EVxRAVE: Incorporating Neural Synthesis in an Augmented String Instrument Platform

Brian Lindgren

Keywords: chordophone, rave, variational autoencoder, neural audio, nime

TL;DR: EVxRAVE integrates IRCAM’s RAVE autoencoder into the EV, an augmented bowed string instrument, using its own recordings to map pitch, amplitude, and spectral features to latent dimensions of timbral expression.

Abstract: The EV is an augmented bowed string instrument that integrates digital signal processing with acoustic performance. This demo presents EVxRAVE, the incorporation of IRCAM’s RAVE variational autoencoder into the instrument. Trained on recordings of the EV itself, the model produces an eight-dimensional latent space that can be shaped in performance through descriptors such as pitch, amplitude, and spectral centroid. The accompanying performance, two tales from the shadows of the grid, demonstrates how RAVE extends the EV’s acoustic expressivity into a neural domain, exploring the interplay between automated modulation and embodied gesture within a hybrid analog–digital instrument.

Track: Demo Track

Confirmation: Demo Track: I confirm that I have followed the formatting guideline and included all author names and affiliations.

(Optional) Short Video Recording Link: https://youtu.be/aVPE8Ac6XU8

Submission Number: 84