The face still looks quite realistic to me in the second picture. The floor texture you are using is not so good and that wont help with the illusion of realism. Try making the background looking more realistic. That camoflage texture on the clothes looks wrong as well, a bit too unnatural looking.

You are also missing any shadows-- this gives the appearance that the clothes are 'pasted' on the body. Shadows at the hem lines, wrinkles/folds in the clothing would help. Even within the figure shadows falling on the face from the hair would help. Not sure how your lighting is set up, but from the image it appears very flat with no direction, kind of washing the image out. This might help in creating a more realistic rendering.

We are used to seeing photos when we see images that are "real" You could try desaturating and adding noise/grain. I great way to do this is with the free Virtual Photography Photoshop Filter... were to get it I can't remeber. Also dark sin is wrong at sholders and knees could be darker. Sharpening the image often helps offset the softness of Poser Renders at mid-distance. Finally the mess in the hair is lost as you pull away... messing it up again in post-work might help, especially were the highlights in the hair are showing a perfect band. Armpits should be quite light and with the suggestion of stubble. The real problem as you pull out is the loss of bump noise giving super smooth skin. I sometimes brush special gray scale skin texture over in post-work.

I think the lack of shadows, as djthomas has mentioned, are one the primary ingredients missing from this image (even the first one). Shadows, however subtle, add that extra "3rd" perception of depth when viewing a 3D image. Without them, the image just seems flat, or more "cartoony".

Another thing might be the textures/materials used. The ground mat doesn't resemble anything realistic that we humans know of in the real world. Perhaps marble, but not quite. Try going for more realistic texturing on the clothes and ground. Use more bump and specular properties in combination with each other. Aside from adding detail, bump maps help break up specularity, which provides more realism by nullifying the smooth, "plastic" look of CG. Before bump maps, everything in CG looked like plastic or metal. ;-)

As additional touches, you may also consider adding some DOF (depth of field) to give it a more 'photographic' appearance, and even try to fake some SSS on the skin (if you have P5 or P6), so that the skin seems to have some life. Without sub-surface scattering, CG skin tends to look lifeless. SSS gives the illusion of light being absorbed and scattered under the surface of the skin, much like the properties of real human skin, which our eyes are so used to seeing in life.

Finally, a much over-looked component to realism is expression. Her face needs some kind of expression on the mouth and eyes to tell us she's "real", and not just a department store prop. ;-) Work on some interesting facial expressions for her, and I think it will make a world of difference.

Hope this helps a little. Good luck.

Message edited on: 08/24/2005 12:00

tbird10, how did you do the flat lighting? I like the effect. maxxxmodelz, I am shy of node stuff like SSS... is there a way of faking this by compositing two essentially conventional renders? I think years ago Staale tried some strange aproach using a double skin arrangement on the figure (a hack on the specular?) but I would like to use conventional maps and DAZ figures... would a compositing of two different lighting set ups of some sort work?

A single image based light without AO or shadows with a light grey square as the 'image' (Poser 6) :-)

A little off topic but, the eyebrows do not match the hair and the eyelashes. I agree with the others about textures,shadows and background.

"I think years ago Staale tried some strange aproach using a double skin arrangement on the figure (a hack on the specular?) but I would like to use conventional maps and DAZ figures... would a compositing of two different lighting set ups of some sort work?" Hmmm. That 'double skin' approach sounds interesting, momodot. As far as compositing two different lighting setups... that sounds plausible in theory, but I suppose it would depend heavily on the postwork/compositing skill of the user. I think it certainly could be a solution in the most narrow circumstance, but the setup you'd likely be required to arrange with the lighting and rendering of both scenes would probably be more work than just doing it in the material room with nodes. You'd also have to take into account that SSS needs to be represented in a particular way on the skin, or you'll just end up with strange skin coloration that could subtract from the realism even more. It would be a tough job, but anything is possible. ;-)

"The face still looks quite realistic to me in the second picture. The floor texture you are using is not so good and that wont help with the illusion of realism. Try making the background looking more realistic.

That camoflage texture on the clothes looks wrong as well, a bit too unnatural looking."

Gee..thanks. Sorry, fashion is not my thing. (Going for a military look. And it did fail)

I know the floor is not helping, and lighting is an issue as well. I realized how bad the floor was after the render. Took a lot to get the render, and I didn't feel like doing it over again, just to get help. Not focusing on the clothes. It is the facial area, skin texture, that is lost in the realism. Looks cartoony.

