Hehe nice to hear it. I
think Vlado understood the problem eventually, but so far, no fix has been applied.
From a workflow point of view, the 2sided material works nicely. If you look at a leaf, the front and the back side will have completely different properties - the diffuse color and reflection/glossiness will be the most apparent ones. Just setting up each side as a separate material, then dropping them in a 2sided mat to add the translucency is a great idea. With front/back map, you will have to do a lot more tweaking within the material. Also, being able to simply switch between different front and back materials is easy when using the 2sided mat.
The main issue with the example in the thread over at Chaos, is that the backlit side isn't shown correctly on the other side. Also, when setting the material to the "correct" level of translucency, the material is significantly dimmed when lit from the front. Alas, this works exactly the same way in Corona. When setting a map in the Translucency slot, the material front side is dimmed, the higher you set the Fraction amount. This is not what goes on in reality. A thick piece of paper and a thin paper will look exactly the same when looked at on the lit side, but the backside of those two papers will differ a lot.
I tried to find articles explaining the way light is dispersed and reflected through leaves. (I made a thread about that on Chaos aswell haha!) I found this article, which I think is rather interesting, though I haven't set out to understand the maths of it.
http://www.photobiology.info/Jacq_Ustin.html I would imagine a perfect double sided material to work in the following way:
• Each side maintains the diffuse/bump value independently from each other
• Each side maintains the reflection/specular value independently from each other
• A translucency value directs how much of the incoming light that transmits through the object - a fill color should also be added, to simulate e.g. chlorophyll
Also, to look real:
• A fully translucent material should look the same as a non translucent one when lit from the front
• A fully translucent material should never look dimmed or darker than its back side
These points are open to debate and improvements. :)
Now, I totally see the issue with energy conservation here, and since I'm no mathematician/programmer, I can't say how to solve it. As an artist however, I can say when something doesn't look right, and so far, the double sided solutions for vray and corona are not right. ;)
(I wonder if what rendering software lacks is correct absorption and transmittance of light within objects?)