Adding to Cart…
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2024 Daz Productions Inc. All Rights Reserved.You currently have no notifications.
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2024 Daz Productions Inc. All Rights Reserved.
Comments
@recovanembden, did you ever make the Minka based character available to others?
This thread is amazing, we need to keep it alive, as so many learned so much from it!
This product: Fancy Motella Character Skin and Render Settings | Daz 3D
A quote from the description: "ancy Motella adds extra realism to your scenes with high-quality features, such as photorealistic skin"
Looking at the screenshots the dual specular is just BLAH to me, I don't see how this can be taken seriously as photorealism skin. Am I missing something?
And if the point is about the skin, why did the PA cover up so much of it in all the shots?
Yeah, the promos make it a little hard to understand what you're even getting in the product, and the quality of the images doesn't exactly sell what's described, either.
Redacted
A lot of people seem to agree that spectral rendering is the way to go when it comes to photorealism. I´m on the same boat and I use it for quite a while now.
With the upcoming changes to the Iray render engine the spectral mode has been changed. It is allready implemented in the beta version of DS and will most likely find it´s way to the general release soon. Maybe it has allready. I´m still on 4.15 so I can´t check for myself right now.
Anyway, with the new way spectral rendering is handled we now have the option to chose from several different color profiles which is a good thing in the first place. Unfortunately none of the existing options seem to match the "old way". At least in my tests all of the skin materials that I created for the characters I frequently use are looking different now.
While that´s nothing that couldn´t be adjusted by changing a few settings, I´m asking myself what´s the way to go now in terms of the color profile.
I did some research on the topic and while I´m by far not skilled enough to fully understand all of it, the main takeaway seems to be that aces - or in this case acescg - seems the way to go if your goal is photorealism. Adrew Price - the Blender Guru - has a great video about it. Of course it is made for Blender but the priciples apply to every render engine. I can´t remember if it has been linked somewhere in the thread before but anyways... here you go: https://www.youtube.com/watch?v=m9AT7H4GGrA&t=1658s
First of all it´s great that we have the option to use this in Iray now. So I did some adjustments to my skin settings and loaded an environment to do a test render just to realise that all of the materials of my environment seem to be messed up as well as my skins have been before. From what I understand this is due to the fact that these materials have been created for the older sRGB color space which has a much smaller color gamut compared to aces. Of course this can be adjusted as well and if you put in the work, you will most likely achieve a more photoreal looking result. But is it really worth the effort to adjust all these materials?
Maybe I´m missing something. As said before, I´m not understanding all of it. Maybe there´s some magical tonemapping setting that could be adjusted to "correct" these premade materials or something like that. Probably someone with more knowledge can shed some light on the topic.
On the other hand there´s the rec709 color space. As I learned from my research this one uses the standard sRGB color gamut and therefore doesn´t seem to mess up premade assets. Of course it lacks the advantages that come with aces but it´s still spectral rendering and therefore comes with a more accurate lighting calculation.
So what do you guys think will be the way to go for photorealism in the future? Is there an easy way to use aces without investing hundrets of hours to ajusting materials? Or may rec709 be an acceptable compromise to easily combine the best of both worlds? Maybe there´s another option that I´m missing here. I´m looking forward to your opinions.
Just my twocents about the topic in general.
There is a great site about ACES. It is very informative but so complex that it does not help to directly get instructions for DS / Iray.
https://chrisbrejon.com/cg-cinematography/chapter-1-5-academy-color-encoding-system-aces/
The author also explains the whole lighting ideas and concepts used in film and cg. If you ever wanted to know what a "Shatner Light" is (yes, Shatner like captain Kirk) or what they call "Rembrandt Lighting" this site is for you.
Regarding REC 709 vs, sRGB.
They use the same RGB primaries - meaning, the color gamut (range of colors) is the same in both. But the gamma curve is different.
So, especially in the shadows you will see differences.
And, REC 709 is used for TV / movie work. So, default implementations use a limited RGB range when interpreting brightness levels.
Good explanation here:
https://www.thepostprocess.com/2019/09/24/how-to-deal-with-levels-full-vs-video/
One can also search the web for: "Full range rgb vs. limited range rgb"
While I am writing anyway, one thing regarding rgb range and your monitor:
If your monitor is attached to the HDMI port it might be that it only shows a brightnes range from 16-235 instead of 0-255 (you mainly loose contrast in shadows by this). This is due to the fact, that HDMI was meant for TV not PC - and TV historically was limited in showing deep blacks.
Nvidia Control Panel (something like that should be in Intel panel as well):
And there is the influence of the DS tone mapping. I am not sure how the relation between tone mapping, rgb range, ACES, REC709 or color spaces and gamma in general is implemented in DS. Maybe some day in the future there is a documentation about this.
I haven't moved to 4.16.. can you show an example of an aces rendering?
With aces applied, the skin is all red.
EDIT:
Blowing up the gamma slider to 4 brings some improvements.
I don't think using spectral ACES-CG will work with DS if you render to direct output (render - new window). It is meant to output linear hdr files to be processed in post. The DS tonemapper does not have the features to do this. ACES works in a wide color gamut which means colors can be way more saturated than in a classic sRGB workflow. You need to map those wide colors to the limited sRGB gamut in post.
