A formant vocoder based on mixtures of Gaussians

Author

Zolfaghari, Parham ; Robinson, Tony

Author_Institution

Dept. of Eng., Cambridge Univ., UK

Volume

2

fYear

1997

fDate

21-24 Apr 1997

Firstpage

1575

Abstract

This paper describes a new low bit-rate formant vocoder. The formant parameters are represented by Gaussian mixture distributions, which are estimated from the discrete Fourier transform (DFT) magnitude spectrum of the speech signal. A voiced/unvoiced classification mechanism has been developed based on the harmonic nature of each formant in the DFT spectrum modulated by the Gaussian mixture distribution. Using a magnitude-only sinusoidal synthesiser, intelligible synthetic speech has been obtained. Vector quantisation of the vocal tract parameters enables this formant vocoder to operate at a bit-rate of 1248 bps

Keywords

Gaussian distribution; discrete Fourier transforms; spectral analysis; speech coding; speech intelligibility; speech synthesis; vector quantisation; vocoders; 1248 bit/s; DFT magnitude spectrum; DFT spectrum modulation; Gaussian mixture distributions; discrete Fourier transform; formant parameters; formant vocoder; intelligible synthetic speech; low bit rate formant vocoder; magnitude only sinusoidal synthesiser; speech signal; vector quantisation; vocal tract parameters; voiced/unvoiced classification; Data mining; Discrete Fourier transforms; Gaussian distribution; Gaussian processes; Resonance; Signal synthesis; Speech analysis; Speech synthesis; Vector quantization; Vocoders;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 1997. ICASSP-97., 1997 IEEE International Conference on

Conference_Location

Munich

ISSN

1520-6149

Print_ISBN

0-8186-7919-0

Type

conf

DOI

10.1109/ICASSP.1997.596253

Filename

596253