Speaker: Dr. Fabián Esqueda
Research Engineer, KORG
October 29th, 2026 | 4:00 pm
DEIB, Conference Room "E. Gatti" (Bld. 20)
Contact: Riccardo Giampiccolo
Abstract
Circuit modeling, also known as virtual analog (VA), is a well-established and extensively researched topic in music technology. Over its 30-year history, it has produced a substantial body of research and products, making it possible for musicians and hobbyists alike to access the sounds of legendary vintage musical instruments on their personal computers. The field of VA is often framed as a divide between white-box approaches, which derive models directly from a circuit's physical structure and governing equations, and black-box approaches, which learn a system's behavior from measured data without reference to its internal design. In other words, it’s the underlying math versus the observed behavior. In practice, most real-world modeling work lives somewhere between these two extremes, blending physical insight with data-driven techniques — a gray area this talk will explore. The field continues to evolve and, perhaps unsurprisingly, machine learning has injected new life into it. In this talk, a general overview of the field will be given, before diving into personal contributions to the research corpus and preferred techniques. Along the way, this work will be contextualized with references to recent KORG releases.Short Bio
Fabián Esqueda is a Research Engineer at KORG, where he works on signal processing and machine learning algorithms for musical instruments.His research interests include generative models, sound synthesis, and virtual analog modeling. He holds an MSc in Acoustics and Music Technology from the University of Edinburgh, UK, and a DSc from Aalto University, Finland, where he specialized in antialiasing techniques for nonlinear signal processing.
Prior to joining KORG, he worked as a DSP and ML Research Engineer at Native Instruments.