Hi I’ve been trying to isolate the vocals of a movie using Elevayta Extra Boy Pro plugin without any success. There always seem to be some background instruments and/or noise no matter what I try.
Is this just a case where Vocal Isolation just isn’t possible or does anyone have an idea what I can try?
Thanks for reading.
When it comes to isolating vocals from a mix, we often hear (or use) expressions like “un-baking a cake”, or “extracting the banana from a banana milk-shake”. Until recent advances in digital audio it was flat impossible. Now with tools such as “Extra Boy” and “kn0ck0ut” it can be done … a bit. The degree of success depends on the type of sounds that are in the mix and how much time and effort you spend - as they say “your mileage may vary”. I’ve heard a few very impressive demonstrations, but these invariably use mixes that offer the best possible chance of success - with most recordings the results tend to be a lot less impressive.
That actually put things in perspective for me. I’m not going to lie, what I’m using is by no means the “ideal” scenario for vocal isolation. I think I’m going to try this for the rest of the night, if by morning no success then I’ll give in.
Thanks for the reply.
Let us know how you get on - whatever the result. If you could upload a short sample to the forum as a before/after demonstration that could be highly instructive for other people attempting to do vocal isolation. You can upload short (20 seconds or so) samples in MP3 format to the forum by using the “Upload” tab below the message typing box.
Well, hope I don’t get sued for this. I did make a decent amount of progress I think, it’s either that or I’ve been at this for way too long.
So you can compare the original…
And here’s the result so far.
What I started doing is using a combination of Audacity to save and convert recordings to .wav sound files and Elevayta Player Boy to run the Extra Boy Pro in visual “2D mode”.
What I was doing wrong is assuming I could completely isolate the vocals in one go. Instead I’ve been continuously using Extra Boy to separate the low pitches and center pan, saving and converting the file with Audacity and then taking the converted file back into Elevayta to seperate the low pitches and center pan again. After 5 “flushes” you have the result. It’s a bit time consuming because you have to compensate for random glitches occurring literally all the time but it seems to work. I’m going to try this a few more times but I thought I would put this here since you asked. So it’s here for whatever vague referencing purposes this could serve.
I’ve used Kn0ck0ut (similar method to ExtraBoy) it sounds 'orrible but the voice is isolated …
First I used the inversion centre-pan removal method to create a track which was everything but the centre-panned Will Smith.
Then I Hyperexpanded this Smith-free track then spectrally subtracted it from the original using Kn0ck0ut to give Big-Willy on his own,
(but horribly distorted).
Thanks folks - interesting to hear real-world examples.