Synthesize text for English shadowing training using Google Could TTS (WaveNet)

cloud_tts.py

# coding: utf-8

def synthesize_text(text, path):
    import os
    from google.cloud import texttospeech
    from pydub import AudioSegment

    client = texttospeech.TextToSpeechClient()

    input_text = texttospeech.types.SynthesisInput(text=text)

    voice = texttospeech.types.VoiceSelectionParams(
            language_code='en-US', name='en-US-Wavenet-D')

    audio_config = texttospeech.types.AudioConfig(
        audio_encoding=texttospeech.enums.AudioEncoding.LINEAR16,
        speaking_rate=0.8)

    response = client.synthesize_speech(input_text, voice, audio_config)

    with open(path + '.wav', 'wb') as out:
        out.write(response.audio_content)

    audio = AudioSegment.from_wav(path + '.wav')
    silence =AudioSegment.silent(duration=1000)
    audio += silence
    audio.export(path + '.mp3', format='mp3')
    os.remove(path + '.wav')

    print(path)


if __name__ == '__main__':
    with open('scripts.txt', 'r') as fi:
        for line in fi:
            path, text = line.rstrip().split('\t')
            synthesize_text(text, path)