20 replies to this topic

[Sparks Murphey]

It's been some time since I shared a 3D project on here, mostly because I haven't been making 'Mechs in a while. But the itch has returned, and I've started work on a Wyvern, based in part from Flying Debris' concept sketch below:

So far, I'm partway through the head. I've made a few small changes from the original concept: I don't much like the idea of a cockpit window that doesn't show you what's immediately in front of you (maybe it's just me, but I don't like walking into things) and there are some mysterious dark regions on the side (in shadow) that needed some detailing.

Anyway, the first two work-in-progress shots. Texturing is placeholder.

[Impyrium]

Looks really good. Do you mind me asking how you got the panel effect on the top of the second pic?

[Sparks Murphey]

AUSwarrior24, on 13 August 2014 - 06:14 PM, said:

Looks really good. Do you mind me asking how you got the panel effect on the top of the second pic?

The easy answer is "separate panels". It's a little more complex than that, to be fair.

I'm using 3DS Max and a technique I've seen referred to as the "double turbosmooth" technique. At the basic level at the bottom of the modifier stack, I'm manipulating a relatively low poly mesh that look like this:

The sides I want curved together, like the outer corners of the hatch, are grouped together in smoothing groups. I've then got an instance of the Turbosmooth modifier set to two iterations and with "Separate Smoothing Groups" ticked. This ensures that the curving effect of the modifier is only applied to the surfaces I want it to, and keeps crisp edges in other places. This looks like this:

Finally, there's another Turbosmooth modifier on the top of the stack, this one also with two iterations but without "Separate Smoothing Groups" ticked. That takes the crisp edges and files them down a little, but because the mesh has already been subdivided twice, it keeps the curves relatively small. This gives the feeling of an edge to the panel, and results in the final mesh:

There are two downsides to this technique. Firstly is the insanely high poly count you end up with. Without the Turbosmooths on, the head is so far 1214 triangles, but with the Turbosmooths on it makes for a whopping 310,784. Not exactly game suitable. Of course, I'm on planning to make local renders, so that's fine with me, and the increased detail is preferable. If I did want to make a game version, though, I could still use the original low-poly mesh (preserved faithfully at the bottom of the stack) with maybe a few extra subdivisions here and there and bake the normals of the high-poly version onto a texture sheet for the low-poly version.

The second issue is that it is tricky to unwrap a UV set for the high-poly version. You can unwrap the low-poly version and it will largely get passed the stack, but the Turbosmoothing distorts that in unwanted ways. Alternatively, you can unwrap the high-poly version, but as mentioned, that's a lot of triangles to manipulate. I tend to go for a middle group here: unwrap the low-poly version, then re-unwrap at high-poly to tweak anything that went weird.

Hope that was informative and useful!

[Impyrium]

Wow, thanks for putting the effort into that reply! It's very helpful- I've been using Blender, so I'll have to see if I can translate that technique into it.

[Max OConnor]

Looking great, I've been hoping to see a better version of this image for a while. Looks like you are on the right track. Will this be suitable for 3D printing when you are done?

[Sparks Murphey]

maxoconnor, on 14 August 2014 - 07:03 AM, said:

Looking great, I've been hoping to see a better version of this image for a while. Looks like you are on the right track. Will this be suitable for 3D printing when you are done?

Not without some clean up, unfortunately. I'll be using parts that are below the minimum wall thickness of most 3D printers (assuming you're printing it at miniature scale; if you're printing full size, best of luck to you and send me pics), the turbosmoothing effect raises the poly count above the limit most 3D printers can handle (Shapeways has a cap of 1 million, for example; this could be circumvented by taking the modifiers off, but that would remove indents and chambers that I have relied on the modifiers producing) and some meshes in the scene are not closed (the cockpit windows being a good example). That said, I'll try to make the final version accessible so that those who want to can do such clean up.

[Heffay]

Sparks, have you played around with Sketchpad at all? Just found out Blender can upload to it through the client, and I might start working on that a bit more.

I wonder if these forums support Sketchpad... they should!

[Sparks Murphey]

Heffay, on 14 August 2014 - 02:01 PM, said:

Sparks, have you played around with Sketchpad at all? Just found out Blender can upload to it through the client, and I might start working on that a bit more.

I wonder if these forums support Sketchpad... they should!

No, I haven't even heard of it. Linky?

[Heffay]

Sparks Murphey, on 14 August 2014 - 02:23 PM, said:

No, I haven't even heard of it. Linky?

SketchFab. My bad.

http://cgcookie.com/...dels-sketchfab/

https://sketchfab.com/

You can upload models and they display in 3D in a browser that supports... WebGL? Like I said, I'm learning, but from what I've seen it's pretty amazing.

[Sparks Murphey]

Heffay, on 14 August 2014 - 02:28 PM, said:

SketchFab. My bad.

http://cgcookie.com/...dels-sketchfab/

https://sketchfab.com/

You can upload models and they display in 3D in a browser that supports... WebGL? Like I said, I'm learning, but from what I've seen it's pretty amazing.

You, sir, are awesome. The start-up art company I'm working with has been looking for a HTML 5 implementation of 3D rendering for a while, so you have literally just saved us a few months work building one ourselves. Now I can claim this thread as work time! \o/

I'll see if I can get some stuff uploaded to it. At the very least, some of my other work will almost certainly make it up there.

Cheers!

[Heffay]

Post your account name when you get it set up. I'm Markemp on there. Only have a tutorial I did from cgcookie of a turret + laser beam so far, but may put up more later. Haven't figured out materials yet. I think they need to be baked to work properly with cycles materials.

[Alexww34]

Pretty neat stuff.

Edited by Alexww34, 14 August 2014 - 08:54 PM.

[Sparks Murphey]

Update: Finished of the head (though I just realised I forget the side antenna) and started on the forward chest plates.

[cmopatrick]

very, very nicely done, Sparks. Plus, I like the little tutorial you include. Thanks.

[Sparks Murphey]

Some updates on the progress of the upper torso.

[Max OConnor]

This is so cool, I can't wait to see the final product. Lord it would be so cool to have this as a figure. Love the work.

[Sparks Murphey]

SRM bays and doors! Some versions of the Wyvern have doors for the SRM (like the 3039 TRO illustration), others don't, including Flying Debris' sketch. I've always like the look of them on the Wyvern, though, so I'm including them.

[Sparks Murphey]

Lower torso mostly complete. The back is still incomplete.

Edit: My fiancée just described it as "the coolest looking PEZ dispenser I have ever seen!"

Edited by Sparks Murphey, 04 September 2014 - 04:39 AM.

[Odanan]

This is very good!

Strange thing: I always thought the protuberance in the back was not centered. But checking now other images and the minis, it actually is. :/

[Sparks Murphey]

Yeah, one of the early images of the Wyvern makes it look almost like that piece is attached to the left shoulder.

If the Wyvern makes it into MWO, that back... thing could make or break it, being a huge target that, if hit boxes aren't handled carefully, could allow you to be shot in the rear from the front.