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
© 2024 Daz Productions Inc. All Rights Reserved.
Comments
I think you can do that at Edit-Object-Transfer Utility
Then select Gen8 and in the scnd column your mesh
select item shape current
then show options
click on morph targets and then from source : shape and pose
Then accept
thanks tirenzi3d, though it doesnt seem to be working when I open the eyelids..
It is the same principle as i think rigging fibermesh. You have to make sure the mesh follows the morphs. you can do that in the parameter section and enter edit mode, select all the morph and click on auto follow. Maybe that will help
However, i think rigging around the eye area is the hardest one can come up with.
This area is pretty anal. Try transferring from the G8F eyelids instead of G8F. That way you make use of whatever Daz had to do to make the default eyelids not distort like crazy.
@Masterstroke - I am trying to think outside of the box a little here....is your icon (black and white) a render? Or is that a real photo...because that looks photo-real. But it is small...no meniscus detail, and no discernable skin detail.
No,that's not a render. It doesn't have anything to do with 3d art. My user name is Masterstroke refering to a Queen song from 1974 called "The Fairy Fella Masterstroke" written by Freddie Mercury. My user photo is a photographi of Freddie Mercury from 1974 taken by Mick Rock.
Hello everyone. I tried @davidtriune 's settings with Edward 8 + AutoFace enhancer and a good HDRI and came with this. There are many things that are not as photorealistic as I would like but I think this is my best effort so far. (tone mapped in PS):
I would really like to see the shaders for this eye.
The wrinkles and hair are neat. However, the skin looks like it has overly burned shadows and the overall color is just off. I suggest you try a different skin.
eye socket is Hann Mei's (by Rarestone) and the eyes are from mousso. not sure but I think it was Jayla's , but all of his eyes look pretty much the same lol
I used a custom cutout map for the eye socket to try to blend it in with the sclera better, see attached.
She's neat but overall quite basic. The hair shaders are really good, but I suggest you try out different opacity maps from other hairs. I suggest using sharper gloss for the lips.
I mainly use canvases for tonemapping, but it's a great way to preview lighting without re-rendering. I recommend Jason Maranto's tutorials:
https://www.youtube.com/watch?v=iaLpdZvbOcc
https://www.youtube.com/watch?v=O14yNn79or4
I have had my share of frustration with HDRI for all of the reasons mentioned. But I was having even more frustration with figures and props as backgrounds. Because they never seemed to have the detail of the human figures. What drew me to HDRI was that they were made from photographs of real places. I had read that some human textures were made from photographs of real people but it seems they don't do that with buildings, outdoor areas, or tables and chairs and whatnot. And the included lighting with the building interiors always seemed to be very dark. I realize all of that could be because I have no idea how they work.
With HDRI building interiors, I did have the issues with shadows not being cast. And I even had one scene in which I was fairly happy with the HDRI of an indoor room, except there was a television as part of the HDRI. And the characters were standing in front of the TV but it wasn't showing their reflection on the screen. I got some product that supposedly fixes such problems but I haven't tried it yet even though that purchase was made on 30 September. I hope it works because if I pay for any product and it doesn't work, it's going to be well beyong the 30-day return window when I figure it out.
So ultimately I settled on outdoor HDRI backgrounds where everything would be too far away to expect a shadow to be cast or a reflection; and depth of field wasn't such an issue either. And I even found a couple of them that didn't seem too "washed out" compared to the female figures in the scene, at least to me. More recently I tried a render inside of a building, with the ceiling and some walls removed so the HDRI could light the scene. I think it worked pretty well. But I'm still sitting here hoping one of you will find a magic key to hyper-realism and tell everybody what you did. Or maybe I will find that key. But I probably wouldn't tell anybody what I did. I would just eternally taunt you all with my ultra-realistic renders for all time.
NylonGirl,
Skin textures are often indeed made from photos. Although this is no necessary anymore with BPR. The underlying materials are much more important, think about derma maps, blood maps etc. Right scale of pores and wrinkles are also very important. If everything is made correctly, then you could even make the diffuse map just a solid pink color and still look convincing. You can even make of course a diffuse skin out of procedural shaders.
Regarding other materials like chairs, tables etc. The same hold true for those things. The underlying materials are most important. They give the feel of materials. You can think about normals, even displacement maps and much more before even thinking about the diffuse color( this doesnt even have to be a map, a procedural color is enough)
The same for clothing materials. I have experimented a lot and not even used a diffuse map, for insance burlap. You can get perfect results with a normal map, displacement map and maybe a bump map and then use a solid color. Wool, wood, plastic, metal, knitted clothes, denim etc. You name it, everything in BPR needs the correct underlying surfaces.
This is the most important thing and why those models look off when other models in the picture look realistic.
