3DTMP pearl effect (multi application thread)

MadKoiFish · February 2012

Ok, opening up this thread for all to contribute your methods to this. My interest and curiosity was piqued when IRML uploaded some examples within Lightwave here.

I will attempt to replicate it if in a simplistic nature in Arch Design materials for Mental ray and 3dsmax. (2010-) Nothing saying others cannot post their solutions. I will also try to convert this to scanline or standard max materials too. I have only attempted one proper TMP era ship so I never really though about how to replicate this effect properly. Always it has been a compromise of complexity.

Right now this is what I have. It is a cobbled collection of mix materials in the reflective color slot in arch design. Each mix contains one mix material and a fall off each containing one of 4 colours. blue/gold gold/blue green/pinkred pink red/green
so

mix

fall off red/green

mix

fall off green/red

mix

fall off

stacked like so. with a checker bitmap to mix them.
I then have a gray scale checker in the aniso rotation slot and have aniso cranked up to 3 or so.

As you can see the animation lacks compared to IRMLs example even considering the lack of texture complexity, just the pearl effect itself lacks. Or seems to. I am by no means a expert on materials.

hah anyone remember the code to make it so youturd doesn't show related videos??? My old method seems broken.

MadKoiFish · February 2012

Ok here is the scanline version using a multi layer standard mat to get access to aniso settings.

What I do not like about both of these is the flip of colour happens on all layers at the same angle. This is where I have not been able to figure out how to get this aniso rotation effect to apply individually to each separate fall off. I could be way off in using the method I am to get the solution.

funny too takes longer for me to set up and upload this crap than it does for scanline to render it all. O_O;'
I dunno why youturd made this look like a low colour image. . . ugh.

Uses that same stacked fall off mix thing.

Chris2005 · February 2012

finalRender has an architectural material as well... looks pretty much the same as the MR version.

I'll see if I can replicate what you did...

MadKoiFish · February 2012

WOW thats a lit clone! hahaha That save and play button intrigues me though HAHAHAH. probably some sort of realtime aniso playback? That alignment slot is not present on my MR version. Probably on mr for 2012.

Chris2005 · February 2012

MadKoiFish wrote: »

WOW thats a lit clone! hahaha That save and play button intrigues me though HAHAHAH. probably some sort of realtime aniso playback? That alignment slot is not present on my MR version. Probably on mr for 2012.

When I push that button this comes up:

MadKoiFish · February 2012

Ah that's the same as the templates in arch design. Just mixed up a bit. So I am guessing alignment is the map channel in arch design too. Wish if they were to use the same stuff they keep terminology intact lol. PITA between apps let alone render platforms.

Chris2005 · February 2012

MadKoiFish wrote: »

Ah that's the same as the templates in arch design. Just mixed up a bit. So I am guessing alignment is the map channel in arch design too. Wish if they were to use the same stuff they keep terminology intact lol. PITA between apps let alone render platforms.

The drop down menu for alignment has Automatic, Camera, Map Channel, Radial, and Radial/Planar.

MadKoiFish · February 2012

I stuffed that mental ray one into a environment shows how off my BRDF settings are.

took forever to render lol. Had my settings a bit low so ignore the "grain"
Now time to try a more complex map and explore if there is some other way to do what is seen on IRML's animation.

Buckaroohawk · February 2012

I hope you're able to solve this little mystery, MKF. I gave up trying to replicate the TMP pearlescent effect in MAX years ago!

MadKoiFish · February 2012

Early chart showing how the last few a+d materials were set up. The grain or noise in the reflection is due to a low sampling setting.
It is all still beta, in a way since I am still blundering about.
KIM settings fine details and even things like the fall offs are all exp, and quick set up to test theory.

The scanline one is similar in process and I should have a chart up soon too.

Then back to experimenting.

IRML · February 2012

can you use a normal map which only affects specular and not diffuse?

MadKoiFish · February 2012

There should be nothing affecting diffuse, though setting might be blown out enough it is looking like that. Can you point out what your seeing? It could likely be my crappy lazy texture map. Base diffuse is white.

Normal maps, I do not know if I can apply it like I think your saying. If I stuck one in that slot it would just use the normal map as a bitmap. I do have a normal bump shader but it does odd things. Something I have to experiment with.

I stuck all the previous stuff in the above flow chart into the shader and got that. lol
Damn ok I know why now, that is to create normal texture maps in ortho to project onto low poly objects for games. durrrr hahah 4am here atm so eh.

Mental ray has no specular. Just gloss and reflectivity which gloss is a part of. So you have to adjust the gloss and fiddle with the aniso bits to get a soft surface. That noise in the last animation is a example of low settings causing "texture" Due to reflective settings being way way too low. I normally set it to 60 and have issues with noise in dark shadowed areas.

{EDIT} Ok I think I know what you mean is how it is so strong or all over, Im doing it overblown to see the effect easier. I am setting up some cheapo maps IE using what I have in my resources and adapting them and pinching it down to a more uh, useable real scale effect.

