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
© 2025 Daz Productions Inc. All Rights Reserved.
Comments
I render in Octane, so haven't bothered with iray yet :D
Just to clarify my post based off this - I do think the standard (ie V7 Ara etc) skins don't look the best by default, since I think they're all aimed at a certain look which isn't one I like. However adjusting it to my own preferences is an easy change and my point was that you can make them look vastly different without ever changing the texture map provided. In terms of PA characters, I've seen quite a few set up the same way that gave the same look, but there are quite a number that come out very differently by default as well.
I think that the people saying they know what is wrong are adjusting it to their own preferences just fine - many of the ones saying gen3 look trash also often say they haven't even bought one or tried adjusting them.
You want skins to look good?
Adopt a linear workflow...without gamma correction correctly set/applied you aren't going to get anywhere near 'good'. And that's without using a PBR. With one, it's essential. The maps (control and diffuse) are most likely set up 'linear' and should be used as such. Excessive 'tanning' is one sure sign of NOT having things set correctly. The same with control maps...if they are over or under corrected (which will happen in a non-linear) environment, then skin can end up being very waxy.
Converting people to a linear workflow. A crusade lost before even starting <head shaking icon>.
Greetings,
I have to admit; I've read almost all the stuff on this that I could get my hands on, and I'm still completely lost.
I have no idea if the diffuse maps have been pre-baked with Gamma (sounds like Dr. Banner is needed), or if the normal maps are gamma-free, or even what settings I'd put into a Surface if I knew. I get that putting the Gamma setting in render settings to 1.0 is supposed to be optimal, but anything below around 1.5 and it starts to look awful... I'm happy to have discovered moving the Gamma down to around 1.5-1.7, as it solves me having to tweak the color balance every single time after rendering, but it doesn't get me anywhere close to what's being claimed around linear workflow.
Now, this may be because the texture maps (skins) are pre-baked at a certain gamme, or because my HDRI is pre-baked, or, or, or...but I have no way of knowing, or controlling that.
...and while Dunning-Kruger is definitely a thing, I don't think I'm particularly ignorant. It's not my job, so maybe I don't have the right impetus to get it exactly right, but if I can't make sense out of it all... Well, it does seem a bit utopian. :(
-- Morgan
No worries I have no idea mostly what they are talking about either. Things tend to veer toward the technical very quickly here. Which isn't a bad thing it just makes it unlikely that a "simple" question ever stays simple. Stick around the forums, read up on as much of this stuff as often as you can and things will (slowly, very slowly) start to make sense. I dont use Iray much I find it very intimidating as a newbie so I am learning 3Delight first, but trying to pick up bits and pieces of Iray as I go. I am starting to somewhat grasp what all the different maps do but the whole baking and gamma stuff just makes me want to go bake peanut butter cookies. Its easier.
I do know that if I keep reading, keep experimenting and don't give up I WILL get it at some point.
Ok, I laughed at the beginning of this... Dr Banner, that was good. And then by the end I was shaking my head. Where did you get that setting Gamma at 1.0 was a good idea? I need to go have a talk with the person who is spreading that horrible information.
Gamma is 2.2.
Period
Very few exceptions.
Repeat after me: I will use gamma correction. The proper gamma setting for my render engine is 2.2. I do not really need to worry about the input like texture maps and HRDI, the vendor and the software worry about those (hopefully). The proper gamma setting for my render engine is 2.2.
Now laugh and smile and go on your merry way, you learned something today.
this was intended to be amusing, so read it with a funny voice or something.
Rule of thumb with linear workflow, if it's a map that affects color, it needs gamma correction. Normal maps are a totally different beast, the color on that is correct. If you mess with the color on it, it will totally mess everything up. Black and white maps shouldn't need any color correction, such as bump or displacement. That is assuming the creator made it in the proper greyscale space. The only reason most people that render to 1.0 do it, is if they want to have more control over everything in nuke or photoshop.If you are just trying to make a picture that looks good as is, and only do minor postwork or paint overs in photoshop, 2.2 is what you probably want to render to.
