Volume Rendering

Volume Rendering Resources

{ 4 comments… read them below or add one }

1 Jayandera Danappal January 20, 2012 at 3:01 pm

Hi Magnus,

I was reading the “Capturing Thin Features in Smoke Simulations” of yours. There a few questions that I would like to ask you if you don’t mind. It starts after the particle’s density transferred to on a high resolution grid.

“Diffusion and other effects are computed on the high resolution voxel grid and changes to the voxel grid representation are applied back to the particles using the FLIP technique (using only the derivate of smoke density). Our technique has very low numerical diffusion, as forward advection of particles is diffusion-free and the FLIP update has very little inherent diffusion (depending on the FLIP blending parameter).”

I am not sure if I understand that correctly. Is it right that there are two FLIP update overall, one for velocity and another for density?…or is the density uses a standard advection method driven by the FLIP velocity?

“…the technique is often 20-30 times faster than semi- lagrangian advection at similar resolution…”

On that, are you comparing the traditional velocity update vs FLIP velocity update?

It will be great if you can answer if you had a chance.



2 magnus January 21, 2012 at 11:36 am

Hi Jayandera,

Yes, we do perform two FLIP-style updates, one for the velocity-carrying particles and one for the density-carrying ones. We don’t use “standard advection”.

Regarding the speed, the 20-30x speedup was a average based on our work on Alice in Wonderland. Because the density-carrying particles are added only where it is needed, and the wispy smoke simulations we were doing were quite sparse, any given volume would be filled only to around 5%, sometimes even less. In these cases, advecting particles was drastically faster than the comparisons we did where the density field was advected on a discrete grid using semi-lagrangian techniques. So it only refers to the advection of the density field, not the advection of the velocity field. Remember, our velocity fields were of much lower resolution than the density fields.

3 Jayandera Danappal January 23, 2012 at 1:30 am

Hi Magnus,

Thank you for explaining. It’s much clear now.



4 Theopisti Stylianou-Lambert September 15, 2012 at 4:25 am

Dear Magnus,

I’m a photographer and academic working on a book chapter about Martin Parr’s book «Small World». In my text I mention the Flickr group you initiated «the Pisa Pushers». Would it be possible to have your e-mail in order to ask you a few questions?

Hope we can manage to communicate.

All the best,
Theopisti Stylianou-Lambert

Leave a Comment