IRML · February 2012

the trick I was doing in lightwave was using the spec map to change the normals of the spec shader, but this can still be done with a normal map if you make it the right way, I didn't do this because I have lightwave's nodal surfaces at my disposal, but if it's the only way to do it in max then it's a little extra effort but still the same result

normal map how you usually expect to see it:

normals 1.jpg

same normal map but with no parallax - this is what my normals trick in lightwave was doing (just using the greyscale spec map and some nodes instead):

normals 2.jpg

MadKoiFish · February 2012

That is looking a lot like the aniso rotation slot in arch+design. Except it only accepts (by what I know) grayscale levels for the rotational value. It affects the rotational pinch of the aniso specular hit for a better word. In max aniso is how round or pinched a specular (matte reflective surface) is. In arch+design rotation can crontrol the direction of the pinch, in standard mats it is called orientation. A number or gray value will define the angle. But if a map is used it will rotate based on that maps value. (See the flowchart above the gray check texture is my rotational map)

Comparing mine to yours it is just mine sucks balls. hahah too "un" random.
mine

yours cropped

Needs some adjustment on the aniso strength. Acually it would be better if I did a map to control how tight it is or loose. More variables? heh

IRML · February 2012

I'm not sure aniso is the right thing, maybe it's different in max than it is in lightwave, let me explain my confusion without trying to sound patronising

isotropy basically means the same in all directions, anisotropy is the opposite - so in lightwave it's used for reflection blurs, anisotropic reflections means the U direction is blurred more or less than the V direction, resulting in the brushed metal or CD effect:

aniso.jpg

that's how I use it anyway, I don't really understand how you're using it in max, for all I know it could be different

if I give you a visual representation of what that normal map I posted earlier is doing then maybe it will be more clear (in reality it's more subtle than this):

normals model.jpg

think about how that is working on the incidence gradients for the colour, as far as they're concerned each tile is rotated differently, so each tile's colour will flip slightly earlier or later (that's what I mean when I say rotation, this could be where the confusion is with your aniso rotation)

incidence gradient:

flip 1.jpg

incidence gradient tweaked normals:

flip 2.jpg

so your way may be varying the specular brightness in a pleasing way, but when it comes to adding the colour tint it won't be flipping like mine

Coolhand · February 2012

As I understand it the TMP paint job was simulating the hull being made of different peices of metal rather than showing how uneven they were.

What the anisotropy direction map is doing, is more similar what the painter of TMP enterprise was doing: Adding gradients that simulate the hull being made of different sections of brushed metal and not that the hull panels are poorly lined up.

So Just adding a global normal map will just highlight the enterprises poor build quality but might look good combined with other things. I see what you're doing with the spec effect and if it works it works, but surely that's infact more of a hack than actually changing the direction of anisotropy effect as max does it?

MadKoiFish · February 2012

That is the issue. it is getting that effect cause max afaik cannot do what your showing directly. Least in regards to that normal map. Only reference I find about normals is creating them from complex objects and using them in displacement slots. I can get the colours to change depending on angle but that angle is always the same. And I cannot seem to get it as deep, 2 or 3 layers seems to be it. Incidence gradient does surely look like the max fall off shader so at least I got that right! lol The fall offs will alter colour based on the view angle light etc and so on. I just do not have much more control than that.

Only other thought is to hack at say things like the car paint shaders. But they often have ugly surfacing outside of "gel coat" look.

Aniso in max seems different since it directly refers to the highlight being "stretched" (hah opposite of what I say pinched

) Least under standard MR :

Anisotropy
Controls the anisotropy, or shape, of the highlight. At 1.0, the highlight is round; that is, no anisotropy. At 0.01, the highlight is elongated. One axis of the highlight graph changes to show changes in this parameter. Default=1.0.

Rotation
Changes the orientation of the highlight. The sample slot shows changes in orientation. This value can range from 0.0 to 1.0

And yeah I have no real training or book taught terms just seat of the pants pushing buttons. So generic terminology often is a puzzle to me etc. I try to do things based on how or what I know of real objects, how the light plays on them etc. So yeah if you have to dumb down the descriptions and crap do not worry. I hope that the thread will help others as well. Maybe learn more than just how to do TMP hulls. >_>

I should probably take a rudimentary course to learn the deeper bits of max. I can just guess I know about 15% of the app.

IRML · February 2012

Coolhand wrote: »

As I understand it the TMP paint job was simulating the hull being made of different peices of metal rather than showing how uneven they were.

What the anisotropy direction map is doing, is more similar what the painter of TMP enterprise was doing: Adding gradients that simulate the hull being made of different sections of brushed metal and not that the hull panels are poorly lined up.

So Just adding a global normal map will just highlight the enterprises poor build quality but might look good combined with other things. I see what you're doing with the spec effect and if it works it works, but surely that's infact more of a hack than actually changing the direction of anisotropy effect as max does it?

have you read the original post in the other thread? what I'm doing is simulating the way the paint works, the base was very smooth and this is left unchanged with my method, the coloured highlights all catch the light differently and this is why the normals are changed for these

Coolhand · February 2012

I realise you're not making any global changes to the normals. what i'm currently thinking you're doing is perturbing the normals but applying that to the specular only.

