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
The thing is, in real life, these are just the flaws that people try to cover up. Magazines show perfectly manicured nails and airbrushed skin. No one wants to see pores and zits and pockmarks and ugly nails. Even on Instagram, people are using filters on selfies to cover flaws. Maybe, in some cases, realism is over rated?
I have been pondering the idea of realism in light of a huge number of photo's i have looked at over at deviant art. I find it all too ironic that most glamour nudes, fashion photography and portrait photos are manipulated to the point that the models have a doll like smoothness. Photographers are trying desperately to hide even the smallest blemish and us 3dr's are trying so hard to put the blemish back in. You dont even have to look for nudes just look at the fashion photo's so many of the models would put Vicky 4's smoothness to shame and yet we take the time to make every nook and cranny standout that they end up looking like they have been dipped in weird opaque cake batter with a coat of skin colored paint. I am am beginning to reallze that were 3d art is falling down is not the models or the props...but the lighting is so flat that has no life to it. We need more lighting solutions especially in iray being unbiased...should be making 3d counterpoints to gobo lights, sock filters, reflectors and diffusers so we can manipulate the light so its more world like.
Ok i have rambled enough.
Daniel
Um.. Daniel, have you looked at the lighting products by Colm Jackson or others from RDNA? Or Painter's Lights by jCade?
There are some great light products.
I have and put them on my wish list. Thats my total point in my rant though. ALthough not all models textures are created equal i just felt in light of what tools we have the "realism" debate is less about the models and more about lighting we have in 3d programs. The biggest issue in my opinion being the idea of light temps in Iray, although useful real lights either indoor or sunlight doesnt just conform to 1 temp across the board especially outdoor light the temp spectrum varies as it spreads out and bounces off other things. One of the reasons that i think 3d lighting in general is flatter that real world photography. How we get there from here i don't know the answer but i have been expermenting with different light rigs that use variations of light temps and colors. I think if we crack that egg we will see 3d artists bumping up agains that uncanny valley more and more.
Daniel
Not quite. The color temp IS 'fixed' in photography...in digital, by the white point setting of the camera or by the film/processing. Now, our eyes, on the other hand...
There is a reason Iray is setup to use an HDRI by default. That is to mimic real world light, as much as possible. The same with the ability to use IES profiles and have all the things like color temperature adjustable.
Also what is technically/scientifically correct may not be artistically perfect. Too often the 'correct' comes out lifeless because it is too 'perfect'.
Iray actually should take care of that for you to an extent. In Iray when light bounces off something it takes its color in the way it would in the real world. A simple example:
the only lights in this scene are white, as is the sphere. obviously this is an exagerated effect, but for a more real world example consider this in my gallery (there is postwork yes but none of it created color that was not in the original image) once again I used white lights but if you look at the shadows, particularly on the underside of the arm and breasts they are strongly purple, as they are lit primarily by light bouncing off the purple floor
The difficulty is mostly that lots of bounces in the light make things take much longer,
I totally agree, i guess what bums me is that we kinda have to work backwards. While true that light temps are fixed in the camera white balance and exposure the problem is that its the camera thats changing. Not the world. While in DAZ though the light temps of the world are fixed based on how we set them. So making variations is trial and error in any given scene thus slowing way down our workflow. What i would like to see is a white point balance slider in the tone map settings of the iray settings based on light temps instead of a palette. I have no idea what a palette number would be for say a light temp of 4000. I have no idea why they did it like that but it is what it is.
Going back to my orginal point though i just feel like the color matching we get doesnt allow for vivid enough lighting to get us to the point of that uncanny valley without huge amounts of post work. Also i have not messed with LPE profiles yet. I am assuming they help in many ways. BUt one thing i have noticed about the iray photometric lights is that the fall off is very steep. I will use my own office as an example. The room is 10 ft tall and the room is 10X14ft. The light in the celing is a 50 watt incandecent light bulb and it fills the room. If i copied the dimensions of this space in Iray and placed a 50 watt bulb at the top it wouldnt light the entire space most if it would be black. So although Iray is "Unbiased" It doesnt actually work like real world lighting.
Daniel
If you test with a white cube and a primitive sized to represent the bulb with the proper emissive settings you would find that it does indeed fill the room with a very diffused light with little or no shadows. It will not be very bright but because of the bounce potential there will be some light in ever corner at floor level. If anything the bigger issue would be that the shadowing is so diffused compared to a single small bulb. Since the room has the single diffused light source the render would have taken forever so this is a pretty grainy test.
