reference-master

Automated Music Mastering through an Optimization-Based Approach.

Can either be used through python or as webapp through Docker or Python.

Samples

A short comparison video between different mastering services:

https://youtu.be/6kvGeVtv1Cw?feature=shared

Usage

Clone Repo

git clone https://github.com/Wieland3/reference-master.git

Move into directory

cd reference-master

Docker

Run shell script to build the image

sh docker-build-tag.sh

Run docker image

docker run -p 8080:8080 reference-master:latest

Navigate to

http://0.0.0.0:8080

Python

Installation

pip install -r requirements.txt

Webapp

cd webapp
python3 app.py

Usage through Code

Put the Song you want to master in the "songs/raw" folder

Put your reference Song in the "songs/references" folder

from reference_master.mastering.master import master

Master and use loudness of reference Track

master("your-track.wav", "Bohemian Rhapsody.mp3")

Alternatively, you can manually specify the target loudness of the song to be mastered

master("your-track.wav", "Bohemian Rhapsody.mp3", -7.0)

The mastered song will be automatically saved in the "songs/mastered" folder.

Depending on how many epochs are set in the settings.yaml the process can take a couple of minutes. You can change the number of epochs and other parameters for the mastering in the settings.yaml file.

During the mastering process the current distance between the songs is printed to the console. A distance somewhere close to 1 is a hint that the mastering was successful.