AN AUDIO DATA LABELING METHOD BASED ON TWO -CHANNEL AND MULTI-PERSON VOICE TALKING SYSTEM THEREOF

摘要

An audio data labeling method based on two -channel includes: obtaining at least one channel audio data with volume; determining a virtual position of each channel audio data according to the obtained at least one channel audio data; adjusting the volume ratio of the left channel and right channel according to the volume and the virtual position of the obtained at least one channel audio data. A multi-person voice talking system is provided.