Quote - Which is why renderers like Maxwell render in spectra instead of RGB (part of the reason for their ridiculous levels of photorealism)

I understand all the theoretical advantages behind spectral rendering, but so far I haven't found any side-by-side comparison of spectral vs RGB rendering that would show a visible difference. Would love to see one if anyone can find one :)

The closest I've seen is http://www.anyhere.com/gward/papers/egwr02/img/ColorPlate.jpg where the difference is measurable, but barely visible in the direct comparison and would go unnoticed without a reference.

Keith showed us a few pictures, with the intent to show us how gold comes in various colors. That's nice, but I can make a defensible argument that those three samples of gold were identical, and that the color variation came from environment differences or camera differences alone.

I believe it is 24K gold, but I'm not sure. I don't really care. Watch.

The disgusting color is from compact fluorescent light bulbs.

THIS IS THE EXACT SAME OBJECT IN ALL ELEVEN ROOMS.

Clearly, the photographic evidence of gold variation is not useful AT ALL unless you include the un-reflected object as a reference in the same photo.

(o) (o)

But even this was too casual and has a flaw. The utensil is reflecting light onto the paper, changing its color. Measuring these two this way, I'd be pulling in measurements from a feedback loop. This would give incorrect results.

One more thing about the "color" of gold. I gave you an sRGB color. I'm assuming (and make sure to use) the linear version of that in my shader. That means that to get the actual ratio of components, you must not use the direct value of x/255. You must convert from sRGB to linear.

For example, if the red is 255 and the green is 230, the ratio is NOT 230/255 = .902. You have to take it to the power 2.2 to remove the gamma correction. So the true ratio is (230/255)^2.2 = .797.

Poser Pro does this automatically.

If memory serves, you've just demonstrated Chevreul's colour theory which itself was a founding theory of Impressionist Painting and taken to extreme by The Fauves in the very early 20thC. Another artistic revolution driven by science, please continue the lesson

Quote - Keith showed us a few pictures, with the intent to show us how gold comes in various colors. That's nice, but I can make a defensible argument that those three samples of gold were identical, and that the color variation came from environment differences or camera differences alone.

Absolutely, for part of it. The first was taken in a small room where I didn't need a flash because the flourescent lighting was sufficient. The second was taken in a big empty space where I had to use a flash (with a different camera) because otherwise it would have been too dim.

However, and in this case you'll have to take my word for it, the first brick really was significantly duller and of a different shade than the latter. It surprised a lot of the people there at the time (who'd never seen raw gold before) that it really wasn't very shiny at all.

Now for pure, polished gold (which you have with the spoon) you've got a pretty optimum situation for seeing it in different situations. If you're comparing different gold objects, say a beaten gold Bronze Age ornament, it's going to have potentially a radically different colour, even in the same environment, than a piece of gold ornamentation coming from a different source or processed in a different way. It's one of the things I run into as a geologist regularly: you can't rely on colour/reflectivity alone as a diagnostic feature except when dealing with samples from the same environment examined in the same environment.

What makes gold look gold is not the color itself, the illumination effects makes gold look as gold.

I did some quick experiments, as I have no gold trimmed shader and no raytracing I did it with what I have.

In this image I used standart Phong for specular channel, a modified not Lambertian for diffuse channel and a modified environment mapping and three directional white lights.

As you can see it doesn't look as gold and as I said before I haven't the appropiate shader for gold. What made look as gold were the reflections even if fake.

The second interesant thing was that in the second image where the gold look much better I used instead of Phong (left) the Lommel-Seeliger shader (right), nothing to do with gold and metals, this shader is made for rendering the Moon!!!

The Lommel-Seeliger shader produce a strong illumination at glazing angles, so if it made gold look better it means that gold has strong reflections at glazing incidence angles.

ohhh, is environment mapping the secret trick? 

my results last night were so awful, doan want to post it.

Environment mapping is a way to make fake reflections. You can make natural reflections with raytracing. The problem is that must exist something to be reflected. A dull room with nothing will not work. You need objects in the room, extended area light sources or windows acting as a light source.

the above is what I got with lo-qual. FFRender/IDL settings and first approx. of schlick approx. of fresnel gold procedural.  the difficulty here is that if anybody posts screenshot of optimal schlick approx. of fresnel gold procedural, then this would present probs in case bill is planning to sell his metals shader pak.

p.s. the inclusion of the ibis-winged demigoddesses on top of the sarcophagus indicates the origin of the hebrew religion was Amun-Ra.

