Please use this identifier to cite or link to this item:
http://bura.brunel.ac.uk/handle/2438/30136
Title: | Spectral network based on lattice convolution and adversarial training for noise-robust speech super-resolution |
Authors: | Yang, J Liu, H Gan, L Jing, X |
Keywords: | speech communication;acoustic noise;speech processing systems;speech recognition;spectrograms;electronic circuits;data analysis;artificial neural networks;signal processing;Fourier analysis |
Issue Date: | 12-Nov-2024 |
Publisher: | Acoustical Society of America |
Citation: | Yang, J. et al. (2024) 'Spectral network based on lattice convolution and adversarial training for noise-robust speech super-resolution', Journal of the Acoustical Society of America, 156 (5), pp. 3143 - 3157. doi: 10.1121/10.0034364. |
Abstract: | Speech super-resolution aims to predict a high-resolution speech signal from its low-resolution counterpart. The previous models usually perform this task at a fixed sampling rate, reconstructing only high-frequency spectrogram components and merging them with low-frequency ones in noise-free cases. These methods achieve high accuracy, but they are less effective in real-world settings, where ambient noise and flexible sampling rates are presented. To develop a robust model that fits practical applications, in this work, we introduce Super Denoise Net (SDNet), a neural network for noise-robust super-resolution with flexible input sampling rates. To this end, SDNet's design includes gated and lattice convolution blocks for enhanced repair and temporal-spectral information capture. The frequency transform blocks are employed to model long frequency dependencies, and a multi-scale discriminator is proposed to facilitate the multi-adversarial loss training. The experiments show that SDNet outperforms current state-of-the-art noise-robust speech super-resolution models on multiple test sets, indicating its robustness and effectiveness in real-world scenarios. |
Description: | Data Availability: Data are available on request to the authors. |
URI: | https://bura.brunel.ac.uk/handle/2438/30136 |
DOI: | https://doi.org/10.1121/10.0034364 |
ISSN: | 0001-4966 |
Other Identifiers: | ORCiD: Hongqing Liu https://orcid.org/0000-0002-2069-0390 ORCiD: Lu Gan https://orcid.org/0000-0003-1056-7660 |
Appears in Collections: | Dept of Electronic and Electrical Engineering Embargoed Research Papers |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
FullText.pdf | Copyright © 2024 Acoustical Society of America. This article may be downloaded for personal use only. Any other use requires prior permission of the author and the Acoustical Society of America (see: https://acousticalsociety.org/web-posting-guidelines/ ). The following article appeared in Junkang Yang, Hongqing Liu, Lu Gan, Xiaorong Jing; Spectral network based on lattice convolution and adversarial training for noise-robust speech super-resolution. J. Acoust. Soc. Am. 1 November 2024; 156 (5): 3143–3157. https://doi.org/10.1121/10.0034364 and may be found at https://pubs.aip.org/asa/jasa/article/156/5/3143/3320008/Spectral-network-based-on-lattice-convolution-and . | 5.27 MB | Adobe PDF | View/Open |
Items in BURA are protected by copyright, with all rights reserved, unless otherwise indicated.