[dmeyer]

dmeyer did u see the presentation? If so mind sharing what u saw?!

Not sure how much we are to share?

I will say that the volumetric examples looked great.

[numerobis]

Not sure how much we are to share?

Why? I thought this was a public presentation. No?
Or did you have to sign a NDA?

Ok, now you can post your video... please!

[krotala]

Volumetrics?? Realflow??? Seriously????....

All that stuff is borderline cool but I think mostly everybody is waiting on the speed improvements.....that is the one feature that Maxwell desperately needs!!!

[dmeyer]

Not sure how much we are to share?

Why? I thought this was a public presentation. No?
Or did you have to sign a NDA?

Ok, now you can post your video... please!

Haha. I didn't take any video or notes. Juan showed a short presentation of only some of the new features.

From memory:

Speed improvements
Volumetrics
Dual sided materials
GPU accelerated multilight/multilens modifications in realtime
Procedural textures
Open exr 2 and deep support
Rendertime Boolean and subd
Scattering / instance tools

Ill try to remember some more. He showed a lot very quickly and it was only a sampler.

They've been busy.

[sandykoufax]

Procedural textures? Wow!

[numerobis]

Wow! this is a nice list - some really interesting and long-awaited features!
Thanks for sharing!

But... no word about the well known engine problems?

And i really hope the two sided materials include a fixed ThinSSS...

[Fernando Tella]

Wow! Nice list already.

deep support

What's that? An output format?

Rendertime Boolean

I guess that could solve the problem with realflow meshes in glass, right?

[Aniki]

please, pretty please show some images I couldnt go to Siggraph, although some stuff coming from me will be presented.

especially looking forward to the volumetrics

[contact7]

Volumetrics?? Realflow??? Seriously????....

All that stuff is borderline cool but I think mostly everybody is waiting on the speed improvements.....that is the one feature that Maxwell desperately needs!!!

Couldn't agree more.
Speed it up, and people will come.

[choo-chee]

I think that the amount of speed improvement will be a major factor in the decision where to invest money - faster hardware or upgraded software.

[moreschipaolo]

I think that the amount of speed improvement will be a major factor in the decision where to invest money - faster hardware or upgraded software.

I agree, I have already 5xDualXeon X5650 ... I do not want to increase my render farm

[choo-chee]

I think that the amount of speed improvement will be a major factor in the decision where to invest money - faster hardware or upgraded software.

I agree, I have already 5xDualXeon X5650 ... I do not want to increase my render farm

... the remark I get the most is "please get rid of the noise in the image"....

[Rickyx]

I think they are preparing the top, long awaited, edge bleeding feature.

The "Do NICE render" button.

[itsallgoode9]

Wow! Nice list already.

deep support

What's that? An output format?

Rendertime Boolean

I guess that could solve the problem with realflow meshes in glass, right?

i didn't think about that glass thing, but that would be amazing if it worked! finally we could do flowing water and condensation correctly!!

The Deep Support thing is to do with masking and have the alphas work correctly to give you a perfect mask without any hairline pixel edges. It calculates the masks using more than just RGBA channels somehow. I'm hoping this will make object and material id much more effective.

I'm not sure exactly how it's implemented nor how in depth, but here's an overview of a deep compositing system
http://www.deepimg.com/

[Fernando Tella]

Wow! Nice list already.

deep support

What's that? An output format?

Rendertime Boolean

I guess that could solve the problem with realflow meshes in glass, right?

i didn't think about that glass thing, but that would be amazing if it worked! finally we could do flowing water and condensation correctly!!

The Deep Support thing is to do with masking and have the alphas work correctly to give you a perfect mask without any hairline pixel edges. It calculates the masks using more than just RGBA channels somehow. I'm hoping this will make object and material id much more effective.

I'm not sure exactly how it's implemented nor how in depth, but here's an overview of a deep compositing system
http://www.deepimg.com/

Thanks!