Message edited on: 08/24/2005 14:22

The camoflage is not really so bad, as others pointed out the lack of shadows on the clothes make them look a bit 2d.

If I took a picture using a camera (facial), like for example outside and not a cloud in the sky at noon. Their is one photo made. I zoom out (full figure), the shadows, lighting doesn't change. Why in poser?

Showing image example;

On the left is a real person and not made in poser. No background used. No one for sure will agree it was made in poser. Realistic and it is. Image on the right is a poser character. Everyone will right away agree it is a Poser character(V3). If we can do close ups so realisticly, why can't we get the same for a full body figure? And using the same lighting? I been playing with this problem all night, and I am now fustrated. Lighting, shadows, clothing, etc. I had to ask this, and get advice. I see it in a lot of renders, Facial...so realistic, but show full body, it all changes. Get my point?

(I purchased these lights here, and not going to say from whom. They are good, as you can see from the top image.)

Could it be my render settings?

Message edited on: 08/24/2005 15:14

The picture on the left has darker and stronger shadows, your picture has weaker and lighter shadows. It might help to make the shadows stronger. It could be the render settings.

Robo, Well if you're going to be too stubborn to take peoples advice about lighting and shadows. Why are you even asking for help? Poser is NOT the real world. It is Virtual and like a theatre set, Lighting HAS to be setup properly for realism. This includes turning on Shadows when and where they're needed. You've been given some excellent advice in this thread by several people. Why don't you try it using some of it and get back to us with your progress.

"Robo, Well if you're going to be too stubborn to take peoples advice about lighting and shadows. Why are you even asking for help? Poser is NOT the real world. It is Virtual and like a theatre set, Lighting HAS to be setup properly for realism. This includes turning on Shadows when and where they're needed. You've been given some excellent advice in this thread by several people. Why don't you try it using some of it and get back to us with your progress." Yeah..your right. Lock this thread. Why go for realism.

"If you look at the very top image. The render, character looks very realistic. But when zoomed out, the realism quality is lost. Why do I have to change my lighting, shadows?"

Robo, when you look at that top image (the close-up head shot), do you really think it would fool someone into thinking it's a photo? Do you see where it falls short?

The same elements that are keeping the top image from looking like a photo is keeping the bottom one from doing the same.

The top image looks "realistic", but that's only due to the details in the textures, since nothing else about the scene is lending to it's believability. When you zoom out to the full body shot, now you're losing the up-close details of the textures, and since you don't have any other elements to maintain the believability, it looks totally unconvincing.

See what I mean? :-)

Realism is VERY hard to master. Even the real pros sometimes get it wrong. Human beings are the hardest of all, because our eyes see them every single day, and we know how they should look immediately. Plus, unlike some other things that are easy to fake in 3D, human skin is still one of the hardest challenges to duplicate. The way real skin reacts with light, the scattering that happens, etc., all these are impossible to get spot-on in an artificial environment. It would take dozens of quantum computers to calculate the way light reacts with skin, etc. So we do our best to "fake" the effect. Even physically calculated SSS in advanced renderers isn't nearly as accurate as the real thing. ;-) Message edited on: 08/24/2005 15:27

Yeah..I am tired. I struggled. Had enough. I thought I was making a point. But it failed.

I see the advice, and will take it.

I am very, sorry for making a pointless thread.
Gone to bed.

Message edited on: 08/24/2005 15:32

"I am very, sorry for making a pointless thread." The thread isn't pointless at all! In fact, I think it's one of the better ones I've seen started in quite some time. Thinking about realism, and how to improve it is one of the finer points to 3D art. Duplication of reality is the "holy grail" for most of us. It's what 3D was intended to do from the beginning. :-) Don't be discouraged, Robo! Keep on it. We all get tired and frustrated. Once I actually threw a monitor from my office window. No lie. Luckily, no one was injured in that incident (except the monitor... oh, and my wallet when I had to buy a new one). I learned some anger management after that. ;-P

I certainly never meant to imply the question was pointless. Only that asking for help and then refusing it once it was offered seemed pointless. Get some sleep. Come back to it tomorrow. It'll all make sense again once the sandman has had a chance to run rampant in your dreams.

Thanks.. :-)

For the top image, you and I will know it is a Poser Character V3. But anyone who doesn't know poser or doesn't use any 3D rendering programs will say it looks very realistic. That image is far from Poser4 Female when it comes to realism. We are getting better at it. Lighting, render settings, a community like here at renderosity, and a whole lot more. I was trying to keep my realistic look when zoomed out, but it always get lost. Will do again in a few days.

