3Delight Surface and Lighting Thread

1272830323352

Comments

  • SaiyanessSaiyaness Posts: 715
    edited December 1969

    Bugger! Oh well! :)

  • ZarconDeeGrissomZarconDeeGrissom Posts: 5,414
    edited December 2014

    you will need this, not so free. I picked up that shader hair thing, yet to get the US2? thing.
    http://www.daz3d.com/ubersurface2-layered-shader-for-daz-studio

    And no, I was not hoping it would just work without it. It was a why not get that to have something to try UberSurface2 on later.
    (edit)
    opps, didn't intend to pounce like that, sorry wowie. :red:

    Post edited by ZarconDeeGrissom on
  • wowiewowie Posts: 2,029
    edited December 2014

    Saiyaness said:

    Is it included in 4.7? If it's in the Gen2 Essentials packages, I'll cry. :p
    Hmm, I'll see whether if just downloading the Default Lights and Shaders reinstalls Uber. I'll have a look! :)


    No, UberSurface2 isn't part of the included lights and shaders, unfortunately.

    UberSurface2 is a standalone product.
    http://www.daz3d.com/ubersurface2-layered-shader-for-daz-studio

    you will need this, not so free. I picked up that shader hair thing, yet to get the US2? thing.
    http://www.daz3d.com/ubersurface2-layered-shader-for-daz-studio

    And no, I was not hoping it would just work without it. It was a why not get that to have something to try UberSurface2 on later.
    (edit)
    opps, didn't intend to pounce like that, sorry wowie. :red:

    No problem at all. Actually, I think it's wonderful everyone is just so happy to help each other out..

    Post edited by wowie on
  • AnIronButterflyAnIronButterfly Posts: 252
    edited December 2014

    Saiyaness said:
    wowie said:
    You will need the latest UberSurface2 shader to use it.

    Is it included in 4.7? If it's in the Gen2 Essentials packages, I'll cry. :p

    Hmm, I'll see whether if just downloading the Default Lights and Shaders reinstalls Uber. I'll have a look! :)

    No, UberSurface2 isn't part of the included lights and shaders, unfortunately.

    Thanks Mike... this covers my problem to--which we were just talking about!

    And per DIM, the latest version of Uber Surface 2 Layered Shader is from May 2013. Is there a more recent version than that?

    Post edited by AnIronButterfly on
  • AnIronButterflyAnIronButterfly Posts: 252
    edited December 1969

    And it looks like an uninstall/reinstall of UberSurface 2 via DIM has fixed things right up for me. It might just be that my shader files weren't playing nicely with newly installed stuffs. All things appear to be better now... :D

  • wowiewowie Posts: 2,029
    edited December 1969

    A relatively quick IPR render with some of the freebies available.

    test2.jpg
    1642 x 924 - 556K
    test1.jpg
    1642 x 924 - 575K
  • Mustakettu85Mustakettu85 Posts: 2,933
    edited December 1969

    wowie said:
    I think omnifreaker should really update UE2 with the parameters the script exposed (ray caching).

    These particular features are not up to him, I'm afraid, but up to the DS developers. Ray cache is a RiOption, and in DAZ Studio, shaders can't set RiOptions. Only the "scripted renderer" or a "shader camera" can.

    And thank you very much for the Christmas freebie! =)

    -------

    Not sure. From what I've read from his docs and the parameters I've seen, I rather think it's a simple occlusion() call added to the flagging stuff. AOA just exposed some the rest of the occlusion parameters that you don't see in UE. And instead of making it like an ambiant/IBL light, he made a point light. Then he derived the principle to get the other he made (spot and distant)

    Advanced Ambient definitely uses occlusion(), but the AdvDistant light does not seem to (I don't have the Spot light, but from what I've seen in other people's renders, it is close to the Distant one).

    I've made directional lights for myself that cast "extra soft directional shadows" via occlusion(), but they give a distinctly different look. AoA's Distant looks like a transmission()-based shadow - besides, you can get multicoloured shadows with AoA's directional lights, right? With occlusion(), I am not sure it can be done without much hacking because while transmission() returns colour, occlusion() returns float.

