A comparative study between linear and nonlinear speech prediction

TL;DR

This paper compares linear and nonlinear speech prediction methods, focusing on nonlinear prediction using neural networks, and examines neural network weight quantization and compression gain effects.

Contribution

It introduces a neural network-based nonlinear speech prediction model and analyzes the impact of weight quantization and compression gain on performance.

Findings

Neural network nonlinear prediction captures speech features better.

Quantization affects the accuracy of speech prediction.

Compression gain influences the efficiency of the model.

Abstract

This paper is focused on nonlinear prediction coding, which consists on the prediction of a speech sample based on a nonlinear combination of previous samples. It is known that in the generation of the glottal pulse, the wave equation does not behave linearly [2], [10], and we model these effects by means of a nonlinear prediction of speech based on a parametric neural network model. This work is centred on the neural net weight's quantization and on the compression gain.