Tweet
Hello,
We have been optimizing large complex architectural scenes over the past few months and have noticed a few things about reflection depth that I am sure allot of people probably know already.
That is - if you want to see reflections on objects through glass you need to set the reflection depth of that object to equal the amount of surfaces the ray passes thorugh. So if you have one piece of glass with 2 sides, with say a reflective timber floor behind, your max depth needs to be 3 to see reflection on the timber. If it is 2 pieces of glass it needs to be 5, and so on.
My issue with this is in other circumstances, like a steel mesh, or a complex organic object like a tree behind glass. Setting the max depth value to 3, or 5 causes allot of rays to bounce around self reflecting. This can lead to significant increases in render times.
Would in not be logical for the max depth reflection to understand if the ray has passed through refraction, and then return a reflection back through the refractive object? So in the case of the timber floor, if the max depth is set to 1, you will always see a reflection regardless of how much glass is in front of it? Then the refraction setting is made on the glass mat itself and as long as it equals the amount of faces that it has to pass through the reflection isn't important.
I think this would certainly return a more expected result and would allow better optimization of scenes.
We have been optimizing large complex architectural scenes over the past few months and have noticed a few things about reflection depth that I am sure allot of people probably know already.
That is - if you want to see reflections on objects through glass you need to set the reflection depth of that object to equal the amount of surfaces the ray passes thorugh. So if you have one piece of glass with 2 sides, with say a reflective timber floor behind, your max depth needs to be 3 to see reflection on the timber. If it is 2 pieces of glass it needs to be 5, and so on.
My issue with this is in other circumstances, like a steel mesh, or a complex organic object like a tree behind glass. Setting the max depth value to 3, or 5 causes allot of rays to bounce around self reflecting. This can lead to significant increases in render times.
Would in not be logical for the max depth reflection to understand if the ray has passed through refraction, and then return a reflection back through the refractive object? So in the case of the timber floor, if the max depth is set to 1, you will always see a reflection regardless of how much glass is in front of it? Then the refraction setting is made on the glass mat itself and as long as it equals the amount of faces that it has to pass through the reflection isn't important.
I think this would certainly return a more expected result and would allow better optimization of scenes.
