Material Reflection %

[glorybound]

Is there a chart showing how much a material reflects? I have found some IOR charts, but nothing about the % of reflection.

[glorybound]

I am referring to the reflection slot; black being no reflection and 255 being a mirror. I am working on aluminum, so I would expect to find a chart showing how much light it absorbs.

[tzaar]

You know about this site, right?
http://refractiveindex.info/?shelf=m...=Al&page=Rakic

Scroll down to the reflectance calculator. How much light it absorbs/reflects depends completely on the angle of incidence. This is a bit beyond my level of understanding of physics, with all the "s-polarised" and so on. But I use it as a guideline some times. And I wish the site had more materials, but I use it to get a more qualified rough idea instead going on my intuition. Especially if I don't have the material at hand and can play around with it and look at it from different angles.

Also http://en.wikipedia.org/wiki/Aluminium says

A fresh film of aluminium serves as a good reflector (approximately 92%) of visible light and an excellent reflector (as much as 98%) of medium and far infrared radiation.

I guess the keyword here is fresh, because as far as I know the surface of aluminium oxidizes quickly: "Aluminum metal will oxidize to Al2O3 in a fairly short time"
which means what you might be more interested in is something like this: http://refractiveindex.info/?shelf=m...2O4&page=Tropf
which as a pretty different reflectance at a zero degree angle 0,06 compared to 0,91. Seems a bit low?

Not sure what kind of aluminium you are working with, and what the properties of the ones we encounter in our daily lifes have. For example aluminum foil (like for cooking etc) doesn't seem to oxidize and go dull, one side is always super shiny almost mirror like. Which I guess means it's been treated in some way?

[joconnell]

Is there a chart showing how much a material reflects? I have found some IOR charts, but nothing about the % of reflection.

The ior number will automatically set the front facing percentage, like an ior of 1.3 - 1.5, typical of what we'd do a lot is maybe about 4% reflective at the very front. Most things tend to go to 100% reflective at the 90 degree glancing angle.

Here's a few colour swatches which will give you a 0-255 rgb value to use as your 0 degree value, so you'd sample one of the colours, divide it by 255 to get a percentage and then pop whatever the percent was into the vray fake fresnel front facing amount. If you're using a falloff for the material instead, you could pop the colour value in directly.

Charts are kind of a starting point to an extent, likewise the same with scanned materials too. It's all very well and good if you want to 100% replicate the material of the sample you've scanned (so it's great for car companies for example) but if you were to use the scanned / measured values for a metal for example and expect it to perfectly match every single image of that metal you've seen then it won't. You can use the scanned values as kind of an ideal of what a pristine, 100% clean example of that material was, but there'll often be plenty of deviation after the fact.

[glorybound]

Great information (if you are a geek). No, really, thank you! I'll read over it and try to make sense out of it.

[glorybound]

Awesome! Bookmarked.

You know about this site, right?
http://refractiveindex.info/?shelf=m...=Al&page=Rakic

Scroll down to the reflectance calculator. How much light it absorbs/reflects depends completely on the angle of incidence. This is a bit beyond my level of understanding of physics, with all the "s-polarised" and so on. But I use it as a guideline some times. And I wish the site had more materials, but I use it to get a more qualified rough idea instead going on my intuition. Especially if I don't have the material at hand and can play around with it and look at it from different angles.

Also http://en.wikipedia.org/wiki/Aluminium says


I guess the keyword here is fresh, because as far as I know the surface of aluminium oxidizes quickly: "Aluminum metal will oxidize to Al2O3 in a fairly short time"
which means what you might be more interested in is something like this: http://refractiveindex.info/?shelf=m...2O4&page=Tropf
which as a pretty different reflectance at a zero degree angle 0,06 compared to 0,91. Seems a bit low?

Not sure what kind of aluminium you are working with, and what the properties of the ones we encounter in our daily lifes have. For example aluminum foil (like for cooking etc) doesn't seem to oxidize and go dull, one side is always super shiny almost mirror like. Which I guess means it's been treated in some way?

[Deflaminis]

I did some experiments a while back recreating the reflectance function as my falloff curve with a falloff map in the reflection slot and was moderately happy with the results but they were difficult to perfect because the scale of the reflectance curve is different than the falloff limits.

[viscorbel]

This is damn confusing... sort of... Those pbr charts and measurements for reflectance give an average value of 2~8% reflectance for dielectrics.

