hard surface modeling excercise

[ctbram]

I wanted to practice modeling complex unified hard surface parts.

I saw some parts that someone modeled using Max and was very impressed. I wanted to compare what I could model in Maya to the Max parts I saw so I picked this motorcycle fork brace thingy.

The basic process was to create the main individual shapes and systematically boolean union them together and then clean the geometry up until I had a single unified piece. This is about 10 primary shapes all merged into a single all quad component. I am guessing I spent nearly 20 hours fiddling with it. The boolean method was used for about 80% of the work. The rest was a combination of things like edge extrusion, curve lofts, and such.

It is pretty close but I am not 100% happy with it and will probably model it again. If you look close you can see the bottom part that has the four torx screws does not have the insides modeled yet. I got to this point and since I am considering doing it again I just stopped here. The brake part that attaches to it looks to be an interesting challenge as well and I plan on modeling that and some of the other interesting shapes.

http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_01_ref.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_01_ref.jpg) http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_01.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_01.jpg) http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_01_wf.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_01_wf.jpg) http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_02-1.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_02-1.jpg) http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_02_wf-1.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_02_wf-1.jpg) http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_03.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_03.jpg) http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_03_wf.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_03_wf.jpg)

[TravisCowsill]

Rad dude.

May I ask how you're rendering that to get it so pristine? And is that a regular software material like a blinn? or is it Mia mat?

[daverave]

Looks good Rick any conclusions with maya and max, Im glad to see I make a screw head the same way you do.............dave

[ctbram]

Thanks Dave.

Travis: I just used standard maya blinn materials. I like to use blinn's because they show surface flaws very well. There is no sight wall in this image. I like to use a simple sight wall to get some ground reflection and simulated shadows from the final gather but I was in a hurry.

I think I did a post on fast render settings without lights. But basically I just set quality to production or production fine trace and turn on final gather. In the camera attributes I set the focal length to 85mm (to reduce perspective distortion), and set the environment color to a medium grey.

[David]

Very nice looks almost like Escher has opened up a copy of Maya

Dave :)

[TravisCowsill]

Thanks C. Appreciate the info.

[ctbram]

Dave: No lie! It I saw a photo of that part that had been modeled in Max about a month or two ago and it has been driving be insane trying to think how I would model it. So this weekend I sat down and set out to see how hard it would be and although it is a chore it can be done in Maya.

Daverave: I can't really compare the modeling to Max as I have not really used max since version 4. I have just seen a lot of hard surface stuff created in Max and you don't see a lot of hard surface stuff done in Maya, so I am guessing there must be a reason. To be honest it baffles me a bit because I have watched a couple Max modeling tutorials but never really thought that there was anything that would make hard surface modeling easier in Max. Although I have never seen anything this complicated modeled in any package.

I only have images of the completed parts (I think I even posted them in the forums because I thought they were examples of some fantastic hard surface work) and it was driving me insane trying to sus out how he built them. In something like alias, or inventor, or solid works where I can simply stick all the bits together with boolean union and click edges and say fillet or round, and Bob's your uncle. With Maya it was a chore but I am happy to say it can be done.

[daverave]

Isnt that the fun Rick, I could have gone with the easy option for my M4 and done one with straight armour..............dave

[ctbram]

Daverave: Thanks for pointing out the screw head. The original one I made was flat and it was optimized. When I got a closer look at the reference I noticed it was rounded and I made it starting with a 12x12 sphere instead of the default 20x20 but it still looked a bit to dense and I intended to go back and optimize it but forgot. This is the optimized one, basically the same process started with half a 8x9 sphere.

http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_screw_01_wf.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=screw_01_wf.jpg) http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_screw_01.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=screw_01.jpg)

[bullet1968]

well dont there Rick....looks like a lot of hard work. Not sure if I personally would want to get that intense just yet.

cheers bullet

[ctbram]

One thing about using booleans to join things is you can just endlessly keep finding bits of the topology you can improve. I tweaked a few bits. Pretty much done tinkering on this guy. I want to make the brake and disc assembly because they also look like a good challenge.

http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_06_wf.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_06_wf.jpg) http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_06.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_06.jpg) http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_05_wf.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_05_wf.jpg) http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_05.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_05.jpg)

http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_04_wf.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_04_wf.jpg) http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_04.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_04.jpg)

[Jay]

Hey Rick

Looks great. Pretty straight forward job for a man of your calibre!!