Right now, my poser(6) is rendering (Making movie "Animation"), 600 frames. Should be done tomorrow. I am studying if full figure character looks realistic enough when in motion, than in a still image. Problem is I have to wait, and the quality can be terrible due to which setting I have it in (Divx, Microsoft...etc). I forget which one I choosed.

Message edited on: 08/24/2005 16:00

Attached Link: http://www.users.on.net/~pkinnane/RealSkinShaderTutorial.html

One of the things I found out is that Poser materials by default have zero ambience. So when a surface is not directly struck by light, it'll be pitch black. In the "real" world, there's a lot of indirect lighting. On a sunny day, the obvious light source is the sun. But objects in shadow are still clearly visible, they're lit by light that's reflected by surrounding objects. A quick and dirty way to simulate indirect lighting is plugging the texture map into the ambience color input in addition to the diffuse color input. Set ambience color to something between the color of your main light and white, set ambience value to about 0.4 and set diffuse value to about 0.6. Voila, your shadows come alive. P.S: the Gather node actually gathers lighting information from objects that are near. Unfortunately you'll have to crank up the amount of data collected to insane values to have a smooth effect, and render times will approach infinity.... And Ambient Occlusion is also a key to realism. Shadow maps just don't get the detailed shadows between fingers, or that slight shadow effect from the edges of clothing on skin. AO does. If you want to see what can be done with Poser 6 regarding realism, search for Mec4D in the forums. Pretty impressive results straight out of Poser.

Mizrael, I think we all got thinking along practical lines but I think maybe the question was intended more along conceptual lines. Tthe technical ability of Robo2010 aside, Robo2010 had an interesting question... what distinguishes CGI from photographic from Real World visual. I would love to see a consideration of that issue. Meanwhile, issues of texture aside, on a practical level, realism would be tremendously improved with an Ambient Occlusion and/or environmental light set up lighting and wacros plus and most important: the tools of face_off which make nodes actualy useful to us ignoramuses, and in my opinion single handed revolutionized the Poser render and accounts for at least half the realism in any given render. Yet there are those texture, pose, symmetry/asymmetry issues. Would be neat if there was a P6 update of Snow Sultan or Geeps scene construction/render 10 tips or intro. I however am most certainly not qualified to contribute this. We do have to deal with the original question, why does a mid shot seem so much less real then a close shot all other things being equal. The comparison to tthe pin-up is most informative... again issues of contrast color balance and noise I thing to some extent. In photography these aspects are independent of distance to object but in renders not so. Finally, I understand having difficulty Posing a question clearly and wanting to understand principals beyond my actual skill level to address. No offense please Robo2010, I say this for the sake of argument here rather than as a criticism of your actual example... I had imagined from the start your question was not taking background, clothes etc. in consideration. Mizrael too, please take no offence as none is intended... I'm just trying to advocate generosity concerning people's posts and the difficulty in communicating in this medium/format :)

No offense taken and I could have been a lot less caustic in my first response. I just found it a bit frustrating to see all those wonderfully great suggestions on how to acheieve what "it appeared to be" that Robo wanted in his image only met with a posting that sounded like "Well I don't wanna change my lights and you can't make me" by the way it was written. I understand now that that was due to a bit of a language barrier as well as a bit of topic drift. That's why I clarified my point. The question is great and discussion is great. I have nothing against Robo and appreciate his/her excellent question.

If I remember correctly, Staale's trick with the double skin was intended to circumvent P4's lack of specular maps. In Poser 5/6, you can plug a texture or other node into specular, so that workaround isn't needed anymore.

Thanks, Mizerael. svdl, You thing some sort of double skin with transparency would do anything? I should download Staale or Stale's figure and try it...I never did understand how the figure worked and I had no memory on my machine back then.

I think that was his trick. But I'm not sure, it's a long, long time ago. The drawback, of course, is the amount of polygons; doubled. Though I think it could be used to fake subsurface scattering. Those muscle maps from DAZ on the inner figure, with some ambience and displacement, a normal skin texture on the outer figure, slightly transparent, could work.

In some 3D apps, you do have to deal with distance to objects, but not really in Poser. Only for things like physical SSS, sometimes modeling, and using Photometric lighting does it become extremely important to use "real world" scaling. :-)

The image above is just a quick example of how a little SSS, Occlusion, and IBL (to simulate indirect lighting) can really make a figure look so much more lifelike, even at a distance.

