Nonnegative matrix decomposition method for speech signal characteristic waveform

Abstract

The present invention relates to a nonnegative matrix decomposition method of speech signal characteristic waveform, belonging to speech signal processing technology. Said method includes the following several steps: firstly, utilizing fundamental tone pitch of speech signal to divide the speech characteristic waveform into 9 classes, for every class of characteristic waveform utilizing iteration method of standard nonnegative matrix decomposition to train out base matrix W, then for given a frame characteristic waveform utilizing its fundamental tone pitch to make subsumption, then taking out the trained base matrix W correspondent to said class of characteristic waveform, and utilizing iteration method to obtain code matrix H correspondent to said frame characteristic waveform, so that said frame characteristic waveform can be approximately decomposed into the product of base matrix W and code matrix H.

Description

technical field [0001] The invention relates to a method for non-negative matrix decomposition of characteristic waveforms of speech signals, which belongs to the field of speech signal processing. Background technique [0002] With the rapid development of wireless mobile communication, secure telephone communication and network VoIP communication system, people's demand for high-quality speech coding technology below 4kb / s is increasing day by day. At present, the international speech coding models for rates below 4kb / s mainly include binary excitation linear prediction model (LPC-10-Linear Prediction Coding-10), mixed excitation linear prediction model (MELP-Mixed Excitation Linear Prediction), multi-band excitation model (MBE-Muliti-BandExcitation) and waveform interpolation model (WI-Waveform Interpolation). These models are all based on the source-system model of the speech signal, that is, the excitation source signal (simulating the airflow from the lungs) is used t...

AI Technical Summary

Problems solved by technology

[0007] In order to solve the above problems, the present invention provides a non-negative matrix decomposition method of the characteristic waveform of the speech signal. The problem to be solved by the method is to achieve high precision and low complexity when decomposing the characteristic waveform of the speech signal in the waveform interpolation speech encoder degree and without additional delay for the purpose of eigenwave decomposition

Images