Will you stop mentioning the 'B' word LOL .....it burns us preciousssss!!!

Jay

[daverave]

Yes Jay Im having B trouble at the moment.......LOL........dave

[Jay]

Take two polygons and see me in the morning mate...LOL

[ctbram]

Thanks Jay. :D

I had to add a bit that connects the tube with the back brace.

http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_09a.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_09a.jpg) http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_09a_wf.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_09a_wf.jpg)

I just need to fill in the inside details of the part that clamps around the axle.

[bullet1968]

Booleans....has it become some sort of genetic anomoly??? I LOVE BOOLEANS...booleans

[ctbram]

More! More! I must join more things on to this Dojigger!

Honestly I think that is the last piece I have to add, but I still need to fill in the inside of the clamping part. I keep starting to think about modeling it and I see another thing-a-ma-bob that needs to get added. lol

http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_strut.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=strut.jpg) http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_13.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_13.jpg) http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_13_wf.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_13_wf.jpg)

http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_14.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_14.jpg) http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_14_wf.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_14_wf.jpg) http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_15.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_15.jpg) http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_15_wf.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_15_wf.jpg)

One advantage to using the Boolean method is that I saved the original shapes and so as I keep sticking bits on, and inevitably get the surface all wonky, I can just bring back the original part, subdivide it 3 times, make it live and then just reposition the verts onto the live mesh and bingo no lumps or bumps!

Oops I used the forbidden B-word again. From now on I'll just call it Bob.

[publicFunction]

Bob. Funny name for a girl isn't it. :P

I use booleans a lot. I don't understand the anti attitude some users have to them. As long as you understand that there is tidy up work to be done after and you are prepared to do the work then use them I say.


publicFunction (Chris) - Sent from my iPhone using Tapatalk

[clearairstudios]

Nice work. Very clean geo.

[Jay]

Going great Rick

Im going to get a 'B' swear box put into the forum somehow LOL. Or a possible electric shocking system put in, so everytime you type it a thousand volts shoots thru your keyboard and up you fingers....be warned LOL

Jay

[ctbram]

I used the drawing on a live surface to retopologize the back side.

It works well. My only gripe is that you have to physically grab each point and jiggle it. It would be a very nice feature if I could pick select verts and just press a hot key to project them down on to the live surface. But it was not that bad. Took about 20 minutes.

The new version added about 19 polygons but that is because I added a edge loop to create proper edge flow.

RETOPOLOGIZED

http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_16.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_16.jpg) http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_16_wf.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_16_wf.jpg)


ORIGINAL...

http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_15.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_15.jpg) http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_15_wf.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_15_wf.jpg)

[Acid44]

"motorcycle fork brace thingy."

Not a great place to start :P

Looks great though man, especially in keeping the shape together, really hard part to model

[ctbram]

public: I use Booleans quite a bit as well. I got my start as a surface modeler in the auto industry in Detroit. Packages like catia, alias, inventor, solid works all rely heavily on Booleans. I agree if you understand them they can be very powerful.

I can understand why many fear them though as Maya's implementation has a lot of limitations and like NURB's they are scary to folks that come from a poly modeling background.

Clair: Thanks. I am finding that the Boolean method works pretty well but you have to be willing to retopologize a bit to get a proper edge flow. You can rig the deck to some extent by setting up the Boolean objects before joining. Also, by saving the originals it is not hard to delete hole sections of the mesh and then use the original shape (subdivided, and made live) to draw in the proper edge flow.

Jay: :giggle:

Acid44: Actually I did not plan on modeling the entire bike and I started with that piece because it appeared to be the most difficult to model. If I did decide to do the entire bike everything else would be easy by comparison. :)

I do plan on modeling of the other complex shaped bits though.

[Acid44]

I meant place as in the "thingy" frame of mind, figured you wouldn't be doing the whole bike :P

And yeah, after this the panels would be a waste of time :P

[ctbram]

Acid: NP I knew you were joking. I just wanted to explain why I started with such a goofy part.

[Acid44]

Ah, gotcha. You should to the whole bike though. And I mean WHOLE like engine and all :P Would be damn good exercise to do all those little bits in there accurately, and since they're mostly the same you'd have tons of refs

[ctbram]

Hi guys,

Finished with the Brace dojiggermabob thingy and tossed the fender together.

