[stiras]

Hi

We just recently started using maxwell here at the office. So the problem I'm addressing might just be us not knowing what we are doing.

We seem to get these really sharp pixels in our renders that doesn't even out over time either. So it must be some kinda settings we have blew to high.



We have tried to turn down down light values, changed scene size since maxwell relay on real world scale.



We have also tried changing on the material. There is were we got our first break trough. Seems to only happen on reflective materials. So we tried to play around with it for some time.
But playing to much with it also ruins the look we were after.

Materials on the cube:









We assume there must be some rules when setting up the materials that we aren't aware off that makes these things blow up like this.

The computers we work on are speced with Nvidia Quadro 4000 cards 32Gig Ram and 6core xeon processors. Windows 7.
In software we use 3ds max 2012 (servicepacks always up to date). We also use the latest Maxwell renderer and 3ds max maxwell plugin.






Even if you don't know for sure what the problem is, we will love to hear your thoughts, who knows maybe its something we haven't tried out and might even work.

PS: If needed we can share the BlueBox 3ds max scene. But doesn't look like this forum support uploads.

Thanks for taking your time.

-Mats

[Half Life]

Using an "additive layer" over a coating will often give these types of errors -- so I would start with either ditching the coating or the additive layer in your material.

There are a number of other issues here with the material setup -- and I have to admit I'm not clear exactly what it is you are wanting to achieve... could you clarify what real world material you want to emulate?

Best,
Jason.

[stiras]

Hi half life thanks for answering.

I see. Thats a shame tho, i really liked coating, was a nice way to add a layer of reflectivity on top. But i guess i try removing it for now
When should we use additive layer if its a risk of this artifacts to happen?

Well we are not totally sure what the real world material would be like, but we believe it is some kinda frosted glass/ice with some sort of glow inside.



Here is the consept image.


- Mats

[Half Life]

OK I see, well first off lets try to find what the settings you have are actually doing -- I'll try to point out some things to help you create materials in the future.

Starting with the base layer here -- you do not have any scattering coefficient -- so there is no point is having a setting for scattering color or asymmetry... it just becomes potentially confusing as to what is actually contributing to the "look".

The same thing goes for the angle setting of 35 -- there is no anisotropy so the angle setting has no effect.

I also do not see any texture map associated with bump here so the setting of 100 has no effect.

What you do have is a material with what appears to be mapped transmittance -- which would make it a dielectric. Since it is a dielectric, the K setting should not be used, and an Abbe value of 160 is essentially the same thing as disabling dispersion anyway so it should just be disabled.

You also have this dielectric set to an attenuation of 1m which is abnormally high -- as a point of reference most colored glass would be set to about 3 cm.

Finally having a Nd of 11 with Force Fresnel turned on means you have something much closer to a metal (which is also borne out by using the K value) -- the Nd of glass should be roughly 1.5... if you are looking to create a metal effect I would have to point out there are not many transparent metals.

I won't continue on with the other layers, but they all have similar issues -- so I would instead just recommend that you watch these free videos and hopefully they will clarify what these settings are meant to do:

http://support.nextlimit.com/display/tu ... e+from+VTC

When I get to the studio I will make a material to match your concept and hopefully that will help as well.

Best,
Jason.

[stiras]

Thanks allot for the information half life.
I'm a bit embarrassed, but I guess we all have to be new in the tools the first time:)

I gonna take a closer look at the settings you wrote about.
That DVD looks detailed, seems like it goes trough allot of the maxwell features, if you recommend that DVD, I will make sure we get it stocked here


Again, thanks allot for taking your time. And for the really quick response.


- Mats

[Half Life]

No reason to be embarrassed, I made similar mistakes in the beginning as well -- part of why I made those tutorials was to help other users with the basics of the Maxwell material system which IMO is the most complex part of using Maxwell.

Best,
Jason.

[stiras]

wait, your THE Jason? Jason Maranto, the creator of the dvd?
Thats so cool

Started looking at the free parts now. Trying to consume as much info as I can.
And its great that people like you actually spend time to create something like that for the community. Hope i can do something like that myself in the future
Maybe not in maxwell but were i got my strengths.


Thanks again. And I make sure I post a updated render when I got the material right.


Cheers
- Mats

[Half Life]

Well, I'm pretty sure nobody has ever referred to me as THE Jason Maranto before -- but yes, that's me in the videos

I really like Maxwell so I tend to talk about it quite a bit -- I was actually banned for a short time from another forum recently because they mistakenly assumed I worked for Next Limit and was attempting to do some free advertising for them

When I find something cool I tend to just talk about it alot, sometimes I get a little static for talking about Substance Designer so much here (in off topic)... but like Maxwell I just thinks it's a really cool program. It's nice that I get paid to talk about things I already talk about anyway...

I look forward to your future contributions

Best,
Jason.

[Half Life]

Here's a Maxwell test scene that has an MXM within it that you can use to get an idea of how I might approach the problem as shown in your concept art: http://dl.dropbox.com/u/41250852/Cube%20Tester.zip

I've never had the need to do glowing blue ice/glass before so it was an interesting challenge.

cubetest.jpg

cubetest2.jpg

Obviously you can tweak the values and maps to whatever suits your tastes/needs better -- this was simply meant to give you a starting point for setting up a material. Please note that the material is set up for the cube to be 10 cm -- when dealing with dielectrics and SSS materials you want to be aware of the thickness of the material. Since this is an SSS Material it will most likely need to render to at least SL 20 to render clear (the sample is rendered to SL 16) however it should be much more predictable/controllable.