For instance, i see a lot of people work on the diffuse map like crazy. The color red shouldnt be much in the diffuse map, rather blood maps or translucency maps. The red color is driven by the blood flow.. Veins etc are under the skin( see, the correct surface) and not on the skin. Another problem with photographs is that pores are on the skin. Pores are also underlying cavities and not on skin. What happens with a scar that is indented is that there is loss of tissue and fibers tear down the skin, and that exlains the indentation. This is not true for pores but officially they should not be on the diffuse map, but on the bump, normal or displacement map. This is why you can paint a diffuse map.
Does included pores on a diffuse map help make a texture look more convincing? Sure, we lack still a lot of things to mimick skin. I try not to use photographs with harsh pores. But in my Charlize Theron model, the photo source has a lot of visible pores. I made sure to correct that with the normal map.
So in short:
It is not important if the diffuse map has a lot of detail. The detail should be driven by the underlying surfaces in BPR. I hope people will get this right one time.
I have uploaded a part of my study materials. So you have an idea how things work. No matter how small the differences on the map ( i.e. veins) they all contribute to the realism.
I think it should be possible to do 2 things in Iray: 1) render a real-looking human, but also 2) render a photo-real looking 'mannequin' - which needn't have any realistic attributes or details at all.
Did you notice that some renders look flat like toons, or paintings...I think these are using painted maps. I thin kwe are finding photo-real ones are using photo-textures.
There's some ingredient, or phenomenon, in photos...what the heck is it!
Photo-real is more like, this could be real but instead a lot of is not accurate or semi- accurate.
Painted diffuse maps are not wrong or less realistic looking. Cartoon textures can also be made out of photos. I love painted maps myself because you have much more control over the outcome. I oftentimes paint alot on my maps to change things( shadow + highlight elimination). Or to add things like freckles and moles.Some discolorations.
Of course there is nothing wrong with photo-real. It can look extremel convincing, and maybe less time consuming. The thing is, which is already said before, you oftentimes have a very realistic subject( be it a human or another organic object) as the centerpoint and background objects don't work together but collide with the center piece.
This could be fixed of course and is also in the hand of the artist.
However, complete realism is hardly possible in 3D packages, a lot is still faked and i think that is good. Photo-real lends itself more for artistic feelings rather than a render that mimics a photograph.
A good combination of both, is probably best, less time consuming and with good results.
What i wanted to explain was, that if you set up correct materials( sometimes, especially skin is really hard) can save you a lot of time.
I like the way your character looks.
What kind of details should I add to Translucency and Transmittance map? Do veins etc go into a translucency map?
I have attached my shader settings, which you can see I have both Translucency and Transmitted Colour maps. Reportedly this shader set-up is supposed to be okay under certain lighting conditions, so is what im using. My Translucency map is just the default Nix8 map, which i think has room for improvement. And transmitted colour map is just the base colour map i made reddish in Photoshop.
I can tell my skin looks bad, just dont know how to fix it.
lil weep, good question!
The veins go indeed on the translucency maps ( or blood maps). You have to experiment a bit. I changed the translucency maps into more of a blood map. I remove the red color in the diffuse and make it more yellow as a base. Up the contrast a bit and start adding red ( red veins for instance) and use the opacity slider in photoshop to make them blend in. The map on the left is my translucency map.
How a blood map or translucency map looks is a bit dependend of the base skin and what you are trying to achieve. Dry skin has obviously more flaking and more red. I myself dont use a transmitted map, i have a color there.
you can use a map there, what i would do for that is, Andor Kollar on youtube has a good tutorial for a transmitted map, he is using knald for this.
One reason for this is that 3d rendered images are too crisp... take a real photo, zoom in at the pixel level and look at the edge of something in the photo, like the arm. You'll notice all kinds of color distortion and blending between the arm and what's behind it. And I'm not talking about anti-aliasing... that of course does smooth out edges by averaging with neighboring pixels... but real photos have a mysterious yet perceivable 'blur' between elements, even very sharp edges. I believe this is one reason eyes are such a dead giveaway as to a rendered image... the lids, etc, just are way too sharp, even when antialiased. Anyhow, I usually run a chromatic abberation filter to help offset this issue... Also, major difference is amount of detail. In the real world, even the most simple scene has incredibly amount of details, little variations, nuances, etc. Hard to capture this in 3D, but can be done. Take a photo a face and blow it up... see all the colors, weird patterns, etc... then take a decent 3D 'photorealistic' image of a face and blow it up... you can see the difference... When you get it to point where the difference is unnoticeable.. then you've hit nirvana ;)
With immensre respect to your accomplishments, Jeff, I'm not sure that's true. I searched faces and zoomed in on them and it's certainly far from universal.
In the first image here, the eyes aren't crisp, but the chin line and the rest is. In the second, very stylized studio portrait, the lines are very sharp, yet it's still identifiable as a human model, not CGI. And the third one shows how much color variation there can be in skin, but the (in focus) lines are all crisp. Maybe I'm not zooming in enough? I picked these fairly randomly to get a good sample.
