Photosounder

[quote quote="djted3000":2lncakfm]Heard about this on the kickass Sonic Talk podcast. Photosounder is a program that converts audio into a picture, which is then editable by normal Photoshop type tricks. The tweaked graphic can then be played back.

http://photosounder.com has some Youtube and audio examples. One of its tricks seems to be vocal (or backing track) extraction. Looks like some kind of filtered difference process, since the tracks are named "full track" and "instrumental" and then it has a paste/level tweak in Photoshop. Hard to tell how artifacty the output sounds, since the youtube clips are in lo-quality. Interesting twist on the old inversion process.

http://www.youtube.com/watch?v=BKodcXMQHZU

Not exactly the killer acapella app, but has potential, and does other weird stuff…[/quote:2lncakfm]

Thanks for the post, djted3000. This video was actually done before the lossless mode was implemented in Photosounder. Now with the lossless mode well it’s lossless, so it yields much better results.

The example MP3s I worked with on the video were a bit lousy because the instrumental version didn’t have the same equalisation as the full song, so even in lossless mode there still was undesirable things left. However if you (or anyone else for that matter) do have a song and its instrumental and that both versions match in quality I’d be happy to try to extract the acapella using that technique.