Brake caliper

[philnolan3d]

I'm still working on my Audi model that I started a while back and getting very close to finished. Instead of modeling the brake calipers with polyons it's such a complex shape that it looked like it would be easiest to sculpt it in 3DC and then retopo. Here's what I've got so far.

[Greg]

Wow, that's really clean looking Phil. Tips on how you did it?

[philnolan3d]

Thanks, I started with this shape I made in LightWave. The rest is all primitives and booleans in 3DC.

It's 3 voxel layers for now, then I will merge them later. Here they are separately:

[philnolan3d]

OK, ready to move on to LightWave I think. I will add some small details like screws there.

[digman]

Nice sharp and clean line work, great job...

[philnolan3d]

Thanks, Problem now is that I'm exporting and it's not doing a very good job of keeping those clean lines. I was thinking that would be easier than retopo but it looks like I might have to retopo after all.

[Taros]

I am very interested, how you will retopo it...

[philnolan3d]

I've already started

[Taros]

One advice: First retopo the round areas, like holes, circles, transitions and such things, without thinking of connections. You can connect them later. Plane areas "forgive" polygon chaos.

Don't begin to retopo in detail, do it later. If you begin too early with subdividing, you will have problems later. Work rough. It is faster to retopo rough first. You will have enough time for subdividing later.

You don't need to subdivide long quads, subdivide them later, subdivide just curved areas.

If possible don't use tris. Use your animation package for optimising the mesh. As I know, you are using lightwave, there it must be possible to reduce polygons, too.

[polyxo]

Phil,

have you ever looked into Nurbs? This, especially when manually retopoed really looks like

crazy wasted effort to me.

Even in a cheap program such as Moi a beginner can model such a Brake within one or two hours

and export it as a precise mesh in arbitrary densities.

[Robert G]

Wow, this looks very good Phil.

But yes, must be quite a challenge to retopolize.

Looking forward to see the exported version of it.

Good luck, Robert

[michalis]

I don't agree with this approach Philnolan3d. How will you UV map this? Why voxels for this? Why this as a whole mesh?

SImilar questions on archi presentations I usually do.

[philnolan3d]

One advice: First retopo the round areas, like holes, circles, transitions and such things, without thinking of connections. You can connect them later. Plane areas "forgive" polygon chaos.

Don't begin to retopo in detail, do it later. If you begin too early with subdividing, you will have problems later. Work rough. It is faster to retopo rough first. You will have enough time for subdividing later.

You don't need to subdivide long quads, subdivide them later, subdivide just curved areas.

If possible don't use tris. Use your animation package for optimising the mesh. As I know, you are using lightwave, there it must be possible to reduce polygons, too.

Thanks, I am actually sort of doing details like holes first. Basically as I get to each area like that I do that spot first and then make the connections to it. I'm not worrying about neatness very much but sometimes I can't help myself, it's so quick and easy to just smooth things out with the Brush tool. I generally try to avoid tris as a rule of thumb, but why here? it's a static object that won't be bending or anything.

On NURBS: Even if I liked them I can't afford a package that uses them. I also don't need a precise mesh in arbitrary densities.

Voxels because they were easier than polygons, I hadn't thought about exactly where I'll place my UV seams but that's easy. Actually I'll probably only do the one outer most face because it has the "R8" text on it, then rest doesn't need a UV map.

The retopo isn't really difficult at all.

[Taros]

...

On NURBS: Even if I liked them I can't afford a package that uses them. I also don't need a precise mesh in arbitrary densities.

...

Give it a try: http://moi3d.com/

[philnolan3d]

I have tried it, like I said I don't like nurbs. Also like I said, I can't afford it.

[polyxo]

I have tried it, like I said I don't like nurbs. Also like I said, I can't afford it.

The approach you are using here easily takes ten times longer than neccesary (a seasoned Nurbs-Modeler

surely doesn't need more than half an hour for these parts) and the result still is

inaccurate - I just thought I'd mention it . I already have a hard time when I see people

doing such parts in SubD, but with Voxels+ Retopo I really couldn't just watch and keeep quiet...

[philnolan3d]

But I'm not a "seasoned nurbs modeler" and don't ever expect to be. As for accuracy it still wouldn't be accurate if I don't have detailed diagrams or blueprints to work from. I basically have one picture.

[AbnRanger]

Phil,

have you ever looked into Nurbs? This, especially when manually retopoed really looks like

crazy wasted effort to me.

Even in a cheap program such as Moi a beginner can model such a Brake within one or two hours

and export it as a precise mesh in arbitrary densities.

+1. This is why I bought the PowerNurbs plugin for 3ds Max, as it makes this kind of work exceptionally fast and clean. It also has the added benefit of being able to export clean IGES and CAT(Catia) models, that manufacturers can immediately work with. That's a selling point for your services...taking concept/marketing models straight to production, so they don't have to pay someone else to re-model it in a Solid Surface program (like SolidWorks for example).

Now, if you didn't have to retopo manually (auto-quadrangulation rarely ever works well for me), that might be a different story. But in the final analysis, different tools have different strengths. Hard surface modeling may be doable in Voxels, but it is not the most efficient method to get from start to finish.

You have NURBS now in LWCAD, do you not?

[philnolan3d]

You have NURBS now in LWCAD, do you not?

The latest version might. I can't afford to upgrade, I'm still using LWCAD 2.5. Of course I probably wouldn't use that anyway.

[Greg]

Reading this, I think maybe Phil just -wanted- to create this in 3d Coat. Of course there are easier/faster ways, but sometimes I personally just want to try to push a package, either for learning, or for fun or both! As a 3D Coat beginner, I like Taros advice about retopo methods though, and gonna apply it to my own model soon.

Anyhow, I think it's still looking good Phil!

[philnolan3d]

That's silly. I was trying to do it in LightWave like I did with the rest of the car. That was complicated so I went with the best option available to me.

[bwtr]

From my point of view I think NOT doing Nurbs modelling for "hard" objects is such a waste of time and effort.

And MoI is spectacularly good and easy to learn.

Brian

[Greg]

That's silly. I was trying to do it in LightWave like I did with the rest of the car. That was complicated so I went with the best option available to me.

Ah, my bad. I need to stop projecting the way I do things onto other people!

Looking forward to updates!

[philnolan3d]

Looking forward to updates!

Just need to attach a couple of screws then I'm done.

[Greg]

Nice. Keep posting!

[philnolan3d]

Thanks, I'm basically done with it now. Here's a couple of shots of it on the car, though it's hard to see with this angle / lighting.

The tire tread was also done in 3DC a while back, I wrote a tutorial on it in the tutorial area of the forum.

[Taros]

Thanks, I'm basically done with it now. Here's a couple of shots of it on the car, though it's hard to see with this angle / lighting.

...

The tire tread was also done in 3DC a while back, I wrote a tutorial on it in the tutorial area of the forum.

You are crazy, Phil...

[michalis]

OK taros, I think you're right, he is crazy after all.

Great model philnolan.

[philnolan3d]

Thanks.

[UnCommonGrafx]

Hey Phil,

Way to go on this. Curious as to your topo flow, or watching a vid of someone going through the suggested paces.

I think this, indeed, was a proof of concept/practice of method and workflow.

I really like the sculpt. And you've hit upon what I hit: meshing from the voxel room isn't quite complete.