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
BTW I did a test with my old indirectdiffuse()-based GI light, and it seems to work the same or even a bit faster in 3DL 12 (with RT SSS). As for UE2, it seems to be the same speed, as long as there is no SSS. When I enabled SSS - the oldschool one, since I was using US2 for the UE2 test, to better match Wowie's tests - even the 11.something build in 4.8 started to drag itself through molasses. Remember my Fiery Genesis rework which used to render in 15 mins with 5 diffuse bounces when I originally made it, either in 4.6 or 4.7? Not anymore. I stopped the render after it barely started to render the hair when the 15 min mark was long past... and there were only 2 diffuse bounces.
So there's something specific to UE2.
I changed the images in the opening post, BTW. The Girl with the bottle may return one day... when I have re-rendered the bottle (I never noticed the liquid was modelled the wrong way there).
Considering that 3DL does understand 'exr', natively, that's a reasonable idea.
I noticed something very unusual, tonight...
I'm making a freebie and I'm doing a set of 3DL, Iray and Luxus materials for it. And for the Luxrender render I only bothered to convert/set up the thimble. I have my SSS shader on G2F...and Luxus seems to have done a fantastic job autoconverting it!
This has only been running about 10 mins...
Interesting! Could Luxus be converting shaders based on their shadeop calls??
Could very well be...here's an excerpt of the lxs file
That would also explain why some 3DL shaders don't auto-convert well.
Which is why I stuck with 4.7.
4.8 offered very little improvement to my setup. I did a test back when it's launched, but i forgot if I used a character with SSS (don't think so).
Thanks for the info though - I'll make sure to note that for the best performance with my sets, DS 4.7 should be used. Well, unless DAZ developers revamp UE2 or ask omnifreaker to do it.
I'll be very surprised if that happens though. Technically, since i'm just using AO, any AO lights can be used (or even the point cloud based occlusion script).
4.7 has the prettiest blue icon, too =) I have it archived.
I haven't tested UE2 with AO only and SSS. I will try.
With just AO, it generally works fine. It's probably why no one noticed the problem - most people don't use the IDL or BounceGI mode (with raycache), even DAZ devs. I'm thinking of filing a bug report of UE2 behaviour with recent DS and 3delight builds, but I don't think it will be fixed.
Yeah, they are probably all using Iray now.
-------------
I didn't like the way gems are turning out in my renders although I'm kinda setting them up right, so I found two useful web pages on jewelry photography.
http://www.gemphotography.com/
http://www.gemologyonline.com/Forum/phpBB2/viewtopic.php?t=6420
A few more Oren-Nayar roughness values: http://wiki.blender.org/index.php/User:Guiseppe/Oren_Nayar
I was just playing, and noticed that well, it's taking kind of long time for this spot-render, lol.
It's running a tad over a minute per 1% at this point, and slowing down, lol.
So, perhaps I could have used a 'Lighter weight' shader on the hair tie, tho it still looks good.
So, MJC1016,
http://www.sharecg.com/v/83017/view/7/Material-and-Shader/Metals
Is this metal shader falling victim to the broken ray-caching thing? Or is this simply my pore choice of highly reflective curved surfaces again?
It's UberSurface presets, so it shouldn't be too horrible without raycaching.
I didn't do anything with the Occlusion override.
But the slowing down seems more like a transmap occlusion issue (the hair) than anything else.
That very well may be it. The hair is kind of going threw the ribbon mesh in a few spots. I'm sure I set it all up with 3DL mats, tho it may be a zone somewhere still on Iray as well. Still, It looks good.
(EDIT) Nope, it's ALL 3DL omUber. Hair and the ribbon. So it's Uber Opacity and Velvet on the hair clashing for CPU power, lol. Still looks good.
On the hair, with omUber, you can speed up the render by turning on the Occlusion Override and setting it to something as high as 128.
Nice article.
http://www.hyperfocaldesign.com/hdri-tutorials/clamped-mdris-vs-un-clamped-hdris/?awt_l=6x33k&awt_m=3wmmv5pZsAWv1IX
So according to those articles, it's 16-18 exposure steps for the "perfect" sun... I'll need to keep that in mind. BTW what program (outside Photoshop) could show HDR pixel values?
Then, I was reading the 3DL changelog, and here's an interesting thing:
12.0.35 - 2015-11-11
Which means: a) the devs are not forgetting about "oldschool" shadeops completely yet (so the UE2 issue is not because of 3DL updates); b) the work on the OSL implementation is being done as we speak...
