
Biography
Axel Roebel is research director at IRCAM and head of the Sound Analysis-Synthesis team (AS). He received the Diploma in electrical engineering from Hannover University in 1990 and the Ph.D. degree (summa cum laude) in computer science from the Technical University of Berlin in 1993. In 1994 he joined the German National Research Center for Information Technology (GMD-First) in Berlin where he continued his research on using artificial neural networks for modeling of time series of nonlinear dynamical systems. In 1996 he became assistant professor for digital signal processing in the communication science department of the Technical University of Berlin.
In 2000, he obtained a research scholarship at CCRMA, Standford University, where he started an investigation into adaptive sinusoidal modeling. In 2000 he joined the Sound Analysis-Synthesis team of IRCAM where he obtained his Habilitation from the Sorbonne Université in 2011 and where he became Research Director in 2013. He has developed state-of-the-art speech and music analysis and transformation algorithms. He is the author of numerous libraries for signal analysis, synthesis, and transformation, for example, SuperVP, a software for music and speech signal analysis and transformation that has been integrated into numerous professional audio tools. His ongoing research focuses on advancing deep learning techniques for tasks related to music and voice processing. This notably includes neural vocoding, the exploration of signal representation and manipulation within latent spaces, as well as the investigation of disentangling strategies in these latent spaces.
Research topics
Voice processing
- speech analysis (F0, voiced/unvoiced, glottal source)
- singing synthesis
- speech transformation - (shape invariant phase vocoder, extended source-filter speech models (PaN), neural vocoder)
- singing voice separation
- deep learning-based speech analysis, processing, and transformation,
- neural vocoder.
Music
- high-quality signal transformation based on the phase vocoder representation
- additive signal models using advanced algorithms for the analysis and representation of non-stationary signals and the development of Pm2, IRCAMs software for sinusoidal analysis/synthesis.
- structured signal models and perceptually pertinent signal descriptors (fundamental frequency, spectral envelope, ...)
- signal decomposition
- polyphonic f0 estimation
Development activities
- Multi-band Excited WaveNet Neual Vocoder (MBExWN)
- ISiS: singing synthesis software written in python.
- as_pysrc: python packages for signal processing
- Deep learning-based signal processing in Tensorflow
- SuperVP: an extended phase vocoder software allowing high-quality transformations of music and speech signals, and implementing new techniques for spectral envelope estimation and transformation. SuperVP is a cross-platform library that is available in form of a command-line application (SuperVP), which is used in AudioSculpt and OpenMusic as well as in form of a real-time signal transformation module, which is used in Max/MSP and SuperVP-TRaX.
- VoiceForger: real-time voice transformation library based on SuperVP.
- Pm2 library and application for analysis/synthesis using advanced sinusoidal signal models
- MatMTL: a matlab compatible c++ template library
- LibFFT: a support library for cross-platform vectorized FFT calculation
Curriculum Vitae
University Degrees
2013 | : | HDR (Habilitation) Computer Science, University of Pierre and Mairie Curie, Paris VI,France. |
1990-1993 | : | Dr.-Ing. Computer Science, Technical University of Berlin, Germany |
1983-1990 | : |
Dipl.-Ing. Electrical Engineering, University of Hanover, Germany |
WORK
01.2017- | : | Research director, IRCAM |
01.2011- | : | Team Leader, Analysis Synthesis Team, IRCAM |
01.2008-12.2011 | : | Adjoint Team Leader, Analysis Synthesis team IRCAM |
10.2000-12.2007 | : | Researcher and Developer, Analysis-Synthesis Team IRCAM |
04.2006-07.2006 | : | Edgar Varese Guest Professor, Electronic Studio, Technical University of Berlin |
04.2000-09.2000 | : | Invited researcher, Center for Computer Research in Music and Acoustics (CCRMA), Stanford University, USA |
01.1996-09.2000 | : | Assistant Professor, Communication Science, Technical University of Berlin |
08.1994-12.1995 | : | PostDoc, GMD FIRST, Berlin. |
Projects
2023-2026 | : | ANR project EVA, Explicit Voice Attributes |
2023-2025 | : | Project DeTOX, Lutte contre les vidéos hyper-truquées de personnalités françaises |
2023-2026 | : | ANR project BRUEL, ElaBoRation d’Une méthodologie d’EvaLuation des systèmes d’identification par la voix |
2023-2027 | : | ANR project ExoVoices, Virtual Story Telling for Kids: Expressive and Cognitive Aspects of Voice Synthesis |
2020-2024 | : | H2020 project AI4Media, Deep learning for media production |
2020-2024 | : | ANR project ARS, Analysis and tRansformation of Singing style |
