语音处理-批量文件处理静默区删除

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
84
import fnmatch
from pathlib import Path

from scipy.io import wavfile
import webrtcvad
import struct
from scipy.io.wavfile import write
import os
import numpy as np
import matplotlib.pyplot as plt

#开始将所有文件进行批处理操作
def find_all_files(files_path):
    """遍历指定文件夹所有指定类型文件"""
    p = Path(files_path)
    files_names = []  # 存储文件路径名称
    for file in p.rglob('*.wav'):  # 寻找所有txt文件
        x = str(file).split('\')[-1]
        if fnmatch.fnmatch(x, '._*.wav'):
            continue
        else:
            files_names.append(str(file))  # 以字符串形式保存

    return files_names


def all_void_void(single_file_path,flag_session):
    #single_file_path = 'F:文件情感数据集未处理IEMOCAP_full_releaseSession1sentenceswavSes01F_impro01Ses01F_impro01_F013.wav'
    #single_file_path = 'E:codingpythonpythonfor_deleteVDSes01F_impro01_F000.wav'
    print(os.path.join(single_file_path))
    sample_rate, samples = wavfile.read(os.path.join(single_file_path))
    file_name = single_file_path.split("\")[-1]
    file_package = single_file_path.split("\")[-2]
    print(file_name)
    vad = webrtcvad.Vad()
    vad.set_mode(3)
    raw_samples = struct.pack("%dh" % len(samples), *samples)
    window_duration = 0.03
    samples_per_window = int(window_duration * sample_rate + 0.3)
    bytes_per_sample = 2
    segments = []
    try:
        for start in np.arange(0, len(samples), samples_per_window):
            stop = min(start + samples_per_window, len(samples))
            is_speech = vad.is_speech(raw_samples[start * bytes_per_sample: stop * bytes_per_sample],
                                      sample_rate=sample_rate)
            segments.append(dict(
                start=start,
                stop=stop,
                is_speech=is_speech))
    except:
        try:
            speech_samples = np.concatenate(
                [samples[segment['start']:segment['stop']] for segment in segments if segment['is_speech']])
        except:
            print('产生异常了第一次')
    try:
        speech_samples = np.concatenate(
            [samples[segment['start']:segment['stop']] for segment in segments if segment['is_speech']])
        new_path = 'F:文件研究生组情感数据集已处理IEMOCAP_full_release\' + flag_session + 'sentenceswav\' + file_package
        # os.makedirs(new_path)
        if (os.path.exists(new_path)):
            print('1')
        else:
            os.makedirs(new_path)
        new_path =new_path+ '\'+file_name
        wavfile.write(new_path, sample_rate, speech_samples)
    except:
        print('产生异常了第二次')
    #new_path = 'F:文件研究生组情感数据集已处理IEMOCAP_full_releaseSession1sentenceswav\'+file_package
    # os.makedirs(new_path)
    # if(os.path.exists(new_path)):
    #     print('1')
    # else:
    #     os.makedirs(new_path)
    # new_path = new_path + '\'+file_name
    # wavfile.write(new_path, sample_rate, speech_samples)
#Session_list =["Session1","Session2","Session3","Session4","Senssion5"]
Session_list =["Session5"]
for senssion_i in Session_list:
    files_name = find_all_files('F:文件情感数据集未处理IEMOCAP_full_release\'+senssion_i+'sentenceswav')
    for file_name in files_name:
        all_void_void(file_name,senssion_i)