accurate value (n and k) for Complex Fresnel shader

[hammerbchen]

Hi,

I been using the Complex Fresnel shader for a while until recently I start to change the n and k value other than the default from here:
https://docs.chaosgroup.com/display/...Fresnel+shader

I assume the default value for this "complex_ior.zip" is for gold material. After checking http://refractiveindex.info/ I start to get confused. There seems no "standard" value for different materials? For example copper or gold, if I choose different shelf or Page I get different n k values.

Even with Siger's free plugin, those preset values are different from http://refractiveindex.info/
https://www.sigerstudio.eu/shop/plug...omplexfresnel/

Which value I should trust? Or maybe I misunderstand something?

Rgds,

Chen

[joconnell]

Nope, you're not misunderstanding anything.

The problem with scanned values and the best thing about them is they're only correct for the one gold object you've sampled. If you scan a slightly different gold object, it might be a slightly different gold and give you slightly different numbers. If you have a look at all the gold objects in your house, are all of them the very same shade of yellow or are some slightly more red or slightly more green? Say for example you were doing iron man and you had to match a physical suit on set,, if you had a physical sample of the suit which you could scan, you'd have a 100% perfect match. For gold though, there's loads of different types of golds in the world so there's no guarantee that the numbers you get from refractiveindex.info are the correct numbers to match the gold that you're trying to achieve!

[hammerbchen]

tks much!

I assume those variance still within reasonable range. Gold still looks like gold.
Or maybe 18k, 24k... kind of differences.

[joconnell]

Yeah - all of the values you're reading have been taken from actual gold samples so they're all valid values for gold, it's just whether you can get values from a similar gold to what you're trying to make. the only difference in the osl shaders is some will allow you to do simple colour corrections on the curve afterwards to change the hue easily, rather than trying to mess with the n and k values which isn't the most intuitive!

[^Lele^]

tks much!

I assume those variance still within reasonable range. Gold still looks like gold.
Or maybe 18k, 24k... kind of differences.

https://en.wikipedia.org/wiki/Fineness#Carat

Nothing's ever easy...
Edit: well, then there's the VRscans...