Quote - aha now i understand. for true realistic renders we need to use HDR maps. but we can not use this for backgrounds. because this is all data.
correct?

Didn't say that. You went too far. I said it is not fair to say it is darker. Because the raw data is there, I can make it darker or lighter or anything I want, because all the true luminence is there, regardless of how super bright it is or how super dark the details are. They are all there. A true HDRI image records the actual luminence even if the brightest is a trillion times brighter than the darkest.

Ordinary JPEG, PNG, etc. do not have that kind of range. They are encoded in a narrow range. Any luminance info that doesn't fit in that range is simply discarded.

Think of the HDR image as being real life. You can take a photo of it at any number of different exposure values and display that as a JPEG. But no single JPEG can accurately depict all the values that are there.

If you're standing in an enclosed room, looking through a doorway at a brightly lit exterior, you can look at that as a JPEG with either the inside looking good or the outside looking good, but not both at the same time. This is not the same as saying you can't look at it. You can look directly at it. But first you must decide which range of values you want to map to your 8-bit illuminance range for your computer screen. That's all.

The full data is there in an HDRI, just as it was in real life, and so when we render an image that utilizes multiple exposure values simultaneously, such as looking directly at it as well as looking at muted reflections, we see detail in both as well as lost information in both.

The direct view of the doorway shows the interior wall clearly and the yard outside is totally white. The indirect view using a muted reflection on the ball shows the outside yard clearly and the interior reflection is totally black (not visible).

bagginsbill how could we make gamma corrected white plastic? plastic is a little reflecting right?

HI  -Moron question  related to something mentioned earlier in the thread -
where/ how do I find a SKY DOME  and apply it to a scene.

http://sites.google.com/site/bagginsbill/free-stuff/environment-sphere

Too cool.

thanks  Ice-boy  and  Bagginsbill.

Quote - bagginsbill how could we make gamma corrected white plastic? plastic is a little reflecting right?

Shiny plastic is easy - but you must respect the simple facts of the Fresnel effect. If you set up uniform reflection intensity across the entire surface, you will fail at realism. Plastic reflects less when facing the viewer, and more when facing away from the viewer, sideways.

There are plastics that are so shiny they are effectively glass. Others are microscopically rough, but still smooth enough to reflect a little bit at the edges, not at all towards the camera. You must tune the Fresnel effect to compensate.

Just like the soft metal reflections we did in the past, ice-boy, we may have to use blurred reflection to get perfect realism.

Also, even white plastic does not perform 100% diffuse reflection, so don't put Diffuse_Value = 1. It should be about .8. I will build you a couple examples and you can work from there.

with edge blend we can get that it reflects more when facing away and less when facing at the viewer right?

Quote - with edge blend we can get that it reflects more when facing away and less when facing at the viewer right?

Yes, but that doesn't make it right. It makes it less wrong. Which is more similar to your head, an apple or a chair? The apple, right? 

So if I draw you with an apple instead of your head, and also I draw you with a chair instead of your head, which is the "better" drawing? The one with the apple, right? Is it a realistic drawing? No. Less unrealistic, yes, realistic, no.

The Fresnel effect is very specific and not directly calculated by the Edge_Blend node, although using the Edge_Blend node (like the apple) is better than nothing. I used to approximate Fresnel with Edge_Blend. I also used to produce non-realism. I now produce 100% realism on Fresnel effects, because I use the real formula.

is this the same formula like in the ORB shaders?

Formula, yes, actual values in parameters, no.

Also, I have discovered an approximation to the 14-node Fresnel equation for glass that is only 3 nodes and is very close to the same value.

interesting.
learning every week something new.

how are the parameters and values  different between the orb shaders and plastic and maybe car paint?

any luck with the plastic shader? :)

can you then maybe explain to me some basics to gamma correct reflections?
when i am using your car shader and gama correct it it looks very bad.

a different question. i can not open poser now so i dont know if this is already possible. can we make an object in poser HDRI? for example i make two balls. one has reflection and the other one is bright red(like a light). can we somehow make objects in poser HDRI? so that in the reflection the ball is super bright? 

If you have been reading the thread "Nodes for Dummies", you'd have heard me discussing hyper-colors and hypo-colors. A hyper-color is one with an RGB component that is more than 1, which is what you're asking for here.

Now there are many ways to produce hyper-colors in Poser, although you cannot directly SEE any of them. The simplest way is to take an existing color and multiply it with a big number. For example, if you start with RED  RGB(255, 0, 0), which is numerically [1, 0, 0] and you multiply by 5, you will get [5, 0, 0], a hyper color. And when you look at that in a slightly reflective surface, such as plastic, with a 2% reflectivity, you will see the red reflection at .02 * 5 = .1.

