Speaker Normalisation in the MFCC Domain

Stephen Cox

Speaker Normalisation in the MFCC Domain

Research output: Contribution to conference › Paper

Abstract

It has been shown in several recent publications that application of vocal tract normalization (VTN) is a successful method for improving the accuracy of speaker independent recognisers. We argue that VTN can be implemented in the filterbank domain and propose a model to achieve this. We show how the model can be implemented directly in the MFCC domain, where it may be viewed as a constrained version of maximum likelihood linear regression (MLLR). The parameter estimates produced by the model are in accord with our ideas about how it should operate to perform VTN. Recognition results on a phoneme recognition task are presented which show a small improvement in accuracy.

Original language	English
Pages	853-856
Number of pages	4
Publication status	Published - Oct 2000
Event	6th International Conference on Spoken Language Processing - Beijing, China Duration: 16 Oct 2000 → 20 Oct 2000

Conference

Conference	6th International Conference on Spoken Language Processing
Abbreviated title	ICSLP 2000
Country/Territory	China
City	Beijing
Period	16/10/00 → 20/10/00

Cite this

@conference{ea69c9cfae534111a321e3f9e915cf17,

title = "Speaker Normalisation in the MFCC Domain",

abstract = "It has been shown in several recent publications that application of vocal tract normalization (VTN) is a successful method for improving the accuracy of speaker independent recognisers. We argue that VTN can be implemented in the filterbank domain and propose a model to achieve this. We show how the model can be implemented directly in the MFCC domain, where it may be viewed as a constrained version of maximum likelihood linear regression (MLLR). The parameter estimates produced by the model are in accord with our ideas about how it should operate to perform VTN. Recognition results on a phoneme recognition task are presented which show a small improvement in accuracy.",

author = "Stephen Cox",

year = "2000",

month = oct,

language = "English",

pages = "853--856",

note = "6th International Conference on Spoken Language Processing, ICSLP 2000 ; Conference date: 16-10-2000 Through 20-10-2000",

}

TY - CONF

T1 - Speaker Normalisation in the MFCC Domain

AU - Cox, Stephen

PY - 2000/10

Y1 - 2000/10

N2 - It has been shown in several recent publications that application of vocal tract normalization (VTN) is a successful method for improving the accuracy of speaker independent recognisers. We argue that VTN can be implemented in the filterbank domain and propose a model to achieve this. We show how the model can be implemented directly in the MFCC domain, where it may be viewed as a constrained version of maximum likelihood linear regression (MLLR). The parameter estimates produced by the model are in accord with our ideas about how it should operate to perform VTN. Recognition results on a phoneme recognition task are presented which show a small improvement in accuracy.

AB - It has been shown in several recent publications that application of vocal tract normalization (VTN) is a successful method for improving the accuracy of speaker independent recognisers. We argue that VTN can be implemented in the filterbank domain and propose a model to achieve this. We show how the model can be implemented directly in the MFCC domain, where it may be viewed as a constrained version of maximum likelihood linear regression (MLLR). The parameter estimates produced by the model are in accord with our ideas about how it should operate to perform VTN. Recognition results on a phoneme recognition task are presented which show a small improvement in accuracy.

M3 - Paper

SP - 853

EP - 856

T2 - 6th International Conference on Spoken Language Processing

Y2 - 16 October 2000 through 20 October 2000

ER -