pyaudio+pocketsphinx

参考例程：https://pythonism.wordpress.com/2013/06/06/pocketsphinx-voice-recognition-with-python/

          
其他参考：http://blog.youkuaiyun.com/u013000753/article/details/46344567
https://github.com/mondhs/lt-pocketsphinx-tutorial/tree/master/impl/demo-py/public_service 立陶宛Lithuanian
功能点：pyaudio录音。音频流。
importsys,os
importpyaudio
importwave

 

hmdir = " / usr / share / pocketsphinx / model / hmm / wsj1"

lmd   = " / usr / share / pocketsphinx / model / lm / wsj / wlist5o. 3e - 7.vp .tg.lm.DMP"

dictd = " / usr / share / pocketsphinx / model / lm / wsj / wlist5o.dic"

 

def decodeSpeech(hmmd,lmdir,dictp,wavfile):

 

     import pocketsphinx as ps

     import sphinxbase

 

     speechRec = ps.Decoder(hmm = hmmd, lm = lmdir, dict = dictp)

     wavFile = file (wavfile, 'rb' )

     wavFile.seek( 44 )

     speechRec.decode_raw(wavFile)

     result = speechRec.get_hyp()

 

     return result[ 0 ]

 

CHUNK = 1024

FORMAT = pyaudio.paInt16

CHANNELS = 1

RATE = 16000

RECORD_SECONDS = 10

 

for x in range ( 10 ):

     fn = "o" + str (x) + ".wav"

     p = pyaudio.PyAudio()

     stream = p. open ( format = FORMAT , channels = CHANNELS, rate = RATE, input = True , frames_per_buffer = CHUNK)

     print (" * recording")

     frames = []

     for i in range ( 0 , int (RATE / CHUNK * RECORD_SECONDS)):

         data = stream.read(CHUNK)

         frames.append(data)

     print (" * done recording")

     stream.stop_stream()

     stream.close()

     p.terminate()

     wf = wave. open (fn, 'wb' )

     wf.setnchannels(CHANNELS)

     wf.setsampwidth(p.get_sample_size( FORMAT ))

     wf.setframerate(RATE)

     wf.writeframes(b''.join(frames))

     wf.close()

     wavfile = fn

     recognised = decodeSpeech(hmdir,lmd,dictd,wavfile)

     print recognised

     cm = 'espeak "' + recognised + '"'

     os.system(cm)