PDA

View Full Version : Your opinions on the most difficult material to create?



prometheus
04-24-2016, 11:37 AM
Just out of curiousity, and perhaps it could serve as a reference guid to the community and the developers of Lightwave, maybe if we get enough postings here, it may show what people struggle with to get it right, so
it is interesting to see if one specific material stands out more than anything else.

Throw it in here, what material do you folks think is the hardest to create? could it be a brushed metal, gold, some type of metal reflection, is it skin, leaf, water, milk, rocks, glass etc?
Any material is of interest...it could be surface shading only or textured.

Thanks in advance for any kind of input.

Michael

hrgiger
04-24-2016, 01:11 PM
Metallics and and transparencies are pretty much a breeze nowadays with physically correct materials. It can sometimes be achievable to get decent skin results but difficult to appear correct under differing lighting setups. The one thing I've found challenging is to make decent snow or ice materials.

spherical
04-24-2016, 02:13 PM
Well, I'm still struggling with that brushed metal... :D

Another is etched glass. I usually end up using a hack consisting of a double-sided poly with some transparency and a clip map, floating just in front of the glass. Works, but the purist in me would like to do the etching on the glass itself. I always figured that I could get the scale of a procedural down far enough that it would mimic the etched surface, but it seems to bottom out at a certain level and remains too coarse.

prometheus
04-24-2016, 02:14 PM
Metallics and and transparencies are pretty much a breeze nowadays with physically correct materials. It can sometimes be achievable to get decent skin results but difficult to appear correct under differing lighting setups. The one thing I've found challenging is to make decent snow or ice materials.

Yes..I think snow and ice are a bit trickier than most common transparancy or metalics.
two snow presets in the preset database? the old one had one nice snow with glitter and some ice too, havenīt checked these ones though...
http://3dxyz.pro/?s=snow

Is it just me..or is the preset 3dxyz website slow?

the arctic snow/ice mix kind of iceberg is probaby a bit hard to get right.

prometheus
04-24-2016, 04:55 PM
Iīm not quite sure if this is going somewhere regarding ice, just made it up quickly without any node materials, trying to keep it fairly simple, though there is a fast fresnel shader, there is surface thickness gradients, there is translucency.

refraction of ice is something around 1.3, depending on some factors(?)
How much transparency? how much diffuse? well that is questionable, ice is indeed complex depending on how packed it is, how much turbidity, but it does spred diffuse light very different depending on ice thickness and how dense it is ..I think, ice is complex in the way it seem to have several states of materials, at least if you look at the iceberg or ice frozen on a mountain and some part of it still is more loose snow.

I would need a better modeled ice peice though, more suitable for that iceberg look, and I still havenīt checked those presets ad 3dxyz, and I will still have to go through dielectric, sigma and the delta node materials later on.

133551

133552

spherical
04-24-2016, 05:32 PM
Those ices are a good start. I'm assuming that the frosted portions are mapped on the polys. We sometimes need to have "stuff" in our cast glass: bubbles, clouds and impurities that alter the refraction/transmission. Have yet to find a good way to do that. With your working with sprites for fog, I thought I'd try that, as it worked great on the exterior night shots of our new studio.

jwiede
04-24-2016, 05:57 PM
Well, I'm still struggling with that brushed metal... :D

Yeah, and while we got close with the galvanized steel, getting realistic anisotropy on the little crystals was still a problem (at least with current LW). I get the impression accurate finishes involving complex anisotropic effects remain problems for LW. That kinda fits with ice/snow being problematic as well.

If the question is "across all renderers", I'm not sure there are any truly problematic materials anymore. Unbiased engines like Maxwell, Arnold, etc. seem generally competent across most materials these days. Getting accurate SSS can still be a bit tricky, even with ray-traced diffusion capabilities, due to perf. cost.

Volumes with embedded materials definitely still represent the most complex sorts of materials, but some of that also comes down to willingness to realistically model the material (with separate inclusion materials) versus trying to generate a material that somehow "fakes" the inclusions. Here again, perf. cost is a huge control factor.

For LW, which materials present problems seems likely to change with the new render engine.

prometheus
04-24-2016, 06:14 PM
Those ices are a good start. I'm assuming that the frosted portions are mapped on the polys. We sometimes need to have "stuff" in our cast glass: bubbles, clouds and impurities that alter the refraction/transmission. Have yet to find a good way to do that. With your working with sprites for fog, I thought I'd try that, as it worked great on the exterior night shots of our new studio.


