The thanks goes to Jeff Patton for creating the scenes and to Jared Martin and Harry Bardak for translating them to Maya and XSI. Just wanted to make sure these guys get their PROPS!

Update: Caustics is disabled by default because more people use transparent shadows than caustics. Everything is configured, all you have to do is enable caustics and that’s all. No need to adjust photons counts, size, etc., it’s already configured.

Thanks, it is good to have end users with fresh eyes comment, we knew it would happen. I am leery to comment on Jeff's scenes just like I would be leery to provide a JPG with gamma settings since he explained options in the included PDF file "3dsMax gamma.pdf". Hopefully, your comments will help us make the documentation more clear.

It depends on your work flow, the Max 9 file it was important to point out the gamma at 1.0 otherwise the scene output would be all wrong. We actually almost didn't realease the Max 2008-9 version because there are so many differences in output in the newer versions, same scene renders differently, and we knew these kind of questions would come up. In the end we figured Jeff had gone to the trouble to set up the Max 2008-9 scene, we may as well use it and realize we would probably get all sorts of question. We have outlined several methods of setting up gamma and included rendered out images so that you can visually compare and make sure your settings are right, we can't control all the changes Autodesk makes and it comes down to asking do we try to keep the settings as consistantly as possible between the versions and let these changes show up or do we fake fixes so that the renderings match?

It really doesn't help IMHO to fake the output, the purpose of these MatLabs is to provide a real world kind of balanced setting to set up materials, of course you may need to tweak the materials in your actual scenes, lighting conditions will very, this is just an averaged environment so to speak to represent as best as we can an average of typical condtions. You will see people saying that the scene should be optimized to render the fastest possible and my opinion and I believe Jeff's would be that the setting should match a more typical scene than just a test lab, the materials should be close in the MatLab as they would in a real scene, optimizing the settings for this little scene doesn't reflect real world conditions.

P.S. You can substitute "Jeff Patton" in a lot of the places I said "We", Jeff and I had discussions, but Jeff created the scenes and applied his expertise.

I wanted to follow up, I had removed this from my response above because I thought I better verify before saying I think you are mistaken. I believe your Max 9 scene above has caustics on and the Max 2008-9 does not, I just got done rendering both default scenes and they are very similar and do not display the discrepancies that yours do. I think you may have turned on the caustics in the Max 9 scene without realizing it.

So as you can see there is the differences we discussed about changes in Max, but not the differences your renderings indicated.

Thanks for the follow up, since you actually stated that they both had Caustics off, I figured you checked and rendered out and so I didn't want to disagree until I actually checked myself. I tend to keep from inserting foot in mouth when I do that