About Me

Clément Gaultier

Clément Gaultier

I work as a Research Fellow at the University of Cambridge, studying with a multidisciplinary approach, new speech enhancement strategies able to cope with noise and reverberation for cochlear implants users. My research topics mainly comprise solving acoustic and audio signal processing inverse problems. During my PhD I had been working on declipping and denoising using analysis/synthesis sparse priors. I also worked on binaural sound source localization with massive regressions machine learning techniques. Additionally, I am really interested in auditory related topics such as augmented hearing, or hearing protection.

My CV


You will find below a short description of my last positions and academic degrees.

MRC CBU, University of Cambridge

From December 2021, I have been working as a postoctoral researcher at the MRC CBU (University of Cambridge) thanks to a \"Fondation Pour l'Audition\" Fellowship. The project relies on a combination of auditory-inspired machine learning, multi-channel signal processing and accurate room acoustics for building an improved speech enhancement strategy able to tackle the joint problem of noise and reverberation for cochlear implants users.

Current position
Research Fellow

Orange Labs

From September 2019 to February 2021, I worked as a postoctoral researcher with Orange Labs (R&D division of the Orange group) on multichannel acoustic echo cancellation for distributed audio systems and interpersonal communication.

Sept. 2019 - Feb. 2021
Postdoctoral researcher

PANAMA team (IRISA - Inria/CNRS)

From February to August 2019, I worked as a research engineer for Inria mainly on audio restoration algorithms (multichannel declipping, pop/burst noise removal...).

Feb. 2019 - Aug. 2019
Research engineer

PANAMA team (IRISA - Inria/CNRS)

In January 2019, I graduated from Université de Rennes 1 where I did my PhD in the IRISA lab (the Panama team - a joint project team between Inria and CNRS). My main research interests were inverse problems in audio & acoustic signal processing.

Nov. 2015 - Jan. 2019
PhD Candidate

Hearing and Balance Centre (ISVR - Southampton University)

During this research stay, I studied auditory adaptation to noise for improved speech perception. The focus was finding noise characteristics and aclimatization time-scales for enhanced speech perception. For that I designed and conducted different listening experiments on 33 volunteers with normal hearing.

Mar. 2015 - Sep. 2015
Engineer in training / Master thesis

French Graduate Engineer (ENSIM)

In November 2015, I graduated from ENSIM, an engineering school in 5 years under the authority of the French Ministry of Education and Research. I specialized in Acoustics, Vibrations, Sensors.

Nov. 2015
Post graduate Student

University of Le Mans Master of Science (Research M.Sc in Acoustics)

In November 2015, I also graduated from Le Mans Université, as I was following a double degree (Research M.Sc specialized in acoustics).

Nov. 2015
Master Student
Download my detailed CV

Publications


You will find below a list of my scientific publications & communications.

2023

  • Perceptual analysis of speaker embeddings for voice discrimination between machine and human listening

    Iordanis Thoidis, Clément Gaultier, Tobias Goehring
    2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) IEEE.

2022

  • Deep learning strategies for compensation of noise and reverberation: single- vs multi-microphone approaches and applications to cochlear implants.

    Clément Gaultier, Tobias Goehring
    Hearing, Audio and Audiology Sciences Meeting, Southampton, UK (Sept. 2020)

2021


  • Double-Talk Robust Acoustic Echo Cancellation Using Partition Block Frequency-Domain Adaptive Filtering.

    Clément Gaultier, Alexandre Guérin, Grégory Pallone & Marc Emerit
    29th European Signal Processing Conference (EUSIPCO), (pp. 171-175). EURASIP, (August 2021) [link]

  • Sparsity-based audio declipping methods: selected overview, new algorithms, and large-scale evaluation.

    Clément Gaultier, Srđan Kitić, Rémi Gribonval & Nancy Bertin
    IEEE/ACM Transactions on Audio, Speech and Language Processing, IEEE, vol. 29, pp. 1174-1187, 2021.
    [arXiv][HAL] [Demo][Code]

2020

  • A Comparative Study of Multilateration Methods for Single-Source Localization in Distributed Audio.