After running a bunch of test renders and measuring the reflectance a 0 degrees it seems that the way to achieve those values is by using a fresnel of 1.33~1.8 with the reflection color set to pure white [255;255;255]
For example, 4% front reflectance is achieved with Fresnel 1.5 and pure white in Reflection slot.
Replicating a fresnel curve as a falloff curve with data from refractiveindex site also gives the correct result with the falloff colors set to pure black/pure white (essentially the same thing as pure white reflection map)

So this means no more reflection maps or guesstimated grey values and just using glossiness for specular detail. Is this the way everyone is/should be working now? I know that some game engines use this approach, specular is more or less fixed and you only do the gloss map for detailing. With vray for years and years a reflection map seemed to be the norm. Now these results make me question the old approach.
I'm trying out this pure white reflection plus fresnel value and detailing in gloss map, and it's working out pretty well.

I know that some guys are just taking those pbr measurements and plopping them in the Reflection slot, for example [59;59;59] for glass and set the fresnel to 1.51... But this is clearly wrong as the resulting 0 degree reflectance is way too low, fresnel equation reduces it to something like 0.01

Am I way off track here?

[joconnell]

So!

Right, with the pbr stuff and people using low values in the reflection slot, it's because the pbr materials are hard wired that the value you put in is only used for the 0 degree value. Every pbr material is wired to always go to pure white at the edge regardless of what the zero value is. For the vray material, it gives us more control and the value you put into the reflection slot becomes the value that is used at the 90 degree edge value. You're right that if you put in a value of 50,50,50 and turned on fresnel, you'd get something very week at the front and then it'd raise to 50,50,50 at the 90 degree faces. There's a new map in vray 3.1 called vrayFakeFresnel that gives you something a bit like what the pbr guys do. You'd use it without fresnel turned on and it gives you three values - the reflectivity value for the 0 degree faces, the reflectivity for the 90 degree faces and then a bias to choose how quickly the curve between the two changes. On the pbr side it'd be like leaving the 90 degree value always at 1.0, putting whatever your measured value is in the 0 degree control and then having the bias as your ior.

With the glossiness map, you'll find with the vray material that it's been doing something for years that pbr only does now - as your reflection gets softer, it looks like it gets duller. It's reflecting back the same amount of energy hitting it, but the amount is spread out more so it seems darker. If you have a single type of material on your object that isn't covered in dust or dirt or anything else, setting a single reflection value and then using the glossiness to control the specular amount is quite valid - having variations of greys in the glossy map will give you variety in the brightness of the reflections, similar to what you'd also get by putting your map into the reflection slot.

Where you start having to use maps in the reflection slot more, is if you have a material that's a blend. Say for example you have a highly reflective material but it has some areas of dirt on it. Then you'd need to do something like have a map that's white for the highly reflective bits, and then a more grey value for the dirty areas. Likewise you'd probably have a similar glossiness map where the highly reflective bits are near white and the dirty bits are more of a mid grey.

[viscorbel]

Ah, so it seems that I've got it pretty much right.

The only difference in my approach is that I prefer to use Blend materials when I need to simulate compound shaders. This seems to work better for me, as I can create shaders for multiple materials and just use b&w blend maps to control the overall look (as opposed to modifying multiple maps to make a small change) There's a little performance hit compared to doing it in a single layer, but since it's not a game engine and I want maximum quality and ease of adjustment it's something I can live with

[joconnell]

Yep you're bang on. Some lovely person here also wrote a script to bake down layered shaders into a single set of maps too which gave a pretty decent speed increase - might be worth searching for.

[viscorbel]

Ok, thanks!
I think the script you mentioned is this one http://dabarti.com/vfx/vraymtl-maps-baking-script/

Yep you're bang on. Some lovely person here also wrote a script to bake down layered shaders into a single set of maps too which gave a pretty decent speed increase - might be worth searching for.

[joconnell]

That's the one!

[Macker]

I always have pure white reflection and refraction for glass - any absorption I deal with using the fog colour, which create far more realistic glass than any greyscale values. I simply adjust the IOR accordingly to replicate more/less reflective types of glass.

As for metals, I set the diffuse to pure black, the reflection to pure white and adjust the IOR until it looks right.

[3dcrgr]

I always have pure white reflection and refraction for glass - any absorption I deal with using the fog colour, which create far more realistic glass than any greyscale values. I simply adjust the IOR accordingly to replicate more/less reflective types of glass.

As for metals, I set the diffuse to pure black, the reflection to pure white and adjust the IOR until it looks right.

But then the metal is all black. How do you adjust the color?