I wanted to play with compositing in PS and AO. So I made a AO, WIRE, and DIF render of each shot and composited them in PS. I did not use FG at all since it just blows out the scene with the all white background.

The renders are without lights. I wanted to examine the affect of the distance value of the MR AO texture node. As you can see from the setup there is a plane that has no visibility in the scene and receives shadows and is not visible in reflection. I believe distance is the max distance from each point can be affected by occlusion. But I am not 100% sure because reading the mental ray documentation is like reading stereo instructions, written in Japanese, and then transcribed to English by a French speaking Chinese person.

Anyway, this is what I have gleaned so far.

The Spread controls size of the hemisphere projecting rays to each point. It has a range from 0.0 to 1.0. Think of a half sphere with a cutting plane. When the value is 1.0 you get the entire hemisphere. As the value gets smaller the cutting plane moves up and reduces the size of the dome, which narrows the cone of rays that affect the occlusion of each point. The documentation says the range is 0.0-1.0 but I have seen people using 1.1 so I am assuming that this might allow the cutting plane to move down below the mid-line of the dome and thus begin to cast rays at an upward angle?

The distance is the maximum distance from each point that rays can affect the occlusion of that point. Think of it as the radius of the hemisphere above each point. I am assuming it is in scene units. I used inches in my scene. With a value of 0 the dome is the size of the entire scene. I tried values of 0, 57 (the distance of the camera from the center of the object), 22 (the diagonal of a cube surround the object in the scene), and 14 (the approximate diameter of a sphere surrounding the objects in the scene).

The first affect I notice is that the greater the distance value the longer the render and the darker and more spread out the shadows. The render times for the same preview section of my scene were as follow...

value time
------ ------------
0............~4 min+
57..........~3 min+
22..........~2 min+
14..........~1:50
7...........~1:50 and some of the parts farthest from the camera had no occlusion

Here is the scene setup....

http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_rendersetup.png (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=rendersetup.png)

These are the renders at a distance of 22 and a spread of 1.0....

http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_27_AO.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_27_AO.jpg) http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_27_WIRE.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_27_WIRE.jpg) http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_27_DIF_AO.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_27_DIF_AO.jpg)

http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_26_AO.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_26_AO.jpg) http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_26_WIRE.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_26_WIRE.jpg) http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_26_DIF_AO.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_26_DIF_AO.jpg)

http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_28_AO.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_28_AO.jpg) http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_28_WIRE.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_28_WIRE.jpg) http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_wheelBrace_28_DIF_AO.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=wheelBrace_28_DIF_AO.jpg)

[Jay]

hey Rick

good findings.....

Spread can be over ridden....we do it all the time in modelling for reviews. Its because some models can have flat surfaces, without the detail where there is nothing to occlude from.....very similar to what dave Rave had on his tank, the body work had no real detail. Hence his whites really burning. The more spread you add the darker the shadows, so thats really choice at the end of the day. Max distance is the distance the rays will bounce to its furthest point, if it is zero then the bounce is actually infinate. So by capping it off with a good value then you should get a good result. Also adding the ground plane really helps with the bounced AO rays, otherwise again the rays are just hitting the object making use of the main creased area and bouncing of the larger spread area int o space giving again burnt whites.

I had a similar prob with an occlusion render on my P51. Because the wings were flat the rays were infinate even with a bit of fall off, but I put in a polyplane above the model so the rays could bounce back. Job done

Higher values do mean longer render times but if you get the balance right then the results are sexy!! Scale is important too...best to model real life objects to real size and take advantage of the power thats on offer this way...a bit like Nparticles, Ncloth etc...work best at proper scales

and as for the documentation on it...I have a bin here for mine LOL. The help is fine generally but really to gain a full understanding of this stuff...working in the right environment really opens the eyes and throws the book out of the window. Thats why Im always happy to pass the knowledge on.

we'll have to do a write up between us mate and post it up

Jay

[ctbram]

Thanks Jay. I have not spent a lot of time in mental ray. I generally just turn on FG and crank the quality to production fine trace, fiddle a bit with the env color and toss in a sight wall.

I have been wanting to replace FG with AO for a while now but each time I look at the MR docs my head starts to hurt. No one I work with uses entertainment modeling sw so I have to figure it out on my own along with help for good folks like you.

