chinese_test.py

if __name__ == '__main__':
    import os
    from RealtimeTTS import TextToAudioStream

    def dummy_generator():
        yield "我喜欢读书。天气很好。我们去公园吧。今天是星期五。早上好。这是我的朋友。请帮我。吃饭了吗？我在学中文。晚安。"

    def synthesize(engine, generator):

        stream = TextToAudioStream(engine)

        print ("Starting to play stream")
        stream.feed(generator)
        filename = "synthesis_chinese_" + engine.engine_name

        # ❗ use these for chinese: minimum_sentence_length = 2, minimum_first_fragment_length = 2, tokenizer="stanza", language="zh", context_size=2
        stream.play(
            minimum_sentence_length = 2,
            minimum_first_fragment_length = 2, 
            output_wavfile = f"{filename}.wav",
            on_sentence_synthesized = lambda sentence: 
                print("Synthesized: " + sentence),
            tokenizer="stanza", 
            language="zh",
            context_size=2)
        
        with open(f"{filename}.txt", "w", encoding="utf-8") as f:
            f.write(stream.text())

        engine.shutdown()

    def get_engine(name):
        if name == "coqui":
            from RealtimeTTS import CoquiEngine
            # ❗ use these for chinese: cloning_reference_wav="female_chinese", language = "zh"   # you can exchange cloning_reference_wav with you own
            return CoquiEngine(voice="female_chinese", language = "zh") # using a chinese cloning reference gives better quality

        elif name == "azure":
            from RealtimeTTS import AzureEngine
            # ❗ use these for chinese: voice="zh-CN-XiaoxiaoNeural"   # or specify a different azure zn-CN voice 
            return AzureEngine(os.environ.get("AZURE_SPEECH_KEY"), os.environ.get("AZURE_SPEECH_REGION"), voice="zh-CN-XiaoxiaoNeural")

        elif name == "elevenlabs":
            from RealtimeTTS import ElevenlabsEngine
            return ElevenlabsEngine(os.environ.get("ELEVENLABS_API_KEY"))

        else:
            from RealtimeTTS import SystemEngine
            # ❗ use these for chinese: voice = "Huihui"   # or specify a different locally installed chinese tts voice
            return SystemEngine(voice = "Huihui")

    for engine_name in ["coqui", "elevenlabs", "azure", "system"]:
        print ("Starting engine: " + engine_name)
        engine = get_engine(engine_name)

        print ("Synthesizing with engine: " + engine_name)
        synthesize(engine, dummy_generator())