It works exactly like real world lighting, Its just real world lighting as interpreted by a camera, not our eyes below os a lovely picture of my room in daylight...
...With exposure settings that match the daz defaults in the render settings panel ;). That is, f/stop of 8 and sutter speed of 1/128. (this is a genuine picture I took with my camera just now)
Please pardon the intrusion!
Do DOrdiales or JCade or Jimbow or Arnold C. have links (or screen shots) to skin, and eye settings you are using? Many thanks.
This is a compelling demonstration. Seems with these unbiased engines that knowing a bit about how cameras operate and generate images is useful. As a general rule I find that more light is better than not enough.
But then with human skin, there is yet another element. Which is that if a skin material has SSS, then it cannot reflect as much incoming light, it must instead absorb some of it, making SSS skins slightly darker than fully diffuse textures. So in general, we need more light for unbiased engines than we do biased engines.
We don't have a "human eye" camera available in any of the unbiuased engines. This is probably because while the physiology of the eye can be modeled well enough the psychology of the way our conscious mind ignores and emphasises certain things under certain conditions is impossible to model into a single neat algorithm.
What I suggest is to increase the gamma, because that alone can fix many of the deepe shadow issues and reveal details that would otherwise remain hidden.
@rashad_carter +10
Well said.
I was watching the vidio from this thread the other day : http://www.daz3d.com/forums/discussion/89951/nvidia-debunk-apollo-hoax#latest ; . It is a bit older but it will help you "see" how things that are not light will impact the lighting in a scene.
The link leads to an empty page.
Well fiddle faddle.. A direct link but the original credit goes to Tengu for posting it https://www.youtube.com/watch?v=O9y_AVYMEUs
That was a very cool article Khory. Thanks for posting the link. NEAT!
So i guess this demonstration begs the question why the heck are we bumping the lumen levels on our lights to huge and exagerrated numbers instead of playing with the exposure values in the tone map settings. If its truly unbiased why are we trying to make it behave like biased based render engines by bumping the "intensity" of a light source. It would seem to me that doing so makes some weird values internally how a specific thing is rendered because we are making the lights SO bright but lowering everything in tone map so it doesnt get blown out.
Daniel
Daniel... because its easier for the average user?
The correct approach isn't always welcomed. People have been pushing for the correct approach since the early days of iRay in Studio. There are some impediments: it takes work and a little bit of study, requires some knowledge/trial and error... and probably the biggest problem: Tone Mapping is computationally more expensive (your render will take longer, sometimes considerably so).
I've moved away from Tone Mapping during the render (I turn it off completely) and do my tone mapping in post. I need to do some experiments on lighting conditions using this method... right now I just take the lights out of the box as it were from various sets and rely on post to get what I want. If/when I get a new gpu, I may go back to tone mapping during the rendering phase, but then I may not... I like the results I get in post much better.
I go back and forth on tone mapping. Sometimes I find it harder to tone map in post than doing it up front.
Which reminds me... does Bloom apply to canvasses?
Yes, Wil, it does. At least I believe so... I don't use it well, so I tend to avoid it. Tried it recently and it didn't work out for me.
I find it incredibly useful for very specific things, like space scenes with a sun (and possibly to do 'atmosphere'... I'll have to play with that)
Because the lumens are not huge and exaggerated; they are physically correct for the size of the emitter. Seems impossible, but it's true. Lumens are a measure of total light output, but if an emitting surface is enlarged, and with no other change, physics dictate that the light covers a larger area. The actual light output remains the same, but is now dispersed into a larger area. The scene looks darker.
There are a lot of things going on in the real world that are harder to render in CG. The main thing is that your eyes and brain don't have to take minutes using floating point math to calculate light paths. Nature does that. All your eyes have to do is receive the photons. Iray has to receive those photons, PLUS duplicate the physics of the known universe. Obviously a tall order, so we do things to help it along.
Tone mapping alone will not create better or faster renders, and speed is the thing. So we assist Iray in duplicating physics by not lighting our scenes in twilight, and expecting that to just magicically work. That's okay for our eyes, which are sensitive enough for this, and which don't have to calculate ray paths. For Iray, we need to give it all the help and hints we can.
