In live multitrack recordings, each instrument is usually captured by dedicated close microphones. Unfortunately, it is also captured by other microphones intended for other sources, leading to so-called interferences, aka cross-talk or microphone leakage.
Reducing this interference is desirable because it opens new perspectives for the engineering of live recordings.
Interference Reduction in Python
MIRAPIE is a python algorithm developed in MULTISPEECH (a joint project-team between INRIA and LORIA) under the supervision of Antoine Liuktus and it the context of the D2YC2 project.
It is based on on the expression of interference reduction as source separation problem within the Gaussian Source Separation framework.
The code is available on github.
- Interference reduction on full-length live recordings, D. Di Carlo, A. Liutkus, and K. Déguernel, in IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2018.
- Gaussian framework for interference reduction in live recordings, D. Di Carlo, Diego, K. Déguernel and A. Liutkus, in AES International Conference on Semantic Audio, 2017
- Gaussian Framework for Interference Reduction in Live Recordings, D. Di Carlo, Diego (supervised by N. Orio, A. Liutkus), Universitá degli Studi di Padova, 2017
The Mixing Matrix