Speech analysis method and device

Abstract

The invention provides a voice analytical method and apparatus. The voice analytical method comprises the steps of S1, receiving an input to-be-detected voice signal; S2, extracting the voice feature information of the to-be-detected voice signal in a framing manner, and generating an identification result of the to-be-detected voice signal according to the voice feature information and an acoustic model; S3, primarily detecting a voice end point of the to-be-detected voice signal according to the identification result and a preset silence detection algorithm; S4, calculating confidence coefficient information of the to-be-detected voice signal; S5, analyzing semantic information of the to-be-detected voice signal; and S6, obtaining a voice analytical result corresponding to the to-be-detected voice signal according to the confidence coefficient information and the semantic information. According to the embodiments of the voice analytical method and apparatus, the effectiveness of voice can be identified accurately; and meanwhile, whether the semanteme is related to the context or not can be judged through semantic analysis, so that the performance of a voice identification system is improved, a more fluent interactive process of multiple rounds of dialogues is realized, and the usage experience of users is improved.

Description

technical field [0001] The invention relates to the technical field of voice recognition, in particular to a voice analysis method and device. Background technique [0002] With the development of human-computer information interaction technology, speech recognition technology shows its importance. In speech recognition system, speech endpoint detection is one of the key technologies in speech recognition. Speech endpoint detection refers to finding the starting point and ending point of the speech part in the continuous sound signal. Whether the endpoint detection is accurate or not will directly affect the performance of the speech recognition system. If there is an error in endpoint segmentation, it will lead to missed recognition or misrecognition, which in turn can lead to inaccurate speech recognition results. [0003] At present, the traditional speech endpoint detection method is mainly to obtain the energy in time domain or frequency domain, and compare it with a...

AI Technical Summary

Problems solved by technology

[0004] However, in the process of realizing the present invention, the inventor finds that the above-mentioned method has at least the following problems: when a segment of speech is detected for speech endpoints, the accuracy of the detected speech endpoints cannot Low, affecting the performance of the speech recognition system, may be misjudged as noise, resulting in poor user experience

Images