Hey everyone. This is my first post here on Polycount. I'm currently in the process of making a game-res model of an Xm8 Machine gun, and would really appreciate some feedback on what i've got so far.
I started by modeling the high-res asset, then modeled the low poly model around my high poly. My next step, after getting the low poly model finalized, is to layout uvs, and produce a transfer map in Max to create the Normal for the low poly. Any suggestions, comments or critiques are pleasantly welcomed
.
Low Poly Model WIP screens:
Replies
Here's a little tutorial from http://racer445.com that's pretty useful.
As far as the cylinder thing goes, I don't think that would actually save enough faces to make it worth it:
I only save two tri's doing it this way, and also lose the 'rounded' look of the scope's glass eyepiece.
I guess I should've mentioned that it's sitting right around 5k tri's at the moment. Any other suggestions?
@Predator - I definitely see your point. I only need the point in two of my cylinders, the ones in the scope, the others can be changed. It does add up, but in this case, not to very much. I appreciate the guidance!
Looks great, hoping to see how well it turns out.
I'm glad you suggested it... I'm so tempted to, I think this is most likely the size I'll go for.
@ Pior - thanks for the tip, I hadn't thought of that before.
I dunno about tvs not being able to use them though. I mean, they can't view a full 2048 texture at one time, at full res, but a 2048's still gonna be crisper than a 1024 when you get real close to something in-game.
BUT, 2048 is pretty big for a gun... maybe I should rearrange the uv's to fit in a full 1024X1024...
If its a FPV weapon, in most games it isn't simply a gun or just another proper, it IS THE character in the game, its how you interactive with the world, and its the most important character in the game in most cases. The thing that will always be on the screen, 100% of the time, why would you ever skimp on the single most important asset in a game?
Yeah, just to elaborate on what EQ said...
I'm guessing what you mean by this is you think that because the resolution of the HD video signal isnt 2048 in either dimension that anything approaching or beyond that is a waste.
However, what you're failing to consider is that portions of this model may fill the screen or be zoomed in on or otherwise be composed in a scene such that higher detail becomes necessary.
Besides, this is a portfolio piece, so aiming for higher specs helps to future proof it.
Additionally, it makes good sense to author high and scale down as needed during production. Besides, those higher-rez maps will always get used in marketing and pre-rendered video.
There is, or should be, imho, a disconnect nowadays between authoring the asset and what actually winds up in the game - technology is changing and memory budgets for a game are in flux until the game ships, its better to be able to easily adapt by crunching some textures down a power of 2 or rebaking to a smaller/new UV layout, etc.
The worst thing in the world is when you wind up authoring an asset and then later specs change and you find you could have made it more detailed.
Here's the final bake from max, plus a quickly crazybumped specular map, with a little hint of AO as well. I'll be overlaying another normal map once I finish the diffuse color map of course, to get all those gritty little details (such as the gun logo, etc).
any suggestions are helpful. I'll be back soon with texture work!
I should probably also say that I'm planning on texturing the clip and barrel similar to the shell ejector door on the side, and that long, scraped up piece just under the handle on the top of the gun.
Thanks!
MY PORTFOLIO
By looking at the normalmap I believe it might be that you're using crazybump to overlay some normalmap shapes or something. Seems like you've been using shape recognition which is causing gradients all over.
Could you post a version with the raw, unprocessed, baked normalmap?
If you send me your Low poly in a PM or email i could take a look at it and see what's wrong.