Multichannel Singing Voice Separation
by Deep Neural Network
Informed DOA Constrained CNMF
This page contains separation results for article "Multichannel Singing Voice Separation by Deep Neural Network Informed DOA Constrained CNMF" submitted to the IEEE 22nd Workshop on Multimedia Signal Processing.
Authors: Antonio J. Munoz-Montoro, Archontis Politis, Konstantinos Drossos and Julio J. Carabias-Orti
Datasets
The
experimental evaluation was carried out on the Demixing Secret Dataset
(DSD100). A subset of 50 mixtures has been used for the training
process together with their corresponding individual sources, and
another 50 mixtures for the evaluation.
Small Array Results
Large Array Results