This repository was archived by the owner on Nov 8, 2023. It is now read-only.
-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathaudio_process.py
More file actions
132 lines (118 loc) · 6.79 KB
/
audio_process.py
File metadata and controls
132 lines (118 loc) · 6.79 KB
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
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
from pydub import AudioSegment
from pydub.silence import split_on_silence
from pypinyin import lazy_pinyin, Style
import os
import re
def combine_audio(path1, path2):
audio1 = AudioSegment.from_wav(path1)
audio2 = AudioSegment.from_wav(path2)
audio = audio1 + audio2
audio.export(path1, format="wav")
def split_audio(path, ts):
audio = AudioSegment.from_wav(path)
audio1 = audio[:ts*1000]
audio2 = audio[ts*1000:]
init_num = path[-6:-4]
num = int(init_num) + 1
position = len(path) - 6
path2 = path[position:]
path2 = re.sub(init_num, "%02d" % num, path2, 1) # 用flags会出错
path2 = path[:position] + path2
audio1.export(path, format="wav")
audio2.export(path2, format="wav")
def gen_multispeaker_datasets(file_path, output_file_path):
speakers = os.listdir(file_path)
j = 0
for speaker in speakers:
if os.path.isdir(file_path + "/" + speaker):
files = os.listdir(file_path + "/" + speaker)
if not os.path.exists(output_file_path + "/" + speaker):
os.mkdir(output_file_path + "/" + speaker)
else:
state = input("文件已存在，请输入y/n确认是否继续")
if state == "n":
exit(-1)
label = open(r"%s\%s\labels.txt" % (output_file_path, speaker), "w")
for audio_file in files:
if audio_file.endswith(".wav"):
text = re.sub(".wav", "", audio_file)
text = re.sub("[「」~。？，…！：—、.]", "", text)
audio = AudioSegment.from_wav(file_path + "/" + speaker + "/" + audio_file)
audio.export(r"%s\%s\%06d.wav" % (output_file_path, speaker, j), format="wav")
pinyin_list = lazy_pinyin(text, style=Style.TONE3, neutral_tone_with_five=True)
for i, pinyin in enumerate(pinyin_list):
if pinyin == "n2":
pinyin_list[i] = "en2"
pinyin_str = " ".join(pinyin_list)
label.write("%06d|%s\n" % (j, pinyin_str))
j += 1
# audio_path = r"D:\clouddrivedownload\nahida"
# if not os.path.exists("%s\\nahida-voice" % os.getcwd()):
# os.mkdir("%s\\nahida-voice" % os.getcwd())
# os.mkdir("%s\\nahida-voice\\reliable" % os.getcwd())
# os.mkdir("%s\\nahida-voice\\unreliable" % os.getcwd())
# os.mkdir("%s\\nahida-voice\\unknown" % os.getcwd())
# label = open(r"D:\paddlespeech_nahida_tts3\input\nahida\labels.txt", "w")
# audio_files = os.listdir(audio_path)
# j = 0
# for audio_file in audio_files:
# if audio_file.endswith(".wav"):
# text = re.sub(".wav", "", audio_file)
# text = re.sub("[「」~。？，…！：—、.]", "", text)
# # print("%s\\%s D:\\paddlespeech_nahida\\input\\nahida\\%06d.wav" % (audio_path, audio_file, j))
# # text = re.sub("[「」~]", "", text)
# # texts = re.split("[。？，…！]", text)
# # texts = [split_text for split_text in texts if len(split_text) > 2]
# audio = AudioSegment.from_wav(audio_path + "\\" + audio_file)
# # if audio.duration_seconds > 1:
# # silence_len = 420
# # thresh = -40
# # chunks = split_on_silence(audio, min_silence_len=silence_len, silence_thresh=thresh)
# # chunks = [chunk for chunk in chunks if chunk.duration_seconds > 1]
# # if len(chunks) > len(texts):
# # silence_len = 500
# # chunks2 = split_on_silence(audio, min_silence_len=silence_len, silence_thresh=thresh)
# # chunks2 = [chunk for chunk in chunks2 if chunk.duration_seconds > 1]
# # if len(chunks2) == len(texts):
# # for i, chunk_audio in enumerate(chunks2):
# # chunk_audio.export("%s\\nahida-voice\\unreliable\\%04d_%02d.wav" % (os.getcwd(), j, i),
# # format="wav")
# # pinyin_list = lazy_pinyin(texts[i], style=Style.TONE3, neutral_tone_with_five=True)
# # pinyin_str = " ".join(pinyin_list)
# # label.write("%04d%02d|%s\n" % (j, i, pinyin_str))
# # else:
# # for i, chunk_audio in enumerate(chunks2):
# # chunk_audio.export("%s\\nahida-voice\\unknown\\%04d_%02d.wav" % (os.getcwd(), j, i),
# # format="wav")
# # for i, chunk_text in enumerate(texts):
# # pinyin_list = lazy_pinyin(chunk_text, style=Style.TONE3, neutral_tone_with_five=True)
# # pinyin_str = " ".join(pinyin_list)
# # label.write("%04d%02d|%s\n" % (j, i, pinyin_str))
# # elif len(chunks) < len(texts):
# # silence_len = 340
# # chunks2 = split_on_silence(audio, min_silence_len=silence_len, silence_thresh=thresh)
# # chunks2 = [chunk for chunk in chunks2 if chunk.duration_seconds > 1]
# # if len(chunks2) == len(texts):
# # for i, chunk_audio in enumerate(chunks2):
# # chunk_audio.export("%s\\nahida-voice\\unreliable\\%04d_%02d.wav" % (os.getcwd(), j, i),
# # format="wav")
# # pinyin_list = lazy_pinyin(texts[i], style=Style.TONE3, neutral_tone_with_five=True)
# # pinyin_str = " ".join(pinyin_list)
# # label.write("%04d%02d|%s\n" % (j, i, pinyin_str))
# # else:
# # for i, chunk_audio in enumerate(chunks2):
# # chunk_audio.export("%s\\nahida-voice\\unknown\\%04d_%02d.wav" % (os.getcwd(), j, i),
# # format="wav")
# # for i, chunk_text in enumerate(texts):
# # pinyin_list = lazy_pinyin(chunk_text, style=Style.TONE3, neutral_tone_with_five=True)
# # pinyin_str = " ".join(pinyin_list)
# # label.write("%04d%02d|%s\n" % (j, i, pinyin_str))
# # else:
# # for i, chunk_audio in enumerate(chunks):
# audio.export(r"D:\paddlespeech_nahida_tts3\input\nahida\%06d.wav" % j, format="wav")
# pinyin_list = lazy_pinyin(text, style=Style.TONE3, neutral_tone_with_five=True)
# pinyin_str = " ".join(pinyin_list)
# label.write("%06d|%s\n" % (j, pinyin_str))
# j += 1
# label.close()
gen_multispeaker_datasets(r"D:\clouddrivedownload\yuanshen_audio", r"D:\paddlespeech_nahida_tts3\input")