Audio samples from "CycleGAN-Based Unpaired Speech Dereverberation"
Authors:
Hannah Muckenhirn, Aleksandr Safin, Hakan Erdogan, Félix de Chaumont Quitry,
Marco Tagliasacchi, Scott Wisdom, John R. Hershey
The audio examples on this page were randomly selected from the sets used
for human evaluation described in the paper.
The reverberant inputs were not seen during training: the Librivox samples come
from the evaluation split of the dataset, while the AMI dataset is completely
out of domain since none of the AMI examples were seen during training.