As David says, I usually preload all samples to buffers in a "samplebank" type of object. In your case as they're single notes it should do the trick.
Then maybe use clone with X table readers (your number of voices) and the poly object to assign them?
Here it is!
It’s based on the [freeze3] patch mentioned in the first post.
It uses some other abstractions of mine so I just bundled the whole thing, check the readme.
I can make a self-contained vanilla patch if this is confusing.
This is my first go at spectral processing - feedback and advice much welcome!
As the title says, I’m looking for help / feedback / tips on how to optimise my patch.
I understand it’s already quite complex, but I’d need to push it another step in order to get where I need: a swiss army patch for live sonic improvisation / exploration, combining bits I’ve built in the past for myself or others, or found on this forum.
I’ve made sure (at least I believe) that guis aren’t updated unless they’re showing, dsp is switched off whenever is possible, I’ve tried to limit buffer space, etc. However it’s still very slow at loading and closing, and begins to crack when all elements are active.
I’m with Ubuntu on an old i5 so I understand there are limits, I just didn’t think they’d be met so fast.
I’d also love to know if someone knows how to remove menu bars on all windows except the Pd window?
The patch requires pd 0.50 and the ELSE library available on Dekken.
Then just make sure you point to the “pudu” folder in the paths and startup.
Using the patch should be quite straightforward, though still undocumented WIP. If I get past current limitations I’ll make it solid: it might be of some help to noobs as it wraps bits from my favorite projects such as else, audiolab, rj-lib, diy2, mmb-master, etc.
Any help appreciated! Thank you forum, as always.