It is starting to make some sense to me though. The one thing that drives me bonkers with MR is the seeming randomness of everything. How are you supposed to figure out how to set things if there are no rules pr guideline or work flows? That is why a set out to systematically see what the distance value and spread really do and if you can use scene measurements to get you to the right numbers faster rather then just pulling them out your ass so to speak. And this is just one of many variables that all affect each other! It's like trying to solve a multi-variable equation with dozens of inputs! Where do you start? What order do you set things?

I mean what is the point of getting a render to 10 secs if you take 600 hours to figure out the settings and then the next render is completely different and you have to start all over again!

I am now going to scale the model from the current real world size and try the renders again with the distance set to the same proportional distance and see if I get a similar render with the model scaled larger and smaller.

Also I have a ao shader, a wf shader, and then procedurals (blinns, mostly) for the painted renders and what I am calling the diffuse render looks milky because of the all white background but if I tried adjust the levels the whole thing gets darker and so then when I composite the AO layer it gets very dark.

Anyway can you dig that crazy ass part! I counted it and there are like 15 shapes all crammed into a single unified mesh! I thing the topology looks pretty good. I did end up with 3 triangles.

[Jay]

Use a surface shader with the ao texture in the color channel unless thats what you are doing.

J

[ctbram]

yep that's what I am doing

[bullet1968]

Wow Rick....looks intensely difficult this project....nicely done though mate..right up your alley with automotive!

cheers bullet

[ctbram]

Thanks Bullet.

There are so many renders of what looks like the same part because every time I was about to declare victory I would find another reference image and see there was another hole, or slot, or pocket to cut into it or a new whatchamajigger to stuff into it and of course in the very end I found my front view proportions to be off a tad. Not enough that I am going to worry about it but it make the ECD in me get a little twitchy!


I have been looking for the next bit to do there are a few parts on the bike that look challenging. Deciding which way to go next.

[Jay]

Cool mate, I wasnt sure (reply for the AO on surface shader)

Jay

[ctbram]

Okay with Jay's help I have started to use render layers. I still have not figured out how to use them to create all renders in a single shot. Primarily because they share a single render globals from what I can see and they all look at the current preview render settings.

So for wire render I want the preview render setting to be 1 and in the render globals contour rendering enabled.

For all the other renders I want preview render to 3 and no contour rendering.

In the DIFF and AO render I want FG off in the globals.

In the FG I want FG on in the globals.

So I have to do a render for the wire, switch settings and then do a render for the AO and DIFF, then switch settings and do a render for the FG.

TEST RENDERS

Here is the DIFF only. The image is milky and colors are muted and do not pop

http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_DIFF.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=DIFF.jpg)

Here is the FG only the colors are more vibrant but the image is to bright

http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_01p_FG.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=01p_FG.jpg)

Here is the DIFF and FG combined (multiplied) better color and balance but no shadows

http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_01p_FG_DIFF.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=01p_FG_DIFF.jpg)

Here is AO shadows

http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_01p_AO.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=01p_AO.jpg)

Here is DIFF + FG + AO To me this looks pretty much correct (possibly a tad dark)

http://i181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/th_01p_AO_FG_DIFF.jpg (http://s181.photobucket.com/albums/x185/ctbram/bmw%20s1000rr%20wip/?action=view&current=01p_AO_FG_DIFF.jpg)

[ctbram]

Ah I just learned I can override render setting for each render layer so each RL can have unique render settings.

So now the only obstacle to rendering all layers at once is the wire frame render that I want to render at smooth preview setting 1 (1-key) and the rest of the layers I want to render at smooth preview setting 3 (3-key).

Is there a way to do that?

[honestdom]

you should multiply the AO by the diffuse.

but your diffuse shouldn't have any specular, likewise with your FG render. For the FG, the pass you want is the "indirect diffuse" this will be your indirect light which is basically FG only. Because you are just using diffuse direct and indirect and AO. you will need the direct specular pass too. you can add these together and you will get your primary.

[THExDUKE]

You could use contour rendering in Mental Ray. But Im not sure if it works for smoothing level 1 !

[ctbram]

I do use contour rendering in mental ray. It works in smooth preview level 1. If you do not choose level 1 you get the actual render tessellation and for level 3 that is 2 levels of subdivision and you get a very dense wire frame that does not show the actual modeling cage.

So the issue for the MR wire frame is I just have to manually set my preview render setting to the 1-key for all the objects I want wires on. For all the other renders I just set it to the 3-key for the default 2 levels of smoothing.

