tts_piper.py

import sounddevice as sd
import numpy as np

if __name__ == "__main__":
    from base import BaseMouth
else:
    from .base import BaseMouth


class Mouth_piper(BaseMouth):
    def __init__(
        self,
        device="cpu",
        model_path="models/en_US-ryan-high.onnx",
        config_path="models/en_en_US_ryan_high_en_US-ryan-high.onnx.json",
        player=sd,
        wait=True,
        logger=None,
    ):
        import piper

        self.model = piper.PiperVoice.load(
            model_path=model_path,
            config_path=config_path,
            use_cuda=True if device == "cuda" else False,
        )
        super().__init__(
            sample_rate=self.model.config.sample_rate,
            player=player,
            wait=wait,
            logger=logger,
        )

    def run_tts(self, text):
        audio = b""
        for i in self.model.synthesize_stream_raw(text):
            audio += i
        return np.frombuffer(audio, dtype=np.int16)


if __name__ == "__main__":
    import torch

    device = "cuda" if torch.cuda.is_available() else "cpu"
    mouth = Mouth_piper(
        device=device,
        model_path="../../models/en_US-ryan-high.onnx",
        config_path="../../models/en_en_US_ryan_high_en_US-ryan-high.onnx.json",
    )

    text = (
        "If there's one thing that makes me nervous about the future of self-driving cars, it's that they'll "
        "replace human drivers.\nI think there's a huge opportunity to make human-driven cars safer and more "
        "efficient. There's no reason why we can't combine the benefits of self-driving cars with the ease of use "
        "of human-driven cars."
    )
    print(text)
    mouth.say_multiple(text, lambda x: False)
    sd.wait()