The VOiCES Corpus
The Voices Obscured in Complex Environmental Settings (VOiCES) corpus is a creative commons speech dataset targeting acoustically challenging and reverberant environments with robust labels and truth data for transcription, denoising, and speaker identification.