Using nn~ on Organelle?


I don’t have an Organelle (yet), so I can’t try this myself.

Has anyone tried to use the nn~ external on an Organelle? It can apparently run on a Raspberry 4, but if I got it right the Organelle M/S has a Raspberry 3?

It is for synthesizing sounds using a learned machine learning model. Check out the links:

There’s also a github for the RAVE: Realtime Audio Variational autoEncodercode that creates the models that nn~ can use (I can’t link it here since new forum users are limited to two links)

I haven’t touched PD in a long while, but I’m thinking about buying an Organelle. If it would be possible to use nn~ with it I’ll buy it right away.

Here’s a video. I couldn’t post it in the original post because of some restriction on the number of links by new forum users.

200 ms latency is unusable, so even if Og-M had a cm-4. still wouldn’t be enough… Anyway, you could use Organelle to control & loop nn running in your pc

Maybe a silly question, but where does that number (200 ms) come from?

From the video. seems like the nn version for raspberry pi is not the full one. We have been testing this model in M1 and takes like 300% cpu and still has a lot of latency., Its kinda cool but still not for use in realtime, I think…

1 Like

Oh, my mistake, I didn’t remember it or catch it in the video. Thanks for the replies! The rave code caught my attention for another reason (not needing low latency), but then I thought that wow, maybe it would be possible to use it on the Organelle. Well apparently it isn’t. Cool stuff anyway!