Gamma correction is garbage. I wish they had proper colour management. Let me decide what profile to use for my render target and actually have profiles in the textures. 2D has had this for a very long time. Why not 3D? Maybe there is support for this that I don't know about?
OP,
I've read the comments so far in this thread and I agree with most everything that has been said. Beneath all the Art is Technology. The artist is not truly free until the technical aspects have been mastered. That said, one of the things we can do more often in the forums (like in the old days a few years ago) is to actually help people out and explain exactly WHAT these terms mean. Yes, indeed those of us with the necessary background find ourselves explaining the same things over and over in different threads, and that can be tiring, but it can be rewarding as well. That said, here are a few things to help speed you along. These are not textbook definitions, I expect you to find those on your own. But these are rough ideals you can begin to think on until you have a chance to shore up your own knowledge.
Unbiased Render vs Biased Rendering
Truthfully, this distinction shouldn't really exist, but the fact is that it does. For decades, physicists have known the proper equations for blackbody curves, and therefore, could and should have written all rendering engines as unbiased from the outset. But due to the calculation overhead, most applications introduce some degree of bias to speed the rendering equation. There are several types of bias, some more detrimental than others. Most Biased engines skip certain steps of light calculations. A few examples of bias to consider.
1. Biased engines break light into two types, direct and indirect. Direct light is traced from the positions of light sources which are often not really there, they are just mathematical representations called "point" lights. When a light ray strikes a surface, the ray is either absorbed and stopped, transmitted via transparency, or refracted via some inner medium which can redirect the rays in any number of ways (SSS). In a biased engine every point on every surface in the scene can "see" all lights and are lit uniformly no matter where they are locatd in the virtual space. To create the impression of shadows and 3D depth of light interaction, biased engines must fire special "shadow" rays to cancel out the otherwise ever-present light from sources within the scene. In this case when the light ray strikes the cube, the cube geometry then fires a shadow ray traveling in the same direction as the original light ray. This shadow ray lands onto a nearby object and removes some of the light that used to be there, creating the impression that the cube has cast a shadow onto the nearby object. Biased engines are not bound by energetic conservation laws. So in biased engines you have tools such as "ambience" which adds uniform lighting at the texture level. Technically this sort of tool violates conservation laws because the texture is allowed to radiate more light than it is receiving from the environment, and worse, the ambient light doesn't affect nearby surfaces as it would if it was glowing in real life. Ambinec eis one of the giveaways of biased rendering approaches and is quite visible once you learn to isolate it. I digress. Direct light comes from dedicated light sources, while indirect light is gathered from the surfaces of the surrounding environment. For indirect lighting, biased engines most often calculate the diffuse inter-reflection between the surfaces in the scene. Such as the blue color of a wall tends to lend some blueness to the white sofa pillow leaning against the wall. But most biased engines do not calculate the indirect Specular (caustic) lighting. Most times the lack of indirect specular contribution is not noticeable when viewing an image in isolation, but becomes more apparent when compared to a real photograph.
Unbiased engines take a thermodynamic approach. The scene starts off fully black without the need for shadow rays at any time. The distinction between direct and indirect is completely removed, because point lights themselves are more or less removed. In an unbiased redner engine shadows rays are not necessary at any time. The behavior of the light is more natural from the onset than in biased engines, because in unbiased engines light almost never comes from a single point, but from a surface with real size and shape. No tools like Ambience from the texture will be found in unbiased engines, instead you have Emission, which works in a real work manner and lends illumination to nearby surfaces. In nature objects do not glow unless they are high in temperature, so Emission requires temperature settings in order to operate. Unbiased engines basically remove all levels of cheating, which requires more processing time and will almost always look better than biased, but in ways that are often too subtle to appreciate which is why we still love our biased engines.
Applications like Iray and Octane can take advantage of the processing cycles of your nvidia cuda enabled graphics card. It turns out that the types of formulations needed to calculate light rays is more efficiently handled by the graphics processor. The normal CPU can perform the same functions, but with much less efficiency. So suddenly, the overly pricey rendering time for unbiased dropped considerably, sometimes even producing images of superior qualty in much less time than biased engines working with CPU. Thus the unbiased craze.
