fresnel reflections, when to use them

[pailhead]

I am really confused with fresnel reflections. I know generally why they appear, and i've been trying to understand the mathematical and the physical aspect of it, but still something escapes me.

Why are we using them on wooden materials like floors, cars, pavement, walls and even metals? Does it have something to do with the coating these surfaces are covered with, so that light rays are actually refracted at the same time or is it something else?

[dimo]

Hello,

They should be used for the clear coat these materials have. For example the car paint has a clearcoat layer above the metallic layer. The clearcoat layer follows the rules of a fresnel reflection.
It's important to use linear workflow with the fresnel effect to get the right kind of effect.

Best regards,

Dieter

[Clifton Santiago]

I use fresnel reflection on almost everything (except mirror materials). I'm no physicist, but it seems to me that most materials even with low amounts of reflection (such as pavement) appear to have fresnel-like fallfoff properties, as in they get more reflective as they approach the horizon, i.e. as they become more parallel to the field of vision. I usually set the appropriate glossy value, set reflection to white, click fresnel, unlock the IOR, and control the reflection strength via IOR value.

[percydaman]

nearly all materials in life follow the fresnel rule of reflection.

[pailhead]

But why exactly? I remember a lesson i learned in elementary school, and the given water surface example. If you look straight at it, you'll see through, but if you go below a certain angle, it will act as a mirror. Was that the fresnel law?


Where are rays refracted on a pavement surface, in order for a fresnel effect to appear?

[glyph]

dont confuse reflection and refraction - a pavement does not refract the light

its the reflections that are behaving in a fresnel manner - ie stronger at glancing angles

[joconnell]

Apparently it does actually refract the light (seems to be called transmission in physics papers) so the part of the light that doesn't reflect back is being absorbed / transmitted into the surface. Here's a wee thingie that describes the effect and in the third paragraph of this page it mentions how when light hits another medium it will reflect a certain part and transmit the rest into the new medium so it's killing off part of the light we can see.

Pretty much everything in the world is reflective to some extent and one of the large contributing factors is how polished the surface is. If it's got loads of tiny pits and bumps, the reflection itself gets broken up and less focused which is what the vray glosiness effect is trying to replicate. A perfectly flat surface such as still water or a mirror has no tiny pits and bumps so the reflection will appear perfect. The amount of reflection that occurs is based on the ior of the surface. That said since we're only making pretty pictures go for what looks good. i'll normally use around 1.333 for water, 1.5 for glass and then 2/3 upwards for more reflective surfaces or in the case where I have really good reference images or something to match, I'll do it by eye until it looks about the same.

[glyph]

surely the refractive index governs how much the object acts like a lens

its the "reflectance index" which governs the reflection falloff

both often behave in a manner where the falloff fits Mr Fresnel's equation (whatever the equation actually is) but the two items are different

surely thats why there's two places for the two different IOR's -
though I guess there's a relationship between the two IOR's for transparent materials

I could be wrong ...!

[VKNT]

There's an interesting article on fresnel here.

http://www.vray.info/topics/t0039.asp

[andrewjohn81]

hey joconnell, or anyone else for that matter

Any idea why we have to use such physically incorrect IOR's to get physically correct looking images? I mean, chrome has an IOR of just under 3 right? Most images I've seen, and even the shaders on vray materials site, need something more like 12 to look at all like chrome.

It's reasons like that I never fully understood having what seems to be a fixed curve for fresnel. Shouldn't the curve end up evening out a bit at higher numbers so there is less difference in the reflection values, or is that what it does now and I just can't tell? It's just that using a value of 3 for chrome just doesn't seem like enough.

[Clifton Santiago]

Why is it suposed to be 3? I like 16 for chrome. About 8 for brushed stainless, and about 4 for less reflective metals like aluminum.

[domi]

thankyou - useful thread

[joconnell]

hey joconnell, or anyone else for that matter

Any idea why we have to use such physically incorrect IOR's to get physically correct looking images? I mean, chrome has an IOR of just under 3 right? Most images I've seen, and even the shaders on vray materials site, need something more like 12 to look at all like chrome.

It could be that the lighting isn't physically accurate in that case - the lights mighn't be bright enough to give the right amount of reflection so people turn up the ior. Admittedly though we work in visualisation so whatever looks correct is correct. If you use the correct ior for a material you'd also have to be sure that all of the other factors are correct - this is what makes something like maxwell look so good, that it's lighting, camera and materials all work in real values all the time, so you can rely on an ior table to give you the look you want. Vray is a bit more flexible but not as easy as a result.

To get a reflection on a surface, you can either have a really reflective object and a dark object to reflect, or a less reflective object but a brighter object to reflect in it - you could theoretically achieve the same level of reflection with the two, much like clifton using a really high value for the ior.

[paulison]

surely the refractive index governs how much the object acts like a lens

its the "reflectance index" which governs the reflection falloff

both often behave in a manner where the falloff fits Mr Fresnel's equation (whatever the equation actually is) but the two items are different

surely thats why there's two places for the two different IOR's -
though I guess there's a relationship between the two IOR's for transparent materials

I could be wrong ...!

I wish someone else would address this point as well who knows the answer and can give a response and be 100% confident that it is correct

Personally I think that the IOR value effects both reflection and refraction and should be the same. So when fresnel is ticked for glass, its using the IOR value of 1.6 for refraction but also in the fresnel reflection falloff calculation. Please someone correct me so that this doesnt mislead anyone - theres a fair bit of confusion in this thread!

If we can get a final answer to these questions it would help a lot of people out. For example, in our office a lot of people still use the oldschool max way of using falloff in the reflection slot with fresnel unticked and they don't really know what the IOR 'stuff' is for. Would be nice to set them on the right path with some conclusive answers because these days, to get anyone to follow what you are saying, you almost need to prove it through examples and a concise explanation otherwise they ignore you!

[paulison]

just had a quick look on the vray help page:

http://www.spot3d.com/vray/help/150R...ial_params.htm

"Fresnel reflections - checking this option makes the reflection strength dependent on the viewing angle of the surface. Some materials in nature (glass etc) reflect light in this manner. Note that the Fresnel effect depends on the index of refraction as well."

I guess that helps clear up my question to myself anyway