2017-2022 | : | H2020/ERC project IRiMaS, Interactive Research in Music as Sound. Conseil and collaboration on signal processing methods for music analysis. |
2018-2021 | : | ANR Project TheVoice, Voice creation for media content production. Supervision of PhD thesis on deep learning-based voice conversion. |
2014-2017 | : | ANR Project Chanter, Real-time controlled digital singing. Coordination of WP2 on text to chant synthesis |
2012-2015 | : | ANR Project Physis, Physically informed and semantically controllable interactive sound synthesis. Coordination of WP3 on low level sound representation |
2011-2015 | : | FP7-ICT-2011 Project 3DTVS, 3DTV Content Search. Coordination of WP4 3D Audio & Multi Modal Content Analysis and Description. |
2010-2013 | : | ANR Project Sample Orchestrator II Hybrid Sound Processing and Interactive Arrangement for New Generation Samplers. Coordination of WP2 Structured Instrument Models and Signal Transformations |
2000-2000 | : | DFG Project Ref RO2277/1-1 : Adaptive additive synthesis of non-stationary sounds. Research scholarship at CCRMA |
PhD Students
Directed or Co-directed
2023- | Maximino Linares, Musical instrument audio synthesis via physics-informed neural networks (co-directed with T. Hélie) | |
2023- | Simon Rouard, Control and Adaptation of Deep Learning Models of Music Generation |
|
2023- | Mathilde Abrassart, Voice Identity Conversion with DNN for the Simulation of Voice Identity Usurpation Attacks. (co-directed with N. Obin) | |
2023- | Théodor Lemerle, Text-to-Speech Synthesis for Expressive Storytelling (co-directed with N. Obin) | |
2021- | Lenny Renault, Deep learning-based generation of high-quality music from symbolic music representation. | |
2019-2023 | Frederic Bous, Voice Synthesis and Transformation with DNN | |
2019-2023 | Yann Teytaut, Speech and Singing Alignment and Style analysis with DNN | |
2019- | Antoine Lavault, Drum synthesis with DNN | |
2019-é023 | Clement Le Moine Veillon, Expressive speech transformation with DNN | |
2016-2019 | Hugo Caracalla, Sound texture synthesis from summary statistics, Sorbonne University, 2019 | |
2016-2019 | Céline Jacques, Machine learning methods for drum transcription (in French), Sorbonne University, 2019 | |
2014-2017 | Luc Ardaillon, Synthesis and expressive transformation of singing voice, UPMC, 2017 | |
2012-2015 | Wei-Hsiang Liao, Modelling and transformation of sound textures and environmental sounds, UPMC, 2015. co directed with X. Rodet (IRCAM) and A.Su (NCKU Taiwan) | |
2011-2015 | Stefan Huber, High quality voice conversion by modelling and transformation of extended voice characteristics, UPMC 2015, co-directed with Xavier Rodet | |
2012-2015 | Henrik Hahn, Expressive sampling synthesis: Learning extended Source-Filter models from Instrument sound databases for expressive sample manipulations, UPMC 2015, co directed wih X. Rodet |
Supervised
2009-2012 | Marco Liuni, Automatic adaptation of sound analysis and synthesis, UPMC, 2012, PhD directors X. Rodet and M.Romito | |
2006-2010 | Fernando Villavicencio, High quality voice conversion, UPMC 2010, PhD director X. Rodet | |
2007-2010 | Gilles Degottex, Glottal source and vocal-tract separation, UPMC 2010, PhD director X. Rodet | |
2003-2008 | Chunghsin Yeh, Multiple fundamental frequenc estiation of polypohnic recordings, UPMC 2008, PhD director X. Rodet. | |
PhD/HDR Jury
2022 | Ajinkya Kulkarni (PhD), Expressivity transfer in deep learning based text-to-speech synthesis | |
2022 | Merlijn Blaauw (PhD), Modeling Timbre for Neural Singing Synthesis | |
2022 | Grégoire Locqueville (PhD), Voks: A vocal instrument Family Based on Syllabic Sequencing o Vocal Samples. | |
2022 | Javier Nistal (PhD), Exploring Generative Adversarial Networks for Controllable Musical Audio Synthesis (Examiner) | |
2020 | Muhammad Huzaifah (PhD), Directed Audio Texture Synthesis With Deep Learning (Reporter and examiner) | |
2020 | Alexandre Defossez (PhD), Optimization of fast deep learning models for audio analysis and synthesis. (Reporter and examiner) | |
2019 | Alexey Ozerov (HDR), Contributions in audio modeling for solving inverse problems: Source separation, compression, and inpainting, (Reporter and examiner) | |
2019 | Alice Cohen-Hadria (PhD), Estimation de Descriptions musicales et sonore par apprentissage profond, (Examiner) | |
2019 | Clément Laroche (PhD), Apprentissage de dictionnaire et décomposition orthogonal pour la séparation de sources harmoniques/percussives, PhD, (Examiner) | |
2017 | Benjamin Cohen-Lhyver (PhD), Modulation de mouvements de tête pour l'analyse multimodale d'un envirnonnement inconnu, PhD, (Examiner) | |
2013 | Ricard Marxer (PhD), Audio source separation for music in low-latency and high-latency scenarios, (Reporter and examiner) | |
2013 | Saso Musevic (PhD), Non-stationary sinusoidal analysis, (Reporter and examiner) | |
2013 | Alexis Moinet, (PhD) Slowdio: Audio time-scaling for slow motion sports videos, PhD, (Examiner) |