Vision-Steered Beam Forming and Transaural Rendering for the Artificial Life Interactive Video Environment, (ALIVE)
Perceptual Computing Technical Report #352

Michael A. Casey, William G. Gardner, Sumit Basu
MIT Media Laboratory, Cambridge, USA
mkc,billg,sbasu@media.mit.edu

Abstract:

This paper describes the audio component of an interactive video system that uses remote sensing to free the user from body-mounted tracking equipment. Position information is obtained from a camera and used to constrain a beam-forming microphone array, for far-field speech input, and a two-speaker transaural audio system for rendering 3D audio.

• Vision-Steered Beam Forming
  • Introduction
  • Free-field speech recognition
  • Adaptive Beamforming
  • Fixed Beamforming
  • A Visually Constrained Beamformer
  • Fixed Beamformer Algorithms
  • Designing the Array
  • Conclusion
• Visually Steered 3-D Audio
  • Introduction
  • Principles of binaural spatial synthesis
  • Implementation of binaural spatializer
  • Performance of binaural spatializer
  • Principles of transaural audio
  • Implementation of transaural filter
  • Performance of combined system
  • Adding dynamic tracking
  • Preliminary results
  • Conclusions
• References
• About this document ...