The figure on the left uses Face_Off's V3 realism kit, with AO and IBL. The figure on the right is lit with the same direct lighting, but minus the IBL, shader system, and light shadows. Both are rendered in P6. Obviously, the figure on the left looks more alive I think. Message edited on: 08/24/2005 18:29

Skin isn`t usually as shiny as that, if you look at your face in a mirror you will see that there is not much shininess, it is the same with bodies. A specular map would look better and be more realistic, where some parts are shiny and some not.

When skin is moist (from either lotion or perspiration), it can certainly get shiny like that, given a particular lighting situation. I think the problem with it is that the bump map should be tweaked a little higher to break up the specular more. Obviously, tweaking settings to get the proper look for a particular scene is the best way to go about it, but that was simply a very quick 5 min comparison render. ;-)

Message edited on: 08/24/2005 20:20

maxxxmodelz, damn that is a good side by side comparison. Thank you. And again... face-off makes it all work. To me the breasts and thighs are most obviously better but the difference in the faces is less dramatic although the face on the left is definately less Poser like... has the look of a MAX render maybe. It is MAX that people round here use as a strong render engine, right?

"It is MAX that people round here use as a strong render engine, right?"

Well, 3dsmax has several really excellent render engines that can be used with it, but some of them are extra/seperate purchases. I think most people on this site are probably using Bryce or Vue as a render alternative. I use 3dsmax primarily for modeling and animation. You can't really compare Poser's animation tools to what they have in 3ds. ;-P

For most stills, however, I find P6 does the job as good as any. In fact, better than most when it comes to skin shading. Face_Off's skin shader system, I believe, is based on the famous Stahlberg skin shader technique for Maya, which is one of the best (and fastest) skin solutions out there. I'm sure Paul will correct me if I'm wrong about that. ;-P Message edited on: 08/24/2005 20:44

Max - the whole idea came from Stahlberg - but only about 30% of the shader network is now based on his system. His research and techniques were ground-breaking stuff. His techniques for human modelling and joint deformation are outstanding too - and feature in a new book - d'artiste: Character Modeling [pls excuse the plug for a book put together in my hometown Adelaide/Australia).

Whoa!..come back and all these posts. :-)

maxxxmodelz;

Post #32

I sure like to know what your render setting are at? To get the character on left. That is what I am fighting for and light settings.

Message edited on: 08/24/2005 21:04

"I sure like to know what your render setting are at? To get the character on left. That is what I am fighting for and light settings." Ideally, Robo, I'd love to send you the actual .pz3 file to examine if I could, but I'm using some commercial products in there that you may not have yet. So instead, I'll just have to take some screencaps of my rendersettings, lighting situation, and AO/IBL settings. From there, you may be able to reproduce something similar. Give me a few minutes and I'll post them for you.

Ok, here's the Poser settings used in the example image above. Keep in mind, however, a lot of the difference you see there is dependant on Face_Off's skin shader system, which I can't really show you, since it wouldn't make much sense without actually examining the shader node tree first-hand.

Anyway, here's what I can show you...

Render Settings
These are pretty common (I think)for final render settings, and is the basis of what I always use for most of my renders.

Light Control Rig
Pretty basic. Only 3 lights were in the scene, and only 1 with actual shadows. Keep in mind, this rig isn't universal. You can experiment with your own setup that works with your scenes.

IBL Settings
This too is not universal. Your IBL settings can vary, depending on the scene. Major thing to keep in mind is that if your IBL light seems to "wash out" your scene, then try lowering the diffuse color value of the light in the material room, or lower the intensity. As far as AO... well, this particular render has the AO turned ON for the light itself (as you can see). However, I suggest using an AO node in the material room only on the objects/materials that you need it to be. In that test render, it was just easier to turn it on for the light. Face_Off also has a product called, Occlusion Master, which is great for adding the AO node to certain materials in the scene (like skin). AO is hard to control when it's applied via a light, and much easier to control via a material node, but in the end it's up to you which way to go.

Really hope this helps you some, Robo. If there's anything else, let me know. ;-)

Message edited on: 08/24/2005 22:27

Will work around on it.

BastBlack..damm that is nice.

Happen to find this render I did many years back. Shelly for the "P4 Woman"

Not so great on morphing. PPro rendered.

Message edited on: 08/24/2005 22:55

BastBlack, can you elaborate on the render in layers supressing the P4 look? How does that work? Not so much how to do it as what it does for an image... I thought people did this only inorder to avoid large scenes crashing or hanging up in the render. That shelly render sure is nice, Robo2010.

BastBlack...please tell me about your 'layers' technique!