Linear Workflow
Linear workflow is the idea of removing as much bias as possible from a known biased rendering approach. Linear Workflow means you adhere to the conservation laws as closely as possible, and that you apply consistent values to settings so that results are predictable not only in your own studio, but in the studios of the other members of your professional pipeline (such as Hollywood). More on this in a moment.
Gamma is a somewhat generic term which can mean different things sometimes. In this case, gamma is referring to the color space that is being used. In simple terms Gamma decides on your display how white the whitest pixels should be, and how close to true black the darkest pixels should be. It also determines how red the maximum redness value can be in a given image, and so forth. With Gamma 2.2, you are working with the texture with the contrast satuarion and brightness the original author intended it to be seen on your screen. But when building shaders in iray and 3DL, often times it is nice to tweak the gamma on the individual textures. Maybe the texture by default is too saturated...well lower the gamma from 2.2 to some lower value. Somewhat counter-intuitively, the lower the gamma setting numerically the closer to white the texture will appear such that 0 means you get a washed out full white image. When gamma is raised above 2.2, a sort of multiplication effect occurs, deepening saturation and darkening the appearance of the image.
Again, these are not textbook definitions, but they should warm you up and whet your curiosity.
As an added note...Iray has a bug in the 4.8 version where if one placed an image file into the Translucency channel they got a negative image result...red skin appeared greenish instead. I have read that this has been corrected in 4.9. Seems relevant to the current discussion.
Fun fun.
Greetings,
Thanks! I often like to cover up my ignorance with humor, thus the 'Dr. Banner' comment... :)
Okay, so I'm going to make some presumptions.
So, ignoring programmatic shaders for a moment, aren't we ALREADY dealing with a linear workflow? The images are linearized on input, all the calculations are done on the internally correct numbers, and then the result is rendered out adjusted to the stock sRGB color profile. So what would be different about 'adopting a linear workflow', as the earlier poster implied? Or is the complexity entirely in the handwaved-away programmatic shaders?
-- Morgan
Well, this could be partially as PC vs Mac distinction. I think I read some time ago that Mac uses a lower gamma setting than PC, so that images appear the same when viewed on each. But I could be wrong I have used a Mac is some 20 years.
Wow, that was technical. More technical than I get, to be honest. I can speak to point 1, only.
In Studio when rendering with 3Delight, GC is a two stage process (0 stage in Iray, because it is on by default). Turn Gamma Correction On and set the Gamma slider to 2.20; that's your two steps.
Provided that the whoever made the materials you are using did NOT set any options via Image Editor, Studio makes a guess (and its usually correct) about what the gamma of the input should be (I have seen vendor's make mistakes with this and set them wrong, but not often). So you don't NEED to worry about setting much else, unless you start to see some really strange things. Then you can check the settings. Diffuse images set the gamma in Image Editor to 0 or 2.2, effect images (bump, normal, etc) set the gamma to 1. I am starting to see correct settings on most things now, rather than them being all set to 0.
Iray has GC turned on by default and the Gamma is already set to 2.2, but the other points still apply and the gamma on input images should be either set correctly or all set to 0 and let Studio make its best guess as to what it should be.
I'm glad you created this thread, thank you for doing so
Some of the info shared here I find interesting and may help me improve my own renders.
Though I do not use Iray, I still use 4.6 so it's just rendering in 3Delight.
sometimes I find I need to tone down the Specular or increase it, not too often, since once I get a skin the way I like it, I save it as a material preset, dunno if the same would be needed in Iray
I don't use a Mac, but my understanding is that up until 2009, they encoded/decoded with gammas of 0.55/1.8 (as opposed to the more standard encoding/decoding with gammas of 0.45/2.2).
BTW, I'd just like to add one thing . . . Personally, I don't believe you need to understand the math behind this stuff in order to have an opinion about how it looks.
- Greg
Your welcome. And come over to the newbie contest threads. We tackle different things every month. It was Depth of Field last month, Materials this month. I have learned a TON of stuff, and we are a super friendly bunch. Good place to pick up the basics and have fun too.