It's not a major inconvenience.

[THExDUKE]

You could use Approximation Editor and assign a Subdivision Approx. for your desired geometry in a new renderlayer. You can also create overrides for the values. Go Window/ Rendering Editors / Mental Ray / Approximation Editor.

So that means you could leave everything at level 1 and drive the SubD only by the Approxeditor. And for the contourlayer you simply change the values in the editor.

[ctbram]

oh I will look into the approximation editor.

Part of the frustration is I do not particularly enjoy rendering or lighting in maya (primarily because I hate mental ray). I find it tedious and boring. Also, there are several technologies that were added to maya that overlap and in some case conflict.

Like render layers vs render passes. If you try to google for info you get a mish mash of tutorials that talk about render layer presets that do not exist anymore and whacky work flows that do not really demonstrate how to use either the render layers or passes properly. So you are left scratching your head in frustration.

[THExDUKE]

Yeah...I know. i never use the RenderPasses Presets since they never work and its more work to set them up as it takes to create your own custom renderlayers.

[clearairstudios]

Hi Rick, Great work.

Are you using Linear Workflow for the renders?

[LML]

hi there

would love to see a tutorial about using booleans like this; any good ones u know of out there?

nice work!

LML

[Jay]

CTBRAM is the Boolean king so Im sure he'll point you in the right direction

J

[ctbram]

Thanks Jay, but I honestly cannot think of a single polygon modeling tutorial that advocates the use of Booleans in any way.

Most of the workflow I use is based on my background in nurbs surface modeling which relies heavily on booleans and rounds and fillets and trims and such. I have been searching for ways to use that nurbs workflow to create clean G1 and G2 surfaces and then convert those to polygon objects usually through retopo.

Sometimes I boolean polygon objects directly and clean the geo. Other times I boolean nurbs objects and then make them live and redraw the polygon surface directly on top of the live nurbs surface (retopo). Sometimes a bit of both. In my experience there is no better way to create clean surfaces with G1 and G2 continuity then with NURB's. The challenge in Maya is that they have really let the stability of the NURBs tools decline and so most poly modelers avoid them because of their complexity and ever increasing instability.

I am not sure what the future of Nurb's is in Maya. I hope they do not become extinct. However, as long as I can import nurbs objects I can always make them in other packages like SW, Inventor, where they are stable and more predictable and then import them into maya and use them as scaffolds to build poly geometry around.

Just depends on the final shape I am shooting for. Part of this exercise is to search for viable work flows to take nurbs surface modeling techniques and see if I can use them to create clean complex unified polygon objects.

[Jay]

Sure.

Theres no harm in them in any way as long as the 'user' is aware of what not to do and how to clean up and tackle the problems like edge filling from holes. We've used them alot this last two weeks because of hard surface models that have required holes in odd places....all good fun and are quick to sort out generally.

My Lead Supervisor has twitched when we mentioned them LOL, but its all in jest....

cheers
Jay

[ctbram]

Thanks Jay,

I agree and as you know I use Booleans a lot. I just cannot think of any Maya tutorials that I have seen that uses them in any significant way to suggest to LML. I guess I could try to come up with something at some point if there was enough interest.

I have seen Booleans used quite frequently in other 3D applications, one in particular that is used heavily in made for TV science fiction shows and and rhymes with bright grave. But they leave geometry that I fear would cause people on Maya sites to have seizures just looking at it. I mean their eyes would literally bust into flames! lol

[ctbram]

Here is an example of a little test project I started a long time ago. The idea was to develop a workflow for taking a simple car surface well suited to nurbs modeling techniques and making a nice G1, G2 NURB's surface model. Then build a clean quad G2, G3 polygon surface using the nurbs parts as a scaffold.

