LibriMix: An open-source dataset for generalizable speech separation

Joris Cosentino (1) , Manuel Pariente (1) , Samuele Cornell (2) , Antoine Deleforge (1) , Emmanuel Vincent (1)

1 MULTISPEECH - Speech Modeling for Facilitating Oral-Based Communication
2 UNIVPM - Polytechnic University of Marche [Ancona, Italy] / Università Politecnica delle Marche [Ancona, Italia]

In recent years, wsj0-2mix has become the reference dataset for single-channel speech separation. Most deep learning-based speech separation models today are benchmarked on it. However, recent studies have shown important performance drops when models trained on wsj0-2mix are evaluated on other, similar datasets. To address this generalization issue, we created LibriMix, an open-source alternative to wsj0-2mix, and to its noisy extension, WHAM!. Based on LibriSpeech, LibriMix consists of two-or three-speaker mixtures combined with ambient noise samples from WHAM!. Using Conv-TasNet, we achieve competitive performance on all LibriMix versions. In order to fairly evaluate across datasets, we introduce a third test set based on VCTK for speech and WHAM! for noise. Our experiments show that the generalization error is smaller for models trained with LibriMix than with WHAM!, in both clean and noisy conditions. Aiming towards evaluation in more realistic, conversation-like scenarios, we also release a sparsely overlapping version of LibriMix's test set.

Mots clés: speech separation; generalization; corpora

Domaines: Traitement du signal et de l'image [eess.SP]

Joris Cosentino, Manuel Pariente, Samuele Cornell, Antoine Deleforge, Emmanuel Vincent. LibriMix: An open-source dataset for generalizable speech separation. 2020. ⟨hal-03354695⟩