____   ____ ____  ______
|    \ /    |    \|      |
|  o  )  o  |  D  )      |
|     |     |    /|_|  |_|
|  O  |  _  |    \  |  |
|     |  |  |  .  \ |  |
|_____|__|__|__|\_| |__|

"Are you transcribing a conversation? BART can help!"

BART (Beyond Audio Replay Technology) aids transcribe tasks by taking a source audio file and creating automatic repeated loops, allowing transcribers to listen to fragments multiple times (with possible overlap between segments).

Installation

Make sure to clone this repository and install Python dependencies using:

pip install -r requirements.txt

BART relies on Pydub for processing audio. Pydub needs ffmepg or libav to be available on your system. Read this carefully.

Installing ffmpeg on MacOS

On my macOS system, I did the following.

  1. Download ffmpeg and ffprobe binaries from: https://evermeet.cx/ffmpeg/
  2. Unarchive downloaded files and move them to your $PATH, e.g.:

sudo cp ffmpeg /usr/local/bin
sudo cp ffprobe /usr/local/bin
  1. Double check they are in your $PATH:
which ffmpeg || echo "ffmpeg not found"
which ffprobe || echo "ffprobe not found"

Usage

BART is very simple and expects all input files in the input/ directory.

Overmore, BART is very naive. It always writes output files to the output/ directory with exactly the same name as the input file. If the file exists, it will overwrite without warning.

BART is very chatty and will tell you when he adds a segment and when he adds a pause.

Example

python main.py --input test.mp3 --segment-length 10 --pause-length 1.0 --segment-step 5 --segment-repeat 2

Usage:

python main.py -h

GitHub

View Github