Reconnaissance automatique du Human Beatbox et application à la compréhension d'un langage musical évolutif // Automatic recognition of Human Beatbox and application to the understanding of an evolving musical language
ABG-131536
ADUM-65500 |
Thesis topic | |
2025-04-29 | Public funding alone (i.e. government, region, European, international organization research grant) |
Université Grenoble Alpes
Saint Martin d'Hères cedex - Auvergne-Rhône-Alpes - France
Reconnaissance automatique du Human Beatbox et application à la compréhension d'un langage musical évolutif // Automatic recognition of Human Beatbox and application to the understanding of an evolving musical language
- Biology
parole, beatbox, reconnaissance automatique
speech, beatbox, automatic recognition
speech, beatbox, automatic recognition
Topic description
Ce projet de thèse de doctorat explore la parole beatboxée comme un langage musical émergent et en rapide évolution qui repousse les limites physiologiques et phonétiques de la parole humaine. Il s'appuie sur les approches computationnelles récentes et les technologies à la pointe en phonétique expérimentale pour i) développer des outils innovants permettant de reconnaître et d'annoter automatiquement la parole beatboxée ; ii) élaborer un modèle linguistique du Human Beatbox ; iii) accompagner son usage thérapeutique dans la prise en soin orthophonique. Du point de vue informatique, ce projet aborde les questions liées aux modèles auto-supervisés (acoustiques et grands modèles de langue) et au transfert d'apprentissage à partir de grands modèles musicaux généraux vers le Human Beatbox. Au niveau des sciences de la parole, il questionne les aspects linguistiques du beatbox et permet l'étude de la culture beatbox depuis les années 80. Il contribue à son usage comme outil de rééducation en orthophonie. A son terme, ce projet mettra à disposition de la communauté i) de grands modèles de langue et acoustiques auto-supervisés permettant de transcrire automatiquement les sons beatboxés ; ii) des corpus annotés permettant de développer la recherche autour de cet art vocal ; iii) des applications libres autour du beatbox (reconnaissance, exercices, suivi de l'évolution du beatbox).
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
This PhD project explores beatboxed speech as an emerging and rapidly evolving musical language that pushes the physiological and phonetic limits of human speech. It draws on recent computational approaches and cutting-edge technologies in experimental phonetics to i) develop innovative tools for automatically recognizing and annotating beatboxed speech; ii) elaborate a linguistic model of Human Beatbox; iii) support its therapeutic use in speech therapy. From a computational point of view, this project addresses issues related to self-supervised models (acoustic and large language models) and to the transfer of learning from large general musical models to Human Beatbox. In terms of speech sciences, it questions the linguistic aspects of beatbox and enables the study of beatbox culture since the 80s. It also contributes to its use as a rehabilitation tool in speech therapy. When completed, this project will make available to the community i) large self-supervised language and acoustic models for the automatic transcription of beatboxed sounds; ii) annotated corpora for the development of research into this vocal art; iii) free beatbox applications (recognition, exercises, monitoring of beatbox evolution).
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Début de la thèse : 01/10/2025
WEB : https://cloud.univ-grenoble-alpes.fr/s/Y7CHNxGDPPFDfqL
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
This PhD project explores beatboxed speech as an emerging and rapidly evolving musical language that pushes the physiological and phonetic limits of human speech. It draws on recent computational approaches and cutting-edge technologies in experimental phonetics to i) develop innovative tools for automatically recognizing and annotating beatboxed speech; ii) elaborate a linguistic model of Human Beatbox; iii) support its therapeutic use in speech therapy. From a computational point of view, this project addresses issues related to self-supervised models (acoustic and large language models) and to the transfer of learning from large general musical models to Human Beatbox. In terms of speech sciences, it questions the linguistic aspects of beatbox and enables the study of beatbox culture since the 80s. It also contributes to its use as a rehabilitation tool in speech therapy. When completed, this project will make available to the community i) large self-supervised language and acoustic models for the automatic transcription of beatboxed sounds; ii) annotated corpora for the development of research into this vocal art; iii) free beatbox applications (recognition, exercises, monitoring of beatbox evolution).
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Début de la thèse : 01/10/2025
WEB : https://cloud.univ-grenoble-alpes.fr/s/Y7CHNxGDPPFDfqL
Funding category
Public funding alone (i.e. government, region, European, international organization research grant)
Funding further details
Concours pour un contrat doctoral
Presentation of host institution and host laboratory
Université Grenoble Alpes
Institution awarding doctoral degree
Université Grenoble Alpes
Graduate school
216 ISCE - Ingénierie pour la Santé la Cognition et l'Environnement
Candidate's profile
connaissance en reconnaissance automatique de la parole; connaissance en phonétique/linguistique
knowledge of automatic speech recognition; knowledge of phonetics/linguistics
knowledge of automatic speech recognition; knowledge of phonetics/linguistics
2025-05-23
Apply
Close
Vous avez déjà un compte ?
Nouvel utilisateur ?
More information about ABG?
Get ABG’s monthly newsletters including news, job offers, grants & fellowships and a selection of relevant events…
Discover our members
ONERA - The French Aerospace Lab
CASDEN
ASNR - Autorité de sûreté nucléaire et de radioprotection - Siège
SUEZ
TotalEnergies
MabDesign
Ifremer
Tecknowmetrix
Nokia Bell Labs France
ANRT
PhDOOC
MabDesign
Institut Sup'biotech de Paris
Généthon
Aérocentre, Pôle d'excellence régional
CESI
ADEME
Laboratoire National de Métrologie et d'Essais - LNE
Groupe AFNOR - Association française de normalisation