5th European Conference on Speech Communication and Technology

Rhodes, Greece
September 22-25, 1997

Wideband Coding of Speech Using Neural Network Gain Adaptation

Cheung-Fat Chan, Man-Tak Chu

Department of Electronic Engineering, City University of Hong Kong Kowloon, Hong Kong

In this paper, a high-quality wideband speech coder is proposed. The coding structure resembles a LD-CELP coder, however, several novel improvements are made. The gain adapter for the stochastic codebook is driven by a neural network and it updates the excitation gain in a sample-by-sample fashion. The purpose of incorporating a neural network is to exploit both the intra- and inter-frame correlation of speech signal in a non-linear manner. A psychoacoustic model instead of a simple perceptual weighting filter is used to shape the quantization noise. Simulation result shows that the proposed coder can achieve transparent coding of wideband speech at 16 kbps.

Full Paper

Bibliographic reference.  Chan, Cheung-Fat / Chu, Man-Tak (1997): "Wideband coding of speech using neural network gain adaptation", In EUROSPEECH-1997, 1507-1510.