AudioSep is a foundation model for open-domain sound separation with natural language queries.

#deeplearning #demixing #ai #artificialintelligence

https://audio-agi.github.io/Separate-Anything-You-Describe/

Separate Anything You Describe

RUBBER SOUL: Drums & Bass only

“De-mixing” A.I. will eventually change music forever. There will be a point in the future when any listener, anywhere, can deconstruct any music recording.

https://youtube.com/watch?v=LSgFPUlhc2o&si=v-56mgsz792urWPb

#TheBeatles #Beatles #ArtificialIntelligence #Demixing #Remixing #Remix #Audio #Sound

rubber soul but it's only bass and drums

YouTube

More adventures in #demixing with #izotope. A 1979 recording of 2 #Cream covers. Wasn't multitracked, but was still a decent quality stereo board feed, so a little easier to separate the stems. The most challenging part was that the guitar was (of course) very loud so not much coming through the board apart from what was coming through the 3 vocal mics. I was able to isolate the guitar a bit and bump it up.

1/

https://youtu.be/gVt5lWePMIE

#mastomusic

Badge and SWLABR (audio only) Gateway at Muhlenberg College 1979

YouTube

Another little experiment with #demixing with #izotope. This is an ancient stereo cassette recording of my band at a frat party in 1979. The B3 player played a massively wrong chord at one point.
First, I went into the incorrect section and used #Melodyne to remove the wrong chord notes. Then I found the correct chord in a later section, isolated the guitar and keys, and copy/pasted it where the wrong chords had been.
https://youtu.be/dxgfMW-fTZA

#mastomusic #steelydan

Gateway: Don't Take Me Alive (audio only)

YouTube