Here is one of the trimmed and booleaned nurbs surfaces and I was building the polygon part with the exhaust holes on the live nurbs surface.

http://i181.photobucket.com/albums/x185/ctbram/retro%20racer%20nurbs%20to%20poly%20workflow%20exp eriment/th_example_2.png (http://s181.photobucket.com/albums/x185/ctbram/retro%20racer%20nurbs%20to%20poly%20workflow%20exp eriment/?action=view&current=example_2.png) http://i181.photobucket.com/albums/x185/ctbram/retro%20racer%20nurbs%20to%20poly%20workflow%20exp eriment/th_example_1.jpg (http://s181.photobucket.com/albums/x185/ctbram/retro%20racer%20nurbs%20to%20poly%20workflow%20exp eriment/?action=view&current=example_1.jpg) http://i181.photobucket.com/albums/x185/ctbram/retro%20racer%20nurbs%20to%20poly%20workflow%20exp eriment/th_example_3.png (http://s181.photobucket.com/albums/x185/ctbram/retro%20racer%20nurbs%20to%20poly%20workflow%20exp eriment/?action=view&current=example_3.png)

Here are the trimmed and booleaned sections of the car...

http://i181.photobucket.com/albums/x185/ctbram/retro%20racer%20nurbs%20to%20poly%20workflow%20exp eriment/th_retro_racer_concept_02.jpg (http://s181.photobucket.com/albums/x185/ctbram/retro%20racer%20nurbs%20to%20poly%20workflow%20exp eriment/?action=view&current=retro_racer_concept_02.jpg)

And here are some various progress shots...

http://i181.photobucket.com/albums/x185/ctbram/retro%20racer%20nurbs%20to%20poly%20workflow%20exp eriment/th_retro_racer_concept_03.jpg (http://s181.photobucket.com/albums/x185/ctbram/retro%20racer%20nurbs%20to%20poly%20workflow%20exp eriment/?action=view&current=retro_racer_concept_03.jpg) http://i181.photobucket.com/albums/x185/ctbram/retro%20racer%20nurbs%20to%20poly%20workflow%20exp eriment/th_retro_racer_concept_01.jpg (http://s181.photobucket.com/albums/x185/ctbram/retro%20racer%20nurbs%20to%20poly%20workflow%20exp eriment/?action=view&current=retro_racer_concept_01.jpg) http://i181.photobucket.com/albums/x185/ctbram/retro%20racer%20nurbs%20to%20poly%20workflow%20exp eriment/th_retro_racer_02.jpg (http://s181.photobucket.com/albums/x185/ctbram/retro%20racer%20nurbs%20to%20poly%20workflow%20exp eriment/?action=view&current=retro_racer_02.jpg)

I stopped here once I knew the workflow was viable. I intended to come back and finish it but other real life stuff popped up and I never got back to it.

At the time I thought this might be a fun little project to demonstrate the proper use of Maya's NURB's modeling tools and workflow and in the end use the nurbs surfaces to create a clean poly model with very precise surface continuity.

Quick note on surface continuity:

What is all this G0, G1, G2 stuff?

G0 G1, G2, ..., Gn are commonly used in surface modeling to describe the continuity of two surfaces that meet at an edge.

G0- positional continuity - means two surfaces that meet along a common edge will have any point along that edge co-resident (they meet at the edge) but they do not have to be tangent/ typically you will have crease where these surfaces meet.

G1 - tangential continuity - means two surfaces will meet at a common edge and the slope of any point along that edge on either surface with be tangent. In other words their first derivatives defining the edge where the surfaces meet will be equal. No crease where the surfaces meet.

G2 - curvature continuity - means two surfaces will meet at a common edge and the rate of change of slope of any point along that edge on either surface will be equal. Or in other words the second derivative of a the function defining a point along the edge where the surfaces meet will be equal. No crease where the surfaces meet and the rate of change in the slope is uniform (think of a spiral vs a circle).

and to G3 would mean the 3rd derivative of a function defining a point along a common edge of two surfaces would be equal and so on. However, although I know packages like alias (formerly studiotools) can handle continuity beyond G2, I have rarely seen them used.

[bullet1968]

I know you cant split edges from a boolean Jay..if thats what you meant..and as you said there is a workaround...I think ct has mentioned it before. As long as you have at least 1 edge on section one that the boolean will intersect...after the boolean you can then split from the edges of the resulting geo.

cheers bullet

[LML]

I am coming from Rhino--so I understand the frustration w/ booleans in Maya;

LML

[tkev]

What is all this G1, G2, G3 stuff?

G1, G2, G3, ..., Gn are commonly used in surface modeling to describe the continuity of two surfaces that meet at an edge.

G1 - positional continuity - means two surfaces that meet along a common edge will have any point along that edge co-resident (they meet at the edge) but they do not have to be tangent/ typically you will have crease where these surfaces meet.

G2 - tangential continuity - means two surfaces will meet at a common edge and the slope of any point along that edge on either surface with be tangent. In other words their first derivatives defining the edge where the surfaces meet will be equal. No crease where the surfaces meet.

