Audio signal processing method, apparatus, device, and storage medium

Abstract

The present disclosure relates to an audio signal processing method, device, equipment and storage medium. The present disclosure obtains local text content from a speech recognition result, and labels the local text content. The labeling result of the local text content is used to indicate whether the local text content is a dialogue. According to the speech recognition result, a plurality of audio segments are obtained from an audio signal, and timbre information of each audio segment is extracted. Further, according to the labeling result of the local text content and the timbre information of the plurality of audio segments respectively, a speech recognition result containing a speaker label is generated. Since the present embodiment combines information of a text mode, such as the labeling result of the local text content, and information of an audio mode, such as the timbre information of the plurality of audio segments respectively, compared with the prior art which only uses speech single-modal information, the speech recognition result containing the speaker label generated by the present embodiment in combination with multi-modal information is more accurate.

Images