What does Absorption color lightness mean?

[vicnaum]

Did another test. Now I fixed the Distance to 5mm, and changed the VALUE on different models with different wall thickness:


As you can see, values like 255 or 10 give very strange results. So you should keep somewhere in-between...

[juang3d]

Hi there.

The results are not strange, first of all you should not divide your thought into the three color values, you just have to decide "artistically" what color do you want to use for absorption.

Now answering your question, the results are not strange, since you use a bright color (in the 255 value) the light ray has more energy so it takes longer to dissipate, since I don't know the measures of your object I cannot say why the distance is behaving that way, but probably this is because the color is not reached within the distance you set, on the other hand, you should not be using sun and sky for those tests, you should test this in a clean environemtn with just one white light so you can observe the effects of absorption over white light, in your scene there is no white light if I'm right, there is a sun (wich is not white) and a mixtures blue from the sky, so this will affect your results also.

Now int he value of 10 the results are also predictible and this is because you are using a very dark color, wich means with low energy, in fact you can see alue as energy I think if you want to separate the three components (you should not do that IMHO, as I said you should pick the color artistically and use distance) but anyways, since the color is so dark the light rays loose it's energy a lot faster so it will give you a dark volume inside the object because you are "killing" light inside the volume(as a manner of speaking)

Anyway, you have to take into account how much does your model measures, because without that info the distance value is useless :)

Cheers!

EDIT: you can see, in your 3mm test and the 255 value that you reach that color and it stays, this is because it has enough energy when it has mutated so it keeps going with that color until it reaches the exit of the object, it all depends on the measures of your object.

[vicnaum]

The measurements of the model wall thickness are written above it, please check. The distance is 5mm.

The sun&sky affection isn't so strong, but giving more natural looking results - on sterile tests you can be fooled more easily.

I completely understand how absorption distance falloff works, and that thicker walls should be darker, etc etc.

What I want to emphasize is that I can't get _predictable_ results - on 255 values I get a strong blue tint without going into the black even if the wall thickness is 15mm, and we have two walls there on that cylinder = 30mm - but we still get deep blue instead of black.

Values of 192 or 128 are more predictable - because now we get black on very thick objects, while still retaining the COLOR on 1mm thickness (although it's more desaturated).

And on very dark values, like 10 - we get black everywhere, with a slight tint of color on thin walls - if I make it even more thin (0.1mm) - then I'll get a very faded and desaturated color.

If you look at my table from top to bottom, how different VALUES affect the absorption - for example, on 1mm thickness - you'll see that the transparency isn't much changing, but just the color becomes more faded/desaturated, dirty, and dark.
Or on 5mm thickness (which is equal to 5mm distance in material) - you can see a blue "halo" around on 255 value, which becomes a strong dark substance on lower values.
Or on 15mm - 255 value has very saturated blues.

But most of all - please compare value 255 on 15mm, and value 10 on 1mm - they're very different.

So the VALUE != doesn't equal DISTANCE. It's another control, which lives on it's own, and does the similar, but completely different thing.




It's like the CoronaLights - you can set a color there, and there is value also - you can set a light to 10,2,8 RGB - muddy pink. But why? You have the intensity to control the brightness of the light. But what's most important - wise people added the Temperature to Lights - when you can control just the HUE, because you already have the Intensity control.

Imagine that you should set the Light color, light color brightness (which works completely different than Intensity), and light Intensity again - isn't this "artistic hell"?

But in material absorption - there is no Temperature, HUE, or whatever. There is just color. And you should pick this color brightness by eye, and pick the distance by eye again (which should control a similar thing, but a bit different). And that drives my mind crazy.

[romullus]

Maybe it helps if you look at this graph this way? To me it seems perfectly normal.

[juang3d]

@vicnaum

What is your usual render engine?

How does this works with that engine?
Can you post the same test?

I say that because I don't understand your way of thinking about colors as three separated values, I think this is an overcomplicated way of thinking but maybe is useful in some scenarios, so I would like to see what are you used to see in your other render engine :)

Cheers!

[vicnaum]

Okay, let me give a more picturesque example:


You see, how VALUE works as saturation?
(and also reminding vray Bias control)

[vicnaum]

@vicnaum
What is your usual render engine?

How does this works with that engine?
Can you post the same test?

I say that because I don't understand your way of thinking about colors as three separated values, I think this is an overcomplicated way of thinking but maybe is useful in some scenarios, so I would like to see what are you used to see in your other render engine :)

My usual render is Corona. Before I used vray (a couple of years ago).

Let me explain that way of thinking. Just answer this question:
When you set up the CoronaLights colors, do you always use BRIGHT COLORS, or maybe sometimes you use DARK COLORS too?
In the CoronaLight color picker.

[juang3d]

I use little variations of white for light, unless I want a color saturated effect, and I tend to reach a level between color and multiplier.

I think I see what you say in your last example, the ineer of the object tends to be red in that case, or an over saturated color, but I think this is due to the energy loss of the light rays and a concentration of them on the inside of the object, but what I just said may be a complete stupidity, I think Keymaster should chime in here, this is over my knowledge of light physics hehe

Cheers!

[vicnaum]

Yeah, I'm trying to say exactly that. We now have a parameter called VALUE which acts as a saturation, and gives dull saturation on low values, and oversaturated colors on high values.

Btw, SATURATION effect is even funnier:


It does a HUE shift! Lowering the SATURATION does a hue shift towards base color (in this case - red), removing the yellow parts.

Will be funny, if the HUE control will animate the launch of statue to the moon :)

P.S. Vray isn't better - it's even worse. If a color is having 255 VALUE in Vray - it will NEVER turn into black, no matter how thick is the object, or how much the fog multiplier is - it's just becomes pure solid red (e.g. in this example). And also vray has this "magic number" fog multiplier. I really like corona having a distance. But I'd like it to have more physically correct and intuitive absorption system of controls.

[Ondra]

well now we are getting to the base of the problem - volumetrics just work differently than surfaces, and as a consequence, are harder to figure out. Because the light is not multiplied by your color, it is exponentiated by it. Which gets weird fast.