Combining features via LDA in speaker recognition

Z.P. Sun & J.S. Mason Department of Electrical & Electronic Engineering, University College of Wales, SWANSEA, SA2 8PP, UK email: [email protected], [email protected] ABSTRACT This paper1 discusses cepstral feature combinations via linear discriminant analysis (LDA) in the context of automatic speaker identi cation (ASI). Two static cepstral features are considered, namely standard MFCC and a sub-band ltered form derived via linear prediction known as RASTA-PLP. These two are compared along with their rst order dynamic forms as both single and combined feature sets. LDA is shown to provide a useful means of combining (dissimilar) feature sets and permitting a direct trade-o between the number of coe cients and ASI performance, particularly when testing under noisy conditions. Also, the importance of pre-normalisation is demonstrated. It is shown that in the case of individual features, the two static forms give the best performance in clean conditions, with a cross-over to the two dynamic forms being better in the region of SNR=15dB. The LDA combination of the two static forms gives the best overall results under both clean and noisy conditions.