What are some ways to make photo realistic people?  What I mean by that is a picture that can be fooled by others into thinking that it is a picture of a real person.  It seems that still lifes are easier to do that renders of people.
 I have been told that something about the eyes make it looks like it is a computer done  person verse a picture of a person.  How can I get better looking eyes if that is the case.  I mean the whole eye not just the textures of the eyeball, but the area around the eye including the eyelids.
Thanks for any help

If your going for a closeup shot. 
Adding a bead of moisture to the lower lid of the eye helps.
And very good textures, lighting and reflections are required.
Good luck in your quest.

Its ALL about the texture and the lighting I think. I've been searching for that holy grail of rendering for ages and no matter how good the texture is, with the wrong lighting it will look fake and vice versa.  The best I could ever manage was with this image (above) but note I had to make the image sepia and add a little blur to achieve it (postwork) . Cararra's HDRI does a pretty good job of skin as well (see some of the images in the gallery at the minute by others).

I say 50% texture and 50% lighting.  See http://www.shonner.com/drafts/realistic_carrara_renders.htm for some examples of Poser imports to Carrara.  The skins are from real photos, which is the best way to go for realism.

All of the above plus a convincing expression. IMO, that is the most difficult to achieve.

It should be noted as well that if you use a photo texture, you need one that has been worked in an image editor to get rid of the "baked in" effects, like specularity and even SSS.  When you take a photo of someone, you are capturing the skin with all of it's natural, real-life components  from whatever lighting situation that person is in at that time.  So the photo of the skin will already have subsurface scattering effects and specularity, which are elements you would want to reserve specifically as properties of your skin shader/texture at render time.

Photo textures need to be carefully and skillfully touched up to reduce those "already present" natural effects, and to separate the color/diffuse texture from detailed and well-constructed specular, bump, and subsurface scattering maps that could be added to your skin shader to help enhance realism.  Regardless of the application you work with, these basic principals allow you to utilize whatever lighting you might require or desire in a scene.

You should never be limited by your skin textures... having to match the lighting in your scene to make your textures look good means you're not using a well prepared skin shader/texture.  Your skin material should be reusable under almost any environment lighting condition in your 3D scenes with only minimal tweaking required.

another way to do it, is avoid using colored textures at all, just greyscale maps
almost all shading work i ever do is with greyscale maps, usually being used as a multiplier to whatever color i choose, which has the added benefit of allowing limitless skin colors, rather than being bound by only a very small range near the color of the texture

another added benefit is it saves memory used by texture maps, since most skin keeps the same ratio in colors, (ie 200R/100G/100B vs 250R/125G/125B, why not drop the last 2 channels G and B, since we know they'll be 1/2 of R?) so leaves 2/3 of the texture file to be replaced by bump, spec, or something else, then have 1 of the values used to control each thing in shader room, one of many ways to use texture maps to store more useful info :)

Quote - another way to do it, is avoid using colored textures at all, just greyscale maps
almost all shading work i ever do is with greyscale maps, usually being used as a multiplier to whatever color i choose, which has the added benefit of allowing limitless skin colors, rather than being bound by only a very small range near the color of the texture

another added benefit is it saves memory used by texture maps, since most skin keeps the same ratio in colors, (ie 200R/100G/100B vs 250R/125G/125B, why not drop the last 2 channels G and B, since we know they'll be 1/2 of R?) so leaves 2/3 of the texture file to be replaced by bump, spec, or something else, then have 1 of the values used to control each thing in shader room, one of many ways to use texture maps to store more useful info :)

That's actually the correct way to go about it, because it allows the shader, which can be light-independent, to drive the skin color in any giving lighting situation.

I haven't used Carrara since early on in my 3D experience, but the same theories apply to almost any 3D application.

It's good practice, for example, if you are going to use subsurface scattering in your skin shader, that you at least reduce, or sometimes completely get rid of,  any red color channels in your diffuse texture map before hand, since your shader will be used to drive the subsurface scattering, based on the lights in your scene, and provide the adequate skin "reddening" in the final output.  I used to consistantly make the mistake of using full color diffuse textures with SSS in some other apps, and the resulting skin always needed extensive color correction for being far too oversaturated.

It also helps if you are able to use greyscale "control" maps to help drive the scattering depth, etc.  Proper texture maps play a vital role in realistic rendering, especially for something as infinitely complex as human skin.

one downside to carrara, so many plugins are "an end to the branch"
you can't control them with other inputs, unlike poser which allows every variable to be changed by inputs from something else. i find myself creating tons of new plugins replacing existing shaders, but i can only create so many, some are just too advanced for me :(

And I differ from anxcon.  Poser is dreadfull and Carrara miles ahead--and some of the newer plugins available for Carrara open up "Trees" that just keep on expanding!

Quote - one downside to carrara, so many plugins are "an end to the branch"
you can't control them with other inputs, unlike poser which allows every variable to be changed by inputs from something else. i find myself creating tons of new plugins replacing existing shaders, but i can only create so many, some are just too advanced for me :(

 

I 100% do not agree.  But then you are using the word "plugin" incorrectly, so it's even confusing what you're saying.

ok mr shonner :) i want lumber shader to have the "perturbation" value change from the output of a color gradient controlled by Z coords, see? can't be done, lumber mill shader is at the end of the branch, same for tile, brick, etc. in poser i could use brick controlled by 100 math nodes (which thankfully carrara has formulas yay) to control design of the bricks, but carrara has brick at end of the branch, as another example.

i recreated many of them to allow me this control back, but i have limits :(

In Carrara, you start with a brick or a piece of wood and you mash it, grind it, warp it, whatever, into something bigger, better even.  But it sounds like you're trying to re-invent the fractal.  I'd like to see some pics of yours that show "perturbation" value changes done to lumber, brick, and tile.  Maybe there is a simple way of doing it in Carrara without the added number-crunching.

Start a new thread so that we can stay on-topic here.

Quote - I'd like to see some pics of yours that show "perturbation" value changes done to lumber

 

like i said, lumber mill is end of the tree, can't have outside values control it, one of the ones i wasnt able to rewrite, hmm wonder if daz will share source code :)

brick is an example that i could do, but leaving for work in 5 minutes, no time to write a formula so i'll try explain quick. brick you can choose many variables, but they can't be controlled by any shader inputs, say i want a circle patio design? that requires the "width" of the brick and mortar to go from wide at one point, to narrow, before you can warp it to a circular pattern, in poser i could simply have V coords effect it, thus adding more bricks to the row as V coords go from 0 to 1

in carrara i'll admit this isnt impossible to do with workarounds, i could make 1 branch for every row of brick, but that can lead to hundreds of branches, so my solution was to rewrite a new brick plugin to use, that allowed its variables to be changed by outputs of whatever i choose

late for work, hope that helps, gotta run -.-