Full Professor in Acoustics

logo_cnam_losange Laboratoire de Mécanique des Structures et des Systèmes Couplés

Avatar

Éric Bavu

Full Professor

LMSSC

Cnam Paris

I am a full professor at CNAM Paris, working within the AVA Team (Acoustics, Vibrations, and Audio) at the Laboratory of Structural Mechanics and Coupled Systems .

To date, I have supervised or co-supervised 10 doctoral theses (8 defended, 2 in progress), and have led the ANR Deeplomatics project until June 2022.

I defended my HDR in December 2019. My habilitation thesis has been deposited on HAL , and I have made available a video with chapters of the defense as well as the slides in html format.

My research activities focus on deep learning for acoustics and audio, inverse problems in acoustics, and multichannel signal processing.

Interests

  • Deep learning applied to acoustics and audio
  • Inverse problems in the time domain
  • Source localization
  • Reverberant and noisy environments
  • 3D audio
  • Transient elastography (postdoctoral research)

Education

  • Accreditation to Supervise Research, Acoustics, 2019

    Conservatoire National des Arts et Métiers (FR)

  • PhD thesis in Mechanical Engineering, Acoustics Specialty, 2008

    Cotutelle University of Sherbrooke (Qc, CA) and University Paris 6 Pierre et Marie Curie (FR)

  • Master in Acoustics, Signal Processing and Computer Science (ATIAM), 2005

    Pierre et Marie Curie University (FR)

  • Teacher training degree (french Agregation) Physics, 2004

    École Normale Supérieure de Cachan (FR)

  • Bachelor and Master of Fundamental Physics, 2002-2003

    École Normale Supérieure de Cachan / Université d'Orsay (FR)

Selected recent and upcoming talks

Body conducted speech capture and enhancement using generative deep learning

Body conducted speech capture and enhancement using generative deep learning.(Spectre Workshop 2026)

Bringing interpretability to neural audio codecs

Interpreting neural audio codec tokens via analysis–synthesis modeling (Interspeech 2025)

EBEN - Extreme bandwidth extension network applied to speech signals captured with noise-resilient body-conduction microphones

Body conducted speech enhancement using a dedicated GAN architecture (ICASSP 2023)

Recent publications

(2025). Real-time speech enhancement in noise for throat microphone using neural audio codec as foundation model. 2025 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics (WASPAA).

PDF Project Poster

(2025). Bringing Interpretability to Neural Audio Codecs. Interspeech 2025 - 26th edition of the Interspeech Conference.

PDF Project DOI arXiV

(2025). Vibravox: A Dataset of French Speech Captured with Body-conduction Audio Sensors. Speech Communication.

PDF Project DOI Site web Code HuggingFace dataset arXiV

(2025). Donner du sens aux Codecs Neuronaux : Interprétabilité des Tokens discrets produits pour des Signaux Vocaux. CFA 2025 - 17e Congrès Français d’Acoustique.

Project

(2025). Synthèse ambisonique de champs rayonnés par des pavillons, intégrant un modèle de directivité. CFA 2025 - 17e Congrès Français d’Acoustique.

Project

(2025). Interpolation et extrapolation de réponses impulsionnelles par krigeage dans des salles de concert. CFA 2025 - 17e Congrès Français d’Acoustique.

Project

Experience

 
 
 
 
 

Full Professor

Laboratoire de Mécanique des Structures et des Systèmes Couplés

Sep 2023 – Present Cnam Paris

Research Activities: Deep Learning for Acoustics and Field Reconstruction, Speech Enhancement, 3D Audio, Localization and Recognition of Moving Sources


Funded Research Projects: ANR Damier (2024-2027), ANR AHEAD (2020-2026), DGA RAPID Deeplodocus (2025-2027), ANRT contracts with L-Acoustics (2023-2026) and Essilor-Luxottica (2025-2028)


Thesis supervision: Noé Guttierez (2025-…), Louis Reine (2023-…), Argan Verrier (2023-2024), Julien Hauret (2022-2025), Philippe Thorner (2022-2025)


Courses taught in acoustics and signal processing for initial training and CNAM engineering courses in evening classes, in person and via distance learning. Students from Bac+1 to Bac+5.

 
 
 
 
 

Senior Lecturer (HDR)

Laboratoire de Mécanique des Structures et des Systèmes Couplés

Sep 2009 – Present Cnam Paris

Research Activities: microphone array processing, inverse problems in temporal acoustics, deconfinement, localization of moving sources, deep learning in acoustics for source localization and sound source recognition.


Funded research projects: ANR Deeplomatics (2019-2022), ANR PARABAS (2009), FUI LICORVE (2009-2012), LSC-BIOAC Project (2016-2018)


Thesis co-supervision: Hadrien Pujol (2017-2020), Aro Ramamonjy (2015-2019), Guillaume Mahenc (2013-2016), Stéphanie Lobréau (2011-2015), Yacine Braikia (2009-2012)


Teaching in acoustics and signal processing for initial training, apprenticeship diplomas, and Cnam engineering training in evening classes, face-to-face and distance learning. Audiences from 1-year to 5-year Higher Education courses.

 
 
 
 
 

Post-doc CNRS

Institut Langevin, Équipe Physique des Ondes pour la Médecine

Dec 2008 – Aug 2009 ESPCI Paris
Transient and broadband elastography by ultrasound imaging method Supersonic Shear Imaging for the detection of liver fibrosis
 
 
 
 
 

PhD thesis

Institut Jean le Rond d’Alembert / Groupe d’Acoustique de l’Université de Sherbrooke

Sep 2005 – Oct 2008 Paris / Sherbrooke
High resolution focusing and imaging using time reversal in the audible domain. Development a digital time-reversal sink for imaging, with applications to localization and characterization of unsteady acoustic sources.

Research Projects

Contracts and supervised PhDs

*

Vibravox

A general purpose dataset of speech captured with body-conduction transducers

Virtual Sensing for Large-Scale Sound Reconstruction

Louis Reine’s PhD thesis (2024–…)

ANR ASTRID project ‘DAMIER’

DGA funding 2025-2028

3D Sound Field Synthesis and Directional Source Auralization

Philippe Thorner’s PhD thesis (2022- 2026)

Bandwidth extension of speech signals

Julien Hauret’s PhD thesis (2022- 2025)

Sound recognition using Deep Learning

TimeScaleNet, a multi-resolution time domain architecture for speech and environmental sound recognition

ANR ASTRID project ‘Deeplomatics’

Project leader - DGA funding 2019-2022

Acoustic sources in supersonic motion

Guillaume Mahenc’s PhD thesis (2014-2017)

Contact

  • +33 1 40 27 21 66
  • Cnam, 2, rue Conté, Paris, 75003
  • Laboratoire de Mécanique des Structures et des Systèmes Couplés, Accès 31, Bureau 31.0E.02