Attached Link: http://www.renderosity.com/mod/tutorial/index.php?tutorial_id=2295

Quote -

As soon as hardware can do displacement in real time, they'll add it.

They could now give us a preview using parallax mapping since DirectX10 cards and above support that.  Blender's game engine now supports three types of parallax mapping.  I just don't feel that viewport improvements have ever been a high priority to the devs since almost everyone does their final renders in Firefly.

Quote - Here's another experiment, which makes no sense unless the mesh has moved.

My big sphere is now made of glass. There are two small decorated spheres. The lower one is being refracted through the glass of the displaced lobe. How is that possible if it isn't really there? How do you explain why the two smaller balls do not appear identical, if the lobe isn't really there between the camera and the little sphere?

If the renderer was just moving the glass pixels to the right, we wouldn't see the sphere through it because in that case it isn't really behind the glass.

This is not an illusion as you described it.

Hey, bagginsbill, a quick question on lighting if you please. Why is the shadow of the Sphere solid? If it were indeed made of glass shouldn't some light be shining through it?

Also, you said the blinds were made using a one-sided square and a displacement map. Well, if displacement maps actually move the mesh around as you say, why is the shadow solid?

And for the record, I'm not knowledgeable enough about any of this to be calling you out. Just want to know why those two examples have solid shadows? I'm assuming the answer has something to do more with Poser Lights than with displacement maps, but I honestly don't know.

Hey, bagginsbill, a quick question on lighting if you please. Why is the shadow of the Sphere solid? If it were indeed made of glass shouldn't some light be shining through it?

thats Caustics.. something that poser does not support atm.

Quote - Hey, bagginsbill, a quick question on lighting if you please. Why is the shadow of the Sphere solid? If it were indeed made of glass shouldn't some light be shining through it?

thats Caustics.. something that poser does not support atm.

But you can fake it if you make the sphere transparent, right?

ignore me.. I misunderstood (was still half a sleep) and thought you were refereing to something else.

I'm not sure about the glass sphere, but I would imagine that the high refractive index prevents most of the light reaching the ground. 

But on the blind, the reason for the solid shadow is clear. The light is shining from top right, and the slats of the blind are angled such that the light can 'see' no gaps between them. However, if you look closely at the far end of the shadow (through the blind) you can see where the light is shining through the end holes in the slats.

(edit) also in the very bottom left corner of the image.

 Yep, you're right about the blinds. I had to take a much closer look, but I did see the light slipping between the the end.

I'll take your word on the sphere.

No, no...  don't take my word on the sphere, I was only guessing!
:O)

But here, with the light coming from a lower angle, it can go between the blinds.

Think about the magic of this - a single polygon casting shadows on itself, but not everywhere. Pretty cool, IMO.

Regarding the shadow on the glass prop:

Kaibach was right, sleepy or not - Poser does not implement caustics.

See: http://en.wikipedia.org/wiki/Caustic_(optics)

The operative phrase in the definition of caustics is "refracted by a curved surface ". Poser transparency is not refraction - it simply lets light through in a straight line without bending it. In particular, caustics not only bend light rays, but they tend to bring many of them together, i.e. focusing them into a hot spot. Strictly speaking, the lack of caustics is not a shadow issue, although they are related when we look at the effect of a glass object on a table.

Poser's lighitng calculation is much faster the way it is done, but it loses the ability to detect that some light from a light source could be coming from a direction that is not where the light source is. Poser only deals with examining what is directly between any given spot on a surface and the exact location of the light source. It does not take into account the possibility of going to other places (places not directly on the line between the surface point and the light) and bending back toward the light source. So - no caustics, i.e. no focused hotspots of light that should have landed somewhere else if refraction wasn't involved.

Now let's talk about shadows.

Poser shadows deal only with objects interposed between the rendered point and the light source origin. Any object that intersects that line is considered an opportunity to block the light. The "transparency" of the object will determine what percentage gets blocked.

When we use a Refract node, we're asking Poser to show us what is behind the surface and in a new direction (bending our viewpoint). This is not at all the same as modelilng the bending of light - we're just bending our viewpoint and adding some stuff from another place in the scene to the currently rendered point. So, technically, shaders that use Refraction are not transparent.

Think of refraction more like looking at the output of a web cam on your screen. The web cam "bends" your viewpoint so you can see things that are not directly in front of you, but you "see" them directly in front of you on your screen. Nevertheless, your screen is opaque, not transparent, and it blocks light.

It would be possible to design additions to Poser that would automatically take caustics into account when a Refract node is used on the material. In such a case, objects behind it would be able to be lit through the glass, so not in shadow, and also they'd be able to actually experience extra brightness due to focusing. But the number of rays necessary to accomplish this is way more than the usual one or two rays we use in Poser-style raytracing.

I have come up with a way to make a refractive glass that let's light through in Poser. It does not accomplish caustics (focusing of extra light) but it does at least let the light through. It involves combining transparency and refraction and it's a little tricky and slow but it does work.

I recently posted it here:

http://www.renderosity.com/mod/forumpro/showthread.php?thread_id=2696787&page=7#message_3461574

Attached Link: http://en.wikipedia.org/wiki/Caustic_(optics)

I noticed the glass prop has smooth shadow edge at the tip of the displacement.  does this mean that it's got substantially more polygons than the poser spheres?

