OK, I converted to mp3. I hope this works.

Back story: we were woodshedding during Hurricane Harvey in 2017. Austin was more or less fine but it was a pretty horrific event for the state. My buddy Larry, who is a legit artist and art professor suggested we try something akin to the "exquisite corpse" surrealist art exercise where a piece of art is created sequentially by different people, without knowing what the previous artists had done. We couldn't really figure out how to do that musically so we just built a track from the ground up, starting with an improvised drum part, and then a bass part, and then a couple of guitar parts. Nobody had heard anything that came before, everything just one take. Don't blame the bass player, he set the progression, but Larry and I went off the rails. It's basically Sonic Youth meets John Cage on acid, or maybe not. Anyway, I like it in a perverse way. It is 100% antithetical to anything else I have ever done or have been a part of. I added some weird white Christofascist shit that happened to pop up as I scrolled through TV channels looking for news about the hurricane. (No endorsement of that message, for sure.) No drugs were involved. Plenty of beer, though.

Edited September 24Sep 24 by jimmyjazz

12 minutes ago, jimmyjazz said:

No drugs were involved.

Well that's a shame.

This is fun though man, it actually works entirely better than you would think

2 minutes ago, G650 said:

Well that's a shame.

This is fun though man, it actually works entirely better than you would think

To me it's weirdly soothing. Anyway, we took a shot. It was better than the overdubs we were doing on garage rock tunes.

24 minutes ago, jimmyjazz said:

It is 100% antithetical to anything else I have ever done or have been a part of.

I really dig that and would encourage you to be antithetical more often.

One AI thing related to that last video. I finished all the video editing and was pretty happy, but the audio track captured from the Zoom Q2N had way too much crowd noise in it to sound like anything other than marginally recorded audio of an inspired performance in a very noisy room. Like, I wanted the crowd and bar noise to be at normal levels at the beginning, end and when I remarked "That sounded like the devil went down to Georgia on that one", but otherwise be brought down relative to the musical performance.

I figured, let me see What Would Claude Do (WWCD is the new WWJD), and it built python scripts that extracted the audio from the .MOV file, used some open source neural network stem extraction lib to do the stem split, apply the requested changes to just the crowd noise track, then recombined it into a new stereo audio file, applied some light compression and an eq curve appropriate for headphones and created a new .MOV file with the altered audio. It did it in about 15 minutes from three prompts while my attention was on my day job.

Dunno if anyone is interested, but I could upload a snippet of the original audio and the altered audio and you can hear the diff it made.

3 minutes ago, Goredho said:

Dunno if anyone is interested, but I could upload a snippet of the original audio and the altered audio and you can hear the diff it made.

I'd love to hear that.

There's all sorts of applications that do that.

23 hours ago, jimmyjazz said:

No drugs were involved. Plenty of beer, though.

Here is the original audio from 7:30 to 8:30


Here is the same audio clip after Claude dropped the crowd and bar noise, added compression and an EQ curve


Its a pretty fucking stark difference to me.

Edited September 25Sep 25 by Goredho