Echo cancellation method and voice recognition equipment for voice recognition process

Abstract

The invention provides an echo cancellation method for a voice recognition process. The method comprises the steps that digital signals are divided into the first digital signal and the second digital signal, wherein the first digital signal is converted into a first analog electric signal through a first DA conversion module and sent to a player to generate a sound signal, and a microphone collects the sound signal and converts the sound signal into a third analog electric signal; the second digital signal is converted into an output digital signal after being filtered by a filter, and the output digital signal is converted into a second analog electric signal through a second DA conversion module; the second analog electric signal and the third analog electric signal are coupled and superposed, and an error electric signal is calculated; the error electric signal is converted into an error digital signal through an AD conversion module, and the filter conducts output digital signal updating through the error digital signal. The method is good in echo cancellation effect, and voice signals for voice recognition are purified.

Description

technical field [0001] The invention relates to the technical field of voice recognition, in particular to an echo cancellation method and voice recognition equipment used in the voice recognition process. Background technique [0002] With the development of people's intelligent life, smart speakers have entered ordinary families as important smart devices. An important function of smart speakers is speech recognition. For a better user experience, it is necessary to perform speech recognition when the speaker is playing music or voice prompts. So at this time, in addition to receiving the user's voice prompts, Mike also received the music or voice prompts played by the speaker. The echo cancellation function is needed to remove the echo, and then perform speech recognition. [0003] When performing speech recognition in existing smart speakers, in one case, speech recognition is performed without playing music or voice prompts. This method sacrifices user experience. An...

AI Technical Summary

Problems solved by technology

Another situation in the existing technology is to perform very good sound isolation on the microphone on the smart speaker. This method has high requirements on the sound insulation performance of the sound cavity material of the speaker, which is not conducive to the realization of the cavity, and for Bass doesn't work well either

Another situation in the prior art is to reduce the music or prompt tone so that the sound collected by the AD is not distorted, and the echo cancellation is performed in the digital domain inside the chip. In addition to sacrificing the volume of the music or prompt tone, this method also It is necessary to reduce the gain of AD, thereby reducing the signal-to-noise ratio of speech and also affecting the accuracy of speech recognition

Images