Archive

Posts Tagged ‘frost’

Frost vs Realflow mesh test

May 6, 2011 1 comment

In this test I’m checking if it’s possible to achieve a flicker free mesh of a high viscosity Realflow simulation using very low res values while meshing with Frost and comparing the result to standard Realflow meshes.

Normally in Realflow high viscosity values (100 to 400) makes the simulation highly unstable (particles explode) and to keep a smooth interpolation between frames you need to increase the min-max substeps significally. If you need the realflow mesh to keep enough details it needs to have at least 1-2milllion particles. To keep such a high count from exploding the min-max substeps need to be raised accordingly and kept at least at 2000min to 3000max which also translates to days of simulation time…

Such high values will produce a great looking realflow mesh, but if your shot is in close up and rendered in HD there’re good chances the mesh will still flicker heavily and won’t pass any broadcast standards. The cause is due to other friction values, small global scale, etc. Common solution is just to blur it in comp. If that’s not possible then you have no choice but to sim at a higher scale and tweak more values of that long simulation which is a real pain….

Thinkbox’s FROST finally makes it possible to mesh a smooth flicker free mesh even from quite a low resolution simulation which means a HUGE time saver in production time.

Categories: Portfolio, RND