Oh anybody who can figure out what to do is free to post it. Believe me, you're not going to match what I have. I have the same shader doing several golds, several coppers, several brass, bronze, silver, platinum, steel - in all of these I have various finishes (smooth to blurry/rough).  I have grain and mottling where needed. I have the specular node perfectly tuned to match the blur and IOR so that reflections of Poser lights look just like reflections of a Poser glowing sphere. I have several hammered versions that are really beautiful. I don't know of anybody outside guys at Pixar who can match what I'm doing. Not even at SM.

Looking at yours, Miss Nancy, I see:

There is diffuse there in your shader. I measured it on real gold - polished gold diffuse response is less than .1%. Get rid of that.

You don't have the right change in color as a function of viewing angle. That was really tricky and took me a long time to get right.

Did you use the color I gave for straight on? It seems to look a bit too green. That looks more like brass.

Quote - Oh anybody who can figure out what to do is free to post it. Believe me, you're not going to match what I have. I have the same shader doing several golds, several coppers, several brass, bronze, silver, platinum, steel - in all of these I have various finishes (smooth to blurry/rough).  I have grain and mottling where needed. I have the specular node perfectly tuned to match the blur and IOR so that reflections of Poser lights look just like reflections of a Poser glowing sphere. I have several hammered versions that are really beautiful. I don't know of anybody outside guys at Pixar who can match what I'm doing. Not even at SM.

Now you're just making me impatient LOL

Release that already, will you? (Please?)

Yeah BB, hurry up and finish the pack already, you'll have an instant #1 product!

So get off yer arse, perfect what you believe to be incorrect, and give it to us already! lol

Just kidding, don't rush it cuz us buttheads are impatient hehehe

o.k., here's one where it's (255, 215, 147), more colour change and less diffuse.  it looks less cartoonish in this one, but they may expect cartoonish gold if they haven't got a few pieces lying around to compare to their renders.

MN - 100 times better!

Are you using GC or no?

no, it was HSVTM=1.67, gain =1.33, envdome/hdri and one inf lite at 75%, 4 IDL bounces (quickie).

it's gonna look too dark to many, as this sponza building is like an elevator shaft - even with 16 bounces there ain't much to reflect, excepting the floor.  it would look better in something bright with alotta sky, e.g. the free OBL compound and envdome.  or I could put some lites in the sponza atrium.

That looks really good.

Are you thinking to include some rusted and dented metals?

Over time - initially I just need to get this basic/essential shader done and out the door.

ricardo, to do one of these in poser 4 or earlier, place a camera with focal length = 1 mm at location of gold object, delete gold object, render the scene with that camera, then use that render as the reflection map for the gold object.

what one notices when holding a large flat gold surface next to various coloured surfaces: it reflects red, orange, green and purple rather well, but blue looks almost black, hence one may need to decrease blue value in head-on colour of procedural.

Some critics. The romm in the scene looks very good, but the covenant ark looks very bad. The poles or handles are completely black without any relief and un-natural. The ark itself doesn't look as gold. It can look as metal or painted metal that has suffered severy corrosion or have been exposed to the environment for some centuries. The object in the back has gold reflections, but looks more as a black magic mirror.

Quote - Some critics. The romm in the scene looks very good, but the covenant ark looks very bad. The poles or handles are completely black without any relief and un-natural. The ark itself doesn't look as gold. It can look as metal or painted metal that has suffered severy corrosion or have been exposed to the environment for some centuries. The object in the back has gold reflections, but looks more as a black magic mirror.

In which image? The one just above your post? The handles show up brown on my monitor and look like wood. Might need to have a look at your monitor, maybe? :biggrin:

Quote - In which image? The one just above your post? The handles show up brown on my monitor and look like wood. Might need to have a look at your monitor, maybe?

In the image above. The poles doesn't look as wood and look dark and flat 2D surface (RGB 48.8.0) no matter how adjust the monitor. But the room continue to look very good in my monitor

Okay, I'll take your word for it, Kawecki. However, on my monitor, the handles look like a rose-wood, with a wood-grainish texture. Might need to check my monitor, then. :blink: Wonder what the original intent was of the image maker? Black or rose-wood? Or something else?

Well, not all the monitors looks the same, the famous gamma story. But, forget the handles, we are not discussing wood that is other story. Our focus is gold.