    And surface-to-light tracing, I think he mentioned it on the forums somewhere himself. Transmission() is the most immediate solution for that.

  • StrixowlStrixowl Posts: 301
    edited December 1969

    I own UberSurface2. Still not working :long:

  • SaiyanessSaiyaness Posts: 715
    edited December 1969

    I didn't even notice that you needed that extra product! My bad! :p I forked out the $20 for it. I hope it has other uses (that I will actually use.) o_o I'm going to blame it on the new website - I always overlook the Required Products bit now. It kind of disappears. >_>

    Haven't had a good play with the shaders yet. I wanted to see how they played with Ali's hair products. :)

  • AnIronButterflyAnIronButterfly Posts: 252
    edited December 2014

    Strixowl said:
    I own UberSurface2. Still not working :long:

    Strixowl, I had to uninstall and then reinstall mine. It might be worth a try....reinstalling, UberSurface 2 that is...

    Post edited by AnIronButterfly on
  • wowiewowie Posts: 2,029
    edited December 2014

    Saiyaness said:
    I didn't even notice that you needed that extra product! My bad! :p I forked out the $20 for it. I hope it has other uses (that I will actually use.) o_o I'm going to blame it on the new website - I always overlook the Required Products bit now. It kind of disappears. >_>

    Haven't had a good play with the shaders yet. I wanted to see how they played with Ali's hair products. :)

    I think you'll find a lot of uses for it. UberSurface2 is a very useful shader.

    I did note the MATs require US2 when I submitted, but seems like they didn't put that info in the product page. The readme has some usage tips, so be sure to read that. Fastest way to access the page is by clicking the Info button in the DAZ Install Manager (DIM).

    Just in case it's not available, here's the notes:

    Hair MATs Usage:
    Load your hair prop as usual. Apply the hair MATs so the hair opacity maps are loaded.
    With all hair surfaces selected, remove any texture loaded in the diffuse color, specular color, and specular strength channels.
    With all hair surfaces still selected, Ctrl+Click on the the 'Hair.duf' MAT preset. This will bring up extra options. Make sure 'Selected' and 'Ignore' are used. This will load the MAT but retain the textures used.
    Optionally, you can use the 'Hair without anisotropy.duf' MAT preset to use the non-anisotropy version of the MATs.

    By default, the MATs comes with the Blonde Light colors. To apply a different color, select the hair surfaces and double click on the color MAT preset you want.

    Hair Geometry Shell MAT usage.
    Select the hair prop and create a geometry shell. Leave the default options so the geometry shell will load as child node of the hair.
    Select the geometry shell you've just made. Apply any of the supplied hair MATs that came with the hair prop. This will load the opacity maps including other maps as well. We only need the opacity maps, you can remove any other maps if you like.
    With the geometry shell selected, go to the Properties tab. Go to the General section. Make sure that the Offset Distance is set to 0 (zero).
    To avoid having the scalp cast shadows, go to the Shell section. Make sure the scalp (it will either be called Scalp, Skull Cap or Base), is turned off.
    Select all the hair surfaces on the geometry shell. WIth the surfaces selected, apply the 'Hair Geo Shell.duf' to make the hair cast colored shadows.
    Last step, go to the original hair prop and open the Parameters tab. Go to the Display section. Make sure 'Cast Shadows' is disabled.

    Known Issues and Workarounds:
    The skull cap with some hair props may appear not entirely correct near the hairline. To minimize this problem, you can change the skull cap's tiling offset. Usually a change of <1 (ie, 0.9 or 1.1) will fix the problem.</p>

    Some extra notes I forgot to add:
    Generally though, you need a hair prop that has sufficient amount of hair surfaces/layers. If a hair prop relies on the diffuse map to add those details, you won't get the xame look, but the MATs can still be used. If I recall correctly, you'll need to insert the bump or diffuse map in the specular color channels and transluency channels. To add some additional brightness, you can enable the 2nd diffuse channel and insert the opacity map to the 2nd diffuse strength channel.

