Pick-up method and system based on microphone array

Abstract

The invention relates to a pick-up method and system based on a microphone array; the method comprises the steps of 1, using the microphone array to pick up and output one of multiple speech signals for speech activation detection, and judging whether a speech activation signal occurs or not; if yes, executing step 2; if not, repeating the step 1; 2, performing sound source positioning on the multiple speech signals to obtain a sound source positioning direction; 3, subjecting the speech signal of the sound source positioning direction to speech augmentation to obtain an augmented speech signal; 4, subjecting the augmented speech signal to speech wakeup detection, and judging whether speech wakeup is detected or not; if yes, executing step 5; if not, repeating the step 1; 5, allowing the microphone array to pick up and output the multiple speech signals; 6, processing the multiple speech signals as one signal that is output finally as picked speech. Speech signals can be better picked up in far field environment, particularly accurately in high-noise environment, and an excellent solution is provided for remote speech control.

Description

technical field [0001] The invention relates to the field of signal processing, in particular to a microphone array-based sound pickup method and system. Background technique [0002] Recording high-quality speech signals plays a vital role in speech analysis methods such as speech recognition. The traditional single-microphone recording method has a sharp drop in recording quality in a long-distance high-noise environment, which greatly limits the application scenarios of speech analysis methods. Therefore, applications such as voice input methods and voice search on mobile phones must ensure that the distance between the speaker and the microphone of the mobile phone is close enough. This kind of sound pickup environment is classified as near-field sound pickup. [0003] The use of microphone array recording can use multi-channel speech signal data for post-processing, thereby suppressing noise and enhancing the target speech signal. Therefore, in the far-field pickup, t...

AI Technical Summary

Problems solved by technology

[0004] However, the above-mentioned method has the following problems in the actual use process: (1) there are speakers speaking at all times, and the voice at all times does not need to be picked up. The recording key can be easily solved, but it is not easy to deal with in the far-field environment; (2) when there are multiple speakers, it is difficult to determine which one needs to be picked up

Images