AmbiSystem

A modular system for 3D-audio rendering in SuperCollider

Dirk Beckmann, Master's Thesis (German Diplom), supervised by Till Bovermann, Thomas Hermann and Helge Ritter, Faculty of Technology, Bielefeld University, Germany, 2007

Download (pdf, 2.5Mb),

Summary

In this thesis the AmbiSystem for 3D-audio rendering in SuperCollider has been developed. The design of the system allows multiple sound sources to be calculated in realtime for one listener position. Apart from positioning sound sources in three-dimensional space, the reflections of these sources within a virtual room can be taken into account. The soundfield descriptions use an extended Ambisonics-format (The AAB-Format, which was developed in this thesis) and therefore can easily be played back on different speaker setups or - via headphones through convolution with the appropriate HRIRs. Furthermore a visualisation of the 3D-scene can be used to navigate the representation of the listener for whom the soundfield is created.

Screen Shot

This screenshot shows the basic signalflow of the AmbiSystem and a SuperCollider session script demonstrating its typical usage. An instance of the AmbiSystem class and a synth definition (lines 10-20) is created, so that sound sources can be instantiated with a given location inside a virtual room (line 24). Additionally in this session the AmbiSystem is used for selecting the required AmbiDecoderSynth for headphone-playback (line 3), adjusting the virtual room (lines 28-35) and setting up an AmbiOSCReceiverFunction (line 41) to control the listener representation from an external application. The ambiControlBus (aCB) depicted in the signalflow graph informs all AmbiEncoder/Decoder-UGens about such changes in real time.

Examples

The two videos below show demos of the AmbiSystem. In the first video three sound sources inside a virtual room can be heard and seen, the second video is a simple sonification of the IRIS data set where the radius of emitting sound sources is limited to focus the listeners attention to the nearest data points. The accompanied sound is binaural and therefore should be listened to via headphones.

Example 1: (Video (mp4), 3500kB)
Example 2: (Video (mp4), 1600kB)

References

Thomas Hermann, (Sonification), 2007
James McCartney, (SuperCollider)
Introduction to (Ambisonics)

Contact

For questions and comments please contact Dirk Beckmann, Till Bovermann or Thomas Hermann.