At present, this is impossible.

Audio files may also be recorded in separate tracks.

But most video files only have one audio track, which means that music and dialogue are recorded in the same audio track.

When you open it with audio software, all sounds are superimposed in the same track. In this state, it is impossible to separate the vocals and instrumental sounds.

Using an equalizer can only amplify or reduce the overall sound of certain frequency bands.

The frequencies of human voices are roughly as follows:

Bass reference range, midrange and treble

Male 82～392Hz 64～523Hz 123～493Hz 164～698Hz

Female 82～392Hz 160～1200Hz (Alto: Same as above Soprano: 220～1.1KHz)

The span is not small, even if only this part of the frequency is retained, the rest will be reduced. There are also other musical instrument sounds within this frequency.

Therefore, vocal purification in video files cannot be achieved.

Who knows if there are any high-end instruments that can do this, but the equipment or software available to ordinary people on the market cannot do it.