There are several programs you can use to remove a voice from a song. One such program is Adobe Audition 3.0 or higher with the Center Channel Extractor VST plug-in built into it.
Instructions
Step 1
It will not be possible to remove a voice perfectly with this method. After removing the vocal, a subtle echo of a sort of distant voice may remain.
Step 2
Launch Adobe Auditions, load a track (wav, mp3 or any other format) with a voice inside by simply dragging and dropping into the editor window.
Step 3
Go to Effect menu -> Stereo Image -> Center Channel Extractor. Here you can make additional adjustments so that you can remove vocals more completely, so that the phonogram remains natural sounding.
Step 4
Extract Audio From - you can specify in the drop-down list where the extraction will take place: right, left, center, in the surround file, or selectively. If the voice in the phonogram is in the middle, it is enough to indicate Center, if on the left - Left, on the right - Right.
Step 5
Frequency Range is the frequency range to be cut.
Step 6
Male Voice, Female Voice, Bass, Full Spectrum - respectively, male, female voice, bass range and the whole spectrum. Set your own values in the Start and End fields before the Custom option (the beginning and end of the range of the selected center channel). Knowing the range of this or that channel, you can apply more precise settings to remove it, keeping the quality of the original phonogram file in the best possible way.
Step 7
Center Channel Level - a slider that allows you to set the level of the center channel (or vocals) intended to be removed. In the position from -40 dB all the way to the left - the recommended settings for suppressing the center channel.
Step 8
The next settings are Discrimination Settings. In them, you can tweak the sound of the mix itself, slightly improving the effects of cutting out the voice and cleaning up the remnants of the phase processing of the vocals.
Step 9
Crossover is a slider that sets the main audio level as a percentage. This level is set at 93-100% when the voice is removed from the song.
Step 10
Phase Discriminations is a phase discriminator. It is recommended to set the value in the range from 2 to 7.
Step 11
Amplitude Discriminations is an amplitude discriminator. Recommended values are 0.5-10.
Step 12
Amplitude Bandwidth - it is better to set the value in the range of 1-20.
Step 13
Spectral Decay Rate. With it, you can smooth images in a track. The best values are 80-98%.
Step 14
FFT Size - settings between 4096 and 10240 work best.
Step 15
Overlays - the recommended value is from 3 to 9.
Step 16
Now figure out the Interval Size - set the value between 10 and 50 milliseconds.
Step 17
Window Width - Values between 30% and 100% work well.
