|
||||||
|
|
|
|
|||
|
|
|
|||||||
| Register | FAQ | Forum Rules | Members List | Calendar | Mark Forums Read |
![]() |
|
|
LinkBack | Thread Tools | Display Modes |
|
|
#1 (permalink) |
|
New Artist
![]() 5
- 0
|
few questions about game art and workflow
Heya!
Im new to this forum. I've decide to try and make some game models, will probably start with something easy like weapons, props etc. Anyway I got some questions about workflow and things that can be problematic. To make things easier for you Ill try to discribe problem and write a possible solution that might work so you can wirte if Im thinking right or not, whats the better solution etc. Oh and Im using 3ds Max, Zbrush, Photoshop, also got xNormal and maybe Im gonna buy CrazyBump beacouse it seems to be cheap and super usefull. 1. Hard edges. Ill split that into to parts: a) organic models (humanoids, creatures, flora and fauna but also I consider fantasy weapons, armors etc as organic if you know what I mean) So obvious for most parts "pinch" brush in zBrush will do the trick but for example what with weapons like swords (blade), hammers when hard edges but be very harsh. Possible solution: Duplicate base mesh in 3ds Max and make a little bit higher model with additional edge loops so when imported to zbrush and subdivided it will keep shape and nice crisp edge. b) hard surface like sci-fi stuff, real world stuff etc. For flat surfaces like walls, doors and overall environment stuff making normal maps from texture in CrazyBumb, Photoshop or xNormal (if possible) should be fine right? But again what with some kind of futuristic range weapons for example? "Sculpting" in zbrush hard surfaces is kinda wierd and confusing. It can be done and I know that but maybe there are other solutions, maybe something like: Possible solution: Duplicate base mesh in Max, make highpoly version in Max as well. Then export both low and high and make normal map in xNormal. Eventually paint some smaller details in zbrush on flat plane, make normals and combine both normal maps in Photoshop. 2. UVs. Good UVs are very important but when is a good moment to unwrap your model? Answer seems to be simple: after base mesh is ready/before importing to zbrush but... what if my model will change proportions a little bit while sculpting so my low poly model with "perfect" UVs will be a little bit diffrent? Should I tweak UVs or this small changes wont be noticable? Also what if Ill have to optimize my low poly mesh after sculpting? I need quads in zbrush but when sculpting is done I dont need quads anymore right? So if I start to removing/adding edges in my low poly mesh will that change UVs as well? Will my baked normal maps works fine on otimized mesh? Or Ill have to unwrap again, bake normal maps again etc.? 3. Baking normal, cavity, ao maps. Just a 2 quick questions here: Lets say my low poly mesh cannot be changed at all but of course it will change by sculpting, moving points etc. So how do I solve that? Bake all maps on changed low poly mesh and hopefully it will be fine or re-import unchanged model on lowest subdivision level and then bake maps? Where do you bake your maps? Z-mapper, xNormal? xNormal seems to be good, very simple and free soft but maybe there are some cons of baking maps in it? What with cavity, ao maps? Are they OK based on your experience with other possibilities? 4. Color, specular, emissive maps a) Color map Any golden rules for making color maps? I heard somewhere that color map shouldnt be to saturated for some reason, not sure why and if thats true. b) Specular map I saw many peoples are making specular maps colored, why? What it exactly do? Are this colored maps used in both "specular level" and "specular color" so grey values of colors controls how much specular we will see? c) Emissive map These are just self-illumination maps right? Do they have to be perfect or they can be brightened/darkened in game engines if needed? 5. Presentation Of course after hard work you want to show your finished model with all maps, fireworks and stuff So what we got is: - render - pretty straightforward, assign maps, add lights, hit render - viewport screenshot - How to do that? Ive tried with simple plane with some textures on it and it didnt work. I have Direct 3D as a display driver, got latest DirectX 9, Ive tried DirectX shader but it didnt worked. Which *.fx file are you using? What kind of lights are working fine (if any)? - xNormal viewier - I saw that Diablo model from examples and it seem to be good, what you think about it, did you tried it? Thats it for now. I know most peoples dont like text walls but I hope Ill find answers here. ![]() Regards marlev |
|
|
|
|
|
#2 (permalink) |
|
Founder of Statspaddling
|
1. Hard edges.
(a) Not quite sure what you mean I'm afraid.. but to get nice hard edges without a lot of segments, simply use Smoothing groups ![]() (b) Aye, sculpting things like weapons etc. with hard edges usually isn't a good idea. As you said, creating a highpoly and painting in/generating normals from a heightmap will give the best result. 2. UVs. Good uv's is indded important. Personally while sculpting a model, I start off by doing the lowpoly and import it to ZBrush and start sculpting right away. Once the sculpt is done, I go down to the subdv my final model will have and export it, then I take it into max and unwrap it. After I've done that, I export it from max with the UV's, and import it to ZBrush at the same subdv. This will work perfectly and adds your UV to the model while still containing all the sculpt-detail. It'll however fail if you edit the model before importing it to ZBrush again, as it won't recognize it. As it comes to optimizing, you could optimize it by creating a new model in your desired 3d app, and then take it into ZBrush and retopoligize. 3. Baking normal, cavity, ao maps. Not sure what to day about that really, never used XNormal either. But I usually bake all my maps in ZBrush if I'm sculpting something. Dunno about AO and Cavitymaps though. 4. Color, specular, emissive maps (a) It all lies in pratice and recieving feedback, aswell as comparing and learning from others textures. A strong saturation isn't bad at all, depends on what you're aiming for. In a catroony texture for example, you'd want your texture to be very saturated. (b) It's as you said. The cooured specular map is used both in the Specular Level and Specular colour slot. The brightness defining the amount of specular in the specular level. And the colour adding colour to the specular. As for a copper texture; it wouldn't look like copper if your specular were to be white. (c) They're almost the same as an self-illumination map.. the only difference is that from an emissive map, light will actually radiate. If they have to be perfect or not comes down to the engine itself. For UT3, they don't have to, as there's a lot of options to edit and animate your maps in the engine. It'd be better if it's as good as possible from the very beginning though. 5. Presentation I'd just go with a screengrab, using Doylle's shader, or the standard DX Display. Preferably Doylle's shader as this handles the maps much better and has some filtering which the DX Display lacks. There's a description on how to use it aswell. And about xNormal, as I mentioned earlier I never used it. Hope this helps, and you better be grateful, cause my fingers are aching already! ![]() Cheers!
__________________
|
|
|
|
| The Following User Says Thank You to Tiros For This Useful Post: |
|
|
#3 (permalink) | |||
|
New Artist
![]() 5
- 0
|
Thanks a lot! Your post gave me answers for most of my questions
but also born a new ones (not many I promise :P)Quote:
Quote:
Quote:
I checked your tutorial about creating specular maps. I learned something new beacouse I didnt know that colors in specular map should be inverted. That kinda explains why I saw blue specular map for human skin :P Anyway I thought Ill thanks for that tutorial here as well ![]() |
|||
|
|
|
![]() |
| Bookmarks | |||
Digg
|
del.icio.us
|
StumbleUpon
|
Google
|
| Thread Tools | |
| Display Modes | |