WSJ-0

WSJ-0 is a dataset containing clean speech clips used to simulate speech signals of moving speakers in reverberant environments for evaluating multichannel speech enhancement algorithms.

Papers using WSJ-0 (3)

Coarse-to-Fine Recursive Speech Separation for Unknown Number of Speakers2022 · 2 citesGLD-Net: Improving Monaural Speech Enhancement by Learning Global and Local Dependency Features with GLD Block2022Attention-Driven Multichannel Speech Enhancement in Moving Sound Source Scenarios2023