Audio Editor

Automatically cuts silence, splits audio into phrases, allows you to remove failed takes, and applies minimal professional audio processing.

Features

• Silence Detection — automatically finds long pauses between phrases
• Speech Recognition — displays the text of each phrase (offline, via Vosk)
• Phrase Splitting — divides the recording into separate meaningful fragments
• Selective Deletion — interactive selection of failed takes for removal
• Compressor — evens out volume for better sound
• Gain — automatically raises volume to maximum without clipping
• Smart Trimming — trims phrase edges at zero crossings for silent junctions
• Export — saves the result in stereo WAV 48kHz 16-bit

Requirements

• Python 3.8 or newer
• pip (Python package manager)
• Vosk speech recognition model for Russian language

Quick Start

1. Clone or Download the Script

git clone https://github.com/ncojam/audio_editor.git
cd audio-phrase-editor

2. Download the Speech Recognition Model

Download one of the language models from the official Vosk website

Extract the archive into the "model" folder next to the script.

3. Create a Virtual Environment

Windows:

python -m venv venv
venv\Scripts\activate

Linux:

python3 -m venv venv
source venv/bin/activate

4. Run the Script

Install dependencies, place a .wav file next to the script, and run it:

pip install -r requirements.txt
python audio_editor.py