Thanks for taking the time to answer BB. That was very informative.

Question: Is all this information anywhere in the Poser manuals, or is it found elsewhere? 

Quote - bb, that first link of yours somehow missed out the closing bracket, and doesn't find the right page.

Thanks - I accidentally touched the delete key and I put the paren back but it didn't become part of the link. Doh!

Quote - I noticed the glass prop has smooth shadow edge at the tip of the displacement.  does this mean that it's got substantially more polygons than the poser spheres?

No - that smoothness is because of the displacement, which was procedural, done with math nodes. Poser executes procedural displacement exceedingly smoothly.

Quote - Thanks for taking the time to answer BB. That was very informative.

Question: Is all this information anywhere in the Poser manuals, or is it found elsewhere? 

Most of the deeper-workings stuff I write about Poser is not written down anywhere, until I write it. Some of it is, of course, but most isn't.

If you ever get round to writing the definitive book of Poser BB, you'll be able to retire and go buy Hawaii :)

"Zen and the art of Poser" would be a good title

"Zen and the art of Poser" would be a good title

If you have read "Zen and the art of motorcycling" and considering the fact that poser can completely drive you nuts, I totally agree with you.

best regards,

Bopper.

read a bit of it which is why I thought of the title :biggrin:

Hi BB, soory to bug you but i had a quick question on a theard were you talked about displacement.

How exactly would I use the math node to subtract from a disp map where 50% grey is 0...
black is -100
white is +100

disp map is from zbrush, I'm not getting how to use it in poser...50% grey is inflating my mesh.

Could you help? I'm just not used to this in poser, I've been working in maya till now and I dont get these nodes so well

In Poser, the numerical value of whatever you plug into the Displacement channel is used to decide the displacement.

In an image, all the numerical values are positive.

But many applications use mid-gray to represent neutral, no change. We would like Poser to treat that value as numerically 0.

This screen shot shows the problem. I have a simple displacement map, with Photoshop's 50% gray as the common color. I have written some things in white and black and I have drawn some things and also placed a smooth gradient on the bottom right.

I have arranged two squares that are in exactly the same position, one slightly larger. The larger square is white and mostly transparent. The smaller square is orange and has the displacement map on it, just plugged straight in.

Because all of the values are positive, except the black values, which are 0, then everything that is not black is moved toward the camera, displaced in the positive direction. No part of the orange square is behind the white square. Our goal is to change that.

The places that were just touching the white square are now behind the white square by 1/2 inch. (Because I'm using inches and my Displacement number is 1, which means the difference from black to white is 1 inch.)

So over the depressed areas, we can now see the white semi-transparent square covering the dips.

Now this is a big improvement, but not quite precise.

It turns out that 50% gray is not precisely .5. Depending on what your software thinks is mid gray, it could be RGB 127 or it could be RGB 128. In either case, this is on a scale of 255, so the correct offset is either 127 / 255 or it is 128 / 255.

But rather than try to figure that out, I have a better solution, which I'll show next.

Use a Color_Math:Subtract node, and for Value_2, click on it, then click on a neutral gray area of your displacment map in the preview. This will copy the exact neutral value into the subtract node.

So it is automatic this way.

You can make any value you see be the neutral value just by pointing at it. Easy peasy.

Because the two squares in the neutral areas are perfectly aligned, we see a goofy pattern where Poser cannot tell which one is closer to the camera. That is good - we did it right.

Of course I found my previous explanation after doing it all over again.

Here it is:

http://www.renderosity.com/mod/forumpro/showthread.php?message_id=2949080&ebot_calc_page#message_2949080

I appreciate the quick reply BB!

I found this while I was trying to figure out that math node
http://poserpros.daz3d.com/forums/viewtopic.php?t=44342

I uses the math and component nodes, but your looks simpler!

...and he's done it again.

sticks this in his Evernote database

Quote - I appreciate the quick reply BB!

I found this while I was trying to figure out that math node
http://poserpros.daz3d.com/forums/viewtopic.php?t=44342

I uses the math and component nodes, but your looks simpler!

Not only simpler, but Stonemason's node setup is actually sort of wrong, but he got an OK result by accident.

First of all the Comp node he used pulls out one color channel from an RGB image. But if all the colors are shades of gray, then the numerical value of the RGB color is identical to the numerical value of any of its channels, so that is a waste of time. The image itself is the same value as the Comp node, as long as the color in the Comp node is white, which it was in the first one. But the second one he had RGB 38,38,38 in it, which actually scaled the numerical value down by 38/255.

So, if we call the numerical value of the displacement map M, then mathematically his node setup calculates:

2.5 * (M + (38/255 * M - 1))

This equals:

2.5 * (1 + 38/255) * M - 1

which is approximately

2.873 * M - 1

Whereas what he really wanted was:

2.5 * M - 1.25

Since this isn't far off, he didn't notice the difference.

For proof that

1) Poser colors are floating point numbers

and

2) mid gray is 127 / 255 or 128 / 255, but neither is equal to 1/2

see:

http://www.renderosity.com/mod/forumpro/showthread.php?thread_id=2761668&page=5#message_3396544

Quote - [

Not only simpler, but Stonemason's node setup is actually sort of wrong, but he got an OK result by accident.

I agree totaly,I havent used that method in years..now I just do the one math node for displacement  :)

:)