It seems that the middle photo should be relatively simple to recrate convincingly in 3D, since the skin is obscured and the lighting is fairly simple, but very few renders I've seen would pass as a similar photo. My suspicion is that the telling details aren't tiny things like velus hairs, eye moisture, or little skind details because even in at a small size, it's usually apparent a render (at least a DS render) isn't a photo.
Taking the airbrushed studio photo in the middle, I wonder how realistic that would look if you removed all of the high frequency detail areas, like the hair and eyes. Would the thumbnail still look indisputably real? Only when you zoomed in and saw all the micro details would you probably be able to confirm that it was real if you were just assessing the skin alone.
In the other two photographs, there is such an insane amount of detail in their 'texture' and 'polycount' and 'shader' settings, to say nothing of the morphs that respect natural assymetry etc. I dont think it's even possible for an artist to input that much data without spending an extraordinary time micromanaging every detail of a project.
It would be great if we could get assets with Ian Spriggs's portrait-level of detail in Daz. But who would make them?
Very interesting discussion! My 2 cents: It's not information missing from renders, it's rather exformation;)
Before zooming into the picture, did you have any doubts that image was fake? My guess is not, and that very few DS renders leave you with any doubt—you know they're fake even with small picture sizes. Before you get to renders where you have to zoom in and check details, you have to get to where you pass the first look test.
Look at this picture—it's small, so her face is only ~50-100 pixels. If you zoom in, you can't see any details, but it's clearly a real photo. I've never seen any CGI that would pass for this. There's something more readily apprent than the small details going on.
implying you need to zoom in to see the insane amount of detail in that picture.
I'm not going to go into all the things wrong with my render, but I can tell you that just a straight up add figure, add textures, using my materials, and pose renders a so so looking image below using an HDRI lighting setup only. In order to get realism you have to do lots of work and every detail is important. I hate to say it be we are a bit of a ways off before you can just simply add a realistic figure and render away getting great photorealism in any lighting scenario.
Very very nice. Congrats!
I think you're missing my point—even in a picture where there are no details to zoom in on, we can still tell it's a photo, not CGI. That suggests that when we spot a render as CGI, the thing that's tipping us off isn't some small detail, but something visible even in small pictures without details. If you have a render that people mistake for a photo until they zoom in, then it's probably the details. If they can tell without zooming, then it's likely something else.
To use @Siicliano1969's very good render above, when I have my screen set for typing, I can only see the thumbnail, and not any details. Even there, despite it being a quite good render, I can tell it's not a photo. When I look at the full size, I can see the fabric isn't quite rright and the teeth are too grey, but I that's not what's telling my brain it's a render in thumbnail view. I'm trying to get down to what that telling thing(s) is(are). It seems it must have to do with lighting, but I can't tell what.
We humans do have a lot of that high-frequency detail and extra bits artists put into their models, but not all cameras can capture them and the photos that do have them get photoshopped. I think the way we approach a photo-realistic model should be the same, by adding in as much realistic detail as possible and then removing some in the render or in post. Just a thought.
Btw, has anyone looked through the Texturing XYZ gallery? Some of the work in there look pretty dang close to photo-real.
Edit: Another thing I've noticed that breaks the realism is the gaze of the character. The majority of character work I've seen tend to have a thousand-yard stare, even when looking at the camera. Real eyes aren't parallel when they focus on an object. There will always be one eye that's more angled toward the object in question than the other, depending on where the object is in the fov of the person. That focus is what allows people to create a connection between each other. When a 3d character lacks a focused gaze it's impossible for the viewer to create a connection, making even the most realistic looking characters feel a bit uncanny. So, maybe adding that in might help as well???
Thanks aararibel caado for the kind comments. I totally agree..... there are things like hair coming through the cap, teeth not looking right, and the uniform looking like paper mache (too thin scarf), but there is just something that isn't crossing that line making it look like a photo. I guess my frustration is that I can't put my finger on it. I think Jeff pulls it off great. Even if you look hard enough at some of the best renders you can eventually tell. I really believe our eyes are subconciously trained to know when something looks as we expect it. We might not be aware of it, but since we see ourselves and other people everyday if something is missing or not correct it throws us off. When I set up poses I try to find real people posing to make sure I am getting the correct look. I also try to take a real studio shot and try to duplicate it and see how close I get. Some of the things that will send your mind the "Its CG" right off the bat....
Hard - human beings are organic not plaster statues. Skin is soft not like stone or stucco.
Sharp edges - humans are soft and nothing is sharp....I see lots of bump, normal, and displacement maps that have edges that are way too sharp. The details are too crisp.
Overdone SSS - we have seen them... SSS is subtle not overdone. Unless you have a strong light behind ears or strong sunlight its not as strong as some SSS I have seen.
Too much gloss - unless you are oily or wet keep the gloss to a minimum.
While it is important to have skin tone variations abrupt changes in skin tone or too little will make the figure look incorrect. I have seen some textures that in direct light look like the figure has severe skin issues. Skin tone variations need to blend smoothly.
Asymmetry - human beings are not perfectly symmetrical!
Its frustrating but its something we do to keep moving forward.