Warning: It can get wild, lower volume... Tried out a bunch of samples and by slightly tweaking some parameters you can end up in overdriven white-noise hell depending on the "brain" + samples.
Sounds like it fits in the family of concatenative synthesis. I've played with this technique before but never got any nice results, this tool is good motivation to give it another go. Some other cousins:
What would these images signify?! I can't seem to find any readme files or docs using them.
Also..: The date puts this sometimes just around or after the release of Syro (which I was really disappointed with by the way but that's another discussion). Does anyone know of any tracks that were made using this tool?
The multicolor image with the wav files being displayed vertically, if I were to guess. The left side shows columns of your sample bank “ samplebrain”. Right side is your target. It shows how the “samplebrain” selects the most similar “ sample window “ of the target. So, I haven’t the knowledge to drill deeper, but it’s either a visual representation of the process, or maybe the code uses an actual image matching process?
You might not want that. I was going to post some quip about RDJ's (aka Aphex Twin) code, so I'll do it here.
I could tell this app wasn't written by RDJ, because it's fairly well laid out, and his code has always been that of a messy tinkerer.
He's released several bits of code on the web, most of it pseudonymous and it's always been a hacked up mess. I've "uncovered" him a few times, just by recognizing his chicken scratch code.
Edit: I won't mention any of his pseudonyms, except one that's semi-public and long-retired: eric_hard_jams ("richard james" with the "e" moved). He used that one around the turn of the millenium. If you Google it, keep in mind that some fans later co-opted the handle for their own use.
Edit 2: Also, hi Richard! I'm sure you're reading, and I hope life and fatherhood are treating you well.
Max/MSP, some C/C++. All sound related, that I know of. He used to spend a lot of time online. Much more than he lets on. I don't know what he's doing now -- immersed in family and normal life stuff, I reckon.
I don’t think I ever came across any code that was accredited to Richard. Definitely remember some max/msp patches throwing the Autechre name around. Still love the supercollider bbcut libraries, Nick Collins definitely knew what the rephlex folks were up to :D
I don’t have experience coding, but I’m keen to try this program. Would someone please give me a very broad overview of what I’ll need to run this program on an Apple Silicon comp after downloading it from Git? I just need to be pointed in the right direction.
Wow, that page really blew up in 2020. Thanks for the tip, has led me to some great stuff. It seems , to me, like these programs are taking an opposite approach to generative music , vs openAI jukebox style. From my view, openai is depending heavily on descriptors ( band names ) , and genre labels. But these samplebrains don’t sweat that stuff, no baggage.
Sounds about right! Can’t wait to play with this tomorrow.