G3 - curvature continuity - means two surfaces will meet at a common edge and the rate of change of slope of any point along that edge on either surface will be equal. Or in other words the second derivative of a the function defining a point along the edge where the surfaces meet will be equal. No crease where the surfaces meet and the rate of change in the slope is uniform (think of a spiral vs a circle).

and to G4 would mean the 3rd derivative of a function defining a point along a common edge of two surfaces would be equal and so on. However, although I know packages like alias (formerly studiotools) can handle continuity beyond G3 I have never seen a case where it was used.

This came out exceptionally clean. I'm still trying to figure out how you were able to build and possibly test for G3 continuity building curves in maya. Even modeling the initial surface it's difficult to get the curves just right while matching a reference. Does anyone know a good resource on this? I understand the part where it's being chopped up with trim surfaces, but maya's curve curve tools leave a lot to be desired.

[ctbram]

There is no way to check for for continuity in Maya using polygons. There are crude tools in maya Nurbs for fillets, rounds, and sewing surfaces, the square tool, and project tangency tool where you can specify tangential constraints (G2).

You can estimate G3 by eye but judging rate of change of curvature without comb plots and other continuity tools is impossible.

G2 curvature is easier to achieve by simply building the poly cage on top of a nurbs surface. But then when you cut the surface to create a seam you have to once again just play it by feel when placing the supporting edges on either side of the seam as the overall surface curvature will change when you cut it. I always have to re-snap the verts to the nurbs surface and then manually make small adjustments to maintain Tangency and overall surface curvature along the seam (and once again this is all very much by eye).

G1 is trivial and is the only form of continuity that does not require any additional tools to verify and so all you can really say with certainty of any two surfaces that meet at an edge in maya is they are always at least G1.

You can get a very good approximation of G2 and using a chrome environmental shader you can get a similar visual check as with a chrome shader in an actual surfacing application. But once again there is no zebra shader or comb plot with Maya so it's all touchy feely.

NOTE: The above applies to trying to capture Nurbs surface continuity when converting to polygons. Maya does have crude and I mean very crude Nurbs tools to create fillets and rounds and sew surfaces with G2 continuity with tangential constraints.

NOTE: read up on Nurbs surfacing tools. - specifically....

- project tangent (pick a curve and a surface) as far as I recall this is the only tool that you can specify curvature (G3) continuity in maya.
- square tool This tools options can specify tangency and the tool will actual display if tangency is achieved
- and of course the fillet, rounds, and stitching tools

Correction: I have been using Cx notation. The correct Gx notation for positional, tangential, and curvature continuity is G0, G1, G2 respectively. The x value implies the degree of the differentiation of the vectors tangent to the common edge between two surfaces. So just subtract 1 from all the Gx references above. Sorry.

[tkev]

There is no way to check for for continuity in Maya. You can estimate G3 by eye but judging rate of change of curvature without comb plots and other continuity tools is impossible.

G2 curvature is easier to achieve by simply building the poly cage on top of a nurbs surface. But then when you cut the surface to create a seam you have to once again just play it by feel when placing the supporting edges on either side of the seam as the overall surface curvature will change when you cut it. I always have to re-snap the verts to the nurbs surface and then manually make small adjustments to maintain Tangency (and once again this is all very much by eye).

G1 is trivial and is the only form of continuity that does not require any additional tools to verify and so all you can really say with certainty of any two surfaces that meet at an edge in maya is they are always at least G1.

You can get a very good approximation of G2 and using a chrome environmental shader you can get a similar visual check as with a chrome shader in an actual surfacing application. But once again there is no zebra shader or comb plot with Maya so it's all touchy feely.

NOTE: The above applies to trying to capture Nurbs surface continuity when converting to polygons. Maya does have crude and I mean very crude Nurbs tools to create fillets and rounds and sew surfaces with G2 continuity.

Thank you :). I'm going to try that. Bleck some of those limitations make me wish I used a different software for modeling, but I can't find any plugins or anything that truly build upon the base tools in a technical manner. Maya can be an incredibly annoying program to model in. I picked it up due to its popularity over other packages here, but I wish I had something with tools that provide better accuracy from a technical standpoint combined with the ability to import to maya without a huge amount of artifact issues. Alias is cost prohibitive unfortunately.

Edit: edited version noted, I will take a look at those