Wouldn't it be a better way to achieve the effect by making the material have anisotropic highlights and using a map to control the direction of those anisotropic highlights...? Which is the approach that the others are taking - isn't that much closer in theory to what the painter actually did to the real model?

IRML · February 2012

I don't get it, you said my way was making the panels 'uneven' but the anisotropic way is doing a very similar thing, the colours may not be flipping unevenly but the highlights are still uneven on both, and this is correct as long as it's kept subtle

pearl paint doesn't behave anisotropically, it's an incidence effect, so I don't think that's the right way to go

imagine this is a microscopic cross section of the way the pearl paints went on, this is what the normal map is trying to copy:

paint micro.jpg

(the effect is actually more within the paint, but the principle is the same)

Coolhand · February 2012

i didn't say it was making yours uneven, only if you tried to apply a normal map globally in other applications. You already know what you're doing, i'm talking about everyone else;)

MadKoiFish · February 2012

Ok I think it is bed, 6:30am and I keep making mistakes and max keeps crashing on render cancellation. I will set this sample to rotate in a animation render and look to see if it does anything close to what is wanted.

All way to overdone and well obviously re-purposed textures.

I will sift about on the vomit that is google's results on normals to see if I can figure out how to do it within mental ray and maybe standard materials.

Anyhow if anyone has input on anyway to get better TMP hulls feel free to inject tuts or whatever in this thread. Idea is for it to be a resource in the end.

IRML · February 2012

Coolhand wrote: »

i didn't say it was making yours uneven, only if you tried to apply a normal map globally in other applications. You already know what you're doing, i'm talking about everyone else;)

ah I see, well this is why I wanted to know if madkoifish could do it in max, I know the surfaces can be limited compared to lightwave's nodes

there could be more than one way to achieve it though, if you can't add a normal map to the specular maybe you could use a global normal map and then a modifier on the diffuse to make it smooth; or perhaps you can use material mixing with a matte white material for the base then a glass like layer added on top of that, fully transparent except for specular, if there was no refraction a normal map could be global here; or are there clear coat modifiers you can add and mask with a map?

there's usually tonnes of different ways to do the same things

Coolhand · February 2012

you're absolutely right, your normal trick just sounds completely whacky to probably most of us here. though if you consider the particles in pearl paint, one of the effects they have (colour being separate, angle of incidence, probably being separate) is that the should have an influence on anisotropy and anisotropic orientation, like the brushed lines in brushed metal... its just a way of approximating that type of surface without the overheads of actually modelling it, so these guys i'm sure are on the right lines.. but there is more to it of course.

Coolhand · February 2012

Image 1, map control ONLY of reflection colour, based on alternating red/green and blue/yellow, as angle of incidence based, or falloff maps.

Image 2, map control ONLY of reflection colour as above PLUS angle of incidence based control of anisotropic orientation - basically copying the entire reflection colour tree to the orientation slot. so essentially the direction of the highlight will change as the colour changes as the angle of incidence changes.

everything else is just a single value, no other maps are used in any other slots, not "spec" (more properly, reflection) or gloss or anything.

Not saying that exactly nailed the effect or anything but just wanted to illustrate it.

Coolhand · February 2012

same but with incidence based gloss control - essentially the same thing used in the orientation slot copied to control gloss.

Tralfaz · February 2012

Is this the type of effect you are looking for? Sorry about the ghosting edges, this was done really quickly in Cinema 4D.

Al

MadKoiFish · February 2012

Tralfaz wrote: »

Is this the type of effect you are looking for? Sorry about the ghosting edges, this was done really quickly in Cinema 4D.

Al

Yes, that is stage 1 of it. I want to get that and have it be controllable in placement like IRML had within his video in the link. Also avoiding the white blow out or washing out of the colours. Being able to control each layers amount of strength etc.

MadKoiFish · February 2012

Here is the video set to render earlier on. Waste of half the frames though

720p btw, probably just as ugly as lower rez.

Effect is overdone to make it easier to see what it is doing. Still no control over what layer has a stronger effect over the others.

Tralfaz · February 2012

Okay, let me play with it some more.

Al

[Edit]

I guess I don't know about the material system in C4D to actually come up with something like you or IRML did. Just more stuff to learn when I get the time...

[/Edit]

MadKoiFish · February 2012

Another test this time using a fall off then mix mats to introduce the other shades. Seems to work better. Also tried dumping the fall off stacks into the aniso rotation but I had to add of course a bitmap to add the panels but I do wonder if I just put the panel map into the gloss or aniso level then let the falloffs only control the rotation (assume it is the same thing refered to as orientation >_>)
Anyhow gonna take a break from this for a while and do some painting. Likely have at it again later this evening. I think I will take what I have tried and apply it to a final material and see how it looks. I still want to mimic what IRML did with the same grainy reflection and intensity of colour.

Greetings!

3DTMP pearl effect (multi application thread)

Posts

Greetings!

Quick Links

3DTMP pearl effect (multi application thread)

Posts