I'm beginning to think that the whole omnifreaker line is more Renderman than 3DL internally.
Given all the references to Renderman I keep seeing all over the place, that would not surprise me. Especially if it was an attempt to make them more compatible for porting surface settings from one to the other.
I fussed with UE2 maps briefly, and my biggest issues I had was solely with capturing a map with the proper 'Latitude' angles in it. So while that was a good read Wowie, it mostly went over my head, lol.
Not to say, GIMP dose not have a 'Cylindrical' distort of any kind, so this was the only way I could do that.
So I kind of gave up on that for the moment. I kind of have allot on my plate to be wasting my time on a single 'Side order' item, lol.
Quite possible. Just check out the mysterious US2 bounce depth controls. These aren't 3DL attributes (those would go to the rendertime script, and as it's not encrypted for US2, it's clear they aren't there), these are something else. They _add_ IDL brightness when set higher than the overall scene max diffuse depth. There's another thing about UE2 - it tries to rotate the coordinate system in some way, and again, it's not the 3DL way like the one I'm using (define a new named coordsys and just pass it to the shadeop).Yes, the HDRI information is very nice. Shows just how much you're missing with low range/medium range HDRI.
But what is generally looked over is the fact that lighting has a very wide range. In an outdoor environment, most of the light will come from the sun or parts of the sky where the sun shines through. On a clear day, the sun will be able to cast somewhat sharp shadows, while with cloud cover, it should be less defined mainly because the clouds pretty much acts as a diffuser.
And what about the ground? I bet that is not pure black with no reflection of light back up? lol. That's why I was fussing with a UE2 map for my chamber, the floor is reflective, lol.
Sorry, I couldn't understand exactly what are you saying? If you're referring to making the ground reflective, the diffuse should be very low in strength. Not necessary black.
Zarcon, UE2 won't help with your mirror-like floor. Mirror bounces in a "specular" (sharp) way, not a diffuse sort of like UE2 can do. UE2 would make the scene look as if the whole floor were a light source.
Besides, there is no way to reliably figure out the coordinate system UE2 uses. So, well, you shouldn't bother.
I have seen people in this community not being able to recognise this simple fact. Maybe they all live in the tropical zone and hence have never seen an overcast sky IRL, I don't know. The difference in exposure levels for a clear sun and a cloud-covered one, as shown in the articles, is staggering.I live in an area with an average of 211 cloudy days/year...I know ALL about exposure levels of cloud vs cloud-free sunshine.
There were serveral very long threads on the misalignments of UE2 and basically, it's a real head scratcher as to what, other than a much different coordinate system (yes, even converting maps to it's system still leaves some misalignments). It's almost as if the original coordinate transforms were not 90 deg, but more like slightly under and slightly over.
Well, I live in the equator. Must be in the minority.
I've yet to find a HDRI with range great enough to have a somewhat strong sharp shadows though. Technically, if you have one, you can use it to have a rough idea of how a pure HDRI light rig should be in terms of intensity. Even if you have tinting on the non-sun areas, you can always dial down the saturation with UE2. Once you've worked out what UE2 intensity should be, it's easy enough to figure out what the intensities for other lights should be.
I haven't done much with UE2 re:shadows because of the rotation issue. Here is a test quickie map made in Krita: a boring grey gradient sky with a bright "sun" - which is too large, I think. There is only 10 exposure steps I managed to squeeze in, though (I don't know Krita well enough yet to use the gradient tool with various exposures, and by default it paints at zero).
http://www.mediafire.com/download/9tttd9yffpzorrr/test_10steps.exr
// in case anyone likes it, I release it under CC0 - public domain //
And here are renders with my shaders and this map at 64 GI samples and 8 pixel samples (no other light sources - the surfaces use glossy reflection and get the map from DelightGI)
The hair on M4 is one of Kozaburo's older freebies - it's too simple to look convincing without a diffuse map =( No layers = no depth.
I'll check it out later.
Oh yeah, got an email from The Foundry. They're doing a behind the scenes look of Jurassic World with Image Engine. :)
https://www.thefoundry.co.uk/events/sessions/anatomy-of-a-blockbuster/
17th December, 7 PM GMT.
Got around to testing your HDRI and the freebie from Hyperfocal. At 0 contrast and saturation, I can drive UE2 intensity scale to 500%. With contrast and saturation at 100%, the Hyperfocal one behaves better (I think). I didn't get any tinting until I hit 50% contrast. Intensity can be lowered to something like 150%.