Maxwell Render | Official public forum

Hi!
I´ve imported a terrain object into MS with corresponding maps at 4096x4096 resolution and tried different subdivision settings for the disp map, from 10, 100, 200 subdivisions. Even tried "Adaptive" setting after reading the documentation, which forced me to abort the render as MR was still calculating the scene after 3 hours, had not even got to voxelization stage! So "Adaptive" setting is ruled out as an option, takes too much time, but if that setting -"locks the subdivision value to the given texture detail (at half pixel accuracy)"- at what subdivision level would that be, and how do you calculate the value at whole pixel accuracy? Only want to know how to get the most detail out of the maps without wasting time on MR calculating "infinity"!?

Bo

It's a good idea to start with small displacement subdivision and work your way up to get the detail you need.

A good result also depends on your geometry being tessellated/subdivided nicely.

The terrain object is a >750K polygon mesh, hopefully dense enough to make nice displacements!? But at a subdivision level of 200 it takes MR about an hour before starting rendering on my machine, so there´s a lot of waiting time for results! That´s why I wondered if there´s a mathematical relation between subdivision and map size!?

If highest quality is reached with the "Adaptive" setting it just might be too time consuming to consider, but I haven´t a clue to how long time it would take for MR to complete all calculations and voxelisation at that setting! I aborted that render after more than 3 hours as I didn´t know if it would go on for days or just some minutes more??The only way to "know" is to be able to estimate a time needed for MR and then decide if the time needed is worth the wait?

Something is not working right. Post the scene so it can be diagnosed. I've never waited for displacement to process for more than a minute.

I don´t think Maxwell or the file is faulty in any way! At low subdivision levels, ie 10, it performs the calculations and voxelization in minutes but with increased levels time increases also! My goal is to understand the logic behind subdivision level and pixels, to be able to choose an optimal setting for time versus quality in displacements!

The hard way to do it would be to set subdivision level to "Adaptive" and just leave the machine to do it´s work, not knowing if it will take weeks or 5 hours, then I could do the math myself, but this alternative is my last resort!

a disp map of 1000 x 1000 pixels would need amax resolution 1,000,000 polys to get you the most of the map

That would be one pixel / one polygon but isn´t Maxwell Studios subdivision a sub polygon displacement tool??? My mesh terrain is 750K polys and the displacement map are about 16,7 million pixels so to get a ratio of 1 to 1 pixel/polygon MS would have to "divide" the mesh further, about 22 times but what subdivision level should I use in MS to get to that point?

with pretess, each subdivision will give you 4 x the polygons amount .

so a 1 million polygon object would need pretess set to 2

1.000.000 x4 = 4.000.0000 x4 = 16.000.000 polys

about

this is with pretess, I tend not to use on the fly coz it takes much more time to render;

h:

Ok, I see! Then it would be sufficient to use a sub division level of only 3 to get my 750K mesh fine enough to use every pixel of the disp map!???

I´ve been using On The Fly, maybe it´s a better idea to use Pretessellation!? Have to check it out!

Thanks all for helping me out with this issue! Will prevent me from over doing tessellations

Bo

yes pretess is a better bet, especially for a ground

h:

If you have enough RAM.

I´ve got 32Gb RAM, so I´ll see if it´s enough or if my machine will choke on that terrain???

It will be enough