    If you're using an ambient occlusion light, youyou may need to disable occlusion on some of the hair prop surfaces (not the geometry shell). If you find the hair loses most of its dark areas when you disable occlusion, you can add a geometry shell as an additional shadow blocker. The shell should have a mesh offset of -0.1 (so it's slightly under the hair). Just remember to apply the original hair MATs (to load the opacity maps) and apply the geoshell MATs to that geometry shell.

    If you want to speed up renders, you should change the Occlusion Shading Rate Mode to 'Override' and Occlusion Shading Rate to '128' for the hair surfaces.

    Post edited by wowie on
  • ZarconDeeGrissomZarconDeeGrissom Posts: 5,414
    edited December 1969

    really nice render wowie. I've been ripping out my hear over the "Turbo Encabulator" needed to make Dynamic cloths work, and have yet to even install any of the free stuff this month. I still have a million things I need to dredge up answers on making that "Turbo Encabulator" work as I need it to. lol.

    I've been burnt a few times by the requirements the past nine months, and have gotten so pessimistic about that, I dig threw everything I can find, that states what is needed for each application and why. Just so I know that I am getting something that will actually work, and I have all the necessary stuff to make it work.

    Unfortunately I often end up trying to understand threads and posts that read as follows;

    For a number of years now, work has been proceeding in order to bring perfection to the crudely conceived idea of a transmission that would not only supply inverse reactive current for use in unilateral phase detractors, but would also be capable of automatically synchronizing cardinal grammeters. Such an instrument is the turbo-encabulator.

    Now basically the only new principle involved is that instead of power being generated by the relative motion of conductors and fluxes, it's produced by the modial interaction of magneto-reluctance and capacitive diractance.

    The original machine had a base plate of pre-famulated amulite surmounted by a malleable logarithmic casing in such a way that the two spurving bearings were in a direct line with the panametric fan. The latter consisted simply of six hydrocoptic marzlevanes, so fitted to the ambifacient lunar waneshaft that side fumbling was effectively prevented.

    The main winding was of the normal lotus-o-delta type placed in panendermic semi-boloid slots of the stator, every seventh conductor being connected by a non-reversible tremie pipe to the differential girdle spring on the "up" end of the grammeters.

    The turbo-encabulator has now reached a high level of development, and it’s being successfully used in the operation of novertrunnions. Moreover, whenever a forescent skor motion is required, it may also be employed in conjunction with a drawn reciprocation dingle arm, to reduce sinusoidal repleneration.

    It's not cheap, but I'm sure the government will buy it. [unfurls $750,000,000 price tag]

    You know, a diagram or a screen-cap would have been nice with allot of the posts like that, lol.
    Hoping y'all had a good holiday. :coolsmile:
  • Mustakettu85Mustakettu85 Posts: 2,933
    edited December 2014

    Happy New Year everyone!

    Here's a render I made of V6 (materials based on my shader with raytraced SSS) and Ali (Mankahoo) Hr-074 with Wowie's Lumina presets - because this hair geometry does not have many layers, I added the bump map back and set it to 200% at +- 0.25. Makes the hair look somewhat damp, hence the beach theme =)

    The swimsuit is this: https://amarijun.wordpress.com/2014/05/10/aj-hydrangea-beachwear-for-genesis-2-female/ (I hid the flower decorations because there seem to be issues with normals).

    Render time is 5 minutes on my laptop, there are two area lights and a GI light with an HDR map (hair excluded from actual GI - but of course it is lit with the IBL). The glass on InaneGlory's candle holder uses my shader that combines raytraced reflection/refraction with the same envmap; mesh lights on the candle were turned off and replaced with a small custom point light with soft shadows.

    The only downside about raytraced SSS right now is that it is sensitive to area light quality. Larger, curved mesh lights with dense geometry always need more samples to resolve cleanly, and raytraced SSS makes it more evident: you need more light samples to decrease SSS noise. Simple flat planes work best for meshlights. The key light here is a concave sphere at 160 light samples - which may slow the render down when using hair with more layers (Mairy&3Dream;'s Gregoria hair takes about 10 minutes to render with this light).

    v6_artsy_5mins.png
    700 x 800 - 399K
    Post edited by Mustakettu85 on
  • wowiewowie Posts: 2,029
    edited December 1969

    Happy New Year everyone!

