Realtime sound rendering 

My name is Raphael Sebbe. I am electrical engineerfrom both FPMs and Supelecas of June  2000.
I am currently working for BabelTechnology SA, a voice technology company based in Belgium. There,I am responsible of the Macintosh and Linux products.

You can reach my website at

This work was first initiated by MetacreationsEurope team and me, in the form of a Carrara (their latest 3D software)extension, under Prof. Chapelle control. As Metacreations retired fromthe 3D market, closed its European R&D centers and sold most its softwaresto other companies in December 1999, the work has been re-targeted towardsmy custom computer graphics project, ThreeD at the time (now Sphaera),to provide it sound capabilities.

Faculty members in charge of this project :


The goal of this thesis work is to conceive a real-time acoustics simulationsystem. The elements of the simulation come from a virtual environment,i.e. an environment stored as computer data.
The system computes the sound perceived through a receiver, the resultbeing a function of the location of the receiver and sources, their characteristicsand the environment.

The work has been done as an extension of an existing computer program,ThreeD, that has been developed as a spare-time hobby. The programminglanguages used are C / Objective-C and the project runs on MacOS X operatingsystem by Apple. It is based on the state-of-the-art object oriented environmentnamed Cocoa (OpenStep successor).

The project deals with both theory, through the development of the soundrendering algorithms and the object oriented design, and practise, throughits implementation.

Emitter and receiver modeling takes directional gain into account. Thedeveloped method allows the rendering of distance attenuation, Dopplereffect and room reverberation components. The real-time constraint meansthat interactive experience is made available to the end user, but requiressimplifying hypotheses leading to fast algorithms.

Graphical user interfaces have been realized too. Their goal is to alloweasy manipulation of the parameters. These integrate transparently insidethe host (ThreeD) interfaces.

Virtual reality is, of course, a direct application of this system.

Here are examples of sounds computed with the system. They are mostlygiven in the form of quicktime movies; the images provide a better understandingof the various configurations.

These examples show directivity considerations.

Next, an example illustrating the artificial reverberation algorithm(s)implemented.Last but not least, a configuration showing Doppler's effect.
 text / presentation

You can find the text here (in French). Both Postscriptand PDF versions available (theyactually look better printed than on screen).

last updated : December 16th, 2000