Phone Backpanel Shader

[JonasNöll]

fixeighted: Checked out your file, thanks again!, I think it's kind of working, in some cases in my situation. I had some situation where it looked exactly like the effect from the reference with those rainbow patterns, and then I changed a little bit about lighting are shading and it totally changed the look again. It's super sensitive to the hdri and the scene. So I personally prefer the first approach with the gradient in the bump because that is just so efficient and controllable! Somehow the material previews in compact Material Editor in the new V-Ray also crashes my Max all the time because it tries to calculate the dispersion in the previews swatches, only way to not make it crash is to enable the classic material swatches for me. Did you also run into this problem or is that hardware specific for me? It basically freezes like 10 minutes to load the preview of the BlendMaterial with the 3 abbe enabled layers...
aandronov: I think that approach really works quite well in my case. In the end I found the easiest way would be to make one BlendMtl for one rainbow pattern that only contains 3 Layers with red, green and blue reflection, all slightly offsetted with the gradient ramp trick to create one rainbow. Then create a new blend mtl and load this rainbow gradient in multiple layers and offset all of them with an addition BumpMtl on top of each layer where you can position each rainbow seperately. Like this its actually no big problem to create lots of rainbows and they are all instanced to the same base rainbow BlendMtl!
piotrus3333: Cool, I will check that out, so far never used this PluginNodeTex, seems interesting. But my guess it is just a falloff ramp with different colors depending on the viewing angle, so it would just tint the existing reflections but not really offset these multiple rainbow reflections around my main one. But would need to try out.

[piotrus3333]

just follow real life optics. easier that way in this case.
bit of info about the node:
https://docs.chaosgroup.com/m/mobile...n#page/7145142

edit: you did ask for no-fake approach after all.

[fixeighted]

Yes Jonas, that material is a massive preview headache LOL. It doesn't crash on my machine, so maybe just hardware related. It does take a few mins to update though, which is why I added the instanced separate materials that are a bit quicker, if you have that blend material hidden.
I did try it with different lighting and it seemed to work fine - I will have another look in a while. Meanwhile I'd love to see what you achieved with the bump setup

Piotrous - I did try with the film shader but it just wasn't giving those multiple dynamic reflections/refractions, as Jonas noted.
Please share if you have managed to get it to work in this way.

What I tried to do was to imagine the real surface and I believe it may be made of several bonded layers, each with a different ior, that interact with the base layer.
Without a physical sample it's impossible to say in which conditions the effect is most prominent or even what triggers it.

[fixeighted]

Ok, I agree with you Piotrus
Thin film works well and seems to give the desired result after a bit of fiddling.
At least it is the simplest and quickest method and works in any conditions as far as I can tell.

[JonasNöll]

That looks cool, is that with this thin film? How were you able to get those parallel lines? Or is this combined with the other approach with the bump in the gradient?
In my testing I wasn't able to get these parallel lines, so would like to know if I'm missing something. For me it seems to be just generating random colors based on different thickness values where I could map for example a falloff inside. So in case that was done with the thinfilm would be curious to know how.
Sorry, I got nothing to show as of now as the client later specifically requested to not include this effect, so I ended up not having do do it.

[piotrus3333]

it’s not random. things are starting to happen when thin film is thinner than wavelength of particular colour.

[fixeighted]

It was actually ridiculously simple and meant I did not have to touch that horrible slow blend material
Just a clear plastic with film shader in refl slot - tweaked the parameters using a simple gradient for the strips.
Thanks to Piotrus for forcing me to look at it again really...couldn't resist figuring it out
Here's the scene https://www.dropbox.com/s/pxgs66475u...y%202.max?dl=0

It's an odd client request but then they are a bit odd, in my experience

[Slazzo]

Would like to push this topic a little further, thin film doesn't seem enough for this one, in this case diffraction must be simulated,
one would fake it by baking spectral flares inside environment map but let's be honest, is 2020 we should be able to render such effects already

[fixeighted]

Can you explain that scenario please? That is, what is the casing made from...any cover...thickness etc.
Also what lighting in this particular room.
Can try to see what's achievable...all good fun

[Slazzo]

Can you explain that scenario please?

The lighting set doesn't really matter although is quite easy to deduce it from the above picture, it's not about reproducing a still image.
In last 3 years or so phone manufacturers started using diffraction grating casing in a variety of patterns, the example above is using a pattern similar to next example

[fixeighted]

Yes, I realise it's not about a still image but rather the precise behaviour.
I just hadn't come across that material before. It's interesting and after a bit of research I see what's happening.
In the lens effects options you can set lens scratches to achieve the splitting I see in the video.
So possibly the scenario you posted is just a variation of these settings

[fixeighted]

And in the interests of the original post, this does work perfectly, as in the reference.
So this is just a blend with a glass coat, so no abherration and no thin film (which as it turns out is not at all the same thing as this sort of diffraction). All based on the lens effect
parameters. Very interesting stuff.
Many thanks to Slazzo for intriguing me enough to delve deeper to discover it

[Slazzo]

Unfortunately the lens effects will solve only for highlights visible in screen space, pathtracing diffraction is a nighmare and hard to control

[JonasNöll]

Oh alright, interesting, I wasn't thinking about using the lenseffects . True, that should be able to get you similar effects like in the original picture that I posted on page 1. Downside would be that you always would need to have a bright enough reflection to trigger the treshold. You can see on the right phone on page 1 the main highlight is actually quite dull and it still has this spectral effect, so probably in this case it would still require a shaderbased solution.

[fixeighted]

Yeah, just realised that, so in that case....close, but no cigar, as they say

We'll just have to wait until they find a way to implement this in Vray.....

EDIT: Not sure if my thinking is correct but would a map for abherration do it? Obviously it's not mappable right now, so just a theory.