This is the code repository for the ISMIR 2022 paper Music-STAR: a Style Translation system for Audio-based Re-instrumentation by Mahshid Alinoori and Vassilios Tzerpos.
- The audio samples and sumplementary material can be found here.
- The trained models can be found here.
- The StarNet dataset used in the paper can be found here.
- The fork of Demucs project used for musical source separation in the paper can be found here.
- The fork of Universal Music Translation used for finetuning the single-instrument translation model in the paper can be found here.