Though you probably won’t be able to remove the entire vocal track (technology is just not there right now, without using a neural network and even then it’s not guaranteed to work 100%), you can put a filter over a track in the mid range. Somewhere between 1k -3k is the human voice I [i:29qrvyu5]believe[/i:29qrvyu5]. This is assuming the person singing is singing ‘normally’ and does not have any kind of distortion or filters on their voice.
You won’t get the whole voice out because of harmonics, little "echos" of a sound that extend well beyond it’s origin frequency, but it will get out most of it. However, this also means you will take out any instruments that might be occupying that frequency range (usually guitar, etc.).
Hope this helps! If you want to learn more, just Google ‘harmonics’ and ‘voice removal’.
- BobBobson108 answered 5 years ago
Like BobBobson108 mentioned you won’t be able to remove vocals completely and even worse you won’t be able to remove any part without altering the rest of the music.
One way would be to simply EQ most of it out like BobBobson108 mentioned but I think you would need to start a little lower than 1kHz : try 600Hz-3,5kHz
Also some older tracks have vocals on one channel and instruments on the other (mostly tracks from the ’60s or older) so those would be easy 😉
But if you want a more precise filtering with more recent tracks you might try something else.
Vocals are mostly centered in the "stereo image" while some instruments (like guitars) are partly (sometimes completely) panned to the left or right.
The idea would be to :
– convert your track to mono and then substract the resulting mono track from the left and right channels of the stereo track.
– filter as mentioned above the mono track (watch out to not filter too much of it as most LF instruments are centered)
– put all together again (re-add the mono track to left and right)
This is still not a perfect result but it will keep some more instruments a little alive in your track.
Don’t forget when you try to filter out the voice, don’t try filtering it all!
Filter to reduce speech intelligibility as much as desired and keep the rest.
- Dani1973 answered 5 years ago
Please login first to submit.