split.py

import subprocess
import random
import os
from pathlib import Path
import librosa
from scipy.io import wavfile
import numpy as np
import torch
import csv
import whisper
import argparse

a="character"

def split_long_audio(model, filepaths, save_dir="data_dir", out_sr=22050):
    if isinstance(filepaths, str):
        filepaths = [filepaths]

    for file_idx, filepath in enumerate(filepaths):

        save_path = Path(save_dir)
        save_path.mkdir(exist_ok=True, parents=True)

        print(f"Transcribing file {file_idx}: '{filepath}' to segments...")
        result = model.transcribe(filepath, word_timestamps=True, task="transcribe", beam_size=5, best_of=5)
        segments = result['segments']

        wav, sr = librosa.load(filepath, sr=None, offset=0, duration=None, mono=True)
        wav, _ = librosa.effects.trim(wav, top_db=20)
        peak = np.abs(wav).max()
        if peak > 1.0:
            wav = 0.98 * wav / peak
        wav2 = librosa.resample(wav, orig_sr=sr, target_sr=out_sr)
        wav2 /= max(wav2.max(), -wav2.min())

        for i, seg in enumerate(segments):
            start_time = seg['start']
            end_time = seg['end']
            wav_seg = wav2[int(start_time * out_sr):int(end_time * out_sr)]
            wav_seg_name = f"{a}_{i}.wav"
            out_fpath = save_path / wav_seg_name
            wavfile.write(out_fpath, rate=out_sr, data=(wav_seg * np.iinfo(np.int16).max).astype(np.int16))

whisper_size = "large"
whisper_model = whisper.load_model(whisper_size)

if __name__ == "__main__":
    parser = argparse.ArgumentParser()
    parser.add_argument("--filename", default="linghua.wav")
    args = parser.parse_args()
    split_long_audio(whisper_model, "./" + args.filename, "./custom_character_voice/character/")