Max9 mental ray shader thread?


#21

exellent idea Jeff, this could be a breakthrough for mental ray users!
you have my support.


#22

Here’s the current Max9 scene:
http://jeffpatton.net/Materials/MR-matlab.zip

I’m talking with several people & companies about how (and where) this should be hosted to best serve the community. So please don’t just take the file and start generating materials to post in this particular thread. I’m posting it to get your thoughts as to any changes that could be made that would make the file better, easier, etc…

Also interested in all the ideas on how this should be organized, etc… This is something that should benefit us all, so let’s keep the creative juices going and get this setup just the way we want it.

If you feel this particular test object/file is just completely wrong…feel free to share your concerns/thoughts also…it won’t hurt my feelings.

Caustics…ehh, I suppose we could set this up for caustics but it would require more work to change the light type/focus settings. If there is a demand for photon caustics in this scene, then we need to address it now. My honest thoughts…I purposely avoided photons (of any kind) in this scene. Why? Well, we can easily setup the scene for caustics/photons in general and it would work fine. But when people take that into their own scenes…chances are it’s going to look/work completely different. There are just way too many variables to consider (light types, settings, focus, decay, etc).

Just my opinion of course, nothing more, nothing less.


#23

Jeff, great idea mate!

Personally l would like cm not inches, but thats just because where l am located. Not a big deal though, easy to change . I like the steps too. In fact the whole thing you’ve designed shows everything you would need to know about the material.
I hope l’ll be able to add a few materials too. (I just need my max 9 to arrive! They are always late here in Japan…)

keep up the good work.
cheers


#24

At last!
Well I am going to compare with my own test scene and see if materials look the
same way in both or not, also rendertimes, also number of characteristics tested.
I personally think the scene must have GI and FG on and light must be near real
so I want see what type of lights were used.
I post my opinion about all of this in two days.

Thanks Jeff for take your time on this, I really want this also. I would like a scene all people
think is the better one so I want people doing their modifications and saying why they did them. No modification to the test scene or test object because it looks pretty: all modifications must be because a problem is addressed better that way and you must say what it was.


#25

If we make a version for Max6-8 then maybe photons can be introduced into those scenes. But for this Max9 scene, I say no photons. With Max9 most people will only use FG. It’s faster/easier to setup than dealing with photons for the majority of users…and I really don’t want to stray from that.

Maybe that’s the way to go, Max6-8 scene will have photons setup (caustic & GI). Then if someone is using Max9 and they simply must have photons enabled for their material test (probably caustics)…then they can use the Max6-8 file.


#26

Well, I actually wanted to test photons more for caustics, and less for GI. But thinking more of it, caustics depend too much on scene settings…
Maybe if someone wants o test their caustics for a material, they should do it in their modified test scene:shrug:


#27

It is easy with scene states and mentalray presets. I will post the scene file modified. No testing caustics or color bleed seems an error to me.


#28

Hey All,

I understand why a material may need to be properly tested for
its refractive properties and that caustics are certainly desirable
for product and Arch Viz work, but wouldn’t it be best to get any
material in the ball park using the generic ‘All purpose’ test scene
before getting into finer details that can be tweaked later? Certainly
that is the pipeline we always use for working up materials.

Maybe material creation and testing can go through two levels, basic
diffuse/relective/refractive setup and then a more advanced level
that puts the material through its ‘physical’ properties workout.

Just thoughts

Regards
Bri


#29

Changed the ground plane texture to the smaller, lighter stuff in the bottom right corner…MUCH better, thanks.

Now, another issue that I can really use some advice on…FREAKIN’ mapping coordinates. How should I map this thing where I have a soft edge? (see attachment). Should I look into pelt mapping this thing? I just want it to look uniform and work properly (which the mapping scale I have setup now is completely wrong…)

So I would greatly appreciate any tips/help on the mapping issues here.

Thanks.


#30

R u using box mapping?
I think using UVW Box mapping and equal settings for length,width and height in uvw mapping , can solve the problem.