In utterly dark surroundings, you can see a candle flame across the Grand Canyon.
The human visual system is extremely filtering and complex, and hard to replicate in images.
And, what's more, our eyes and brain 'adjust' for things within frame, as our eyes dart around focusing on individual elements. The camera can't do that. So what may be properly lit in one part of an image results in another part being either blown out, or too dim.
The basics of photography can be learned in an hour. Shutter speed, film sensitivity (ISO), f/Stop, apeture size, focal length.....just learning the 'around this' settings for using a manual camera, not the equations and details that professional photographers use, is pretty simple. Higher ISO film indoors or at very fast shutter speeds. Lower ISO outside in bright lights with slower shutter speeds. f/Stop settings based on both, apeture adjustment to get the lighting consistent and pleasing.
Also, I believe when the render window is open with Iray running, you CAN adjust the tone-mapping settings while it is going. There is a sidebar you can click to expand in the render preview window that allows you to adjust the tone-mapping as it runs. So if after the first few iteration refreshes the image isn't right, you can adjust and wait a few more iterations to see if it improves it. Once it is looking like what you want, let it just run to get clear and eliminate fireflies and grain. (Or am I thinking of something else? I use so many darn different things for different parts of the render process I can't always keep them all straight.....)
I think it restarts now, at least to some extent. The (possible) reason could be that the tone mapper has some influence in setting nominal luminance for the scene. That helps set up Iray for the render, so it can interpret various brightness levels. The majority of the tone mapper works after the rays have been traced.
I recall that in some pre-release versions of 4.8 the renderer didn't restart, or at least didn't have the appearance of restarting. Hard to know what was going on internally. When 4.8 was released, that behavior had changed.
I'm starting to think too much reliance on photography basics is a hindrance, rather than a help. The film in a camera couldn't care less what f/stop or shutter speed you use. It has X sensitivity, and will be exposed by whatever light is coming into the camera according to that sensitivity. When Iray first came out, there was a lot of misinformation about it, with folks suggesting things like raising the f/stop while decreasing the shutter speed. Um, no. That doesn't help in Iray land, at least not the variation we have in D|S. Reciprocal changes in camera settings simply lead to the same EV, and therefore, same exposure.
None that I know of. MDL might be a language that's too specific for common use.
No Screenies atm. And screenshots aren't IMO that helpful anyways, since the DS interface reduces the shown digits of a value on only two. That's not that much, but can give some different results.
That depends on the Generation of DAZ Characters. The newer Gensis 3's mostly come with pre-prepared Iray materials. So there I just correct wrong Transmission values, and decrease Glossy Roughness a bit to get a more decent shine.
For Genesis 2's I'm using (two-layer-approach resembling Stratum Corneum [Base] and Sebum (oily film on skin) [Top Coat] skin layers):
- "Color Effect Mode"s and "Base Color Effect": "Scatter & Transmit".
- the same Diffuse Texture in the "Translucency Color" texture slot with a 1.00 1.00 1.00 (final translucency is already baked in in most if not every of those old 3Delight textures, so you won't need any addition, but you need translucency for the "Volume" settings to work) and a "Translucency Weight" between 0.45 and 0.50.
- "Glossy Reflectivity" of 0.581507 on Base (that's the value for the common Index Of Refraction of 1.55 for the outermost skin layer we see, the stratum corneum). "Glossy Roughness" of 0.547723 (since that parameter is internally squared you need to use the square root from the commonly used value of 0.30 for that). "Glossy Color": always plain white. "Glossy Layered Weight" at 1.00
- "Top Coat Weight" between 0.35 and 0.45. "Top Coat Layering Mode": "Fresnel", "Top Coat IOR" 1.50 (the sebum layer on skin has commonly the same IOR as plastic). "Top Coat Roughness" 0.547723 (the same as that on the Base).
SSS settings:
Transmitted Measurement Distance: 1.00
Transmitted Color: (0.615847, 0.283032, 0.156186)
Scattering Measurement Distance: 0.10
SSS Amount: 1.00
SSS Direction: -0.65
When using the default 3Delight Specular Maps in the "Glossy Color" texture slot (to make the skin's shine more irregular), you need to unlock "Glossy Layered Weight" and set it to about 3.25(!; no kidding) since they're way too dark for PBR use. Finally, adjust "Glossy Roughness" and "Top Coat Roughness" to your likings/needs, since all those old specular maps come out different some way.
