Model made: Need help and C/O/C

[Brian Damage]

Okay, first bit of help I need:

I want to take a screenshot of the model inside the program... I know that I have to do something with the "PrintScreen" key, but I can't remember which key combo to press ... or even where to find the piccy afterwards...

So, once someone's told me that bit, I can upload some shots...

You see, this is my first ever weapon model, and I need some help and input from the immensly talented modellers on this forum... so I thought I'd take some shots in wings3D and put them up here, in order to point out what I want to ask about...

Oh, heck, I'll stop waffling now, shall I?

 

[Korgoth]

All you have to do is hit print screen, then just go into photoshop or whatever (ms paint if you have to) and hit ctr+V to paste and it'll pop right up, then just save it as a jpeg, bmp whatever and upload it here....

 

[Neutrino]

To take just a general screenshot just hit the button "Print Screen". Then go to your clip board and retrieve it. But isn't there an option in the modeling program to save the render as an image? I don't know much about modeling but I'm sure there must be.

 

[Brian Damage]

Ta for that... I'll have the pics in just a minute...

I don't want to take a shot of the render, I want a pic of the model in the editing area, to point out anything I need to ask about...

 

[Brian Damage]

Okay, here it is dudes...

It's a model of the disintegrator weapon I drew for my mod...

Took me a few hours in total to put together...

Okay, first two questions:

1. Does it all have to be one object? Presently it's fifteen. Do they all have to be joined together in one mesh?

2. It has to be transformed into all triangles, doesn't it?

 

[The Dark Elf]

Heh thats pretty cool actually! nice one Brian

 

[Korgoth]

It looks pretty cool man, As far as it being one solid object.. I think it does, but I could be wrong. If it needs to be seperate for animation I'm sure there is a way to do it. But 15 seperate pieces should probably be joined together (easier to process in game) And as for triangles, once you import it in game it will automatically be triangulated. What you need to do for the absolute best results is to keep it 3 or 4 sided when modelling. That way you know how the lighting will affect it in game and you will get better results. And to add to that, video game engines can't compute anything over 4 sided (they actually don't even compute 4 sided, as i said they automatically triangulate) So if you think about it a 4 sided poly is just two triangles together. hope that helps

 

[Stormy]

it dont have to be one object. all the different moveing parts should be different objects so that can be animated easerly but i dont think there should be 15 tis bit over the top might get abit complext.

it is most probly in triangles but only showing up as 4 sided pollys. they will all be made up of 2 triangles

 

[Brian Damage]

Thanks, fellas. So, does anyone know for sure if it all has to be one mesh (bar animated bits)?

I can easily turn it all into triangles or quads just by using the mega-useful tesselate function .

Hold on, I'll give it a go so that I can get a triangle count...

 

[Brian Damage]

Okay, it says that I have 2048 faces... that's within allowable limits, isn't it?

 

[Brian Damage]

Oh, hey, I didn't see your post fishymumma.

So, now I need to know how to merge all the immovable bits into one mesh...

 

[Nostradamus]

faces is not trienagles as far as i remember! you need to double the number.... i think that's logic....just look at it..that's far more than 2k tri's...

but nice model tho...however it needs a ton of optimization

 

[Brian Damage]

Well, when one object is selected it calls them polygons, when two or more are selected it calls them faces...

So, first I need to merge it, then I need to optimise it...

Any suggestions on how to do either are welcome...

 

[Mr. Coffee]

I believe the final result must be one object, but it doesn't have to be a cohesive mesh. You could simply select all 15 objects and merge them into one. The only thing you need to watch out for is any internal polygons, which should be deleted.

Animation is done by attaching the mesh to a bone structure, just like a character. Moving parts like a trigger or slide are rigidly assigned to a bone, and the bone is moved between keyframes.

Here is a tutorial on weapon modeling for the orginal HL. It should give you an idea of what it takes to go from a skinned model to a usable game model. http://www.planetquake.com/polycoun...ls/scarecrows_tutorials/weapons/weapons.shtml

And 2048 tris, before backface deleting, is perfectly acceptable. Just remember to factor hands into the polygon budget of 2000 - 2500.

BTW, a polygon (or face) is merely a multisided shape. Triangles, quads, and any other closed shapes are all considered polygons. In the case of HL2, however, a polycount should be taken after the model is tesselated into triangles.

Hope this helps

 

[Korgoth]

A triangle or a 4 sided poly could be considered a face I suppose... Actually the number of vertices means more than anything else, but thats a whole other can of worms. You really shouldn't worry about that to much. You do look like you could use a lot of optimisations. Any more than two faces on a flat surface is useless, if it doesn't help the shape and silhouette then get rid of it. To merge all the pieces simply weld the vertices in the right places, that should do the trick.