I've also quickly made a Substance that will facilitate experimentation to find the exact right colors/ gradient falloff for the mapping -- this Substance will allow you to adjust:

1. The Center Color
2. The Edge Color
3. The Gradient Falloff
4. The Surface Texture Amount
5. The randomization of the surface texture
6. The size and format of the output images

You will need to use the free Substance Player found here to use this: http://www.allegorithmic.com/products/player -- although if you are using a newer version of 3DS Max or Maya Substances are supported directly.

cubetest.zip

Hopefully this gives you a good starting point for a new direction on this material

Best,
Jason.

[stiras]

Haha well now you have THE Jason Maranto!
funny how I read your text with your video voice in my head now.

I know exactly what you mean, i have bin rambling around here on the office and with friends about maxwell myself, i kinda love and hate it right now, love it because its cool and have alot of nice features, and hate it because its a new tool and i suck at it. Haven't got myself banned from anny forums yet tho nice job!


Glad you found the material challenging I cant see your attachments tho

"You do not have the required permissions to view the files attached to this post."

SL20... hmmm this is gonna be used in a animation with a bunch of boxes. so might take some time. I just hope we ain't gonna hit the wall to hard on this.
Guess we need to buy another stack of node licenses then hehe

We run max 2012 and latest service packs. so i think we might have substance. Never used it tho, so gonna be a bit interesting that also.

Cheers.
- Mats

[stiras]

Happy easter everyone
Back from holiday and time to get the feet's back in to maxwell.

Bin working with the example Jason was so kind to share.

and here is what we got after some tweaks and talking with the concept guy.


The image is rendered aprox 1h and a quick comp in nuke.

the grain doesn't bother us that much in this still image, we are more concerned about how that will turn out in an animation.

so we did a short animation render to get a better idea.
http://dl.dropbox.com/u/57681365/AnimTe ... 1_0340.mp4

As you probably can see. 1h ain't sufficient enough to get a clean result.
So my question now, is it a bit to ambitious to think, that we can optimize this enough, so we can get a crisp result at 1h rendering each frame?

Scene changes:

The boxes is now a box primitive with chamfered edges and a shell modifier. Seems like the sphere with the emitter inside the box went crazy if we dint make a shell on the box.

We also did some changes on the material. We sorta fell in love with the sss, because of the bleeding effect it gave us from the light inn side the box. But from experience with other sss shaders and render engines. Sss tends to increase render time. I assume maxwell is no exception?



We also made the decision to added a coat again, because of the nice highlights it gives.

The environment hdri is also changed. Since the scene is setup with studio planes around the room we needed to get the most interesting parts of the hdri image to emit from the top down inn to the scene. So we created a sphere around the whole thing and applied a maxwell shader with emitter controlled by the HDRI. Is it a big no no to do it that way? Because on the sphere we could get a visual appearance of the hdri, so it was easy to get the part of the hdri we wanted looking the way we wanted. Something we dint get from the maxwell render-setup.



So theres is the update on how its going with us and maxwell


Cheers

- Mats

[Half Life]

That looks much more interesting to me now as well -- you are correct in that SSS may be one of the most render-time intensive materials types in Maxwell... although it is much more viable now in 2.6.10 than it has been in the past.

Definitely having an emitter inside the SSS object could be a big issue(which is why I put the glow on the surface)... This is somewhat easier to do with ThinSSS (see this thread for an example: http://www.maxwellrender.com/forum/view ... 48&t=38427 ) but is still less than ideal.

Render time optimizing will be dependent on more than just the materials -- depth of field(out of focus) effects and motion blur could also contribute to longer render times.

I'll look more at this when I get to the Studio... but I'm sure there are others here who can help more with the animation specific considerations.

Best,
Jason.

[Half Life]

OK, there are a few material issues I found by testing your settings over here:

- K (extinction coefficient) is a parameter that is meant only for metals, and if I enable it within Maxwell Studio (as per your settings) it disables transmittance, which in turn disables any SSS effect.

- Also your scattering color should not be a value greater than 225 to avoid excessive noise.

Best,
Jason.

[stiras]

Hehe so we have bin tricking our self beliving we had sss, but its accualy just a fake because we cancle the true sss with the k value
Well that's a good thing to know.

About the motion blur and depth of field, we usually do all those things in comp. So the motion blur is off, and as far as i know you cant turn of depth of field, except increasing the f-stop am i correct?

I have never bin a photograph type of guy so a bit tricky getting the correct camera settings.

We also turned off dispersion in the latest renders. Seems to be another thing we manage to turn on that we dint really need.
http://support.nextlimit.com/display/ma ... Properties


Cheers,
-Mats.

[Half Life]

Well, I can't say for sure since I don't have access to your plugin -- so it could be that the way it operates allows for K and transmittance/SSS to coexist... it's just that I cannot reproduce the material here in Maxwell Studio 2.6.10.

For the record I think the render looks good already -- just time consuming to complete an animation... increasing the F-stop will change the look but may also shave off some render time, so its a balancing act. Just be sure to lock exposure to EV before making F-stop adjustments, so you can keep things consistent.

Dispersion is definitely an effect to avoid in your situation because it makes the noise much more noticeable, being that the noise becomes multicolored rather than monochromatic.

Best,
Jason.