News

École polytechnique researchers develop 3D sound spatialization software

October 18, 2017 - Media of the future

The X-Audio team at the Centre de Mathématiques Appliquées de l'École polytechnique (CMAP, co-supervised with CNRS and part of the Carnot Télécom et société numérique) has developed free binaural rendering software for sound professionals and the general public.

The result of a collaboration with the Conservatoire National Supérieur de Musique et de Danse de Paris, the software developed by X researchers enables sound recordings containing up to 48 channels to be mixed and listened to using only headphones. By manually defining the position of the sound sources, and while tracking the position of the head, the MyBino software enables the user to perceive the movement of sound in space (a plane flying over the listener, for example).

To achieve this, CMAP researchers use complex mathematical calculations to model precisely the functions involved in the subjective construction of a sound space(Head-Related Transfer Function and Room Impulse Responses). By numerically calculating the circulation of sound waves in a defined environment (mixing studio, auditorium, historical monument, etc.), CMAP sets out to reproduce the "sound field", i.e. the characteristics of acoustic waves at every point in the space, particularly in the hollow of the auricle.

The sound rendering quality achieved by MyBino opens up a wide range of applications. Firstly, it offers an affordable alternative to sound engineers who can work on their multi-channel mixes with headphones, making up for, but not replacing, conventional multi-speaker setups. It also gives listeners easier access to these mixes. Secondly, this software is ideally suited to the audio rendering that can accompany virtual reality devices. It makes the virtual universe more realistic, while limiting the effects of cybersickness (nausea caused by the arrival of contradictory visual and auditory information in the brain). Finally, when combined with a spatial localization device, MyBino can guide people using sound signals.

Experiments carried out in partnership with the "MixHandi Cap sur la vie" association have already enabled Stéphane le Sueur, a blind athlete, to run and rollerblade following a spatialized sound source. This research could enable the visually impaired to take part in sporting activities with greater autonomy, thus responding to a major social challenge.

Creation of a "software" consortium

Ultimately, the software developed by X researchers, in conjunction with the Conservatoire National Supérieur de Musique et de Danse de Paris, can be used in a wide variety of applications. This is why the X-Audio team is proposing to potential industrial or academic partners to join a software consortium currently being set up, in order to finance the maintenance and evolution of the software, thus defining the development priorities for the X-Audio team.

Latest news

,

[BELLE HISTOIRE] AI to optimize robot-assisted knee osteoarthritis surgery

As part of a thesis conducted with Ganymed Robotics and LaTIM (a laboratory under the joint supervision of IMT Atlantique, a component school of the Carnot TSN institute), Anna Gounot is developing AI models capable of predicting the state of knee cartilage from scanner images, in order to improve the precision of prosthesis fitting to treat osteoarthritis.

[VIDEO] Hadaptic Evident: an experimental platform at the heart of digital health

At the crossroads of digital technologies, healthcare and applications, the Hadaptic Evident platform at Télécom SudParis, a component school of the Carnot TSN institute, is a unique experimentation and co-innovation facility. Awarded the label of the Carnot institute Télécom & Société numérique, it is a concrete illustration of the ability of academic research to respond to the major societal challenges linked to autonomy, ageing and well-being.
, ,

[BELLE HISTOIRE] Using graphs and deep learning to make recruiters' lives easier

As part of a CIFRE thesis involving the Easy Partner recruitment agency, the SAMOVAR laboratory at Télécom SudParis, a component school of the TSN Carnot institute, and Efrei, Éric Behar has developed a tool to help recruiters. Thanks to a graphical representation and a deep learning model, his recommendation system can identify relevant candidates for a job offer and, conversely, offers corresponding to a candidate.

Need more information?

© 2022 Carnot Télécom & Société Numérique | Legal Notice