Spherical Distributions

Check out this animation of optimization for a near-uniform distribution of nodes on the surface of a sphere. Nodes are modeled as charged particles constrained to the surface of a sphere that mutually repel one another. Each frame is an iteration of the optimization process that minimizes the total potential energy in the system. Iterations are stopped after convergence or 200 iterations.

This is an implementation of the solution in M. R. P. Thomas, “Fast Computation of Cubature Formulae for the Sphere,” in proc. Hands-Free Speech Comm. and Microphone Arrays (HSCMA), San Francisco, USA, Mar. 2017.

Feeling motivated after WASPAA and AES!

I just returned from visiting the 2017 WASPAA workshop and 143rd AES Convention in New York. Two very different vibes: the former is a small workshop focused deeply on the theory on topics including source localization, source separation, acoustic scene analysis, spatial audio, dereverberation and de-noising. AES is a huge gathering of audio professionals from music producers to mixers to hardware manufacturers to researchers like me. I’m back at the office with renewed enthusiasm for my profession!

Pictures here.