Speech enhancement method and device, equipment and storage medium

Abstract

The invention relates to artificial intelligence, and provides a speech enhancement method and device, equipment and a storage medium. According to the method, the pure audio can be expanded to obtain audio samples, the audio samples comprise audio with noise, the audio with noise is preprocessed to obtain a plurality of sequence features, each sequence feature is analyzed based on a time sequence processing network to obtain a plurality of output features, and time frequency features are generated according to the sequence features and the output features, frequency band information is extracted from the time-frequency features, the frequency band information is analyzed based on a frequency band processing network to obtain frequency band features, predicted audio is generated according to the frequency band features and the time-frequency features, network parameters are adjusted based on the predicted audio and pure audio to obtain an audio enhancement model, request audio is obtained, and enhancement processing is performed on the request audio based on the audio enhancement model, and the target audio is obtained. The enhancement effect of the target audio can be improved. In addition, the invention also relates to a block chain technology, and the target audio can be stored in a block chain.

Description

technical field [0001] The present invention relates to the technical field of artificial intelligence, in particular to a speech enhancement method, device, equipment and storage medium. Background technique [0002] Speech enhancement involves extracting target speech sources from reverberant and noisy speech environments. When performing speech enhancement on long-sequence audio, the traditional audio enhancement method will have the problem of gradient disappearance and long-term dependence when dealing with long-term information, making it impossible to effectively model long speech sequences. At the same time, in When the receptive field is smaller than the sequence length, it cannot perform sequence modeling at the utterance level, so it has a certain impact on the modeling accuracy of the actual speech sequence, thereby affecting the effect of speech enhancement. Contents of the invention [0003] In view of the above, it is necessary to provide a speech enhanceme...

AI Technical Summary

Problems solved by technology

When performing speech enhancement on audio with a long time sequence, the traditional audio enhancement method will have the problem of gradient disappearance and long-term dependence when dealing with long-term information, making it impossible to effectively model long speech sequences. At the same time, in When the receptive field is smaller than the sequence length, it is impossible to perform sequence modeling at the speech level, so it has a certain impact on the modeling accuracy of the actual speech sequence, thus affecting the effect of speech enhancement

Images