Close menu


Sunderland Repository records the research produced by the University of Sunderland including practice-based research and theses.

Augmenting Telephony Audio Data using Robust Principal Component Analysis

Mo, Ronald K. and Lam, Albert Y.S. (2020) Augmenting Telephony Audio Data using Robust Principal Component Analysis. In: IEEE Symposium Series on Computational Intelligence, Dec 1, 2020 - Dec 4, 2020, Canberra, ACT, Australia.

Item Type: Conference or Workshop Item (Paper)


Audio augmentation (e.g., corrupting audio data by noise) has been shown to improve the performance of Automatic Speech Recognition (ASR) systems for low-resource languages. In light of this, we are interested in understanding whether corrupting speech data with telephone channel characteristics (e.g., background music, artifact caused by down-sampling) improves the performance of ASR systems as well. In this work, we investigate the possibility of applying Sound Source Separation (SSS) approaches to capture the telephone channel characteristics. We are in particular interested in Robust Principal Component Analysis (RPCA), which is an unsupervised approach used for various SSS tasks. Our results show that augmenting clean speech data corpus with telephone channel characteristics yields a more robust ASR system, with 7.8% of Word Error Rate reduction. We also find that the characteristic, which has the lowest spectral features, improves ASR the most.

Full text not available from this repository.

More Information

Depositing User: Ronald Mo


Item ID: 15847
Identification Number:
Official URL:

Users with ORCIDS

ORCID for Ronald K. Mo: ORCID iD

Catalogue record

Date Deposited: 22 Mar 2023 16:02
Last Modified: 22 Mar 2023 16:02


Author: Ronald K. Mo ORCID iD
Author: Albert Y.S. Lam

University Divisions

Faculty of Technology > School of Computer Science


Computing > Artificial Intelligence

Actions (login required)

View Item (Repository Staff Only) View Item (Repository Staff Only)