Returning to the gold question.

We must understand the behaviour of metals and the behaviour of metals can be very different depending on the roughness of its surface.

1-

Metals with very low roughness behaves as a mirror. In this case if we use Phong, Blinn or Torrance it will not make any difference. Just set the reflective color and use raytracing. With gold you set the reflective color to some yellow, with copper to some orange and with silver to white.

A metal with very low roughness it will act as a mirror and you will not see the metallic object, you will see distorsed reflections from the environment where the distorsion will depend on the geometric shape of the object. With planar surfaces there is not disrtion at all and you see a perfect mirror image on the object's surface.

2-

The other case are very rough metals. Very rough metals don't produce mirror-like reflections. Phong and Blinn are inadequated for this case. Torrance works pretty well. The main characteristics of very rough metals is that have a big specular area and illumination falling quickly beyond this area. You can approximate using Phong with very low exponent values (1 to 4), but is not the same as Torrance and don't produce off-specular peaks. This is what I used in my picture posted aomewhere above.

3-

For intermediate roughness metals there is no simple solution, there do not exist a simple model for this.

4-

In all the cases the diffuse compont of metals is very small or zero.

5-

For making images of metallic objects things get very complicated, unless we use simple geometric shapes as the famous metallic spheres.

The behaviour of a metallic surface changes dramatically with the roughness and roughness can be changed polishing the metallic object. If the object is not a simple geometrical shape and many times has a complicated shape the polishing is never perfect and so, we will have areas that have been polished very well, for eample the borders, and other areas where we were not able to polish.

The result is that we have a metallic object with areas that act as an almost perfect mirror and other areas that have a very rough surface.

Quote - > Quote - In which image? The one just above your post? The handles show up brown on my monitor and look like wood. Might need to have a look at your monitor, maybe?

In the image above. The poles doesn't look as wood and look dark and flat 2D surface (RGB 48.8.0) no matter how adjust the monitor. But the room continue to look very good in my monitor

The poles look brown on my monitor.

on some monitors the gold is gonna look too dark, hence the wood will also look very dark on those monitors.  but looking at gold pieces, I realised it's difficult to even photograph them properly to show them off, the same as trying to photograph the dispersion fx of faceted gems.

The poles look brown to me as well (tops average ~ 90,30,0); ISTR they were originally made of wood of that colour (the nile thorn-tree that giraffes like to eat), but they may need to lighter and more specular to show up well, to show any displacement on them. there also were several errors in the last render in re: lighting and bad welds/smoothing on the lion and wall/floor in the background.  needs more IDL bounces/samples/shading rate.  bill explained how to do blurred reflections in an earlier thread, to give reflection fall-off.

Quote - I realised it's difficult to even photograph them properly to show them off, the same as trying to photograph the dispersion fx of faceted gems

especially when our view of how they "should" look is at least in part tainted by other post-worked images in the media?

Tried experimenting a little more, didn't think I made much in the way of improvements though.

wint, for optically accurate gold renders, we may await the debut of bill's metal shader pak. given the interest shown here, one would suggest including it (even if unfinished) in the next poser release (default runtime, procedurals) in order to obviate any marketplace issues.

p.s. somebody has got a gold sphere which one might compare to our renders, and some guy in ethiopia said he had an ark. when asked if we might see it, he went into his building and did something for a while. then he came out and averred that the ark was still extant, but wouldn't let us see it. :crying:

Quote -

This looks f(*¨&¨$$&¨$%$ amazing.

And here I am, still failing miserably at getting anything to look anything near gold.

o.k., I'll give it up: bill actually posted the basic shader in his floor-tile thread.  it's  ... oh, fab, I can't even find it now.  the one with the blue tiles and the procedural with the schlick approximation inside a red rectangle.  all one really needs is to take a few fotos of some gold items, preferably as large, as flat and as solid gold as poss.

p.s. the coloured reflections may look fake, but I tried it with some flat gold surfaces and it really does reflect most colours quite well.

Is it this thread?

yes, wint.  if, in P8 or later, you can simplify that by deleting some of the tile nodes to get a gold-reflective surface, have 4 ray-trace bounces, use IDL if possible, use GCrender or similar etc., it may look more like gold.  bill gave the gold colour in the first post. you know you're getting close when the bottom half of your sphere is gold, assuming a light diffuse floor colour and IDL.