    Here's a render I made of V6 (materials based on my shader with raytraced SSS) and Ali (Mankahoo) Hr-074 with Wowie's Lumina presets - because this hair geometry does not have many layers, I added the bump map back and set it to 200% at +- 0.25. Makes the hair look somewhat damp, hence the beach theme =)


    Happy New Year.

    I think the skin it lost too much diffuse details, but the SSS looks very good.

    The only downside about raytraced SSS right now is that it is sensitive to area light quality. Larger, curved mesh lights with dense geometry always need more samples to resolve cleanly, and raytraced SSS makes it more evident: you need more light samples to decrease SSS noise. Simple flat planes work best for meshlights..

    I don't think that's a problem with raytraced SSS, but rather area/mesh lights themselves. In some extreme cases, where there's visibly more dark/shadowy areas, you have to raise the samples to 48 or 64 even with a very simple primitive (ie a plane with one polygon).

    Instead of a sphere, you can use several planes to form a ring around the scene and scaling up the planes to cover a larger area.

    I don't know if its possible to implement adaptive sampling for SSS ala AoA's light, but even without it, raytraced SSS should still be faster if you have lots of surfaces with SSS and different SSS group IDs.

  • Mustakettu85Mustakettu85 Posts: 2,933
    edited December 1969

    wowie said:

    Happy New Year.

    I think the skin it lost too much diffuse details, but the SSS looks very good.

    Thank you!

    wowie said:

    I don't think that's a problem with raytraced SSS, but rather area/mesh lights themselves. ... I don't know if its possible to implement adaptive sampling for SSS ala AoA's light

    What I mean is that raytraced SSS needs more area light samples to get clear of noise than non-SSS surfaces do, when lit by the same area light. It should be less of an issue when there is a diffuse channel layered over SSS. So it all depends on artist preference.

    And no, unfortunately adaptive sampling AoA-style does not work for "oldschool" mesh lights - light samples value is a RiAttribute, it is set once per light-emitting mesh.

    The new "trace"-strategy area lights (the ones that make use of path tracing and MIS) should eliminate this issue completely (the surface drives the samples), but we need _all_ surface shaders to support this. Implementing this in my shaders is on my to-do list; the devs might also be working on something (they apparently added the "trace" option to shader mixer mesh lights, so that may be a sign); let's see if "Kettu soon" or "DAZ soon" wins =D


    Instead of a sphere, you can use several planes to form a ring around the scene and scaling up the planes to cover a larger area.

    That's true.

    What I do, though, is letting the GI shader do IBL for the "fill". When the HDR map does not have sun-like bright spots, it gives that overcast look that works well in conjunction with directional lighting, be those area lights or delta ones. For the key light in that render, I used a small half-sphere (the concave side emitting light); it's in the air, similar to a studio light source. The rim light is an elongated plane (it would have been visible, but it's transparent).

    I think I prefer mesh lights for rim lighting - they are easier to place than a spotlight rig.

    but even without it, raytraced SSS should still be faster if you have lots of surfaces with SSS and different SSS group IDs.

    Huge closeups with high RT SSS samples will be slower than oldschool, but other cases are faster, in my experience. No huge prepass pauses - I only notice a little pause with area lights sometimes.

    I made a better version of that render, with a different hair model and a different Lumina colour =) This took about 13 minutes on my laptop:
    https://www.daz3d.com/gallery/#images/51532/


    For inspiration... A description of light rigs supplied with PRMan:
    http://renderman.pixar.com/view/DP25750

  • wowiewowie Posts: 2,029
    edited January 2015

    Here's what I mean: 4 Uber area light planes around the center, with 4 distant lights (2 rim lights, 2 key lights). For the second render, i disabled one of the rim lights (the one on her right) and dimmed the Uber area lights on that section to about 1/3 intensity. I slightly rotated all the lights to her left so the light are more focused on her left/upper side.

