VMD-L Mailing List

From: John Stone (johns_at_ks.uiuc.edu)
Date: Tue Oct 12 2010 - 16:24:42 CDT

Hi,
  Indeed, it would be nice to have real-time AO for VMD, but as you
might imagine, it is MUCH harder to do so for a program that allows
trajectory animation (e.g. VMD) than for a program that only displays
static structures like QuteMol. QuteMol does a pre-processing step to
compute it's AO, and from that point on it can re-use the precomputed
result. This only works if the atoms aren't allowed to move. Also, the
specific techniques used in QuteMol only work for spheres, and not for
arbitatry triangulated or volumetric geometry as is the case in VMD.

Although it isn't practical to apply the QuteMol techniques to a program
like VMD, there are still some AO approaches that can be made to work,
with some significant effort. Screen-space ambient occlusion (SSAO)
might be one way to implement real-time AO for VMD. SSAO has varous
flaws vs. true world space ambient occlusion, but it can be made to
run fast even with dynamically changing geometry. SSAO is used in a
number of video games now, so that demonstrates that it can be made
feasible with work... The main issue with SSAO and related techniques
is that they require multiple rendering passes, which adds significant
computational cost and complexity to the VMD rendering code beyond what
people are used to now. I will likely implement multipass rendering
in VMD for some other features like transparency based on depth-peeling
and possibly for some volume rendering techniques, so once that's in
place SSAO becomes much easier to implement.

Anyway, I'm interested in supporting real-time AO in VMD, we just have
to pick an algorithm that works reasonably well and doesn't have such a
high performance cost that it isn't usable on mid-range GPUs. It has been
on my TODO list for a while, but I have a long TODO list presently...

Cheers,
  John Stone
  vmd_at_ks.uiuc.edu

On Tue, Oct 12, 2010 at 10:54:00PM +0200, Ajasja Ljubeti?? wrote:
> Dear VMD developers,
> Just out of curiosity -- are there plans for real-time ambient ambient
> occlusion lighting along the lines of QuteMol?
> I was playing around whit this program today and
> IA immediatelyA wishedA for the power of selectA macrosA and different
> representations of VMD:)
> Best regards,
> Ajasja LjubetiA:*,
> youngA researcher,A
> Department of solid state physics,
> Institute JoAA 3/4ef AA tefan, 

-- 
NIH Resource for Macromolecular Modeling and Bioinformatics
Beckman Institute for Advanced Science and Technology
University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801
Email: johns_at_ks.uiuc.edu                 Phone: 217-244-3349
  WWW: http://www.ks.uiuc.edu/~johns/      Fax: 217-244-6078