Yes that is interesting with the bubbles, clouds, impurities inside of a material, or marbled glass..will have to take look at that.

For that ice, the frosting portions goes all over, and is just the reflection blurring, and you now what?... I actually forgot the refraction blurring..so no such thing in there.
in transparency wich is around 88% I got a surface thickness gradient, reflection has a gradient set to bump and in the bump channel the fbm noise procedural at 200 um, quite a lot of bump strength at a value of 300, the so called iceberg is 20 meters uniformly scaled, refraction index at 1.3, glossiness 50%, spec 88%, diffuse 150, color channel also has a thickness gradient with some slight blue tone gradient...there is however a balance with what the actual fresnel shader(15% glancing angle) does too.

The metal stuff, well I have that specially reserved for you only..if you mail me.

I suspect the light penetrating the ice getīs more blueish becose of the density? not sure ..and I am not sure if some blender artists are approaching it with translucency ray length? and where the material turns to more snow,..it also gets more loose and get much more crumpled detail, it will get less glossy... as well as starting to pick up glittering from the sun.

http://www.creativeshrimp.com/create-realistic-ice-lighting-book-02.html

this one ..for blender has really good snow, though take a sneak peak of it at 55 min in the vid, it is a bit long, and mostly covering creating track in the beginning.


https://www.youtube.com/watch?v=82zWmOqE0Nc

hrgiger
04-24-2016, 07:57 PM
Nice Tutorial Prometheus.

prometheus
04-24-2016, 09:18 PM
Nice Tutorial Prometheus.

yes it is..but it sounded like you credited me..it wasnīt me doing it...nice tutorial link perhaps.:D
I will have to check this with blender as well..when time allows, though I wouldnīt go and buy me the pro skies, so I would have to get some other images, illumination is important.

hrgiger
04-25-2016, 02:52 AM
Well no I knew it wasn't yours, I should have said nice find.

adrian
04-25-2016, 03:06 AM
Realistic snow for me is definitely the most difficult thing to create. There are a couple of really good threads on it here but even following the tips/tricks in those I've still come nowhere close to realistic snow (for close-ups anyway).

Fabian-Esh
04-25-2016, 04:30 AM
Realistic snow for me is definitely the most difficult thing to create. There are a couple of really good threads on it here but even following the tips/tricks in those I've still come nowhere close to realistic snow (for close-ups anyway).

Hi Adrian,

I made most of the difficult shaders like snow, water, hot liquid glass, isotropic, layered materials etc. The biggest issue I have is that LW does not have micro-poly displacement or subpoly displacement. There are some 3th party setups and fake nodes, but the problem should be fixed in LW itself. For snow, you can use a cook-terrance shader for specular input as a work around, but you need to displace your snow on very small level to make the flakes. Its not about how the shader works, its about how the geometry is displaced. You can use bumps, normals, etc but thats not usable for close-ups like you mention.

If you are looking for an optimised isotropic shader, SOnny made one on the kray forum that is LW native: (I just think he needs to upload the surface setting)

http://www.kraytracing.com/joomla/forum/index.php?f=4&t=4704&rb_v=viewtopic&start=15

I think we should make a list of materials and submit shader setups to first have an overview and see where the problems are. (So we can give LW-ers an updated library of materials) Then request micro-poly displacement for new LW. My small list is:

1. Reflective non-transparent materials (that use isotropic shading - metal for example)
2. Oren-nayer materials (stone, rough wood, etc)
3. Transparent materials: transparent - opaque (subsurface) - translucent (clearcoat, glass, leaves for example)
4. Organic materials
5. Layered materials (material blending, optimised, renderman style: https://www.google.be/search?q=renderman+layered+materials&tbm=isch&imgil=p_SLRQTumG2xMM%253A%253BxhZyGT-jmVMT_M%253Bhttps%25253A%25252F%25252Frenderman.pi xar.com%25252Fresources%25252Fcurrent%25252Frps%25 252FPxrLMMaterials.html&source=iu&pf=m&fir=p_SLRQTumG2xMM%253A%252CxhZyGT-jmVMT_M%252C_&usg=__qSP6iDKeC4PXNv80jpAmaVPB6Ek%3D&biw=2144&bih=1234&dpr=0.9&ved=0ahUKEwj_2sTMyKnMAhWLCsAKHdwhAcoQyjcIag&ei=nu8dV_-TMouVgAbcw4TQDA#imgrc=p_SLRQTumG2xMM%3A)
6. Displaced materials (snow, dirt, clouds, dust)

If you want to be able to do some of these without work arounds , the "make material" should be able to also accept shading nodes for reflection/specular/diffuse- so this node should be updated. To make notes would also be handy in node-editor (fusion and nuke have this in their nodal workspace) Povide choice between pong/blinn/oren-anyer in the material node would be handy (vray has that) And just get rid of that layered surfacing system, should go direct to nodal (like maya) And many more.

