Adding to Cart…

Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2025 Daz Productions Inc. All Rights Reserved.You currently have no notifications.
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2025 Daz Productions Inc. All Rights Reserved.
Comments
Redshift3D will be my 1st choice if I ever went with a GPU renderer. :) Out of core, biased and ridiculously fast - from day one.
How fast? Scenes that would iray take 3 minutes to render with two GTX970 is only 40 seconds (or less) with a single 780 (not 780Ti). That's with brute force GI too, not just direct lighting and AO like with Octane.
I haven't jumped the boat yet because they're still using point based SSS. They are working on a raytrace SSS solution, but it's not out yet. There's also FireRender from the Corona Renderer devs, but that's still in alpha stage, I think.
Those two are only available in pro apps though (Maya, Max, XSI). I'm already thinking about switching to Maya since 3delight for Maya sees more development and tools than DAZ Studio. :)
You got the two fandoms mixed, Zarcon =) Mollari's name was Londo, Lando is the Calrissian guy from SW =)
Actually I'm looking at the original DNA plugins, too, but in my case it's looking longingly, and nothing more =) I don't know if I'll ever be able to afford even Maya LE (and I'm not sure the 3Delight plugin will work in LE... there is no MEL scripting in LE iirc)
Just looked at the new 3delight for Maya materials. I really like what they're doing. Separating transparency/opacity color with refraction is the right thing to do. And they enabled texture slots for scattering and transmittance color, IOR and anisotropy slots. :).
One thing that isn't obvious though is that you can use different IOR values for reflection (reflectivity) and refraction. Unless I misunderstood the description.
Kettu, even the 3delight devs are using colors for subsurface scales :D And it's probably a good idea to stick to the same shader arrangement too. It's very well organized and very logical.
The problem is that the Maya shaders, while possible to use in Studio would be nearly impossible to port over...so all the nice, new shiny stuff would have to be done from scratch. I've done a couple of simple Maya ports (one light and one surface) and figuring out what you need to add/change to get to work in Studio is really shooting, blindfolded, in the dark. It's somewhat disappointing that the RSL editor in Studio can compile them, but Shader Builder throws errors out the ears...and refuses to even try. Going the editor route means having to create all the support scripts by hand...and my DS scripting just isn't up to that task.
I think a lot folks don't understand that the real difference between Iray and 3DL is that all the 'missing' things that Iray has that 3DL in Studio doesn't are because they are implemented into the main code of the renderer, itself, as opposed to being 'true' shaders. And Studio just doesn't have the shaders needed. Because really, there isn't anything in Iray that, with the right shaders, 3DL can't do...not that it is easy to implement in 3DL, or at the main 3DL level is even 'wanted' (IES profiles are a good example...), but it is all 'doable'.
Ya know, on that thought of map slots. I've noticed sometime in the past year, the maps can now go two different spots for each channel, and the maps can be different, for different things.
For example, Glossiness. There is now the capability to put a map in the 'Intensity' slot to control that separately from the 'Color map'. I remember in the past, that was not possible to do in the 'Daz Default Shader'.
The biggest thing this adds, is in making shader sets, one color map can be used for multiple different cloth patrons, opening up many more possibilities. I've only had time to do a basic dose it work tinker with it over the past several months, because of so many other things going on.
I was generally thinking about having the same controls in a ds shader. Doesn't have to be the same shader. And I think some of the parameters are already available in Kettu's shader.
That's the easier thing to do...but there's still a 'hard' part to it, if using the DS default code...breaking it apart to be able actually use those separate inputs. Much easier to add them with your own code/ported code.
Scales, you mean scatter/absorption values? I only see the code side of things, not the Maya interface because I don't have Maya. But from what I understand, you can use Maya colour boxes to specify arbitrary 3-tuplets, and not just 0-255 integers per channel. I don't think that a ShaderBuilder-generated integration script can do that.
It probably can be done by hand (the Iray base seems to have "linear" color boxes), but that will mean extra installation files... and I'm not sure there are scripting docs up yet that cover this.
The internal logic is also a bit different: I didn't put proper pathtraced area light support in yet (that's a paradigm shift for DS, that will need a different release with a whole load of new documentation), so specular and reflection will be separate in mine; and as my speculars are simply additive and not absorptive, they are both best called simply speculars, none is a top coat.
The order right now is SSS, diffuse, two specs, bump, displacement, reflection, overlay diffuse, velvet (including edge cheat), backscatter hack, RT controls. They all have those boolean on/off buttons. I wish we could have nice little checkboxes in the DS interface for the on/off thing. Something that could go in the same line as the strength slider.
Here;s the link:
https://3delight.atlassian.net/wiki/display/3DFM/3Delight+Material
It's generally pretty similar to DS shader arrangement, except for bump and transparency/opacity (further down, instead of after ambient/incandescence).
Thanks! I see they have updated the Maya docs, cool.
I still wonder how the colours are represented in the interface. Or maybe the DS RGB/linear colour conversion thing that I ran across back then only extended to setting attributes, and it will work okay for straight-on passing the values to the shader? You won't be able to see the "linear" values in the interface, but maybe it's not necessary... I'll think about how to do this best.