#31

If you are mapping it then you are saying the object is finished. And I really don’t think so, I want try modify some things in it to add more test zones. So please wait and look at my modifications and only then we can start to think in mapping it like we want. I hope do them today at night when back in home. The mapping after could be done by people and we choose the best with the less stretching.


#32

I’ve tried all the mapping types. Box mapping (with equal lengths) works fairly well, but it produces a nasty seam along the top edges of the object. I’ve tried moving the uvw box gizmo and it just won’t go away.

What I ended up doing was cutting the object into smaller parts, map those, then reassemble and weld the thing back together. But no matter what options I use, there’s always a seam or stretching somewhere that I have to deal with.

I’m asking about the mapping coordinates because I want to know how to solve it. Good gosh man, cut me some slack here. Am I not allowed to ask questions to learn something… :argh:

I plan on looking at your changes to see what you’ve come up with. Although I will say now that I’m still not keen on adding photons to the Max9 scene, but with the scene states (good thinking), it might not be an issue. But again, I just don’t think the majority of Max9 mental ray users will be adding caustics in their day to day workflow. Unless I’ve completely underrated the use of caustics photons these days? I figure only more advanced people will “tinker” with the photon side of things and as such…they will (or should) already know how to setup a photon scene. Obviously, my thoughts on this scene are similar to what Brian mentioned a few posts back.

Just so I’m perfectly clear here, again…I purposely left out the caustic and GI photons. That’s not an “error” on my part. And Final Gather alone does infact produce color bleed. Not as rampant as when one uses photons, but then again I know quite a few people that wish there was no color bleed at all. Aahhhh, those waky arch-viz people and their plot to kill color bleeding :slight_smile:


#33

First, pardon if I sound rude but I didn’t learnt english (I am self-taught in everything and english of course too) and perhaps I lacks of more grammatic. I am seeing you mapping and you know it is near impossible (some tricks available but also time wasting) fix a mapping if you are modifying the mesh. I think the mesh of the test object has some lacks and I was worry you map it because it is time waste.

When we know the mesh is “perfect” and nobody seems see any enhance in it then I think
is time to everybody try map it the best they do. I myself will create a video of the mapping process showing how I did. I like a continuous mesh with the less stretching possible. If we need place a seam we have the back side of the mesh object to use.
Today back in home I hope made the changes to the mesh and submit to your opinion, yesterday I was tuning the caustics and photon thing and it is quicker I was expecting, but I have today to test more.
If my modifications don’t please tell the truth because I really like the way the object is now, only need in my opinion some dark zone and some rebound zone. Today night I upload the modifications and we can comment about. And then start mapping if no one has more suggestions.


#34

Look forward to being able to thrash this out and put
Jeffs idea into action. With a few heads together we
should be able to pull out a good mapping solution that
will do justice to the material/shader creations going
onto the test object.

Regards
Bri


#35

Jeff,
Do you happen to have the unsmoothed version? This would definitely make texturing a lot easier.

-Eric


#36

I’m afraid not Eric. I accidentially collapsed the modifier stack on this one. :banghead:


#37

Well, here my six proposals.

And [this was my test scene](http://212.51.33.74/cb/cgtalk/Tres.jpg) but too complicated perhaps. Look the circle area and you
can see what am I saying about the dark zone for test SSS. These renders were
using precalculated photon and finalgather maps so the quality is bad (for the first two materials). The third one is Max9 with one of the default glass in the new Arch&Design material.
My test scene has areas to test: fresnel, SSS, flat mirror, shadows, refraction...

But I think the Jeff Patton scene is lighter and I like it more. The mapping of mine was more

difficult than I think the Jeff Patton’s will be.


#38

I personally like Jeff’s test object. I agree that photon mapping and such should be done on a per scene basis. I mean Mental Ray isn’t Maxwell. I just want a place a can go to download some materials to apply to my scenes. Once I apply them I’m going to tweak them anyway. I don’t think there is a perfect material that doesn’t require some tweaking after the fact. Chrome for instance would be a nice material to start off with, but in order to match into a particular scene, I may have to tint it slightly, or add a slight bumpmap, or dirty it up. At the end of the day the materials are jumping off points, like the templates in the new A&D material.