Hope this helps.

jwiede
04-25-2016, 05:43 AM
I've found close-up-suitable woven cloth materials can be quite tricky as well -- for example, replicating the subtle sheen of a cotton oxford cloth. C4D recently added a shader specifically for wovens (goes into reflectivity) that produces excellent results, substantially better than most such cloth I've seen from LW.

rustythe1
04-25-2016, 10:00 AM
I have just been doing a big project with ice and snow, its a real headache, one setting it looks rubbish, another setting kills render times, hopping "NEXT" will fix a lot of it, com on hurry up LW3DG
one of the main problems with ice is the fact its really a volumetric as it has a lot of internal elements that are near impossible to fake, there is also the way it changes colour just beneath the water, in photos it looks blue above and green bellow, the same as the water, its bluer on the surface than bellow,

prometheus
04-25-2016, 10:18 AM
I have just been doing a big project with ice and snow, its a real headache, one setting it looks rubbish, another setting kills render times, hopping "NEXT" will fix a lot of it, com on hurry up LW3DG
one of the main problems with ice is the fact its really a volumetric as it has a lot of internal elements that are near impossible to fake, there is also the way it changes colour just beneath the water, in photos it looks blue above and green bellow, the same as the water, its bluer on the surface than bellow,

I suspect the water itself may have something to do with it, since water medium filters/slows down the red spectrum of ligth, so even blood may turn a bit greenish due to the red light not penetrating the water enough...check this...
http://blog.sciencescore.com/did-you-know-that-your-blood-is-green-underwater/

So the effect of that could probably be acheivied by setting a good water surface, I had some sucess before with that when using surface thickness gradients, both on transparency and color of the water, probably not physicly correct... but I thought it worked decently.

Then again it may be extra filtered by the ice itself, one can see that sort of effect on large ice with more snow on the top and at the bottom turning glossy icy and more greenish, so I suspect that may be the same principles..filtering out the red light etc, otherwise I would suspect the ice itself has some pollution/ turbity...maybe a combination?

RebelHill
04-25-2016, 11:33 AM
Ice and snow are a mix of transparency, SSS and reflection. Sigma 2 is your friend for these fellas all used together.

THIBAULT
04-25-2016, 11:50 AM
For me, this year, i've a difficult with polycarbonat.

rustythe1
04-25-2016, 12:06 PM
Ice and snow are a mix of transparency, SSS and reflection. Sigma 2 is your friend for these fellas all used together.

Nice Nice baby, stop, runs off to do some Ex-peri-mentin.

prometheus
04-25-2016, 12:30 PM
Ice and snow are a mix of transparency, SSS and reflection. Sigma 2 is your friend for these fellas all used together.

True..and refraction..or do you equal that to the function of sss?

4 cm of icecubes are one thing, and easier to do than iceberg with materials that transcends to alternate states though, so yes..works fine with sigma for that, and one can also check matt gorners video tutorial "exploring the sigma node shader" and the "exploring ss2 node shader"

There is no reason though to not check out how well it should do for big iceberg with muti material, so mixing sigma ss2 with other materials or just change two sigma materials to get that special colored look might do it as well.

http://www.pixsim.co.uk/

Direct link..
http://www.pixsim.co.uk/video_tutorials/Exploring_the_SSS2_Node_Shader.zip

Direct link2...
http://www.pixsim.co.uk/video_tutorials/Exploring_the_Sigma_Node_Shader.zip

MonroePoteet
04-26-2016, 03:59 PM
RE: the ice internal detail, Fracture can be used with a low number on Cell Count and Exploding the pieces slightly, like 1mm for an ice cube.

If I remember correctly, I made this ice cube (cloned and rotated in the scene) using a 3x3 segment cube, jittering the points slightly, tripling the quads, executing Fracture with 10 Cells and 1mm Explosion offset, jittered the result slightly, constrained it back to a cube shape with Boolean Intersect, and then used a 2mm Magnet to bring any unruly or too-straight points / edges under control.

Works OK, and potentially could be fine-tuned further. The surfaces are just Fast Fresnel, so nothing fancy. White for the ice surface, a very light cyan for the interior.

133580

mTp