The separate scatter/absorption scales is what I call "multipliers" in my shader, BTW. They will be more useful than they are now, if we try to represent the 3tuples via a colour (they will then multiply the 0..1 ranges from the colour picker so as to get proper Jensen values).
Here's a great article by Jeremy Birn that some may not have read yet...
http://www.peachpit.com/articles/article.aspx?p=2165641
I so want that to replace the current default surface...
And this is for those of us who need it clean the drool off our keyboads.
Btw, if anyone ever wanted to try out rendering human skin with other than DAZ figures, you can find a link to the Lee Perry Smith scan from Inifinite Realities in the 3delight Skin Material page:
https://3delight.atlassian.net/wiki/display/3DFM/3Delight+Skin
Using that scan, I did some test and fine tuned my SSS settings. Finally got proper red ears now and nothing like those iray SSS red ears people are coming up. Only recessed parts exhibits backscatter and not the entire ear lobe. So thank you 3delight devs and Lee Perry Smith/Infinite Realities. :D
I've read it before, a long time back. And of course, it all starts with linear workflow.
Sure, that would be neat. :D
Btw, why not add links to those Oren Nayar values and IOR indices to your sig?
Okay...be easier to keep track of them, that way.
Mr Smith's looking great! What did you light him with - same setup as your girls in blue in the GC thread, or some other?
As for Iray, I'm sort of frustrated at my being unable to figure out if the SSS direction in the IrayUber base really means what it means. Sources say skin exhibits strong forward scattering. Why are DAZ Genesis presets using a negative value, then...I can't understand how to convert the Jensen presets properly into Iray colours either. I can't find the main scale anywhere. I guess if I ever do release any character sets, they won't come with proper Iray support. Everything else is easy.
PS I hate the way the new forum software won't let me use cursor keys or type a line break seamlessly when I'm on my android tablet (Firefox browser). And in "source" mode the text has no wordwrap =(
Pretty much, though I did some additional tweaks to the lights - mainly intensities.
Don't know if its the shader, renderer or artist(s), but I simply don't like the SSS in DS iray renders. From what I've seen so far, the backscatter covers the whole lobe - which is incorrect. It's like using the wrong scale with SSS in 3delight. :)
And I guess the issue is the scale indeed. It's very difficult to control the SSS in Iray. It's technically based on an extension of its general volumetric model (the one that does fog etc), and while it does have some basis in reality... it seems like "organic" scatter in solids could benefit from a different algorithm.
I also tried using SSS with not "translucency" (diffuse transmission, in MDL terms), but glossy refraction (specular transmission) - as one Iray dev blog post suggested. It works, but does not look good for skin. And as hard to control.
I'm not so sure that it's an Iray scale problem...it may be a Studio scale. I've been thinking about scaling everything in a scene up by 10x to see what that looks like.
DS to Iray export scale, that's right. I can't find any docs on this. What's Iray's native unit? If it's millimeter like 3Delight, then what you suggest can work (supposing there is no hidden conversion anywhere).
Iray uses the m, DS uses the cm
It seems to me that most PBRs use the meter as the native unit. In some ways it makes a lot of sense, because typical units used for measuring data (light, volumes, etc) are in m^2 or m^3.
Thank you Richard! Do you happen to know if there is a scaling coefficient somewhere in the plugin, or does it get sent the way it is? Because if there's none, then 500 "units" become a hundred times larger...
Now, 3Delight. I made my glass shader behave much, much better (especially because I killed that extra mega stupid "tint reflection the absorption colour" line, what was I thinking?!?!)
There is a choice now to use absorption as-is in the surface shader, and I also made a volume shader to go with it instead of the surface-based algorithm (it's not "real", no raymarching): the volume will look better in some cases, when surface-based tracing misses something.
But neither of these approaches is suitable for stuff like turning a plane into water - they need the other side of the object to work.
When you stick stuff inside, it may look acceptable when it's not the focal point of the image, but it won't get darker with depth.
I guess I will have to write a real ray marcher to get that volumetric water look. =( I can actually think of a surface-based hack to approximate that, but it doesn't seem to work with refraction. It's stupid to have non-refractive water, right?
The render attached shows the volume with cylinders sticking through... there are black splotches in the shadows because there is no GI, just a spotlight and caustics.
Sweet looking glass...getting better.
Thank you Mjc!
Another thing I did, while polishing the RadiumSolid shader now, was swapping the Blinn and Cook-Torrance specular models for GGX at last... the distribution is called "glass-ggx" in the docs and I skipped checking out its bsdf() version because of that "glass" word, but boy am I glad I finally did try it out. It's awesome. The falloff is nearly perfect, paticularly at glancing angles (check out the attached image). There are two layers, "top" and "base" now; the base is attenuated by top (not really a physically based top coat, the top just tones down the base specular when it's visible, to better conserve energy).
Newer 3Delight builds add anisotropy to GGX if I'm not mistaken; I guess this is the one I'll be using all over everything from now on.
PS The more I work with refractive shaders, the more I understand how crucial proper geometry is for the result to look real... check this out: http://adaptivesamples.com/2013/10/19/fluid-in-a-glass/
GGX is a better model. Glad to see you're using it.
Added reflection to the eye reflection surface and skin. Plus a whole lot of tweaks to the fresnel (on the skin).
Thoughts?
Looking good....