REVERB Challenge

The REVERB Challenge is a dataset used to evaluate speech dereverberation methods in distant-microphone scenarios, containing recordings that simulate real-world reverberation and noise conditions.

Papers using REVERB Challenge (3)

Deep Filter Estimation from Inter-Frame Correlations for Monaural Speech Dereverberation2026Magnitude or Phase? A Two Stage Algorithm for Dereverberation2022Speech enhancement with frequency domain auto-regressive modeling2023