Index of refraction and colors

[ArgleBargle]

I'm at a bit of a loss understanding why I'm getting the results I am with the index of refraction. I know all the theoretical aspects of it (that is, learned from college physics). What I don't understand is the rather profound deepening of color I see with applying even the smallest index of refraction (1.01).

 

[attachmentid=16239]

 

The two objects in the image have the same color and 95% transparency. The item to the right has 1.01 index of refraction a it gained a deeper-blue color. In another model I made, a small bit of transparency and a bit of refraction produced a profoundly deep-colored plastic that I liked. However, this time around I want to keep a soft color like the block on the left, but also having an index of refraction. I've played with the color numbers after turning on refaction, but could never get a color match.

 

Can anyone give me a clue about this?

[johnl3d]

I think you are seeing the refraction mixing the the color of the cube with the ground color.

 

 

I did a quick render giving the model a slight blue color and got this with a higher refraction

 

 

[attachmentid=16240]

 

Notice the difference when not seen against the ground.

 

 

Just guessing though and turned up lighting

 

changed ground color

 

[attachmentid=16241]

[ArgleBargle]

I think you are seeing the refraction mixing the the color of the cube with the ground color.

 

Well, it's not that. Notice in my example the two cubes are dealing with the same lighting and ground. Using quick-render when editing the model and with models in situ, there is still a profound difference in two objects I really only expect to differ in how they refract, not in how they are colored.

[johnl3d]

Right

they refract differently one picks up the ground color because it has refraction the other does not without refraction mine was the same color

too

[ypoissant]

Since an IOR of one does not make much sense in reality, it is considered a special case and handled differently in A:M to allow the user to add some special visual effects or to render transparent object in an illustrative style.

[robcat2075]

Since an IOR of zero

 

he really meant 1, right? there is no zero IOR in A:M.

 

To eliminate some variables I set all the lights to white. What I find incongruous is that even though the sky and ground are white behind the cube on the left, they don't get the same coloration from the transparency.

 

[attachmentid=16253]

IORZerotest.zip

[C-grid]

I'm at a bit of a loss understanding why I'm getting the results I am with the index of refraction. I know all the theoretical aspects of it (that is, learned from college physics). What I don't understand is the rather profound deepening of color I see with applying even the smallest index of refraction (1.01).

 

[attachmentid=16239]

 

The two objects in the image have the same color and 95% transparency. The item to the right has 1.01 index of refraction a it gained a deeper-blue color. In another model I made, a small bit of transparency and a bit of refraction produced a profoundly deep-colored plastic that I liked. However, this time around I want to keep a soft color like the block on the left, but also having an index of refraction. I've played with the color numbers after turning on refaction, but could never get a color match.

 

Can anyone give me a clue about this?

 

 

Hello ArgleBargle,

 

Did you try a grid instead of a cube?

 

Niels.

 

ps.

Color matching with, just a white light at 100% in the scene will probably give you a better clue, this for gathering 'adjusting information' in other scenes.

[martin]

The cube on the right looks like it's taking "density" into account: the farther a light ray has to travel through a transparent material, the more of the material's color it assumes. Do larger cubes get even darker? That would be a good test.

[robcat2075]

The cube on the right looks like it's taking "density" into account: the farther a light ray has to travel through a transparent material, the more of the material's color it assumes. Do larger cubes get even darker? That would be a good test.

In my example we're looking across the diagonal of the cube so the edges would be the thinnest and the center would be the thickest. The coloration of the cube doesn't indicate that, however.

[C-grid]

Hello ArgleBargle,

 

Just thinking out loud:

 

Is a workaround changing the transparency, since the IOR function builds his solid pixel from own calculation?

 

Niels.

 

ps.

In other words... figure out the transparency differents between IOR 1.0 and 1.01...

[ArgleBargle]

The cube on the right looks like it's taking "density" into account: the farther a light ray has to travel through a transparent material, the more of the material's color it assumes. Do larger cubes get even darker? That would be a good test.

In my example we're looking across the diagonal of the cube so the edges would be the thinnest and the center would be the thickest. The coloration of the cube doesn't indicate that, however.

 

OK, I went back at this with some of the previous comments in mind.

 

[attachmentid=16255]

 

Here are the cubes again. All have the same color and 95% transparency. Cubes 2 and 3 have an IOR of 1.05. Cubes 3 and 4 have a density of 4. It almost seems that density tends to push the color back to the less saturated color of cube 1.

 

My understanding of IOR (as I recall it from college) is that the angle of a beam of light changes as it moves from one transparent medium to another. I had assumed that density might provide the effect I'm seeing in cube 2, but apparently not.

 

I guess I'm no closer to getting cube 1 with just refraction added, not any color saturation change.

 

 

Is a workaround changing the transparency, since the IOR function builds his solid pixel from own calculation?

 

I've actually moved on from the original problem and did a work-around. However, what I'm trying to fix now is my own understanding of how it's supposed to work. I don't think I've yet arrived.

 

I never really had problems with this until now, mainly because I've always been using something whose base color was white. This time around, I was trying to model a glass with some murky water. I didn't quite get what I wanted because I had a bit of color to the water.

[C-grid]

Hello ArgleBargle,

 

Thinking out loud, part 2.

 

Maybe working with a lightlist, keeping the object white and changing the color through a special assigned light is an option (for a next time...(?)).

 

Niels.

 

 

ps.

I think the nolidge you're looking for should be asked at Hash, the unusual big offset between 1.0 and 1.01, no accusation made from this side of what so ever.

[martin]

I think the nolidge you're looking for should be asked at Hash, the unusual big offset between 1.0 and 1.01, no accusation made from this side of what so ever.

This is not surprising. No refraction processing is used if IOR is 1.0, but anything other than that uses refraction code.

 

There's another possibility. There are two ways of calculating transparency: additive and multiplicative. Multiplicative transparency has the disadvantage of making things darker behind the transparent object, but it has the advantage that it produces an image that more closely approximates real life. Additive transparency is definitely what is needed for transparency maps and cookiecut maps, and for single surfaces with transparency on them. It could be that refractive objects use multiplicative transparency, otherwise additive transparency is used.

[ypoissant]

Since an IOR of zero

 

he really meant 1, right? there is no zero IOR in A:M.

Oups. Right. 1.0. Not 0.0. I'll corect that post. Thanks.

[robcat2075]

If you were open to a compositing solution, you could render the refraction (with no color, 100% transparency) as one pass, another pass to create a mask and use those two to composite you r refraction with some tint onto your scene.

 

I've done this is the past when volumetric lights created impossibly bright intersections.

[C-grid]

Thinking out loud, ArgleBargle:

 

It's the offset of the starting color brightness and the recursiveness, that make a darker starting color gets more darkened than a brighter starting color, travelling the same distance within an object.

 

Niels.

[ArgleBargle]

If you were open to a compositing solution, you could render the refraction (with no color, 100% transparency) as one pass, another pass to create a mask and use those two to composite you r refraction with some tint onto your scene.

 

I've done this is the past when volumetric lights created impossibly bright intersections.

 

Ah, now there's a neat idea. Thanks. I'm pretty sure I've got some things where that'll be handy.