    Skin is my materials for V4 Amy, with SSS scale set to 0.2 and backscatter boost set to 2.0. I added some blue diffuse to the 2nd diffuse channel of the iris to change the color and changed the color of the 2nd diffuse channel of the lips (so it looks more red/pink).

    Tweaked the Lumina Hair mats so it relies more on specular rather than velvet (which is disabled totally) and changed the glossiness/fresnel a bit..

    NYE1.jpg
    800 x 1040 - 406K
    FakeGI1.jpg
    770 x 1001 - 388K
    Post edited by wowie on
  • Mustakettu85Mustakettu85 Posts: 2,933
    edited December 1969

    That's neat. The fabric looks really nice, too.

  • SzarkSzark Posts: 10,634
    edited December 1969

    Can some kind soul point me to Mustakettu85's GI script. I thought I had it but it appears not.

  • wowiewowie Posts: 2,029
    edited December 1969

    Szark said:
    Can some kind soul point me to Mustakettu85's GI script. I thought I had it but it appears not.

    http://www.daz3d.com/forums/viewreply/642793/

  • SzarkSzark Posts: 10,634
    edited December 1969

    Thank you wowie

  • Mustakettu85Mustakettu85 Posts: 2,933
    edited December 1969

    There's also this small edit you might need if you ever use DSM on your lights:
    http://www.daz3d.com/forums/discussion/21611/P570/#650436

    This post tells how to kill the red background that appears after the 4.7 update (and also how to render to EXR):

    http://www.daz3d.com/forums/discussion/21611/P810/#719649


    I'm still in the process of writing documentation and creating presets for the kit, which will include polished render scripts, the raytraced SSS shader (which can also be used without SSS - but it does not support transmapping for a handful of reasons, there will be a dedicated non-SSS shader with mapped opacity a bit later), the GI/IBL light shader, and transparent shaders that combine raytraced reflection with environment maps (and physically-based anisotropic highlights) for non-shadow-casting stuff like corneas (with refraction) /eyesurfaces (no refraction), and stuff like light presets, surface materials for some base figures etc.

    The non-refractive transparent shader can also be used as a "coating" for any shader via the geometry shell trick.

    Here, the test scene is lit with Omnifreaker's KHPark HDR map and a specular light only; the materials are all DS Default shader (mostly "skin" mode with finetuned colour parameters - no "real" SSS), but most are overlayed with the "coating" with various settings: hair, eyes, lips, belts, buckles, hair decoration.

    gaia-bright-closeup.jpg
    622 x 800 - 112K
  • SzarkSzark Posts: 10,634
    edited December 1969

    thanks Mustakettu85 I will have a read and a play when I get a chance. So scripted rendering again...been a while since I did any.

  • Mustakettu85Mustakettu85 Posts: 2,933
    edited December 1969

    Szark said:
    thanks Mustakettu85 I will have a read and a play when I get a chance. So scripted rendering again...been a while since I did any.

    You're welcome Szark, enjoy =)

    I found an interesting guide by Paolo Berto: https://3delight.atlassian.net/wiki/display/3DFM/How+to+Eliminate+Sampling+Noises

  • Mustakettu85Mustakettu85 Posts: 2,933
    edited December 1969

    ...and then this: http://www.allegorithmic.com/pbr-guide - I haven't read it yet, but Allegorithmic make excellent software ( MapZone is old but amazing, and free: http://www.mapzoneeditor.com/ ), so I guess it must be awesome.

    I found this link on 3Delight forums, of course =) The free standalone has been updated to 11.0.131, BTW!

  • wowiewowie Posts: 2,029
    edited January 2015


    I found an interesting guide by Paolo Berto: https://3delight.atlassian.net/wiki/display/3DFM/How+to+Eliminate+Sampling+Noises

    Thanks for this. From a very quick scan, it seems like there's some things that simply needs to be set higher with DS, though of course, it will be different depending on scene and your preference for noise, or lack of.

