Critique => 2D & 3D => Topic started by: Indigo on December 17, 2008, 05:58:19 am

Title: Gluggernaut 3d
Post by: Indigo on December 17, 2008, 05:58:19 am
time to start learning 3d I suppose.  I'm attempting to take my pixel gluggernaut image and convert him to a 3d model.  So far all work has been done in wings 3D (i just seem to like the workflow) started from a cube primitive.  I plan to UV and texture him via 3ds max

Pixel reference:

current progress:
516 poly's - 808 tri's

C+C is very welcome as I go through this process.  This is a bit higher of a poly count than what would be accepted for mobile gaming - but I'm fine with that.  I just want to make a relatively low poly model.
Title: Re: Gluggernaut 3d
Post by: madPXL on December 17, 2008, 08:33:42 am
Very nice Indigo.
For your Gluggernaut, you should down the tris to 500 max.
For the UV mapping and texturing, maybe you should use (and learn if not) blender if you want to be on the way of the "free" softwares. It's very easy, nearly as 3ds max.

And I hope that you will use a pixel art texture!
Title: Re: Gluggernaut 3d
Post by: Dr D on December 17, 2008, 09:05:08 am
I think the arms are a bit too long, even if he is already a bit stubby, and yes I mean longer than your original pixel art work piece, because they're long there, too.
Title: Re: Gluggernaut 3d
Post by: TrevoriuS on December 17, 2008, 09:25:19 am
I can already see some small optimization issues but if you go over the model again you'll probably find them as well. If you have 3Dsmax, there's NO reason to step over to blender. In fact, I hate working in blender >< (I admit, gave up after 30 minutes, but still, most other modelling software feels natural).
Title: Re: Gluggernaut 3d
Post by: huZba on December 17, 2008, 09:32:31 am
If you view your model in tris you'll probably find quite a few easy to optimize spots. There are a few faces with too many edges, like the forearm area. Also some non-planar quads with wonky normals that cause shading errors.
Title: Re: Gluggernaut 3d
Post by: Indigo on December 17, 2008, 09:03:52 pm
thanks for all the feedback guys

downing the tris to 500 will be virtually impossible with this piece (at least for what I'm intending) - The tri count is already at 808.  As for the texture, sorry to disappoint, but I will most likely do a cg texture for this.  I already have my pixel sprite for all the pixel art yumminess.  ;)

Dr D:
I see what you're saying - but I think its actually an illusion of his torso being so short in comparison to the original.  I've fixed that now.

TrevoriuS and huZba:
yeah, I see the optimization issues you're both speaking of.  I've fixed most of them now, but I'll be extra cautious as I continue.

I'll post progress later tonight


Title: Re: Gluggernaut 3d
Post by: TrevoriuS on December 17, 2008, 09:21:53 pm
When using so many triangles and not just quads only, please mention the tri-count and not poly-count -> that's probably what confused madpxl as well as me.
Be sure to animate it breaking something >=D
Title: Re: Gluggernaut 3d
Post by: Indigo on December 18, 2008, 05:26:57 am

fixed a lot of the wonkey polys, fixed/tinkered with proportional stuff, and added the backpack.  Almost ready for texturing.  952 tri's including the backpack
Title: Re: Gluggernaut 3d
Post by: snader on December 29, 2008, 01:22:08 am
personally, i feel 3 sided fingers would be plenty for this.

then stuff those polies in the head/face because that looks really boring atm and i dont think you're gonna get away with just texturing the glasses/eyesockets
Title: Re: Gluggernaut 3d
Post by: snakesqzns on December 29, 2008, 01:56:56 am
Is this a single mesh with the backpack?   Maybe it's a Blender specific limitation, but when I try to UV Wrap a texture on a model that doesn't have  all the vertices linked ( ie backpack here would need to be connected to the back of the gluggernaut).
Title: Re: Gluggernaut 3d
Post by: ptoing on December 29, 2008, 07:00:12 am
Eh? Then you are doing something wrong. Blender can perfectly fine UV unwrap multiple objects into one UV. Gogo read manual time \o/

Indigo: coming along nice, looking forward to the texture.
Title: Re: Gluggernaut 3d
Post by: Indigo on January 06, 2009, 07:40:22 am
starting to texture...
Title: Re: Gluggernaut 3d
Post by: Souly on January 06, 2009, 07:58:14 am
Could you post the map?
Pretty please.
Title: Re: Gluggernaut 3d
Post by: Indigo on January 06, 2009, 08:02:08 am
Title: Re: Gluggernaut 3d
Post by: ndchristie on January 06, 2009, 08:02:36 am
same critique of the leg/knee structure as i had for the sprite when i first saw it :


that's what you have to work with.  it is asymmetrical.  :-X
Title: Re: Gluggernaut 3d
Post by: thedaemon on March 08, 2009, 09:28:02 pm

Ok, just pointing out some of the poles. (Lots of edges coming from 1 vertex.) When deformation occurs this will cause an unwanted effect. Try to keep poles down to the minimal (5 edges MAX) for best deformations. I realized you have to have poles to discontinue lines but try to keep them down to 5 as I said. Also when you have two poles next to each other, as is on the hand, it can really look bad. But the pole on the top of the head may be ok, since I don't think that part will deform. Most of this stuff really applies to a higher poly character more, but the lesson and practice should be the same across both platforms. Hope this helped.  Also, remember the silhouette is most important with low poly modeling especially.
Title: Re: Gluggernaut 3d
Post by: Gil on March 08, 2009, 10:51:10 pm
The one on the foot won't really deform much, same with the one connecting the knuckles to the hand.

The lower leg ones shouldn't deform much (though slightly more) either. I don't foresee any real problems if the animation is carefully done.

The one in the knee is uncertain. I would try to simpify that part a bit, but it seems like it'd deform more or less correct.

The one on the back will deform beautifully, as it lies on an actual real muscle point, so while it will deform a lot, it'll also look correct.

The shoulder might be problematic, same with the wrist one. I don't see a clear fix for either though.