Michal Iwanicki

We know Biorthogonal Bases and Parseval Tight Frames allow us to break rules of basis design and still get good results. What else can bend?

Good questions :) Though i think of BiOB and all the PTF math as just consequence of projection minimizing L2 error rather than bending ;)

But should we really focus on looking for basis with more locality than SH and more smoothness than Haar? I'm not convinced

At some point it's less about just reconstruction, more about flexibility

We can reconstruct irradiance efficiently and in practice we don't really need to *reconstruct* radiance

I'd much rather have an efficient convolution with arbitrary kernels than just better reconstruction of generic signals

That's the issue I have with AD - reconstruction is nice, but doing anything other than that loses the local support.

I see zonal SH as a hack when better radial bases possible, Triple Products have stalled on SH, and we forgot all about nearfield GI bounces