Pickup method of microphone array, processor and storage medium thereof

Abstract

The invention provides a pickup method of a microphone array, which is applied to long-distance voice recognition. The pickup method of the microphone array sequentially performs delay compensation, gain compensation, beam forming, post-filtering and gain control on a target voice received in the microphone array, wherein through delay compensation, the time point for the microphone array to receive the target voice is ensured to be consistent; through gain compensation, signal amplitude difference caused by inconsistency of microphones is reduced; through beam forming, coherent noises in a target voice signal are eliminated; through post-filtering, diffusion noises in the target voice signal are eliminated; through gain control, a part with relatively small remote recording volume is amplified, so that recognition by a recognition engine is facilitated. By the pickup method provided by the invention, the recognition rate of the voice signal can be significantly increased. The invention further provides a processor capable of executing the pickup method and a storage medium thereof.

Description

technical field [0001] The invention relates to the technical field of audio signal processing, in particular to a microphone array sound pickup method for long-distance speech recognition, a processor and a storage medium thereof. Background technique [0002] With the increasing maturity of speech recognition technology, its application is becoming more and more extensive. The existing speech recognition technology performs well at short distances. However, when the speech distance becomes farther, the speech recognition rate drops sharply, making speech recognition difficult to achieve practical results. [0003] In order to solve the problem of long-distance speech recognition rate decline, microphone arrays are often used in the prior art, that is, certain beamforming and adaptive filtering methods are used to eliminate noise, but the general beamforming method for noise elimination is not ideal. , and in the process of actual use, due to the differences in the characte...

AI Technical Summary

Problems solved by technology

[0002] With the increasing maturity of speech recognition technology, its application is becoming more and more extensive. The existing speech recognition technology performs well at short distances. However, when the speech distance becomes farther, the speech recognition rate drops sharply, making speech recognition difficult to achieve practical results.

[0003] In order to solve the problem of long-distance speech recognition rate decline, microphone arrays are often used in the prior art, that is, certain beamforming and adaptive filtering methods are used to eliminate noise, but the effect of general beamforming methods for noise elimination is not ideal. , and in the process of actual use, due to the differences in the characteristics of different microphones, the deviation of the microphone placement position, and the inaccurate direction of the target voice, the effect of processing noise effects is further reduced

Images