Skip to main content

Bandwidth Extension is All You Need

Author(s): Su, Jiaqi; Wang, Yunyun; Finkelstein, Adam; Jin, Zeyu

Download
To refer to this page use: http://arks.princeton.edu/ark:/88435/pr12f7jr1j
Full metadata record
DC FieldValueLanguage
dc.contributor.authorSu, Jiaqi-
dc.contributor.authorWang, Yunyun-
dc.contributor.authorFinkelstein, Adam-
dc.contributor.authorJin, Zeyu-
dc.date.accessioned2023-12-23T22:24:32Z-
dc.date.available2023-12-23T22:24:32Z-
dc.date.issued2021en_US
dc.identifier.citationSu, Jiaqi, Yunyun Wang, Adam Finkelstein, and Zeyu Jin. "Bandwidth extension is all you need." In ICASSP 2021-2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 696-700. 2021. doi:10.1109/ICASSP39728.2021.9413575en_US
dc.identifier.issn1520-6149-
dc.identifier.urihttps://gfx.cs.princeton.edu/pubs/Su_2021_BEI/ICASSP2021_Su_Wang_BWE.pdf-
dc.identifier.urihttp://arks.princeton.edu/ark:/88435/pr12f7jr1j-
dc.description.abstractSpeech generation and enhancement have seen recent breakthroughs in quality thanks to deep learning. These methods typically operate at a limited sampling rate of 16-22kHz due to computational complexity and available datasets. This limitation imposes a gap between the output of such methods and that of high-fidelity (≥44kHz) real-world audio applications. This paper proposes a new bandwidth extension (BWE) method that expands 8-16kHz speech signals to 48kHz. The method is based on a feed-forward WaveNet architecture trained with a GAN-based deep feature loss. A mean-opinion-score (MOS) experiment shows significant improvement in quality over state-of-the-art BWE methods. An AB test reveals that our 16-to-48kHz BWE is able to achieve fidelity that is typically indistinguishable from real high-fidelity recordings. We use our method to enhance the output of recent speech generation and denoising methods, and experiments demonstrate significant improvement in sound quality over these baselines. We propose this as a general approach to narrow the gap between generated speech and recorded speech, without the need to adapt such methods to higher sampling rates.en_US
dc.format.extent696 - 700en_US
dc.language.isoen_USen_US
dc.relation.ispartofICASSP 2021-2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)en_US
dc.rightsAuthor's manuscripten_US
dc.titleBandwidth Extension is All You Needen_US
dc.typeConference Articleen_US
dc.identifier.doi10.1109/ICASSP39728.2021.9413575-
dc.identifier.eissn2379-190X-
pu.type.symplectichttp://www.symplectic.co.uk/publications/atom-terms/1.0/conference-proceedingen_US

Files in This Item:
File Description SizeFormat 
BandwidthExtension.pdf1.38 MBAdobe PDFView/Download


Items in OAR@Princeton are protected by copyright, with all rights reserved, unless otherwise indicated.