Keynotes

For over 94 years, Abbey Road has pushed the boundaries of how sound is created, captured, and experienced. From the invention of stereo sound, to pioneering early recording techniques and shaping the foundations of modern audio, innovation has always been the beating heart of the studios. In this Keynote, Abbey Road’s Head of Audio Products, Mirek Stiles,  will explore the studios’ evolution in innovation, from the early days of audio recording, to Abbey Road’s latest pioneering music production techniques, and the reimagining of its sonic DNA for consumer audio enhancement.

Mirek Stiles started at Abbey Road Studios as a recording engineer in 1998, working on a diverse range of projects including The Lord of the Rings trilogy, The Beatles Anthology, Muse, Nick Cave and the Bad Seeds and Paul McCartney. Mirek also worked as a Digital Producer in the Abbey Road Interactive department on a range of Blu-ray titles including Roxy Music, Ringo Starr and Reservoir Dogs. 

Today, Mirek is part of the studio’s innovation arm, Abbey Road REDD LABS, and runs the Audio Products department, leading development of a range of software and hardware releases. These are based on Abbey Road’s historic recording IP, acoustics, and engineering expertise and created with partners Spitfire Audio, Waves Audio and Chandler Limited. Mirek also sits on the board and acts as an advisor on the Abbey Road REDD incubation programme and is currently exploring and experimenting with Spatial Audio over headphones, ambisonics, spatial microphone arrays and game engine workflow. Recent projects include collaborations with film and video game composer Stephen Barton (Jedi: Fallen Order), joint research with the University of York and University of Huddersfield and chairing the Abbey Road Spatial Audio Forum, bringing together members from gaming, broadcast, VR, music production and film. 

Mirek’s expertise spans over 20 years within the studios, from analogue and vintage recording workflows through to the latest immersive medias including 6 Degrees of Freedom Virtual Reality and Dolby ATMOS capturing and rendering options. More recently Mirek designed, in collaboration with Bowers and Wilkins, the Abbey Road Sound Mode featured in the Volvo EX90 and Polestar 3 cars. This means the same sonic DNA from the studios, as used by artists and producers worldwide, is introduced to the consumer electronics world for the first time.

This talk summarizes different methods used for subjective evaluation of  automotive audio systems, with emphasis on perceived sound quality, spatial impression, distortion spectral balance, and listener preference inside the vehicle cabin. It highlights the challenges of measuring audio in a complex acoustic environment, the strengths and weaknesses of each approach and shows how controlled evaluations can connect human perception with engineering metrics. The presentation also discusses how these insights guide system tuning, design validation, and the creation of more engaging in-car listening experiences.

With a background spanning music, psychoacoustics and product innovation, Sean Olive has shaped the field of sound quality research for four decades. After early research in collaboration with Floyd Toole at the National Research Council in Ottawa, Olive joined Harman International in 1993, ultimately serving as Senior Fellow while leading groundbreaking work in loudspeakers, headphones, automotive audio and room acoustics. Since May 2025, he has worked as an independent consultant. A Fellow and Past President of the AES, Olive is also the author of more than 55 research papers, and co-author of Floyd Toole’s Sound Reproduction: The Acoustics and Psychoacoustics of Loudspeakers, Rooms and Headphones (4th Edition).

Minsuk presents a neural vehicle sound synthesis framework based on differentiable digital signal processing (DDSP), conditioned on driving signals collected from the CAN bus of an internal combustion engine (ICE) vehicle, and demonstrates the feasibility of realistic and coherent vehicle sound synthesis within this framework. Three design choices are investigated for the proposed framework: the definition of the fundamental frequency (F0), the configuration of driving signal inputs, and the conditioning representation.

Specifically, a comparison is made for crank-based and firing-based F0 definitions, multiple driving signal combinations constructed from engine RPM, gear level, accelerator pedal position, vehicle speed, and longitudinal acceleration, and two conditioning representations: direct and encoded conditioning. The framework is evaluated using objective and subjective measures together with qualitative spectrogram analysis. The results show that the crank-based F0 provides more accurate synthesis than the firing-based F0 in the present four-cylinder four-stroke vehicle setting. Driving signal configurations with more complementary signals generally improve synthesis quality, while the contribution of each signal depends on its relationship with the other inputs. Encoded conditioning yields better objective performance, especially when the available driving signals are limited, whereas direct conditioning achieves the best perceptual results under full driving signal configuration and offers practical advantages in simplicity and efficiency. These findings provide practical guidelines for DDSP-based neural vehicle sound synthesis and suggest that conditioning DDSP on driving signals is a promising approach for automotive audio applications such as vehicle sound design and driving simulation.

Minsuk Choi is a Ph.D. candidate in the Music and Audio Computing Lab at Korea Advanced Institute of Science and Technology (KAIST). His research focuses on neural sound synthesis, differentiable digital signal processing (DDSP), and vehicle sound modeling. His current work explores DDSP-based vehicle sound synthesis conditioned on real driving signals, with a focus on controllable and efficient sound generation. He is interested in signal-processing-informed neural audio models for controllable, flexible, and interactive sound synthesis. His broader interests include neural audio generation and interactive audio systems.