The random model thread!

[Bunglo]

Hmmm, you've given me something to think about, Philip.

I've only attempted a few environments (think I pussed out on all of them) and when I first tried them, they were areas you could walk around in and look at from any angle. I started looking at what others were doing (guys who were in the industry) and they all, or nearly all, seemed to be doing these camera based pieces. I personally didn't understand why, other than maybe it was easier. After a couple more failed attempts at doing the walkable pieces, I figured the camera based ones were simply quicker (and perhaps simpler?) to do and if it was getting people in, why not just do what's quicker.

Maybe that's not the right mentality? I don't know, I certainly enjoyed trying to make a piece you could "play" a lot more than trying to stick to the camera angle thing, but the biggest killer for me is I just have terrible ideas for what might look cool.

Perhaps I should just have a go at it again anyways.

[Minos]

Hmmm, you've given me something to think about, Philip.

I've only attempted a few environments (think I pussed out on all of them) and when I first tried them, they were areas you could walk around in and look at from any angle. I started looking at what others were doing (guys who were in the industry) and they all, or nearly all, seemed to be doing these camera based pieces. I personally didn't understand why, other than maybe it was easier. After a couple more failed attempts at doing the walkable pieces, I figured the camera based ones were simply quicker (and perhaps simpler?) to do and if it was getting people in, why not just do what's quicker.

Maybe that's not the right mentality? I don't know, I certainly enjoyed trying to make a piece you could "play" a lot more than trying to stick to the camera angle thing, but the biggest killer for me is I just have terrible ideas for what might look cool.

Perhaps I should just have a go at it again anyways.

Just make game environments! Good looking environments rendered in real-time engines are one thing, and videogame environments are another. Even if you are making an environment just for eye candy, try to imagine how it could be used in a game and build some simple gameplay spaces around that (different levels, stairs, cover pieces etc).

A production artist is a lot more useful than someone who knows only how to make cool shit but knows nothing about smart texture reuse, modularity, optimization etc...

[Armageddon]

I made a pillar, and a light model. Not happy with the textures at all.

[mjens]

[Mr. Happy]

FINNALLY FINISHED MY FIRST GUN MODEL!!!!

The Magnetonator

RATE RATE RATE COMMENT COMMENT COMMENT THANKS THANKS THANKS

[Rick_D]

indexes and draw count?

is that actually that optimised? cause that seems excessive :S

[GregBoffins]

To even have 12k draws you'd need 12k unique meshes or 12k unique textures.

[Rick_D]

To even have 12k draws you'd need 12k unique meshes or 12k unique textures.

imaghine having 12k tetxures, haha even the thought of it makes me sweat

[GregBoffins]

To even have 12k draws you'd need 12k unique meshes or 12k unique textures.

imaghine having 12k tetxures, haha even the thought of it makes me sweat

Yeah totally, a consoles worst nightmare

[Mr. Happy]

I will admit it could be better. Everything I do I look at the next day and say oh damn it could be better...but must ship submit it sometime.

"Optimized" is relative. TF2 is known for having (or have had) weapons that are shall we say...excessive. The number of vertices per triangle is lower than several in-game weapons with fewer triangles. Of course triangle count is not the best measure. Split vertices are better measure. My understanding is that indices are best measure because that is an actual measure of the number of elements in the triangle lists sent to the renderer for all drawcalls associated with the model. There is also phong, so that is multiple fills and my understanding is that is where the higher number of the draw count comes from. Most verts in the model are overlapping in UV. My understanding is that these verts are batched. I do now know how that affects the counts displayed or actual in game performance. Of course all engines are different with slightly different terminology:

The two numbers shown are taken from the TF2 model browser and are lower than comparable weapons. This has a 50% efficient LOD. There are weapons in the game that are comparable in their stats but with NO LOD. There are no weapons in the game with an LOD less than 50% efficiency.

If my understanding is wrong please enlighten me.

It could definitely be better and I may go back and optimize further. I would also like to darken the metal slightly.

[GregBoffins]

That's all well and good but I think you're reading your statistics incorrectly, or they're giving you the wrong numbers.

To have 12,000 draws is amazing. I don't even think Crysis on maximum settings has that many draws, no matter how complex your shaders are.

To have that on one weapon would make TF2 run at 20fps, it's insane

[Rick_D]

i'm guessing indexes is vertices and drawcount is triangles?

even that is pretty high for a weapon though?

[penE]

A couple of days ago I saw this pretty lame documentary in German TV about some of the most dangerous roads in the world. What caught my attention tho were those colorful Tata trucks they were driving around with. So yeah, thats teh story of why Im modelling an Indian Tata 2515 truck right now.

I'm almost done with the highpoly, I'll probably gonna take it into ZBrush next for some extra detail.

[cincinnati]

can't wait to see the paint job.

[Dejavo]

I'm no expert in modeling but that looks insanely awesome to my eyes. Love the amount of details and there is a lot of character.