Quick question. In nature, objects can and do glow without being high in temperature (see fluorescence, chemiluminescence, shrimpoluminescence, etc). How do unbiased engines deal with this sort of thing, since Emission as you describe it doesn't sound like it would do the job?
Wow... I think I actually understood most of that. Thank you!
Ahh but I only know this because we are learing some of it (in easy to understand words) over in the newbie contest on materials. You should really come check it out. I did say most of that, not all of it lol. They lost me in the next post completely.
Greetings,
Don't worry about any of it. Honestly, 90% of the technical side of this stuff is irrelevant. There's important stuff about lighting, and composition, and all kinds of stuff that doesn't require understanding the math behind the renderers, and will have 1,000x the impact on your render results. That's where you should focus, not on these details.
I've just heard folks ranting about how we should all follow a linear workflow (professional digital animation jargon), and as far as I can tell...most DAZ Studio (Iray) users automatically already do. And if you're using 3Delight, if you turn on Gamma Correction and set it to 2.2, then you are also, except in certain edge cases that you REALLY don't need to care about for years and years, if ever.
The original poster is seeing waxy skin, and wants it to be flatter. There's maybe a dozen different things that can cause that effect, but there's no way to know unless you're looking over his/her shoulder.
I even disagree somewhat with how Rashad Carter laid some of that out, and feel that some of it is handwaving over areas that they might not completely understand, so don't feel bad that you don't understand it from their description. That Sonja11 got some useful tidbits out of it is great, and it's okay for you to not get them yet.
We all learn, and it takes time. But I reiterate my handwaved statistic: 90% of what you want to do in DAZ Studio doesn't require this kind of inside-baseball knowledge.
-- Morgan
Funnily enough, I picked up Iray really quickly, but I still have no idea how to do anything in 3DL beyond the most basic 'set up some lights and hit the render button' for those 'just need a reference model for drawing quick' moments.
Hopefully the technical giants won't shoot me down for being completely wrong here, but the simple way I understand the maps is:
- Diffuse (sometimes listed as M in the filename) is the actual colour map which is the main thing you see.
- Translucency as a map is like diffuse but it sits under the skin. A lot of the time it'll have appropriate glows or blood vessels etc that will partly show through the skin (depending on settings.)
- Specular is a black and white map that defines how shiny the texture is (white = all of the shinies, black = no shiny friends.) A lot of the time in Iray the Spec map is used in the Translucency weight slot to control how much shows through. This probably isn't ideal, but it saves making a whole new map.
- Bump makes it look bumpy (white being further out, black being further in towards the middle of the model.) Seems to be preferred for when you have many wider, less impactful bumps.
- Displace is coloured the same way but actually makes the model physically move in or out (displacing!) at those parts. You don't usually see this on Daz models except for a few random cases.
- Normal maps are weird but my understanding is that they bend the model achieving similar but different effects from bump and displace. They need to be perfectly made though so I just stick with what's provided.
I think that's all the ones you're likely to see. I may be missing a few of the more technical details but sometimes it's easier to get the job done by making things simple.
Greetings,
A few small notes...
I think @Sonja11 is referring to this thread, which might be helpful.
Yeah, that's part of what Rashad kinda handwaved over. The term 'temperature' is referring to a theoretical 'black body emitter', which at very specific temperatures emits certain specific wavelengths of light which correspond to colors.
It's a useful theoretic construct, but it doesn't have much to do with actual temperatures.
-- Morgan
Oh this is a perfect explanation!
That's the thread. thanks Morgan no idea why I didn't think to link it.
I know you are getting frustrated with this thread as there is a crap-load of technical jargon being thrown around with little to no explanation or definitions.
I would strongly suggest the book, Digital Lighting and Rendering by Jeremy Birn. It is software agnostic, written in an easy to understand manner, has definitions and explanations on how and why a biased renderer and an un-biased renderer do things the way they do, has a whole chapter that explains linear work flow, gamma correction, lighting concepts and many, many other topics. There are also many working and project files (on-line).
I can't emphasize enough how much help this book has been to me.
Amusing but wrong. It does help to know what is going on with your image maps if you use them.
It's also wrong to assume that you need gamma correction in all circumstances.