U-DREAM: Unsupervised Dereverberation guided by a Reverberation Model
By: Louis Bahrman, Mathieu Fontaine, Gaël Richard
Potential Business Impact:
Cleans up echoey sounds without needing perfect recordings.
This paper explores the outcome of training state-ofthe-art dereverberation models with supervision settings ranging from weakly-supervised to fully unsupervised, relying solely on reverberant signals and an acoustic model for training. Most of the existing deep learning approaches typically require paired dry and reverberant data, which are difficult to obtain in practice. We develop instead a sequential learning strategy motivated by a bayesian formulation of the dereverberation problem, wherein acoustic parameters and dry signals are estimated from reverberant inputs using deep neural networks, guided by a reverberation matching loss. Our most data-efficient variant requires only 100 reverberation-parameter-labelled samples to outperform an unsupervised baseline, demonstrating the effectiveness and practicality of the proposed method in low-resource scenarios.
Similar Papers
Déréverbération non-supervisée de la parole par modèle hybride
Sound
Cleans up echoey voices without needing perfect recordings.
Dereverberation Using Binary Residual Masking with Time-Domain Consistency
Sound
Cleans up echo in voices for clearer sound.
Is Phase Really Needed for Weakly-Supervised Dereverberation ?
Sound
Cleans up echo in recordings without knowing original sound.