Sin2pi commited on
Commit
de1426a
1 Parent(s): f11512f

Upload remove_silence_from_samples.ipynb

Browse files
Files changed (1) hide show
  1. remove_silence_from_samples.ipynb +66 -0
remove_silence_from_samples.ipynb ADDED
@@ -0,0 +1,66 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "cells": [
3
+ {
4
+ "cell_type": "code",
5
+ "execution_count": null,
6
+ "metadata": {},
7
+ "outputs": [],
8
+ "source": [
9
+ "from pydub import AudioSegment\n",
10
+ "import os, librosa, soundfile as sf\n",
11
+ "from pathlib import Path\n",
12
+ "from tqdm import tqdm\n",
13
+ "\n",
14
+ "top_db=20 #in decibels\n",
15
+ "threshold=-20.0 #in decibels\n",
16
+ "\n",
17
+ "def milliseconds_until_sound(sound, threshold, chunk_size=10):\n",
18
+ " trim_ms = 0 # ms\n",
19
+ " assert chunk_size > 0 # to avoid infinite loop\n",
20
+ " while sound[trim_ms:trim_ms+chunk_size].dBFS < threshold and trim_ms < len(sound):\n",
21
+ " trim_ms += chunk_size\n",
22
+ " return trim_ms\n",
23
+ "\n",
24
+ "def remove_silence(input_file, output_file, top_db=top_db):#, target_sr=None):\n",
25
+ " sr = 16000\n",
26
+ " y, sr = librosa.load(input_file)\n",
27
+ " # if target_sr and target_sr != sr:\n",
28
+ " # y = librosa.resample(y, orig_sr=sr, target_sr=target_sr)\n",
29
+ " # sr = target_sr\n",
30
+ " intervals = librosa.effects.split(y, top_db=top_db)\n",
31
+ " y_trimmed = []\n",
32
+ " for start, end in intervals:\n",
33
+ " y_trimmed.extend(y[start:end])\n",
34
+ " sf.write(output_file, y_trimmed, sr)\n",
35
+ "\n",
36
+ "def process_directory(input_dir, output_dir, top_db=top_db):#, target_sr=None):\n",
37
+ " if not os.path.exists(output_dir):\n",
38
+ " os.makedirs(output_dir)\n",
39
+ "\n",
40
+ " for filename in os.listdir(input_dir):\n",
41
+ " if filename.endswith(\".mp3\"):\n",
42
+ " input_file = os.path.join(input_dir, filename)\n",
43
+ " output_file = os.path.join(output_dir, filename)\n",
44
+ " audio = AudioSegment.from_mp3(input_file)\n",
45
+ " # if audio.dBFS < -30:\n",
46
+ " start_trim = milliseconds_until_sound(audio)\n",
47
+ " if start_trim < len(audio):\n",
48
+ " remove_silence(input_file, output_file, top_db)#, target_sr)\n",
49
+ " # else:\n",
50
+ " # (print(\"file removed\"))\n",
51
+ "\n",
52
+ "if __name__ == \"__main__\":\n",
53
+ " input_dir = folder_path\n",
54
+ " output_dir = folder_path + \"_trim/\"\n",
55
+ " process_directory(input_dir, output_dir)#, target_sr=16000)"
56
+ ]
57
+ }
58
+ ],
59
+ "metadata": {
60
+ "language_info": {
61
+ "name": "python"
62
+ }
63
+ },
64
+ "nbformat": 4,
65
+ "nbformat_minor": 2
66
+ }