Research Repository

The robustness of speech representations obtained from simulated auditory nerve fibers under different noise conditions

Ju?rgens, T and Brand, T and Clark, NR and Meddis, R and Brown, GJ (2013) 'The robustness of speech representations obtained from simulated auditory nerve fibers under different noise conditions.' The Journal of the Acoustical Society of America, 134 (3). creators - Meddis=3ARay=3A=3A. ISSN 0001-4966

[img]
Preview
Text
1.4817912.pdf - Published Version

Download (459kB) | Preview

Abstract

Different methods of extracting speech features from an auditory model were systematically investigated in terms of their robustness to different noises. The methods either computed the average firing rate within frequency channels (spectral features) or inter-spike-intervals (timing features) from the simulated auditory nerve response. When used as the front-end for an automatic speech recognizer, timing features outperformed spectral features in Gaussian noise. However, this advantage was lost in babble, because timing features extracted the spectro-temporal structure of babble noise, which is similar to the target speaker. This suggests that different feature extraction methods are optimal depending on the background noise.

Item Type: Article
Subjects: B Philosophy. Psychology. Religion > BF Psychology
R Medicine > R Medicine (General)
Divisions: Faculty of Science and Health > Psychology, Department of
Depositing User: Jim Jamieson
Date Deposited: 14 Aug 2014 08:27
Last Modified: 17 Aug 2017 17:49
URI: http://repository.essex.ac.uk/id/eprint/9961

Actions (login required)

View Item View Item