Daz 3D is part of 
Connect
DAZ Productions, Inc.
7533 S Center View Ct #4664
West Jordan, UT 84084
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2024 Daz Productions Inc. All Rights Reserved.
Comments
So, after reviewing the whole thread I decided to go back and finish some things I had started. (The skin issue totally got me sidetracked) And since pwGhost now is working in both 4.9 and 4.10beta, I fixed the spotlight. Have yet to try the glowshader, but the ghostshader works perfectly well for the light beam;) It's just a simple matter of finetuning opacitymaps and shadersettings...granted it's a cheat, will not allow for godrays, but it's a nice cheat.
I remember having problems with the gel I made for the arealight. Figure the update will fix those issues.
Btw, are normal mapping enabled on those renders with faceting issues?
AoA and UberSurface too, hmmmm.
So if you take any of these scenes and render it through someone else's DoF camera, what will happen? Here's a camera.
No normals!
Tks, will try!
Tested the G2F Beau (AoA SSS) with your camera in vanilla/progressive. No artefacts!
My glitchy cam, use at own risk;)
Testing now with G2F basecharacter in raytracer final/progressive. Indeed there are some weird things going on even with Kettu's cam! Will update when render is finished.
edit: Ok With Kettu's camera, RTfinal/progressive, are the artefacts I'm seeing just a result of mipmapping/ texture optimization?
Manually copied the camera settings and translations/rotations to my camera. Looks identical to the one above. FL was at 800.
Moved my camera closer and set FL to 250, still same render settings RTfinal/progressive
I can see some subtle faceting on the forhead + those weird patterns on the collarbone area.
Manually copying values to Kettu's cam produces identical results!
So, new scene, loaded G2F, changed to vanilla/progressive, loaded Kettu's camera and my camera. Reset both cameras,set FL to 250, identical placement/angle. Both cameras give identical results, so...still in the very dark with this...(lines on the sides of her neck and underarms)
Testing some more brute force pathtracing:
I went back to see if I could find what caused the black streaks I got in my US2 renders and indeed I did! Depth of Field!
RTfinal/progressive through a camera with no DoF: (FL 48)
and with DoF
and with DoF and bump off for the road tiles: (made no difference really, slightly different look)
Depth of Field is the root of all evil (Sven 1:1)
Mighty weird. I have rendered basically "everything" over the last few years, including a good load of ancient Poser models falling apart with or without sub-d... but I've never seen anything just like this. Bad shadow terminators with shader mixer lights - yes. And that's literally all. Always the raytracer, always DoF.
No luck locating a Mac installer of 12.5.9 yet?
These actually look like UV seams. The others don't, though. =D
So that was just G2F, starter content? Would you please upload these scenes?
Yeah, you've seen a good number of my renders done in vanilla with or without progressive. Always DoF, progressive only with a lot of reflective/refractive stuff. But then again, I very rarely do those large size portraits with a high FL, so... maybe I've had those artefacts, only didn't spot them? But now when using the raytracer, they seem to pop up where you least expect them to?
I looked around a bit, still no luck, will try some more!
Hmm didn't think of that. Could be true, I mean some of those camera settings I've been testing with are pretty abnormal to say the least: /
Yup straight out of the box. Hmm gotta admit I ragequit, didn't save the last test
Shouldn't be too hard to recreate, though
Started to experiment with some gems.. trying to understand what's affecting what.. this is my first attempt
I guess I should have to dig up the actual refraction- and absorbtion values for these things, but for now just testing the various transmission channels...
Trying to figure out if "using base values" is the way to go or not...I have that enabled on these renders.
Increased IoR slightly. Still don't know if it's better to decrease base transmission (I used 100%) if I want the center stone darker, or if it's better to increase absorbtion? Or use a base transmission of 50-60%, which also affects the transmission shadow? Setting base color to a darker shade seems to dampen the transmission shadow surprisingly much. Any tricks up the sleeve, anybody?
IOR values should transfer well, but I doubt you need to fiddle with absorption/transmission values. Just pick a transmission/refraction color and adjust scale/absorption depth as necessary. I'd love to add diffraction/dispersion, but bug fixing and optimizations comes first.