The easiest way to get a self-lit hyper-colored ball is to put the color or node you want into the Ambient_Color, and set the Ambient_Value to a high number. (Because the output of that channel is the product of color and value - just what you need.) The effective color is Ambient_Color * Ambient_Value.

You did not say whether this ball is to be self-lit or lit by other things. If self-lit, the red ball should have Diffuse_Value = 0 and Specular_Value = 0.

As for how to handle reflections the GC way, that is simple. If we assume that all surfaces in the scene are GC, that means they are sRGB and not linear. So you need to convert to linear for the calculation.

Add a Reflect node. Connect that to a Color_Math:Pow.Value_1. Connect Value_2 to a gamma number, such as the 2.2 I usually have. This two-node combo gives you linear reflections, even though everything in the scene is using non-linear shaders now.

Once you have that, you add it to your diffuse and specular as usual, using Color_Math:Add nodes, or by using a Blender. Your choice, how to connect. (When doing the Fresnel effect, remember that we need reflection_strength + diffusion_strength = a constant less than 1)

Once the Diffuse + Specular + Reflection has been combined in the right ratios, you gamma-correct the result (power 1/2.2) and plug into Alternate_Diffuse.
 
The general form looks something like this:

( Blender(Diffuse(...), Reflect(...), fresnelValue) + Blinn(...) ) ** (1 / 2.2)

i did something wrong

whait a minute. dont tell me yet. i think i am closer.

Quote - If you have been reading the thread "Nodes for Dummies", you'd have heard me discussing hyper-colors and hypo-colors. A hyper-color is one with an RGB component that is more than 1, which is what you're asking for here.

Now there are many ways to produce hyper-colors in Poser, although you cannot directly SEE any of them. The simplest way is to take an existing color and multiply it with a big number. For example, if you start with RED  RGB(255, 0, 0), which is numerically [1, 0, 0] and you multiply by 5, you will get [5, 0, 0], a hyper color. And when you look at that in a slightly reflective surface, such as plastic, with a 2% reflectivity, you will see the red reflection at .02 * 5 = .1.

The easiest way to get a self-lit hyper-colored ball is to put the color or node you want into the Ambient_Color, and set the Ambient_Value to a high number. (Because the output of that channel is the product of color and value - just what you need.) The effective color is Ambient_Color * Ambient_Value.

You did not say whether this ball is to be self-lit or lit by other things. If self-lit, the red ball should have Diffuse_Value = 0 and Specular_Value = 0.

thanks for this. so if i now make a matte in photoshop of a 360 image i can make fake HDRI?
sometime on the internet i find a good 360 image that is not HDRI. so with some knowledge you know where the brightest parts are. so if i could make a matte i could then in the shader writte what is super bright.

correct? 

Quote -

bagginsbill what do you think? 

Perfect! You did it exactly right. That should work pretty well, eh?

are you serious? i did everything right?

i must admit i cheated a little. when i read your post i didnt know what you are talking about. i connected the nodes like you said. of course it didnt work because i didnt do it correctly. then i remebered that you said that te ORB shader is GC reflection but is using different settings. so i was looking at the shader. and i noticed that the only difference is the fresnel effect. correct?
so i noticed that two nodes were missing from my shader and i conncted them. then i copy/paste the blender and edge nodes.

this ifof course not real plastic but for the demo i rendered without blury reflection.
i am thinking of using this shader for the eyes maybe? with the trick that you told me about hyper colors in reflection i could do now fake eye reflection without HDR images.

Yes the only difference is how the Fresnel effect is calculated. Here you're using a simple Edge_Blend which is fine for many purposes. It's only when the dominant effect on the material is Fresnel reflection that the math of the FE really matters.

This would certainly work for the eyes. Just remember to put some bump on the eye white part.

i will. for the cornea i can use your cornea shader from your apollo shader. right? 

You could, but the Apollo shaders don't implement GC, or not correctly anyway.

The issue with GC reflections is that the non-GC version doesn't produce the right colors.

Let's stick to grayscale for this discussion, so I don't have to deal with RGB triples. The math is the same, just done 3 times, once for each color channel.

Suppose we consider two points on the cornea. One point has the pupil behind it and another has the iris behind it. Suppose the pupil is black (0) and the iris is reflecting (without/ GC, i.e. linear value) mid-gray (.5).

Suppose further that there is a reflection of a window and the brightness there, with GC, is .5.

And suppose that the Fresnel effect says the reflection coefficient here is .2.

