Method and device for extracting accompaniment and human voice from stereo music

Abstract

The invention provides a method for extracting accompaniment and human voice from stereo music and a device of the method. A method for extracting the accompaniment includes that signals of a left soundtrack and a right soundtrack are respectively converted into a frequency domain from a time domain, normalization related values of corresponding frequency point pairs of the signals of the left soundtrack and the right soundtrack are calculated, weighting accompaniment gains are respectively added to the corresponding frequency point pairs of the signals of the left soundtrack and the right soundtrack, values of the accompaniment gains are in inverse proportion to the normalization related values of the current frequency point pairs, the signals of the left soundtrack and the right soundtrack are converted from the frequency domain to the time domain after being added with the weighting accompaniment gains, and the accompaniment of the left soundtrack and the accompaniment of the right soundtrack are respectively extracted. A method for extracting the human voice includes that human voice gains are weighted on mean signals of the corresponding frequency point pairs of the signals of the left soundtrack and the right soundtrack, values of the human voice gains are in direction proportion to the normalization related values of the current frequency point pairs, the mean signals of the left soundtrack and the right soundtrack are converted from the frequency domain to the time domain after the human voice gains are weighted to the mean signals, and the human voice is extracted. The method and the device can effectively extract the human voice and the accompaniment, and tonal quality effect is improved.

Description

technical field [0001] The invention relates to the technical field of audio processing, in particular to a method and device for extracting accompaniment from stereo music, and a method and device for extracting human voice from stereo music. Background technique [0002] At present, some audio playback software or audio processing software already have the function of extracting accompaniment music from songs. For example, if a user wants to record a song he sang, but cannot find the accompaniment music for this song, he can use the above-mentioned function to separate and extract the accompaniment music from the original song. [0003] Existing methods for extracting accompaniment often utilize the fact that the human voice in most songs is basically the same in the left and right channels, and adopt the method of directly subtracting the signals in the two channels to eliminate the human voice. Such as figure 1 Shown is a schematic diagram of a method for extracting ac...

AI Technical Summary

Problems solved by technology

[0004] The above method has the following disadvantages: in some songs, the human voice is not accurately aligned in the left and right channels, and the audio signals of the left and right channels are directly subtracted, which often cannot effectively eliminate the human voice, and some human voices will remain. ;In addition, if the same part of the accompaniment music appears in the left and right channels, part of the accompaniment will also be eliminated by subtraction, making the extracted accompaniment music less accurate, thereby causing the sound quality of the accompaniment music to decline

Since the sound quality of the extracted accompaniment is not good, the human voice obtained by eliminating the accompaniment music from the stereo is also of poor quality

Images