Fresnel node

Thomas M.

New member
O.k., Denis. I played around with the node a bit and for air-blablabla surfaces everything works fine.

But if the ray exits the object (second fresnel node, IORs switched), there's no total reflection going on. Compared the results (preview and render output) with a gradient I created by using data from an excel sheet.

In comparisson to the thinner to thicker medium node, the thicker-thinner fesnel node's preview should have a much thicker white area and the transition from 0 degree to the angle when total reflection occurs is much harder. As far as I remember for glass-air total reflection occurs somewhere around 45 or 50°. In my excel sheet that's the border after which the math function puts out an error message. But in reality the amount of reflection is then 100%.

Please try to fix this. Otherwise the node is top notch.


Cheers
Thomas

P.S.: Are there any explanations availabe for your metal and velvet node?
 

Thomas M.

New member
Thought an image might help a bit to make clear what it's supposed to look like. The lower right gradient is my inner reflection gradient. Your lower fresnel node is like the upper one, only the two IORs are exchanged.

Could you please fix this, Denis? Then the node would rock.

Thanks
Thomas
 

Attachments

  • ScreenShot001.gif
    ScreenShot001.gif
    33.4 KB · Views: 535
Last edited:

dpont

Member
Ok, I will study this and try to fix,
please could you post the entire surface preset,
with both examples, this could help me.

Denis.
 

Thomas M.

New member
Hi Denis,

finally managed to save the nodal surface with the most important nodes. Threw away Specular to "clean" it a bit.

For the inner reflection and inner transparency your Fresnel node should work in a way that you just switch the IORs and at least for flint glass the preview should be similar to my Fresnel gradient. Hope it helps a bit.

A problem you might run into is that the Fresnel equation from the total reflection angle on won't work anymore (spits out an error message) as there isn't any more than total reflection.

Would love if there's a way to fix the node in this aspect.

Cheers
Thomas
 

Attachments

  • Glass for Denis.zip
    1.7 KB · Views: 391

dpont

Member
Thanks for sample, so yes it doesn't match your gradient,
and the first question in this thread about reversibility
could have here a (disapointing) reply, honestly I can solve
directly why this Fresnel formulae could fail, but may be
I could try to find a more accurate one.
I Know only that the total reflection
occurs in in some case in a refract function and calls a reflect
function instead, but for a raytrace vector (shader node to refraction shading)
not for transparency scalar output.

I'm afraid, because I'm not a glass specialist.

Denis.
 
Top Bottom