Let:

Fe = Fresnel Effect reflection coefficient
Dc = Diffuse color (linear) after illumination is taken into account
Rc = Reflection color after illumination is taken into account (the gamma corrected color of some other object in the scene)
G = Gamma correction factor (usually 2.2)

The general formula for linear reflection + diffuse is:

(1 - Fe) * Dc + Fe * Rc

And the gamma correcting formula is:

((1 - Fe) * Dc + Fe * (Rc ^ G)) ^ (1 / G)

So, the pupil section of the cornea would be according to the Apollo (linear) shader would be:

.8 * 0 + .2 * .5 = .1

while the GC shader version would be:

(.8 * 0 + .2 * (.5 ^ 2.2)) ^ ( 1 / 2.2) = .24

The Apollo shader would produce a much lower and inaccurate reflection.

Now let's consider the iris section. The Apollo (linear) shader would be:

.8 * .5 + .2 * .5 = .5

while the GC shader would produce

(.8 * .5 + .2 * (.5 ^ 2.2)) ^ ( 1 / 2.2) = .69

And again we see the Apollo shader does not mix the reflection with the diffuse correctly.

wooow. i thought that for glass we dont need GC since the reflection should be linear

i can not make it to work. i dont understand all this technical stuff :)

Can't make what work? Your plastic shader above looks fine. Change the SimpleColor to white and plug in an image map for the colors of an eye.

For the eye cover or cornea, you can get away with a non-GC shader such as I did for Apollo - you'll just have to guess how much extra reflection to push to compensate.

Or you can wait for me to release the VSS Pro shaders which will take care of all this this stuff in the most sophisticated way possible.

if i can use the old cornea shader then perfect.
i am happy with the plastic shader.

will the VSS Pro be realesed this year? 

I hope. I perpetually have no time to work on it.

I got another contract this week that's very interesting but I can't talk about it. All I can say is I'm booked 6 days a week now.

thanks for the answer.

Attached Link: Raytraced reflections on a sword blade?

You might find some inspiration from the materials I did for Vince Bagna's sword blade.

i think for the best effect we need HDR enviorment maps. because then the bright parts really pop up on the surface.

i am trying to do now a fake way of blurrying teh reflection. if the enviorment map is blury it looks better. but the reflection or objects around the metal would still be sharp.
i was reading a paper from pixar about how they did soft reflection in Ratatouille
graphics.pixar.com/library/SoftReflections/paper.pdf

we dont need details and it doesnt need to be accurate when its soft. but i dont think this can be done in poser.

Raybias too high.

I can't explain why the renderer has to work this way. Makes no sense to me. All I know is given the point from which you are ray-tracing a refraction, any geometry closer than the ray bias is ignored. You are getting the back of his head, not his iris, because the iris is closer than the ray bias.

You said even if you use low raybias there is still a hard edge. That is correct. The person who modeled the cornea made it coincide with the edge of the iris instead of being a layer above it. This is where physics matters. A real cornea surface is never touching the iris - it is above it. But in the model, they touch, i.e. the distance is 0. So you can never set a ray bias low enough.

Instead of decreasing ray bias, you can try to apply a tiny displacement to the cornea and eye surface. Do not try a negative displacement on the iris. Poser does not "see" displacement too well when performing a refraction or reflection, so if you move the iris, it will still appear as if not moved and the raybias problem will remain. On the other hand, in Poser the origin of a refraction can be displaced so moving the cornea might help.

this explains why you had displacement on the cornea in the apollo shader

Yes it does. :) I have done many obscure things without explanation. I've often found it interesting and very enlightening that I'm rarely if ever asked about these things.

you made a fantastic bulge with nodes.

Quote -

You said even if you use low raybias there is still a hard edge. That is correct. The person who modeled the cornea made it coincide with the edge of the iris instead of being a layer above it. This is where physics matters. A real cornea surface is never touching the iris - it is above it. But in the model, they touch, i.e. the distance is 0. So you can never set a ray bias low enough.

can you show a primitive painting how th cornea on apollo eyes should be?

i will try to do a morph to fix this.

i already fixed the problem. i made a morph

was readings something about Tranformers this week.
they didnt raytrace everything on the robots. only where we would noticed the details.
they raytraced only the glass,pipes and the parts that are near the head for close ups.

so i was looking at some screencaps. and i found this. look at the pipe. it doesnt have 2 raytrace bounces. in the mirror it looks like a white pipe.

-one bounce
-2 bounces

• 1 bounce and the diffuse node connected to raytrace
  

so if you have carpaint or something similar you can make a cheat. connect something like the diffuse node to the raytrace node. that way you can use 1 raytrace bounce in the rendersettings.

and the reflection will not be just plain black.

what do you think?