examples/sound_speed/by_librosa.py

#!/usr/bin/python3
# -*- coding: utf-8 -*-
import argparse
import os.path

import librosa
import numpy as np
from scipy.io import wavfile

from project_settings import project_path


def get_args():
    parser = argparse.ArgumentParser()
    parser.add_argument(
        "--filename",
        default=(project_path / "data/voice_clone_audio/e2_tts/audio_0_3_clone_from_audio_0_2.wav").as_posix(),
        type=str,
    )
    parser.add_argument(
        "--output_file",
        default="temp.wav",
        type=str,
    )
    parser.add_argument("--speed", default=1.1, type=float)
    args = parser.parse_args()
    return args


def main():
    args = get_args()

    signal, sample_rate = librosa.load(args.filename, sr=None)
    signal_ = librosa.effects.time_stretch(signal, rate=args.speed)

    # max_wave_value = 32768.0
    max_wave_value = 1 << 15

    signal_ = signal_ * max_wave_value
    signal_ = np.array(signal_, dtype=np.int16)

    wavfile.write(
        filename=args.output_file,
        rate=sample_rate,
        data=signal_

    )
    return


if __name__ == "__main__":
    main()