Download A hybrid approach to timbral consistency in a virtual instrument The aim of this work is to make an instrument that is timbrally consistent over pitch and loudness. This particular work is not attempting to reproduce an existing instrument’s timbre, but to produce a timbrally dynamic virtual instrument that can be designed by the user. In this paper there is a brief introduction to timbre and synthesis methods, followed by a proposal on how to make timbrally consistent virtual instruments out of given timbres.
Download Interpretable timbre synthesis using variational autoencoders regularized on timbre descriptors Controllable timbre synthesis has been a subject of research for several decades, and deep neural networks have been the most successful in this area. Deep generative models such as Variational Autoencoders (VAEs) have the ability to generate a high-level representation of audio while providing a structured latent space. Despite their advantages, the interpretability of these latent spaces in terms of human perception is often limited. To address this limitation and enhance the control over timbre generation, we propose a regularized VAE-based latent space that incorporates timbre descriptors. Moreover, we suggest a more concise representation of sound by utilizing its harmonic content, in order to minimize the dimensionality of the latent space.