Tks, yup IOR values are easy enough to find, then I figure it also comes down to having accurate angles for a physically plausible model...
I'm with you there
, still, started drooling my morning coffee, when I read this
Converted the WP without issues:) Even dared to use DoF;)
You'll also need a proper model, which can be difficult for actual gemstones/jewelry.
Always liked that set. Though I generally used the clean version.
So it has been confirmed that the DoF/skin problem is not Mac specific! Still don't know what is causing it, but it seems clear that extreme camera settings are to be avoided:) If you happen to get this thing, try adjusting focal length and/or distance and camera placement a little, that may do the trick. Still considering filing a bug report...
I've messed around for a while, and yeah, I too managed to reproduce the issue with M4, G2F and G3F on Windows, even with my own shaders.
The seams issue on G2F is worst with AoA, noticeable with UberSurface2, not noticeable with my shaders.
G3F is less "problematic" in that regard - if there are seam issues, I can't see them even with US2; but with focal length over 200, her shoulders or collarbones start to get messy as well, with those weird lines, but only for certain camera angles.
M4 is the worst offender, with or without sub-d.
I don't think it's such a huge issue, though, since it's only triggered by focal lengths well over 100. And do we absolutely need them for what is essentially portrait shots?
IRL, focal length is tied to depth of field, vignetting and the like. In 3Delight, it's not. Moreover, we can always move our camera close enough.
In theory, supposing we were to match some specific plates shot with a telelens, then yes we would find ourselves using high focal lengths. But would a CG human figure then fill the screen? Unlikely.
...the interesting thing is, why we notice these issues on human figures but not clothing?
Tks for testing it out, very interesting, haven't tested the gen4 stuff, the differences between generations are linked to Uv-mapping? My G1 character uses the V4 corrected Uv. And it was a mess. G2F, much harder to mess up=)
Yes, as I said, those camera settings were far from normal, sure there are workarounds.
This!!! Farout, don't know what to think of it:/
Interesting... have been playing with Garibaldi (converted to aweS) and aweSurface. It actually works in vanilla/progressive, but if I switch to scripted, the Garibaldi hair is ignored:) But it works with scripted rendering and IBLM. Rendertimes are not exactly blistering fast, though;)
Are you sure it works properly?
I tried scripted render with IBLM. In the viewport, the hair moves with the model and takes style changes - but when I hit render, the hair remains in default position and ignores any style changes made.
Oh well, at least it is rendering in scripted.
First time I tried, it suddenly became invisible
Seems to work now, have a render going;)
Yeah, but frankly, in vanilla/progressive with AoA lighting, it renders in a fraction of the time it takes with awe in scripted, not sure if it's worth trying...but in the name of science...
It will definitely render faster with AoA lights.
Under lighting control in the light properties you can set a flag for Garibaldi hair and use "primitive ht mode and alt samples" and watch it go!
Interestingly, for me, with AWE shaders applied default progressive is much faster than no progressive. Normally, it is much slower for me - even though it is supposed to be faster.
...OK this is not encouraging considering I am looking at using AweSurface when skinning my SF characters.
Curves should be faster to render than a comparably realistic-looking transmapped model (OOT recent stuff as a baseline? Could work). If they don't, then some optimisation is in order =)
Make sure you aren't using any of those awe hair presets that use SSS - you don't need it on curves. And I'd say you can safely dial diffuse bounce depth dial on the hair down to 1. Or, unless the environment is your only source of lighting, exclude the hair from GI whatsoever. It will still "light itself", just won't get diffuse-traced by other surfaces.
In case any of the presets enables more than one layer of specular, you can try turning one off and see if it makes any difference.
Then see if irradiance samples can be decreased.
I'd also try to get the hair look nice with 100% opacity.
Vanilla "progressive" won't be faster unless there is heavy raytracing going on. Same as using our render script to render oldschool stuff - unless you have refraction, raytraced shadows through transmapped layers and/or AO (especially through transmapped stuff).
Both "progressive" and our script are calling the raytracer, and telling the raytracer not to raytrace offends it =D