    Srđan Kitić, Clément Gaultier & Grégory Pallone
    27th FRUCT Conference, International Workshop on the Internet of Sounds (Sept. 2020)[arxiv]

2019

  • Design and evaluation of sparse models and algorithms for audio inverse problems.

    Clément Gaultier
    Université de Rennes 1 (Jan. 2019)[TEL][Code]

2018

  • Evaluation of an open-source implementation of the SRP-PHAT algorithm within the 2018 LOCATA challenge.

    Romain Lebarbenchon, Ewen Camberlein, Diego di Carlo, Clément Gaultier, Antoine Deleforge & Nancy Bertin
    2018 IEEE-AASP Challenge on Acoustic Source Localization and Tracking (LOCATA)
    International Workshop on Acoustic Signal Enhancement (Sept. 2018)[arxiv]

  • CASCADE: Channel-Aware Structured Cosparse Audio DEclipper.

    Clément Gaultier, Nancy Bertin & Rémi Gribonval
    2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP),. IEEE. (April 2018) [HAL]

2017


  • AUDASCITY: AUdio Denoising by Adaptive Social CosparsITY.

    Clément Gaultier, Srđan Kitić, Nancy Bertin & Rémi Gribonval
    25th European Signal Processing Conference (EUSIPCO), (pp 1305-1309). EURASIP. (August 2017) [HAL][Code]


  • Cosparse denoising: The importance of being social.

    Clément Gaultier, Srđan Kitić, Nancy Bertin & Rémi Gribonval
    The Signal Processing with Adaptive Sparse Structured Representations (SPARS) workshop. (abstract, June 2017) [HAL][Code]


  • VAST: The Virtual Acoustic Space Traveler dataset.

    Clément Gaultier, Saurabh Kataria & Antoine Deleforge
    International Conference on Latent Variable Analysis and Signal Separation (pp. 68-79). Springer, Cham. (February 2017) [arXiv][HAL]


  • Hearing in a shoe-box: binaural source position and wall absorption estimation using virtually supervised learning.

    Saurabh Kataria, Clément Gaultier & Antoine Deleforge
    2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), (pp. 226-230). IEEE. (March 2017) [arXiv][HAL]

2015

  • Acclimatization to different English accents for enhanced speech intelligibility in noise in individuals with normal hearing.

    Rakshita Gokula, Clément Gaultier, Jessica J. M. Monaghan & Stefan Bleeck
    Basic Auditory Science Meeting, British Society of Audiology. (abstract, September 2015)

Working papers

  • A modeling and algorithmic framework for (non) social (co) sparse audio restoration.

    Clément Gaultier, Nancy Bertin, Srđan Kitić & Rémi Gribonval
    [arXiv][HAL] [Demo] [Code]


My Projects


Below are listed some of the projects I am working on. Click on the name to visit the associated webpage.

Clément Gaultier

Ci-Fi 🌀 Experience

A popularization webpage/app demo about cochlear implants, initially designed for the Cambridge Science Festival (2022)

Clément Gaultier

Deep Hearing Lab

Dr Tobias Goehring's research lab with a focus on speech perception with hearing devices (cochlear implants, hearing aids and hearables) using techniques from auditory neuroscience, engineering and deep learning.

Clément Gaultier

VAST

VAST (The Virtual Acoustic Space Traveler) is a research project aming at providing tools for sound source localization using virtually supervised leanring.

Clément Gaultier

A-SPADE

A-SPADE (Analysis SParse Audio DEclipper) is a research project that developped an audio declipping procedure using analysis sparse modeling priors.

Clément Gaultier

JSM

JSM (Journée Science et Musique) is a science popularization open day about science and musique organized every year by the PANAMA team (IRISA research center, Rennes (France)).

Clément Gaultier

SPARE

Spare (SParse Audio REstoration) is a research project aming at proposing a new modeling and algorithmic framework to tackle audio reconstruction issues like denoising and declipping.

Clément Gaultier

Moderncv Template

A latex CV template using the moderncv class with Font Awesome 5 icon support.