Enhanced samples from PERL-AE model (https://arxiv.org/abs/2010.11860) Samples are chosen from VCTK validation dataset for demonstration purposes only.
Bibliography for the ICASSP 2021 paper:
@article{kataria2020perceptual,
title={Perceptual Loss based Speech Denoising with an ensemble of Audio Pattern Recognition and Self-Supervised Models},
author={Kataria, Saurabh and Villalba, Jes{'u}s and Dehak, Najim},
journal={arXiv preprint arXiv:2010.11860},
year={2020}
}