Master Internship on Machine Learning For Audio Augmented Reality H/F

INRIA | 25 Oct 2024


Détail du poste

Le descriptif de l'offre ci-dessous est en Anglais

Type de contrat : Convention de stage

Niveau de diplôme exigé : Bac +4 ou équivalent

Fonction : Stagiaire de la recherche

Mission confiée
Audio Augmented Reality (AAR) is receiving increasing interest in various application domains, such as artistic creation and cultural mediation, communication and information, and the entertainment industry. Embedding computer-generated or pre-recorded auditory content into a user's real acoustic environment creates an engaging and interactive experience that, for example, can BE applied to video games, museum guides or radio plays. A key challenge is to spatialize sounds in 3D in the user's headset in a way that is consistent with the current environment of the user. For instance, a sound source placed 2 meters away in an office environment should not BE rendered the same way as a source placed 10 meters away in a large museum hall. To achieve this, acoustical information about the user's environment is needed. A key task of the collaborative research project HAIKUS is to retrieve such information solely using microphones embedded in the user's headset.

In this context, this master internship will focus on developing machine learning methods that estimate the volume, reverberation time and/or absorption properties of a room based on multiple binaural recordings of a sound source from different user's positions. The problem of estimating the distance of the sound source will also BE considered.

Principales activités
The intern will perform the following tasks :
- Review of the existing recent literature on related topics, e.g. [1, 2]
- Creation of a simulated audio dataset corresponding to the considered audio augmented reality scenario, using available acoustic simulators such as [3]
- Development, training and testing of a deep neural network for the estimation of acoustical properties from these data, following previous works such as [4] and [5].

The intern will benefit from the strong expertise of the team MULTISPEECH in deep learning, acoustics and audio signal processing methodologies.

Bibliography :

[1] Genovese, A. F., Gamper, H., Pulkki, V., Raghuvanshi, N., & Tashev, I. J. (2019). Blind Room Volume Estimation from Single-channel Noisy Speech. In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2019).

[2] Etienne Thuillier, Hannes Gamper and Ivan J. Tashev. "Spatial audio feature discovery with convolutional neural networks." In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2018).

[3] Steven M. Schimmel, Martin F. Muller, and Norbert Dillier. "A fast and accurate shoebox room acoustics simulator." In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2009).

[4] Diego Di Carlo, Antoine Deleforge and Nancy Bertin. MIRAGE : 2D sound source localization using microphone pair augmentation with echoes. In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2019).

[5] Saurabh Kataria, Clément Gaultier, and Antoine Deleforge. "Hearing in a shoe-box : binaural source position and wall absorption estimation using virtually supervised learning. In "IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2017).

Compétences
Bachelor degree in signal processing or machine learning. Good programming skills in Python and/or Matlab. Previous experience in audio or acoustics is a plus.

Avantages- Subsidized meals
- Partial reimbursement of public transport costs
- Leave : 7 weeks of annual leave + 10 extra days off due to RTT (statutory reduction in working hours) + possibility of exceptional leave (sick children, moving home, etc.)
- Possibility of teleworking (after 6 months of employment) and flexible organization of working hours
- Professional equipment available (videoconferencing, loan of computer equipment, etc.)
- Social, cultural and sports events and activities
- Access to vocational training
- Social security coverage

L'adresse du poste

Localisez l'entreprise et calculez votre temps de trajet en un clic !

Infos complémentaires

None

Finalisez votre candidature

sur le site du recruteur.