[-Adrian]

Very cool, i like it.

[polynurb]

Hello again,

I`ve completed the Rough_Wall material and uploaded it to the MXM gallery;

you should be able to get it here:

http://mxmgallery.maxwellrender.com/new ... =0&id=3464

I noticed that adaptive displacement voxelizes with good speed on flat surfaces (2-4 poly only!), but it won`t be useable on curved high poly objects. (for the simball preview I used precision 12 only!)
In such a case precision has to be tweaked individually for the basemeshes.

It`s actually the first material I "dare" to upload; I hope you like it!

cheers,

daniel

[tom]

I noticed that adaptive displacement voxelizes with good speed on flat surfaces (2-4 poly only!), but it won`t be useable on curved high poly objects. (for the simball preview I used precision 12 only!)
In such a case precision has to be tweaked individually for the basemeshes.

Adaptive mode will pick the value according to your base mesh. For example, if adaptive mode internally set the optimal precision as 256 for a 512 x 512 map on a 1 x 1 poly base, it will internally set the precision to a lower value when you subdivide the base mesh. So, as long as you subdivide the initial mesh, adaptive will adapt the required precision internally and it will speed up your render while keeping the same quality.

[Bubbaloo]

It would be good if there was a way we could see the final number that adaptive used for the render. Or a seperate "precision calculator" that will give you the optimum number, and you can use that for other renders where the base mesh doesn't change.
I have found that using "adaptive" creates longer pre-calculation time for the render.

[Fernando Tella]

Is there a way to tell more or less how much precision is need for a specific mesh? What does the value mean?

[polynurb]

Thank you for the explaination Tom!

what I noticed when I did testrenders with the material,
was that when I only have a simple wall(box) voxelation took
about 3-4 sek. on 640x480.
I played around with the fillet of the edges (I work in Rhino) to get rid of
displacement artifacts on the edges; welding over 90deg. in Rhino, corrupts
the texture mapping completly, so I tried avoiding sharp corners with edge fillets.
Generally the Wall with fillets took a lot longer to voxelize.
To me it seemed to be linked to the tesselation of the nurbs surfaces.
During that period the CPU usage was actually 25% on the Quad.

[tom]

I have found that using "adaptive" creates longer pre-calculation time for the render.

It's an illusion. Because, usually there's always an acceptable custom precision value and checking adaptive is asking for more precision, that's why. So, it doesn't mean adaptive is slower. When you catch the same benchmark with adaptive using custom precision, it means that's the matching point.

Fernando, I'll reply with quoting myself from another thread:

This was also requested during the test period but it's logically not possible because the precision is relative to base mesh. Therefore there's no equation for finding the matching amount of precision with adaptive. For example, say you have 1 x 1 plane and 256 x 256 pixel map and the corresponding precision doing the same with adaptive is 512, but if you use 32 x 32 plane and same texture, the adaptive will still do the same (although in less time) while the corresponding precision would be something like 16. This example is more complicated and impossible to compute for an irregular mesh. Although, if we'd done something like that the material should auto-alter this value depending on the mesh it's been applied which makes it impossible to use the material for more than one object in the scene. If you're still looking for the corresponding precision, I'd suggest you checking the initial benchmark and increase/decrease the precision until it's closer.

I know it's boring to read the manual but we've put valuable information and suggestions all about these in those pages. I'd strongly suggest reading the displacement section before you start.

[Fernando Tella]

...if we'd done something like that the material should auto-alter this value depending on the mesh it's been applied which makes it impossible to use the material for more than one object in the scene.

Ok, but the result changes depending on which mesh I use the material unless they have a similar mesh density, right? (I'm talking about precision alone, not adaptive option) Also it has something to do with the number of pixels of the texture that affects each face, right? (wild guess). So... here goes my personal view of how this works: UVW map distributes the texture pixels uniformly all over the mesh; if the mesh is evenly divided I can say each face is awarded with more or less the same amount of pixels, let's say 200; so, if my precision value is set to 50 Maxwell will give a value for displacement every 4 pixels (200/50=4)

Am I close to how it really works?

[tom]

Not exactly Well, it's quite complicated to explain it here. Making some tests on simple plane with simple textures (both having different subdivision and resolution) will quickly give a good idea.

[def4d]

...but it worked

(checker)

Mwaahahhha, excellent

[sandykoufax]

I've installed Mapzone 2.6

but it looks too difficult to use.

[sidenimjay]

Houdini 9.x is out and has always been the leader in procedural texturing and shading as well as procedural modeling.....

VOP networks make shader definitions easy to understand and flexible

now it has texture baking in mantra so you can export your shaders to a useful map.....

one can create every material map needed for virtually any situation, even animated maps are wicked easy

[polynurb]

Hi sidenimjay,

do you know to what extend these features are supported in the Houdini Apprentice HD ?

I have used Houdini for some of my projects at university some time ago,
and it has a really great workflow; steep learnig curve though;

I would like to check it out again.. and the HD is a nice choice for that purpose, as the Master Edition has a hefty price tag

cheers,

p_b

[sidenimjay]

hey p_b

should double check with sidefx, but as far as i know all the features are available with the HD version ...... i doubt you can use the results in production work, however, being that its still a non-commercial license....personal use is ok from what i understand....

there are limitations with the maxwell plugin tho when using apprentice or apprentice HD.... 640x480 max res, 7 samples max, and no writing of mxs files but will give an idea if houdini will suit the workflow...

i know that everything works with the Escape version which is Master minus the DOPs stuff for around $2000.00 US....you get character tools, channel operators, compositing operators, modeling and i think you get particle operators, vex operators , plus 5 mantra licenses.... everything you need to accomplish professional work for a fair price

and the maxwell plugin works great...

[polynurb]

thanks for the feedback sidenimjay!

..back then I was using Houdini for generating ..well..."experimental architecture", and that`s still my main field.
I wouldn`t be making any money with it... `cause nobody can build that stuff anyway

..but you`re certainly right sdfx wouldn`t want exports from HD to be used commercially.

_polynurb