Voice processing method and apparatus, terminal device and storage medium

Abstract

The invention discloses a voice processing method and apparatus, a terminal device and a storage medium. The method comprises the steps of: obtaining noise-containing audio data, wherein the noise-containing audio data comprise a voice source signal; preprocessing the noise-containing audio data, extracting the noise-containing audio features from the noise-containing audio data, and inputting thenoise-containing audio features into a pre-trained voice processing network model to obtain de-noised audio features, wherein the pre-trained voice processing network model comprises multiple cause and effect convolutional layers and at least one recurrent neural network layer, the multiple cause and effect convolutional layers are configured to output texture features of the corresponding voicesource signal according to the noise-containing audio features, and the at least one recurrent neural network layer is configured to output the de-noised audio features according to the texture features; and obtaining an estimated value of the voice source signal according to the de-noised audio features, and outputting the estimated value as de-noised noise-containing audio data. According to thevoice processing method and apparatus disclosed by the invention, real-time noise reduction of the noise-containing audio data is realized by the cause and effect convolutional layers and the recurrent neural network layer, and the voice noise reduction effect is improved.

Description

technical field [0001] The present application relates to the technical field of audio data processing, and more specifically, to a voice processing method, device, terminal equipment, and storage medium. Background technique [0002] With the rapid development of communication technology and smart terminals, people's entertainment, social activities and other behaviors based on smart terminals are more and more free from geographical and time constraints. Not only can live broadcast, calls, chats, etc. Functional activities can also realize various functions outdoors. However, depending on the environment, the smart terminal often receives various background noises while receiving the voice input by the user, which affects the recognizability of the voice input by the user, and is not conducive to the user's various voice activities. Therefore, how to effectively suppress background noise when the user performs voice activities needs to be solved urgently. Contents of th...

AI Technical Summary

Problems solved by technology

However, depending on the environment, the smart terminal often receives various background noises while receiving the voice input by the user, which affects the recognizability of the voice input by the user, and is not conducive to the user's various voice activities. Therefore, how to effectively suppress background noise when users perform voice activities needs to be solved urgently.

Images