"rave_uponRAVE" is an experimental digital sound work that merges the intense energy of rave culture with legendary 'acid-house' basslines, using new electronic sound analysis and transformation tools along with neural network models. This project utilizes the RAVE (Realtime Audio Variational autoEncoder) model developed at IRCAM Paris, trained on a freely licensed audio dataset of 'acid-house' bassline samples. Implemented in Python, the model integrates with Pure Data software, leveraging the 'acids-ircam' [nn_tilde] decoder for real-time sound processing, generation, and dynamic sound coloring. The presentation of 'rave_uponRAVE' involves a Pure Data patch where the [nn_tilde] decoder processes audio signals in real time, generated by modifying random raw files (databending) from the hard disk. The model employs various algorithmic techniques to explore and generate acid-house patterns within latent space. By experimenting with these techniques, the project explores how neural networks can enhance and process audio data. Combining iconic acid-house sounds with neural network technology offers a new interpretation of familiar sonic elements, blending nostalgia with innovative processing techniques. RAVE : https://github.com/acids-ircam/RAVE https://forum.ircam.fr/collections/detail/rave/ nn_tilde : https://github.com/acids-ircam/nn_tilde