基于vad对音频进行切割

最新推荐文章于 2025-04-26 22:36:52 发布

原创

最新推荐文章于 2025-04-26 22:36:52 发布 · 696 阅读

1 ·

CC 4.0 BY-SA版权

文章标签：

#音视频 #python

why?

通过ai生成的音频上传到X音音乐，上传片段超过60s

So

基于vad停顿切分音频，分段上传

"""
pip install pydub
pip install numpy
pip install librosa
pip install soundfile
"""
from pydub import AudioSegment
from pydub.silence import split_on_silence
import numpy as np


def dynamic_audio_split(file_path, output_dir="output", max_length=50, min_length=5):
    """Intelligent audio segmentation with adaptive parameters"""
    # Audio loading and initialization
    audio = AudioSegment.from_file(file_path)
    segments = []

    # Adaptive parameter settings
    base_params = {
   
   
        'silence_thresh': -40,
        'min_silence': 600,
        'keep_silence': 200,
        'step_size': 50
    }

    # Phase 1: Primary segmentation
    raw_chunks = split_on_silence(
        audio,
        silence_thresh=base_params['silence_thresh'],
        min_silence_len=base_params['min_silence'],
        keep_silence=base_params['keep_silence'],
        seek_step=base_params['step_size']
    )

    # Phase 2: Dynamic processing
    buffer = raw_chunks[0]
    for chunk in raw_chunks[1:]