For eyes I'm using the same method DAZ_cjones recommended a while ago. There it's completely irrelevant if one uses the "Glass - Thin" or "Water - Thin" preset, since aside from a different Refraction Index they're absolutely identical and you'll change that value anyways.
Hi all, I want to thank everyone for all of the wonderful help with the Iray settings.
I was also wondering if one of the masters can post some settings for teeth? I have the skin and eyes looking great, but the teeth look horrible compared to the rest :)
Teeth, the final frontier...
The creator of the Yamaki and Marcoor lines... welcome! Well, you're one of the Masters, you should tell us.
Me, I'm no master, just a humble minion trying to decypher the myteries of Physically Based Shading. Teath, like everything else regarding human anatomy's optical properties, are really a pain to implement "near to realism" in Iray. The main component of teeth's outermost shell, the enamel, is the mineral Fluorapatite, a conversion of Hydroxylapatite through the help of flour. A logical approach to make up a teeth shader for a PBR would be to use most of it's optical properties. Over the past year I collected measured data from medical research papers wherever I could find them. That's my early approach so far, using Victoria 7's material settings as base and adjusting some here and there:
Base Color: (1.00, 1.00, 1.00) + default Diffuse Texture Map
Translucency Weight: 0.70
Base Color Effect: Scatter & Transmit
Translucency Color: (0.88, 0.88, 0.98) [RGB 241-241-253] and default Diffuse Texture Map removed
SSS Reflectance Tint: (1.00, 1.00, 1.00) [RGB 255-255-255]
Glossy Layered Weight: 1.00
Glossy Color Effect: Scatter & Transmit
Glossy Reflectivity: 0.718997
Glossy Roughness: 0.20
Refraction Index: 1.631 st... err... borrowed from "Measurement of the refractive index of human teeth by optical coherence tomography"
Zhuo Meng, X. Steve Yao, Hui Yao, Yan Liang, Tiegen Liu, Yanni Li, Guanhua Wang, Shoufeng Lan
Journal of Biomedical Optics, Volume 14, Issue 3
[J. Biomed. Opt. 14(3), 034010 (June 03, 2009). doi:10.1117/1.3130322]
Refraction Weight: 0.14 [originally 0.0070 but at that value very barely noticeable.]
Base Bump: 2.00
Transmitted Measurement Distance: 0.10
Transmitted Color: (0.963580, 0.949424, 0.934915) [RGB 251-249-247] taken from "Translucency of Human Dental Enamel"
R.H.W. BRODBELT, W.J. O'BRIEN, P.L. FAN, J.G. FRAZER-DIB, and R. YU
Surface Science Laboratory, Dental Research Institute, University of Michigan, Ann Arbor, Michigan
[J Dent Res 60(10):1749-1753, October 1981]
Scattering Measurement Distance: 0.10
SSS Amount: 0.99
SSS Direction: -0.35
To get a more natural look we'd need a far better Bump Map, even with bump increased to 2.00 their surface still appear too smooth for my taste. Turning it to 3.00 will only make things to start looking ugly and more unreal. I added a quick (and scary
) test render, broken up at around 8 1/2 minutes. DS 4.9.2.70 seems to have a drastically decreased render speed when Subsurface Scattering features are in use. 
Thank you so very much!
LOL well shucks, haha. (ps. more Marcoor stuff coming soon :D )
I am pretty good with metals and stuff, but people are alot more difficult :)
When I have a bit of time, I will see about creating a high res bump/normal map for teeth. I will also see what I can do about doing some for fingernails as well. I have noticed a major lack of those in the market.
I have been really hoping Mec4D would be releasing some of the amazing stuff shes done :)
I've been away from rendering for about 5 years. So, I'm a bit rusty and trying to learn this new Iray engine. Loving it so far, but for the life of me I can't seem to get the skin to NOT look wet. Any ideas on what settings I should be tweaking to get non-wet skin? This is a test render and the surface options I used were:
Surfaces
- Translucency Color: HSV 9/180/220
- Glossy Reflectivity: .5
- Glossy Roughness: .07
- Top Coat Weight: .5
- Top Coat Roughness: .07
- Top Coat Layering Mode: Fresnel
- Top Coat IOR: 1.333