#39

Nice Stuff,

Looks like some thought and effort went into your
alterations, although the only one that would appear
to help read the material better (IMHO) was adding
the ramp at the side of the stairs.
So my current vote would be to add a ramp to the original
object and then get to the business of material creation.

Bri


#40

The idea of creating a resource of materials for Mental Ray is great and Jeff’s examples are very inspiring. But there are some aspects of the scene that it seems you guys may be overlooking…

Here’s my list of suggestions to consider when setting up the scene:

1- First and foremost: Is the scene being gamma corrected or using any exposure control?

If it’s not, I think it should. First because a correct set up scene will be and for the ones who doesn’t like the idea, be aware that Max 9 is a lot more designed to work in a corrected gamma space. To prove, do a very simple rendering (like a plane with a teapot on it and a photometric light for illumination, just put a very dark color material on the teapot), now turn gamma correction on (set around 1.8 to 2.2, CRT monitors=2.2) and render. Save this image (remember to set the gamma again to the saved image). Now disable gamma correction in preferences, and enable the Logarithmic exposure control and render. Look how close the images are now then they would in prior versions of max. It means that Logarithmic exposure now sort of applies gamma correction too. But the problem with working this way is that the material editor will look darker than in rendering (and if we turn on gamma to affect material editor, the logarithmic exposure will double it in the render), so I still prefer the gamma thing (gamma provides smoother contrast) and a secret, in Max 9 there is a special camera lens shader called “mia_exposure_simple” that is great! It’s a tone mapper and we can use it together with gamma correction to balance the overexposed areas.

CONCLUSION: In Max 9 will be hard not to work in a gamma corrected space with standard tools because if we do not apply the gamma, lights “burn out” very fast, and if we use Logarithmic exposure to avoid that, it will sort of gamma correct anyway. (We could use the camera lens shader as I noted, but I wouldn’t because it’s a wrong procedure to use that without working with proper gamma correction, though a lot of people still use tone mappers that way).

2- Lights shouldn’t generate specular, use a self illuminated plane with same position/scale instead. But it’s very very important to have a correct self-illumination on this plane (proportional to light intensity). A way to do this is to apply an architecural material (yes, the old one) and in the Luminance cd/m² option, we can pick a light and it gets it’s intensity, really handy. For us that will be using those materials, this will give a lot more accurate and consistent results from scene to scene, once they are properly configured. Nowadays it’s very common to use reflection and reflection glossiness (that are the correct way of creating specular) instead of the old specular and glossines of the default materials. But software lights automatically generate these fast (but fake) speculars. They are too bad because as they only react to light sources, a scene mostly lit by indirect illumination, materials will look too flat. Another problem, and this one really matters for the configuration of this scene, is that the specular generated by the light tends to be much more blurred than the actual reflection, so again, in a scene where it’s illumination comes a lot from indirect lighting, these materials will look too shiny, as they only reflect (no fake specular will appear) and the reflection is too sharp. Hope I’m being clear…

NOTE: This must be confirmed by Vormav, Master Zap or anyone that knows what happens. But I wouldn’t use any HDRI, especially on the environment. Because as far as my knowledge goes, glossy reflections, just like area shadows, the farthest an object is to the source, the the blurriest it will appear in reflections and vice-versa. So, how far is an environment map to the source if we cannot get close to it? I have no idea, but the renderer must have a value for it otherwise it would appear completely blurry even with little glossy if treated as infinite distant (am I speaking s****???). Seems to make sense, though.

3- I think that it’s nice to have a grid on the ground, but the text and lines doesn’t need to be completely black IMHO. A medium gray, readable of course, may be enough and probably not so distracting. The idea is not to have a lot of contrast…

4- I still prefer Jeff’s object, it looks really good and provides a bunch of information of the materials.

Well, that’s all I have to say for now :). Hope to help with this project in any way… :thumbsup:

Regards,

Jr.