Split Audio Files by Silence

This script splits audio files based on the length of silence detected in them.

Audio files are split into chunks wherever there is silence for a period longer than the specified minimum silence length.

The bat file is optimized for DDSP-SVC. Rewrite it as necessary.

Dependencies

CUI and GUI

1. At first, install ffmpeg from the official website.
2. Register the ffmpeg bin directory in the environment variable.
3. Install what you need using pip:

pip install -r requirements.txt

Usage

Parameters

Non-required arguments are set to 0.

File Setting

Argument	Description	Default
--input	Input directory path	.\input
--output	Output directory path	.\output
--filename	Output file name	audio
--iformat	Input file format (wav, mp3, etc.)	wav
--oformat	Output file format (wav, mp3, etc.)	wav

Sound Setting

When LPF(lowpass) and HPF(highpass) are used together, they function as BPF(bandpass) and BSF(bandstop).

Argument	Description	Default
--silence	Minimum silence length (ms)	500
--threshold	Silence threshold (dBFS)	-40
--samplerate	Sample rate (Hz)	0
--invert	Invert phase audio(0: false, 1: true)	0
--loudness	Loudness normalization(dBFS)	0.0
--channel	Number of audio channels(0: pass, 1: mono, 2: stereo)	0
--lowpass	Low pass filter(Hz)	0
--highpass	High pass filter(Hz)	0
--reverse	Reverse audio(0: false, 1: true)	0

Output Setting

Argument	Description	Default
--combine	Combine all output files(0: false, 1: true)	0
--skip	Ignore files with less than the specified seconds (ms)	0
--merge	Merge files under a certain number of seconds(ms)	0
--split	Split files longer than a specified number of seconds(ms)	0
--prefix	Prefix of output file name	None

Example

example:

python sound.py --filename "inv_rev" --skip 2 --samplerate 44100 --invert 1 --reverse 1 --loudness -14 --pack 1 --highpass 80 --prefix --merge 5000 --split 10000 "rev_inv_hps"

Thanks

I was inspired to create this site.