watson_tts.py 文件源码

def play_file(fname):
    # create an audio object
    wf = wave.open(fname, 'rb')
    p = pyaudio.PyAudio()
    chunk = 1024

    # open stream based on the wave object which has been input.
    stream = p.open(format=p.get_format_from_width(wf.getsampwidth()),
                    channels=wf.getnchannels(),
                    rate=wf.getframerate(),
                    output=True)

    # read data (based on the chunk size)
    data = wf.readframes(chunk)

    # play stream (looping from beginning of file to the end)
    while data != '':
        # writing to the stream is what *actually* plays the sound.
        stream.write(data)
        data = wf.readframes(chunk)

        # cleanup stuff.
    stream.close()
    p.terminate()