[podro]

[rickyinmotion]

beautiful set podro.
Can you tell us more about render times and SL
Thanks in advance

[tom]

Fantastic!

[Michael Betke]

Awesome! I want to take a seat.

[Maximus3D]

That's a beauty podro good work!

/ Max

[Bubbaloo]

Cool!

[rusteberg]

man podro! that's nice! here's my recent test... (saved at lower bit depth)

[Fernando Tella]

Wow!

[RK_art]

I'm very sorry, but I can't share all this 'V2 is sooo quick' statements.

From my tests with my setups, the speedup compared with 1.7 is not a hammer.

When rendering opaque materials, ok, that's not so bad:



7:33 min.

But when it comes to SSS, V2 is the almost the same bloody slow engine as it ever was before:



36:37 min,

and



1h 6 min, and only one SSS-Layer.

Both renderings have only a hdri, rendered on XP 64 bit SP2 with an Intel Quadcore 2,66 GHz


And here my skinshader-setup with V2, rendered on the same system in > 9 hrs:



Not better or worser than it was with 1.7, and definetely not much quicker.
And these renders used only one SSS-Layer.
When you instead use complex, multi- SSS-layered Materials, V2 breaks totally in, a snail is quicker.

Sorry, but I am really not satisfied with V2 regarding SSS, and not only concerning the slow render-speed.
Also the missing opportunity to use the SSS-presets when creating a SSS-Layer in an existing Material-Setup is a huge drawback to 1.7.
When you instead use this f****ng SSS-'Wizard', you loose your previously created Material-setup.

V2 might be an improvement for architectural needs, but for organical objects like human Skin or complex SSS, V2 has almost no significant speed-increasement compared to 1.7, but some major drawbacks in the SSS-material-system.

[3dtrialpractice]

This another a sss test. Caustic from a caustic inside a SSS volume.

I too am trying to do v1.7 style fog render.. but not nearly same results as mr renato gets.. I even followed his balls in cube example posted on a 1.7 thread:http://maxwellrender.com/forum/viewtopi ... c&start=15


These renders below are the test of the balls in cube..40cm cube w/ 5x1cm light
exept for simplicity no dielectic balls..just ment to test SSS fog look..but really no joy
also found that maxwel engin crashes with nd value of 1.001-1.162 and sometimes at nd 1.000 is any sss is on (i posted in bugs section)
and Belive it or not they are all rendered to 10 samples.. and the cleanes one is the one with the skydone on.. so I guess.. to not get alot of noise skydome reduces the noise???
in the last three examples the nd was acctally set to 1.000 w/o engine crash.. its weird..sometimes it crashes at nd1.000 sometimes no.. but it ALWAYS crashes with nd between 1.001 and 1.162.. and so i set nd to 1.165 and it works..as all other pics in above example r set to either 1.163 (0 asymetry) or 1.165(neg and positve value assymetry - nd1.163 will crash engine if asymerty isn't 0)

[Maximus3D]

rusteberg: Those renderings look very good! nice job

RK: The human head looks fine to me, but i guess you still dislike the rendertimes eh.. looks like Vray is gonna be your new tool from now on. Too bad, you have good Maxwell skills.

A new test of a old mxm..


/ Max

[rusteberg]

max, did your hans solo esque frozen carbonite machine malfunction when you dipped RK's dude in it?...... looks great!

[Michael Betke]

Threw my testscene from 1.71 into v2:



It needed ages (over 24h) to get to SL 20.5 and is not faster. I also get strange white pixels on furniture on the left when it comes to longer renders.
I assume it's a wrong glass material and I didn't made changes in the materials. Maybe wrong scaling i the scene too. Need to get more experience with the renderer.

I did another small scene and really was impressed by the new speed and shading improvements. Will post the other scene later maybe.

[brodie_geers]

Michael, don't forget though that SL20 in 1.7 is different from 2.0 so you can't really go by SL comparisons or how long it took to get to a certain SL. I think for a fair comparison you'd have to run the render for the same amount of time and compare noise levels. Might still show that it's not much different but at least it's apples to apples.

-Brodie

[RK_art]

Here I have one with 2 SSS-Layers:



around 20 hrs !!! on a XP 64 Quadcore 2,66 Ghz, and still totally grainy.

No, NL's proudly announced Maxwell-V2-hyperspeed is a fake when it comes to SSS.
No noticable improvements compared with 1.7 in rendering organical models.
I am bitterly disappointed in V2.
So much money for almost nothing new.

Max:
No, I will work furthermore with Maxwell.
I bought the upgrade so I will use it for sure.
But I hope all members here that need SSS for their work can kick NL some A... that they will improve the render speed also with SSS and displacement as soon as possible and not sit only on their success they archieved with simple opaque materials. If they do, then I must say that V2 is a big and expensive fake for me, sorry.