CMU_try/zh_broadcastnews_ptm256_8000/'
# lm_dir = '/home/wmlab/CMU_try/CMUsphinx-Demo-master/0506.lm'
# dict_dir = '/home/wmlab/CMU_try/CMUsphinx-Demo-master/0506.dic'
lm_dir = '/home/wmlab/CMU_try/try_python/usable_version/0506.lm'
dict_dir = '/home/wmlab/CMU_try/try_python/usable_version/0506.dic'
wav_file = '/home/wmlab/CMU_try/testing_audio/testing003.wav'
speech_rec = Decoder(hmm=model_dir, lm=lm_dir, dict=dict_dir)
wavFile = file(wav_file, 'rb')
wavFile.seek(44)
speech_rec.decode_raw(wavFile)
result = speech_rec.get_hyp()
print(result[0])
def decodepassive():
speechRec = Decoder(hmm=hmdir, lm=lmdir, dict=dictd)
with open(passivewav, 'rb') as passivewav:
speechRec.decode_raw(passivewav)
result = speechRec.get_hyp()
return(result[0])