Realtime neural audio latent seeding w/ RAVE custom model trained on Martsman material in Pure Data

This video is mainly to showcase an updated version of my "Saatgut" component. It generates an array full of randomized values to use as seed bank in realtime inference and latent space tampering for nn~ compatible models. Cycling through this seeds is done via signal value input, e.g. simple oscillators. In the shown setup, I'm achieving repeatable sound sequences by resetting the oscillator phase on a regular basis with the metro object.

The model used in this video has been trained on an augmented dataset consisting of the majority of the tracks I've released in the past 20 years. https://www.martsman.de/showcase/

---
Saatgut component: https://github.com/devstermarts/PD-co...

RAVE is "A variational autoencoder for fast and high-quality neural audio synthesis” created by Antoine Caillon and Philippe Esling of Artificial Creative Intelligence and Data Science (ACIDS) at IRCAM, Paris.

RAVE on GitHub: https://github.com/acids-ircam/RAVE
nn~ on GitHub: https://github.com/acids-ircam/nn_tilde

To train models on Colab or Kaggle, you can use these Jupyter notebooks i've set up: https://github.com/devstermarts/Noteb...