ARUNO/split.py at main · Cmoris/ARUNO · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
from pydub import AudioSegment, silence
import math
import os
from pathlib import Path

def split_audio_by_silence(audio, min_silence_len=2000, silence_thresh=-40, keep_silence=500):
    """
    Splits an AudioSegment using silence as boundary.

    Parameters:
    - audio: AudioSegment object.
    - min_silence_len: Minimum duration (ms) of silence to be considered a split point.
    - silence_thresh: Silence threshold in dBFS.
    - keep_silence: Duration (ms) of silence to retain at the start and end of each chunk.

    Returns:
    - chunks: List of AudioSegment objects.
    - boundaries: List of tuples (start_ms, end_ms) for each chunk.
    """
    # Detect silence intervals
    silent_intervals = silence.detect_silence(
        audio,
        min_silence_len=min_silence_len,
        silence_thresh=silence_thresh
    )

    # Flatten silent_intervals into boundaries
    split_points = [0] + [end for _, end in silent_intervals] + [len(audio)]
    print(split_points)
    # Remove too-close split points (edge case)
    split_points = sorted(list(set(split_points)))

    # Generate audio chunks
    chunks = []
    boundaries = []
    for i in range(len(split_points) - 1):
        start = max(0, split_points[i] - keep_silence)
        end = min(len(audio), split_points[i+1] + keep_silence)
        chunk = audio[start:end]
        chunks.append(chunk)
        boundaries.append((start, end))

    return chunks, boundaries

# 示例：加载音频并保存切割结果
def save_by_silence(audio_file:str, output_dir: str=None, min_silence_len=2000, silence_thresh=-40, keep_silence=500):
    if output_dir is None:
        p = Path(audio_file)
        p.parent.joinpath(p.stem).mkdir(exist_ok=True)
        output_dir = p.parent.joinpath(p.stem)
    else:
        output_dir = Path(output_dir)
        output_dir.mkdir(exist_ok=True)
    audio = AudioSegment.from_file(audio_file)
    chunks, boundaries = split_audio_by_silence(audio, min_silence_len, silence_thresh, keep_silence)

    os.makedirs(output_dir, exist_ok=True)
    for i, (chunk, (start, end)) in enumerate(zip(chunks, boundaries)):
        output_path = output_dir.joinpath(f"output_{i:03d}.wav")
        chunk.export(output_path, format="wav")

def save_by_segment(audio_file : str, output_dir : str=None):
    if output_dir is None:
        p = Path(audio_file)
        p.parent.joinpath(p.stem).mkdir(exist_ok=True)
        output_dir = p.parent.joinpath(p.stem)
    else:
        output_dir = Path(output_dir)
        output_dir.mkdir(exist_ok=True)

    audio = AudioSegment.from_file(audio_file)
    segment_length = 30 * 1000
    total_segments = math.ceil(len(audio) / segment_length)
    for i in range(total_segments):
        output_path = output_dir.joinpath(f"output_{i:03d}.wav")
        start = i * segment_length
        end = min((i + 1) * segment_length, len(audio))
        segment = audio[start:end]
        segment.export(output_path, format="wav")

if __name__ == "__main__":
    audio_file = "/home/muyun/VScode/project/ASR/data/audio/20250323.wav"
    save_by_segment(audio_file=audio_file)