Last year, using a combination of Audacity and Wavosaur (to remove vocals…when I had an older version of Audacity) and inverting, converting to mono, and removing vocals, I took “I Know There’s an Answer” from the Beach Boys’ Pet Sounds album (the stereo backing track from the sessions CD box set) and ended up removing many instruments, leaving only three or four. Attached is a snippet of the original and the remixed version. I promise, it’s the exact same part of the song on each. Now…how did I do it? I have spent the last day trying to figure this out, and I cannot get it! Please, someone get me in the right direction. I am new to version 2.0.5. I think it was version 1.2.6 that I did the above with.
I know how that works. I once got Audacity Noise Removal to suppress echoes in a spoken word piece. You can’t actually get rid of echoes and I did. Once. I was never able to do it again.
I think the key is the music, not your technique. I’ll put significant money that the Beach Boys was “remastered” for the Box Set, which means somebody in a nice studio did some serious shifting and fading and phase management to go from the master tapes (some of which are mono) to the shows you have. Which means the show is half-way to having some of the instruments cancel just by mixing down to mono.
In other words, if you did discover the magic formulae, it would work perfectly on the Beach Boys Remastered Box Set. Full Stop. Probably nowhere else.
Vocal Removal people run into the same problem. Somebody always points to that guy in the YouTube tutorial who got total and clean cancellation of the lead vocal. Anybody can do that just by starting with the exact same song that he did. Other songs may not work out so well and probably won’t.
Koz
Maybe a different part of the same tune , but IMO no-way is “Remixed.wav” derived from “Original.wav”.
It is. My email is ***@gmail.com. Send me an email, and I will send you an mp3.
I’m aware that these are special mixes of the backing tracks, but there should be a way to remake the same thing again. I have always liked “removing vocals” just to hear some of the things going on in the background. A cool way to make special mixes is to remove the vocals and then layer the result back over the original. It flushes out the background.
I removed your email address. Posting one is dangerous on this forum. Send sensitive information via private message.
Koz
Be aware that the vocal removal should you give a stereo output, not a mono.
Those tracks layered (one inverted) gives the center (dual mono).
You’ve probably used pan and gain on one track. This shifts the center of gravity or the stereo band that is cancelled out.
In this way, you can mix and render the result, split to mono and be left with one prominent instrument.
To illustrate this, here’s an excerpt from your song–the instrumental part. It is also two times the same portion.
I’ve used two tracks:
- Original
- Isolated Center (inverted)
Played at the same time, this gives the original - the center
the two parts are the left and right channel in isolation.
There are some techniques to refine the result (applied to the center track): - panning (shift away from center)
- gain (for 2 bands instead of only one dead center)
- rotation (removes e.g. +/- 5 °)
- equalization (for instruments covering a certain spectrum)
- reducing stereo width before center isolation/removal. This affects a bigger portion in the center.
Since there are so many factors, the outcome is hard to predict. I suggest you write down the steps taken for the next time…
However, since you’ve already succeeded with this title, I suppose you’ve another song in mind to tear apart?
I believe I did use pan somewhat when I did the original.
When I get this figured out again, I will write them down. I’ve a mind to tear apart the entire Pet Sounds album, once I get everything like I have on the one song 
In the meantime, instead of using a small clip, I’ve uploaded the entire thing here: http://www.freestateofwinston.org/stripped.mp3
Thanks for the tips!