Another link about the topic.
https://www.premiumbeat.com/blog/aces-workflow-for-realistic-lighting-3d/
The canvas beauty.exr shows the same red results as the Iray png. It is all red. In fact, correcting the gamma in DS is more promising, than in the post in Photoshop.
Did you apply a sRGB RRT as per the tutorial? I would think that would work?
Rendering in Iray spectral rendering (natural 1964)
rec709 left and aces right
Import exr files => Exposure -12.5 = > convert to 8 bit in Photoshop.
Not an expert, but as I understand it, changing exposure just changes the brightness (exposure) levels of your render. What you need to do is remap the colors to the profile of your system (likely sRGB)
In other words, it has to do a translation from the colour space of ACES to sRGB to make it look as intended. That's at least how I understand it.
How do I do that?
My Nvidea card display setting doesn't give me a sRGB option.
You probably use a different colour space.
This article is very readable and might get you further : https://www.toadstorm.com/blog/?p=694
Not by choice. It seems, I don't have an sRGB option.
Thank you for your answers so far!
This actually seems to be a much deeper rabbit hole than I expected.
I was hoping to find an easy answer on how to continue the spectral rendering workflow I am used to without studying color theory for the next 6 months.
I´ll continue to monitor the topic and try to learn more about it but untill then it might be better to just stay on DS 4.15 if one wants to use spectral rendering without getting into too much trouble.
Yeah, it also seems that your diffuse maps need to be in the same colourspace to get the full effect of the new colourspace. Basicly it means you are using colours that your monitor can't handle, so you need to make a compressed version of it when making an output (as you can see in the examples in the articles) That is why @Masterstroke has the reds in his render for example.
Shadows, light and colours become much more accurate it seems in ACES, it has a higher fidelity. But yeah, setting up will be a bit more work.
Does that mean, every texture map had to be converted?
Every map that contains color information would need to be converted. In most cases that would be diffuse and translucency maps.
Other things like normal or bumb maps wouldn´t.
But for a big scene with an environment, several characters, clothing, hair and so on this would still be a huge amount of work.
I´m wondering what color profile the previous versions of Iray used in conjunction with spectral rendering since that worked perfectly fine out of the box.
@M-C
your are right about it being a rabbit hole. My conclusion, after doing the research, is to avoid especially ACES. It is way more complex than just switching a button - and it has its own bunch of build in limitations and issues. And none of the textures and shader settings we use in DS are made with a ACES color pipeline in mind.
Another good read reagarding this:
https://blenderartists.org/t/installed-aces-color-management-to-blender-now-im-confused/1236883/39
or the conclusion in the link I posted earlier
https://chrisbrejon.com/cg-cinematography/chapter-1-5-academy-color-encoding-system-aces/
But you can install the DS 4.16 Public Beta and keep your 4.15 release. Then try the REC 709 preset maybe?
That´s exactly what I´m doing allready.
Maybe someone comes up with a conversion script or something like that.
I´m by far no Blender fanboy but as many times before, this is one of these moments where I´m asking myself why this seems to be indeed a one click solution in Blender for more than 4 years now but still seems impossible for DS.
HDRI mesh lights might help. (Yes, a mesh light with an HDRI texture map as a light source)
Your preview will look a little to bright, but render your scene as an *.exr file and set the exposure in Photoshop.
...and what about the rec2020?
In this case I used a HDRI environment map set to 8.
Exposure in Photoshop: -14
Left to Right
Spectral OFF, rec2020, acescg and aces
Could you try applying a LUT Masterstroke on these renders?
https://www.cinematography.net/ACES-LUT's-for-use-in-a-non-ACES-environment.html
Hm, converted the textures to ACESGG half exr.
Left ACESCG spectral - Right default no spectral. Output via DS default tone mapper.
Still more to learn / understand. If we would just know what DS is doing in its tone mapper (which curves, which rgb primaries, which luts etc.) and whether it does any color or value mapping if exr textures are used.
EDIT:
Ignore the render. I made a mistake in the texture conversion. New one will follow.
Link to conversion tool in case you don't want to do it manually with opencolorIO in your app:
https://mrlixm.github.io/PYCO/standalone/ColorspaceConvert/home/
EDIT 2:
I found out what's causing the "bright eyes". It is an issue with "guided sampling". Seems so, GS is preferring refractive surfaces and sends more samples to them. In my scene the cornea and eye moisture are the only refractive ones and therefore get a lot more light samples than the rest of the scene. It works correctly with GS off or Caustics on. If spectral is off then GS also works correctly.
EDIT 3:
This now ACESCG spectral with converted textures (left), sRGB no spectral (right). Still something in the eyes but I ignore it for now.
Currently reading / watching this, thought others might like it as well : https://filmschoolrejects.com/cgi-skin/
Just curious if anyone has thoughts on why Daz characters are not able to reach this level realism?
Well they can, but with LOTS of work. This scan model has just 1 color texture map and is just using standard CG rendering, nothing fancy.
Is it anatomy, textures, lighting, shader, a combination or something else?
Haha, sorry but this is funny, especially being on p 61