Daz 3D is part of 
Connect
DAZ Productions, Inc.
224 S 200 W, Suite #250
Salt Lake City, UT 84101
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2024 Daz Productions Inc. All Rights Reserved.Daz 3D is part of
Connect
DAZ Productions, Inc.
224 S 200 W, Suite #250
Salt Lake City, UT 84101
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2024 Daz Productions Inc. All Rights Reserved.
Comments
...update...rendertime 1h...
...much better Adding a strong sun really makes it come more alive.
Yes, I struggled a fair bit with the lighting. Tried to use HDRI lighting only but couldn't get the desired result. So added an area light;) Those awePTarea lights
...
...so those come with the AweSurface shader system?
Yes, highly optimized for speed
Kit bashed the Dry Mud Desert and The Waterhole revisited by FB... and the Lost World by AM;) 10x10 ps, 1h
...nice work there.
Tks kk, much appreciated:)
Just another test.
Wanted to render a scene that's completely lit from the outside. 8x8 pixel samples, 2048 irradiance samples, f-stop 1/4, ISO 6400 and shutter time 1/2, so the interior looks very well exposed.
Indoor lit version.
...love the AO in that.
Nice! Looks pretty clean too;)
Run into some problems while testing some skin shader settings. A Genesis figure with hair and some fibermesh body hair, skin SS samples 256, shadowsamples 1024, HDRI lighting, raytracer final with10x10 ps. Rendered in bucket mode, cancelled the render when 26% done after 55 min. Finally found the reason, I used the AshikhminShirley classic on the cornea and fingernails. After changing every BRDF to the default AshikhminShirley it completed in under 20 min. I was under the impression those two BRDFs was pretty similar, only using slightly different ramping, but apparently not:)
Ashikhmin-Shirley Classic, Cook-Torrance and Cook-Torrance Classic are there basically for fallback and debugging. I recommend sticking to default Ashikhmin-Shirley.
I'm actually considering releasing a new shader built of AWE Surface, but with much of the controls simplified (either hidden or removed) and re-ordered in a more logical manner. Internally it would be the same shader.
Proposed arrangement:
The biggest change will be it will be a strict PBR Metalness Roughness workflow shader. So, all your diffuse/base/transmission color textures will go to the 'Base Color' and all roughness will be controlled with the 'Base Roughness'. You can still use the second lobe with different roughness settings/textures if you want.
Other notables:
Some other ideas:
Base color will be the surface color, currently diffuse color. DAZ Studio renders this in the viewport, so you'll be able to at least see/select objects with their color/texture.
I'm considering using Glossiness, so you can actually have some idea of specular highlight in the viewport. It will be converted internally to roughness.
As an added visual cue but isn't used in rendering (at the moment) is the ability to plug an environment map (HDRI or just lat/long jpg) into 'Reflection Color'. DAZ Studio viewport will display this in the viewport as a cubemap reflection. I found setting this to 100% gives a nice 'fake' reflection in the viewport for metals. For dielectric, reflection strength of around 12.5% looks almost like renders.
The biggest problem is that the viewport 'inverts' the reflection map. The viewport also don't allow UV offset for this map, so you'll have to adjust the offsets on the environment sphere to match.
@wowie
A bit unexpected to be honest:) But not necessarily a bad idea, I guess many end users would find that easier to start working with:) A couple of questions: How would transmission shadow be implemented? A fixed value? Since they would be the same shader under the hood, I guess they would convert well one way or the other? Any thoughts on how opacity would be implemented both in the new simplified shader and the next build of aweSurface? (If that's still in the pipeline)
Personally I like to have too many options, rather than too little, but I figure the simplified shader would be enough for most things, and setting it up would probably be a bit faster;) Very interesting!
Internally they would be the same. That includes things like opacity or transmission shadow. There are some adjustments, but shouldn't be nothing major and mostly made since the parameters exposed are slightly different.
I think I've finally got opacity optimization working right. Rather than allowing users to select the actual raw value, opacity filters now mixes between 0 to 0.85/0.9. Previously having opacity filter 2 at 100% will make everything be filtered out and makes everything disappear. SInce now it's limited to 0.85, that won't happen.
Since they share the same code, they will be updated concurrently. As for transferring settings, I think they should, though some stuff with renamed parameters will likely need to be manually transferred.
Looking forward to seeing this new shader;)
https://www.daz3d.com/blue-planet-orbital-view converted to awe:
...very nice.
Tks:)
Did a quick convertion of https://www.daz3d.com/starship-astra, progressive render:
Any tips for making it look smoother in close up? I already converted to SubD, but had to use the bilinear algorithm to avoid deformation, and so it made no difference, still a bit jaggy with this high contrast lighting:(
If you don't mind manual work, spot render troublesome spots at higher res ie 1080p, then use an image editor to downsize with a bicubic filter. The much harder way is to a geograft or replace the troublesome parts with a higher res mesh with proper SubD friendly topology.
If you have several places to smooth out, someone long ago suggested making a plane attached to the camera to block out parts you don't need rendered. Make sure the planes are visible to camera but have everything (diffuse,spec,reflections, shadows) turned off. Here's an example setup.
Pretty effective though not as effective as using AOV (arbitrary output value) mask.
Another set of tests.
For hair, the default settings already works pretty well.
Tks wowie, I'll see how that affects rendertimes, nice tip!
Another Astra render with 16x16 ps, diffuse bounce 6 and ray/specular depth 16, looks cool from this angle without having to postwork. One area light as the main lightsource:
...wow looks almost cinematic.
Decided to rework how subsurface takes into account actual diffuse texture. For one, the rendered brightness/luminance should be roughly equal between 0 and 1.
What do you all think?
Care to ellaborate a little? Not sure what you are doing there...are the numbers diffuse strength % for a fixed SSS value?
Its this - Use Diffuse Texture SSS.
It mixes the diffuse texture used with the result of subsurface. It should help vary the subsurface a bit more. Earlier it would just take into account the texture, but I'm tinkering to make it also vary subsurface a bit depending on the how much light hits the surface.
Aah ok:) Well this says something about the complexity of your shader, I've been playing around with it now for..what..six months, and there is still a whole bunch of things I haven't really looked into yet
Yeah the math is getting a bit comlex too...first set a diffuse strength value, then SSS strength with all those other SS parameters, and finally mix in the diffuse with the result, I need to do some testing before commenting any further at this stage
But I can see it makes fine tuning easier from the renders you posted
So how does the SS know how much light hits the surface? Almost sounds like some sort of dynamic control
It is. Only really noticeable in certain situations though. Not going to look any different with a HDRI environment or really big area lights.
Now, this one should be easier to comprehend. Better opacity optimization code.
Original value without optimization.
100% optimization. Most of the very thin strands are now retained.
Render times are 8 minutes 11.16 seconds for unoptimized, 4 minutes 27.46 seconds for 100% optimization.
With the new code, most of the time you only need to adjust the first filter value.
This is great news:) I've been struggling a bit lately trying to get some complex hair models to both look good and render in a descent amount of time;) So very much looking forward to this update! Nice Job!
This sounds excellent, adjusting opaciy on hair models is a bit hit and miss, so any simpliication will be very gratefully recieved!