Samplebrain: A custom sample mashing app designed by Aphex Twin | Hacker News Reader

Samplebrain: A custom sample mashing app designed by Aphex Twin

gitlab.com

160 pointsdev_snd31 comments

> As we worked on it (during 2015 and 2016) we gradually added more and more tweakable parameters until it became slightly out of control.

Sounds about right! Can’t wait to play with this tomorrow.

I have to try this, see if it fits my work. Just tattooed Aphex Twin logo a couple of months ago.

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:

https://rodrigoconstanzo.com/combine/

http://www.benhackbarth.com/audioGuide/index.html

More info here: https://thentrythis.org/projects/samplebrain/

and in the profile text of RDJ's soundcloud: https://twitter.com/thentrythis/status/1573573106714697729

Is there a demo available of someone using it to create music?

I think the last time I saw Comic Sans used in any software as a main font was when I compiled dogecoin[1]. Can't remember any before that.

[1] https://github.com/dogecoin/dogecoin

Very cool, can't wait to play with this.

https://gitlab.com/then-try-this/samplebrain/-/tree/master/c...

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?

At least some of these images are used for the website.

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?

Shoot, I thought I was going to get to see actual code written by Aphex Twin.

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.

I had no idea he was a coder at all. What kind of code did he release? Is it all music related?

Also how are you so sure he’s reading this?

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

>> chicken scratch code

I guess I would just like to see how musical genius translates, particularly if code is viewed as an art form in itself.

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.

Figured it out. Someone can delete my question, if they wish. (facepalm).

Were you able to get it to work on the Mac silicon?

This is great, but I’m still waiting for the sample to Yamaha dx7 patch midimutant code…

( I’m mostly being sarcastic, I’m grateful for any access to new creative tools. But the MagPi issue #65 really lifted my hopes ). https://magpi.raspberrypi.com/articles/aphex-twin-midimutant

https://github.com/fo-am/midimutant/issues/1 "Main code missing"

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.

Finally had a chance to play with this. Amazing and hilarious.

Love that it uses some sort of Comic Sans as the interface font.

Very on-brand for Richard, somehow.

Reminds me of Sononym sound librarian

https://www.sononym.net

Thanks for the link, looks interesting.

Seems that there are some code signing issues with the Apple silicon version. Intel one works fine though.

I love that it uses Comic Sans

never change!

cool stuff

Reply on news.ycombinator.com