 

[Stormy]

any idear on the origanal HL polly counts ?

 

[Grey]

Ok... my 2 cents.

Leave it untriangulated when in wings. It's ok to once in a while select all of your faces and triangulate to get a count, then undo. Triangulate in whatever app you going to be exporting it for the game from.

As for the seperate objects. You can do this one of two ways. 1) Leave em seperate and combine them in whatever app your going to UV map in. 2) Assign each seperate object a material then while in object mode select all the parts, right click, choose combine. This is kind of a personal preference - but I find doing the basic uv mapping of each sub-object seperately and then when done doing that arranging them on a single uvmap to be easiest. In any case don't combine/attach any parts that will need to be animated(make your animators life easier).

Intersecting parts... this used to be a bigger deal because of low percision z-buffers. For the most part you don't want intersecting parts, but sometimes it's happens to be the best way to do something, and as you get better at modeling you'll be able to do this less and less. Sometimes you can just have the the two parts be as flush as possible with out intersecting which is the way to go if it's possible. What was said about deleting unseen faces is true, you can't do this in wings though, you'll have to do in another app, Wings meshes can not have holes (they must be solid). I'm not talking about holes like a trigger guard, I'm talking about holes that allow you to see inside the mesh.

Any way the model looks good Looks alot better than my first models hehe.

 

[Stormy]

does any one in the world know the polly conts of HL models i wanna make a model for CS!!

 

[Nostradamus]

id say about 700-1000 ...that way it wont be too much anyway.... i think -...try it out!

 

[Brian Damage]

Ok... my 2 cents.

Leave it untriangulated when in wings. It's ok to once in a while select all of your faces and triangulate to get a count, then undo. Triangulate in whatever app you going to be exporting it for the game from.

As for the seperate objects. You can do this one of two ways. 1) Leave em seperate and combine them in whatever app your going to UV map in. 2) Assign each seperate object a material then while in object mode select all the parts, right click, choose combine. This is kind of a personal preference - but I find doing the basic uv mapping of each sub-object seperately and then when done doing that arranging them on a single uvmap to be easiest. In any case don't combine/attach any parts that will need to be animated(make your animators life easier).

Intersecting parts... this used to be a bigger deal because of low percision z-buffers. For the most part you don't want intersecting parts, but sometimes it's happens to be the best way to do something, and as you get better at modeling you'll be able to do this less and less. Sometimes you can just have the the two parts be as flush as possible with out intersecting which is the way to go if it's possible. What was said about deleting unseen faces is true, you can't do this in wings though, you'll have to do in another app, Wings meshes can not have holes (they must be solid). I'm not talking about holes like a trigger guard, I'm talking about holes that allow you to see inside the mesh.

Any way the model looks good Looks alot better than my first models hehe.

Err... heheh, whoops, I kind of went and triangulated in wings, then forgot to save under a different name. Oh well. Looks okay.

Well, I was going to have a go at animating them myself... when I get my new comp, and install that free version of XSI, would I be able to import files from Wings into that? Wings can save in 3D Max formats...

I've still got all the objects separate... so, you reccommend putting a material on each of them separately, then combining them and UV mapping.. right... so, do I still leave the animated bits separate when I do that?

So, is there any particular command or function in Wings that allows me to merge the meshes of two objects into one solid mesh? Or should I just shift their points around until I get as close a match as I can to the connecting surface?

EDIT: Thanks. I feel that I should point out that it was mostly made from a set of really tweaked cubes, with the occasional mauled cylinder and miniaturised cone...

 

[clarky003]

kewl stuff brian, thats fantastic. keep up the good ideas and models.

 

[Brian Damage]

Hey, thanks clarky.

Still working on it, but been a bit busy for the past couple of days.

 

[NoisyMonk]

I have a question, its alittle off topic but I didn't want to start a new thread for it. Its for 3DSM. For some reason whenever I try and use a reference pic, the quality of it looks really crappy in the viewports, no matter how high res the pic is. For example, in this tutorial I'm reading, the pic they used looks like this...

And mine looks like this...

Can anyone tell me how to fix this? Oh, and BTW, nice model Brian :thumbs:

 

[clarky003]

hmm weird, must be the image setting in the program, . but does it really matter, I say aslong as you can follow the lines, no worries,

 

[random.hero]

yeah, theres a way to fix it, somewhere in viewport configuration i think. you might need to change your driver to OpenGL, but im cant really remember.