mental ray 3.9.3.14 release changes list?

[moulder6]

It seems that max users finally got sum updates of the mental ray core. After installing service pack 2 for max, the mental ray version is now 3.9.3.14. Apparently, the mental ray documentation provided by A'desk isn't updated. Is there publicly available list of changes for mental ray, after version 3.9.1.37 ?

[steve]

Should be online at docs.autodesk.com.

[moulder6]

Thanks, Steve! Hope they're gonna update the release notes anytime soon. I am quite curious... But u know how it goes with A'desk... A lot of things that "should be" are actually "not even nearly"
Btw is this the most recent official release of mental ray, or we're left behind, again?

[steve]

Aah, seems 3.9.1 docs are online but not latest 3.9.3 ! Sorry for being blind. Will try to find out why.

[Remydrh]

Lucky Max users. . .

BTW, try Unified Sampling again in 3.9.3 and see if you notice an increase in quality.

I don't think AD integrates the bleeding edge of any stable mental ray release, but I am somewhat surprised that this release went out of the 3.9.1 branch. Possibly because of the improved iRay. (Which I look forward to playing with now. I think this is 2.0 from the SIGGRAPH announcements.)

[moulder6]

I am playing around right now.
First impression is that FG+IP+IrP still produces lovely artifacts; One thing that's noticeable about Unified is what u mentioned in another thread here:

Samples are not a regular pattern, so one pixel may have 7/16 hit and another might be 9/16, etc. So there is a slight variation. However, there has been edge enhancement made since then.

Let's assume that now fur edges seem to be resolved a bit better

Couldn't care less about iRay integration in active shade... This should have been done with mental ray itself 5 years ago... And despite that probably it could be helpful for a realtime preview of scene/material changes (this is case it's somewhat nearly stable) , iRay as a production render is way to limited for what i do, and it won't change anytime soon.

[Remydrh]

IP+FG interpolation artifacts aren't getting fixed in 3.9 so I wouldn't expect that to be better. It's not exposed so it's not high priority I guess.

Unified for small creases etc. might look better if that was included for the Autodesk version. Release notes might mention that when they are posted.

I like iray for doing a still to see what a scene might look like as a reference. Sometimes reveals a caustic or effect I didn't see before.

[rawalanche]

I cannot help the feeling that iRay performance is now worse than before SP2.

[moulder6]

Ghhhhhhhh... no update to the docs, yet... Let's see if there will be any when they release the subscription packs tomorrow...

[moulder6]

From ShaneGriffith@AREA

Talking to a product manager who leaves you with impression that i u r asking about nuclear physics someone who's selling goat cheese...

Me: Shane, r u planning to update the documentation? The mental ray core is updated with SP2, but the mental ray release notes @ docs.autodesk.com doesn't reflect the recent changes.

Shane Griffith: we didn't add any new features and the fixes were documented in the SP2 readme.

Me: Sorry, maybe i wasn't clear enough. I wasn't asking about new features, but rather the mental ray specific core fixes/changes that are lsited here:
http://docs.autodesk.com/MENTALRAY/2.../relnotes.html
The last release note published are for mental ray version 3.9.1.37 - the one that max 2012 originally shipped with. With SP2 the mental ray version is now 3.9.3.14, but the documentation isn't updated accordingly, and no, the mental ray core changes aren't documented in the readme.

Let's see if we get any clarity on the matter at all...

Some thought on Iray 2

Shane Griffith: Please note that iray in SP2 converges faster than before. A single iteration may take slightly longer to compute in SP2 than previously, but that is only because it is now computing more per iteration. So essentially, it takes fewer amounts of iterations to come to a noise-less image. Of course, if you simply set a given a mount of iterations and compare the results, than yes, the actual clock time will be higher in SP2, but only because it made an even cleaner image than it would have previously. I hope this clears things up for everyone.