Robin Doerfler
Portrait of Robin Doerfler

Robin Doerfler

Senior Audio Research Engineer

ML · Neural Audio Synthesis · Differentiable DSP

Audio Research Engineer specialising in machine learning, differentiable signal processing, and neural audio synthesis.

I develop methods for audio generation and processing, drawing on classical DSP, differentiable DSP, and neural approaches. My background spans roughly eight years in audio R&D, evolving from real-time DSP and synthesis algorithms toward machine learning for audio generation, with the first models now entering production on embedded targets. Currently my primary application domain is automotive active audio, where I work as a researcher and as Product Owner of EVx Suite — a DAW for active sound design in vehicles, licensed by major international vehicle manufacturers.

I hold an M.Sc. in Sound and Music Computing from Universitat Pompeu Fabra (ranked 1st in cohort, 9.7/10), where my thesis combined physics-informed inductive biases with differentiable signal processing for controllable neural audio synthesis, incorporating engine-sound physics to improve on general-purpose methods for the task. My work has appeared as papers and presentations at AES, IEEE CASS, and EUSIPCO.

Munich, Germany · Impulse Audio Lab GmbH

Publications & Talks

2026

  1. Gradient-Based Learning of Parametric Engine Sound Representations for Real-Time Resynthesis and Tuning on Embedded Systems Paper [arXiv]

    Robin Doerfler, Matthieu Kuntz, Clemens Zimmer

    AES 6th International Conference on Automotive Audio (Automotive Audio 2026), Detroit, USA

  2. Physics-Informed Neural Engine Sound Modeling with Differentiable Pulse-Train Synthesis Paper [arXiv]

    Robin Doerfler, Lonce Wyse

    34th European Signal Processing Conference (EUSIPCO 2026)

  3. Analysis-Driven Procedural Generation of an Engine Sound Dataset with Embedded Control Annotations Paper [arXiv]

    Robin Doerfler, Lonce Wyse

    34th European Signal Processing Conference (EUSIPCO 2026)

2025

  1. Neural Engine Sound Synthesis with Physics-Informed Inductive Biases and Differentiable Signal Processing Thesis [Code]

    Robin Doerfler

    M.Sc. Thesis, Universitat Pompeu Fabra, Barcelona

  2. Procedural Engine Sounds Dataset Dataset [Dataset]

    Robin Doerfler

    Zenodo

2024

  1. Invited Keynote — AES 2024, Gothenburg Keynote

    Robin Doerfler, Jasper de Kruiff

    AES 5th International Conference on Automotive Audio, Uni3 by Geely, Gothenburg, Sweden

  2. Cycling '74: Max MSP in the Industry Interview [Video]

    Robin Doerfler, Josh Weatherspoon

    Cycling '74 YouTube series

2022

  1. Invited Talk — IEEE CASS RS Talks 2022 Invited Talk [Video]

    Rodrigo Schramm, Robin Doerfler

    IEEE Circuits and Systems Society, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil

  2. Invited Keynote — AES International Conference on Automotive Audio 2022 Keynote

    Rodrigo Schramm, Robin Doerfler

    Audio Engineering Society, Dearborn, USA

  3. AI in Automotive Audio: Approaching Dynamic Driving Sound Design Paper [Paper]

    Rodrigo Schramm, Jasper de Kruiff, Robin Doerfler, Jonas Merkt, Philipp Kampmann, Peter Dowling, Florian Walter

    AES International Conference on Automotive Audio 2022

Experience

Impulse Audio Lab GmbH

Munich, Germany · 2018 — present

  1. Senior Audio Research Engineer

    2022-04 – present

    Research and development of ML-driven audio synthesis and processing — building neural and signal-processing methods that generate and control sound in real time on embedded hardware, applied to automotive audio. Core work spans neural audio synthesis, differentiable DSP, and end-to-end pipelines that combine learned and signal-processing components, including RPM/torque-conditioned synthesis networks, parameter estimation from audio, and timbre-space embeddings. Product Owner of the EVx Suite — an automotive DAW for developing and deploying interactive driving sounds and in-vehicle audio, selected by major international OEMs for production use. I own the technical roadmap and have taken it from early prototype to production.

  2. Lead Audio R&D Engineer

    2020-04 – 2022-04

    Built the real-time DSP and synthesis engines behind procedurally generated engine sound. Developed analysis-and-resynthesis pipelines that decompose recorded combustion-engine audio into synthesis-ready representations — harmonic/noise components, wavetable frames, procedural parameters — for real-time resynthesis on embedded targets across additive, wavetable, procedural, and sample-based methods. Built and annotated a dataset of synthetic engine audio that supported the downstream ML-based synthesis and parameter-estimation work.

  3. Audio Design Engineer

    2018-04 – 2020-04

    Designed and prototyped real-time interactive audio systems that map live vehicle and sensor data to synthesised sound — spanning active sound design and algorithmic prototyping for interactive driving and engine sounds. Built demonstrators and research prototypes in Max/MSP and Python for premium OEM clients including BMW, Mercedes-AMG, Mercedes G GmbH, Rolls-Royce, and MINI.

SEED Mediendesign GmbH

Munich, Germany · 2017 — 2018

  1. Sound Designer | Audio Engineer

    2017-04 – 2018-04

    Audio engineering, recording, and post-production for multimedia applications.

Superstrings Music GbR

Munich, Germany · 2015 — 2017

  1. Assistant Composer | Music Production

    2015-12 – 2017-02

    Score composition, recording, and mixdown for film and media projects.

    Worked with Hans Martin Buff (2× Grammy).

Education

Software & Projects