Ultrasonic-assisted microphone array speech enhancement device

Abstract

The invention discloses an ultrasonic-assisted microphone array speech enhancement device, and relates to voice signal treatment. The ultrasonic-assisted microphone array speech enhancement device comprises an equidistant linear array, a preamplification circuit, an analog-digital conversion circuit, two ultrasonic sensors, a transmitting circuit, a receiving circuit, a position computation module and a weighted stacking module, wherein, the equidistant linear array comprises five microphones; an output end of the linear array is sequentially connected with the preamplification circuit, the analog-digital conversion circuit, data cables and the weighted stacking module; the sensors are installed on extended lines at two ends of the linear array; the transmitting circuit and the receiving circuit utilize the sensors to transmit signals, and the received signals are processed through preamplification and analog-digital conversion and then are sent to the position computation module; in addition, the position computation module acquires return wave receiving signals of the sensors for matched filtering so as to acquire the ultrasonic transit time from a target to the sensors, acquires sound source position information that is input to the weighted stacking module, calculates time-delay compensation values of channels, performs time-delay compensation and stacking on each channel receiving signal of a microphone array so as to enhance voice signals in sound source directions.

Description

technical field [0001] The invention relates to a voice signal processing, in particular to an ultrasonic-assisted microphone array voice enhancement device. Background technique [0002] In voice acquisition and processing systems such as voice recognition, voiceprint recognition, and video conferencing, the signal received by the microphone is usually noisy voice due to interference from reverberation and background noise. This not only affects the intelligibility of speech, but also affects the overall performance of the speech processing system. Therefore, it is necessary to enhance the noisy speech. In complex acoustic environments, single-microphone speech enhancement can no longer meet the needs. [0003] The microphone array can form different responses to signals in different directions, which is the spatial directivity characteristic of the array. When voice and surrounding environment information are collected by multiple microphones, the microphone array can e...

AI Technical Summary

Problems solved by technology

However, because this method needs to align the beam by adjusting the array pointing, there is still a lag in the acquisition of azimuth information, and it is also impossible to obtain real-time azimuth information for moving sound sources.

[0009] Chinese patent ZL200410002739.2 proposes a microphone array sound collection method and its system combined with positioning technology. The sound source orientation is obtained by combining the sound wave signal obtained by the microphone array and the target sound source picture obtained by the digital camera. Since image search algorithms are also needed Judging the size and orientation of the target sound source in the screen obtained by the digital camera to calculate, the system positioning processing has a large amount of calculation and the performance will be affected by lighting conditions, etc.

Since the frequency domain correlation between the speech signals of each channel is still used, the delay estimation accuracy will be significantly reduced in the presence of severe reverberation and other conditions, thus affecting the performance of delay compensation weighted superposition speech enhancement

Images