    ...and then this: http://www.allegorithmic.com/pbr-guide - I haven't read it yet, but Allegorithmic make excellent software

    I posted these two renders in the 3delight vs Octane thread. Even with linear workflow (gamma correction enabled and proper target gamma), you need to make sure your white point is the correct value. Other wise, you'll be adjusting materials incorrectly.

    As in real life, exposure and ISO should be used in postwork, rather than adjusting your lighting (or worse, your materials) to match what you think is right. I do wish we have a camera with such controls in DS.

    The all diffuse ball is lit with 50% ambient light and 50% distant light, so the dark areas should only have 50% brightness. But if you use the wrong value, you'll end up with more brightness. As noted in Disney's BRDF paper, this is an RGB value of 160,160,160 (albedo for most materials is 40%). You just need to convert the sRGB values of known materials normalized to that value.

    2.jpg
    770 x 1001 - 34K
    1.jpg
    770 x 1001 - 38K
    Post edited by wowie on
  • Mustakettu85Mustakettu85 Posts: 2,933
    edited December 1969

    wowie said:

    Thanks for this. From a very quick scan, it seems like there's some things that simply needs to be set higher with DS, though of course, it will be different depending on scene and your preference for noise, or lack of.

    I did that A3 DOF render a few posts above with 6x6 pixel samples. It's much smoother than 4x4. I am going to see if 6x6 makes a difference for the "oldschool area lights + RT SSS" situation. According to this post by Paolo, it may help:
    http://www.3delight.com/en/modules/forum/viewtopic.php?f=4&t=4190#p21462


    I posted these two renders in the 3delight vs Octane thread. Even with linear workflow (gamma correction enabled and proper target gamma), you need to make sure your white point is the correct value. Other wise, you'll be adjusting materials incorrectly.

    As in real life, exposure and ISO should be used in postwork, rather than adjusting your lighting (or worse, your materials) to match what you think is right. I do wish we have a camera with such controls in DS.

    That's so true. I really enjoy rendering to EXR, tonemapping often helps me with getting the image "just right" (adjusting gain in test renders is too tedious, and besides, tonemapping is interesting in and out of itself; I use Picturenaut).

  • wowiewowie Posts: 2,029
    edited January 2015

    Did some experiments today after figuring out new ways of doing things with lights and shader settings. I need to reformulate my material settings, but I think the balls are close enough to physically plausible. Of course, glass still needs the geometry shell hack to cast shadows properly.

    Thankfully, the character MATs looks quite good with the new lights. I probably don't have to fiddle with them.

    Render below is with UE2, KHPark and two distant lights. Took about 50 secs even with US2 precomputed SSS.

    Test1.jpg
    800 x 1040 - 309K
    Post edited by wowie on
  • ZarconDeeGrissomZarconDeeGrissom Posts: 5,414
    edited December 1969

    geometry shell hack? Wouldn't the Refraction settings help at all, or is the angle of the cam just completely wrong to see any upside-down lens effect from stuff behind the sphere?

    Nice render by the way. The other spheres look really good.

  • wowiewowie Posts: 2,029
    edited January 2015

    geometry shell hack? Wouldn't the Refraction settings help at all, or is the angle of the cam just completely wrong to see any upside-down lens effect from stuff behind the sphere?

    Nice render by the way. The other spheres look really good.

    The geo shell hack is simply a hack using a geometry shell to cast the shadows and not the main object itself. This is needed since the shadow pass for US2 is hard linked to opacity settings. It's probably the one thing I wish UE2 did differently. If we can decouple the shadows and opacity, we can have a lot more interesting things like retaining surface opacity, but having more controls of shadows (and shadow colors, for bleedthrough).

    I've set the refraction to 1.0, so there's no light bending. Turning up the refraction value will exhibit the behavior you mentioned.

    Left to right: 1, 1.3 and 1,5. Only IOR is changed.

    Test2.jpg
    770 x 1001 - 224K
    Post edited by wowie on
  • Takeo.KenseiTakeo.Kensei Posts: 1,303
    edited December 1969

    By doing that you can't have caustics. Bad idea in my opinion

Sign In or Register to comment.