Inbkmk

I don't use P4 any more, but I still do the layers thing for some images. I dunno how BastBlack does it, but I do it by setting up the scene and then using the Hierarchy window to make certain parts invisible, rendering so that I have isolated renders of the figures, with and without hair, with and without shadows, sometimes with and without clothes, the background by itself, etc. I render and save as .tif so they've got an alpha channel, so that I can take them all into Photoshop and pull the figures and whatever out with the alpha channels and composite it all together until it suits me. This way I can control how much shadowing is applied, or use blend modes on the layers or paint out or in bits, and so on. I'm also interested in hearing how BastBlack does it, though. I'm always trying to get rid of the "Poserish" look, and any technique to do that is something I want to know about. ;-) bonni

Yep, I do all those things and more. The image above is, if i remember correctly, 15 layers. My machine can't handle big scenes eitherm so i stick to one person scenes and close ups. lol! Rendering with bump maps and shadows will also kill my machine, so I developed this trick to get things done. Now I feel guilty if I do only one render. ;) Good contrast land good composition are most important. Don't do all these painful steps below if your image isn't good to start with, it will not make it "better." These steps are just polish. These are my own lights, 3 in all. One is the blue "moon" backlight. the lower left light is the "campfire" light from below. Because the orange and blue are near opposites, they give a "pop" to your subject. My fake SSS, is the "moon" light. To get it properly on the face and feet, I render nude no hair, with bumps and shadows on. Always save in the highest resolution (final image size) with alpha channel. First render: nude, no hair, with bumps and cast shadows. 2nd render: nude, no hair, with bumps and cast shadows. New higher res eyes (not DAZ eyes) 3rd render: clothes with bumps and shadows 4th render: clothes with bends turned off, or spiral turned off, or fix JPs in the joint editor so you can fix bad bends in post. 5th render: necklace doesn't cast a good shadow, render just the necklace with alpha channel. Renders 5-15: Multiple hairstyles renders, some reposed a few times, each with shadows on, then with shadows off. And if needed, turn body parts on and off if hair intersects body in any places. Hair textures must match, and material setting for each hair must also match. Turning off "cast shadows" on hair is often a good idea, Render 16: Byrce sky with my poser colors roughly in the same position as I had them in poser. I rendered larger with some leeway around the edges so I could correctly align the image in post. Renders adjusted fill lighting: sometimes I change the lighting many many ways and render in separate layers. It could go like this: Backlight, fill 1, fill 2, spot light, fake SS light, bounce lights, etc. In post I get control how the lights will layer on the image instead of trusting Poser to do it for me Make sense? Plus with photoshop, I can set some lights to "lighten" "overlay" "soften" "darken" "color dodge" etc. Generally, I always do 8-12% color dodge on the final image because it gets rid of the "scanned print look." Some times I add noise if there is much "computer gradiant look." Okay, now I'm armed with separated elements to recomposite in photoshop. I build each layer with separated elements. Then I just adjust what get included where and how it gets included. For the hair, some of the layers are "lighten". You can still see the "Wild and Messy" hairstyle, but because there are other styles, it makes the image less boring if you already know what WAM hair looks like. I will often change the alpha channel, and sometimes run filters on a layer's image. What is done in each layer changes by the project, so you need experiment to see what works best your image. I've been using photoshop for over a decade, have worked in photography both taking pictures and developing them, and have more than enough of photoretouching and color balancing for anal model agencies, so some of my pain tolerance if from my work history. lol! ;) bB

Facinating, BastBlack. Thank you.

Other things to consider: Color theme. Use colors that work together. The Clip Art Rule also applies to poser. And try to "own" elements in your image by making with new textures and morphs. (see color theme) Why? Because if elements are easily spottable, it makes the image boring for people who are familiar with each element already. Combining elements in different ways will help people see past the elements used in order to see the image created. Make sense? bB

Also Compsiting in layers in photoshop is doable for animation too, and easiler than rotoscoping in computer too! Render stills with an alpha. Import stack to photshop and run a macro (actions based on the process you did to get the first image). Rebuild the stacks as an animation. I didn't do any hand painting in my image above, so I could animate it. I have the original PZ3 and my photoshop master layers too, so it can be done. (It's a good idea to save often, with different versions, when working in layers.) bB

Ok...realized something while playing with lights. How do you create specular lights? Also for IBL AO, when in material room. Their is and a angle end dial. How far can this angle end go? Anyone ever play with these dials Andle start, angle end? I have my angle end to 450 degrees, thinking all around: 4(3.14)r (Squared). I dunno..been awhile.