Voice processing method and device

Abstract

The invention discloses a voice processing method and device, and the method comprises the steps: responding to an obtained audio, carrying out the voice activity detection of the audio, judging whether there is a voice signal in the audio or not, if the voice signal exists in the audio, judging whether the equipment is awakened or not, if the audio is not awakened, performing beam forming and post-processing filtering processing on the audio, and judging whether the processed audio contains an awakening word or not, and if the wake-up word is included, calculating a direction of arrival basedon beam forming of the processed audio, and performing beam forming based on the direction of arrival. Whether a voice signal exists in an audio or not is judged by performing voice activity detection on the audio, and beam forming and post-processing filtering processing are performed on the audio with the voice signal, so that unsteady-state noise can be effectively processed, and low-power-consumption garbage classification voice processing is realized.

Description

technical field [0001] The invention belongs to the technical field of voice processing, in particular to a voice processing method and device. Background technique [0002] At present, the voice-related garbage classification systems on the market are more focused on recognition than signal processing, so the processing effect is not good in outdoor or indoor noise-containing environments. Therefore, this paper proposes a low-power The advanced voice processing system is applied to garbage classification, and combined with multi-channel post-processing technology to deal with non-stationary noise, it has a good processing effect on outdoor noise-containing environments. [0003] Combined with the garbage classification system of voice recognition, the garbage can be sorted with voice guidance, which can be interacted with, and different garbage can be classified into different garbage bins according to the prompts. [0004] Neglecting the front-end signal processing leads ...

AI Technical Summary

Problems solved by technology

[0004] Neglecting the front-end signal processing leads to low recognition rate in noisy environments, and ignoring unsteady noise scenes. Post-processing filtering uses traditional Wiener filtering, etc., and Wiener filtering is relatively poor in robustness. Based on some mathematical assumptions, it is difficult to effectively suppress the interference of non-stationary noise in real environments

Moreover, it is difficult to meet the conditions for Wiener filtering to obtain all observation data in a semi-infinite time interval, and it cannot be used in the case of a random process with non-stationary noise.

Images