3D-Audio is an interesting for telephone conferencing.
In a development and research project at TU Berlin, we enhanced Asterisk's
conferencing bridge with binaural synthesis.
- Modified internal signal processing to be capable handling interleaved audio (channel, translate)
- Extended confbridge (precisely bridge_softmix) by binaural synthesis (basically added a convolution engine using libfftw3)
Note: confbridge remains capable to mix for mono channels (default mixing) and binaural synthesis in parallel
Practical details (can be changed):
- Required HRTFs are at the moment included as header file
- Confbridge is at the moment only running with binaural synthesis at 48kHz
- Only one 3D-Layout per confbridge is rendered (all participants receive the same spatial representation; one's own signal removed)