Does anyone know if blender is better at handeling models or textures? I’d just like to know which one is more efficient so that I can apply that to my work.
Both I guess?
Thanks for your reply OTO. Does anyone know an answer for sure, or have strong evidence supporting their opinion?
Using a high resolution normal map on a low-poly object would always be much faster than creating all that detail using raw geometry, so in many cases using textures as a way to give an impression of more detail would be a lot faster as seen in large numbers of games dating back to the PS1/N64 days.
However, you always need a model with more or less detail to map the texture to, and while it does seem that the BGE has a fairly high limit when it comes to rendering textures at interactive frame-rates on a single scene on good hardware, the total amount of textures you can use at various resolutions is still dependent on the video memory available in your GPU.
Using polygons for detailing can give a more effective impression in some cases and could save you the need to make some basic normal maps to cover those details, increased polygon density also allows you to make more use of vertex colors to color a material without needing a special texture for it, but that would be for large scale cases like some basic road markings compared to say, graffiti on a wall if you don’t want to push the polygon limit past the point of the scene being real-time.
Your question is in that way curious as it requests to compare things that are strongly bound together.
I guess you want to know how to design models that consume less processing power while running the game ;).
Ace Dragon gave you advice already and I think you will find much more.
Finally it is a balance between artistic demands (high details) vs. technical possibilities (fluid game run).
In my experience, I’ve learned the following:
Geometry will some RAM and rendering power(Rasterizer / Scenegraph++)
Textures will take some GPU RAM and rendering.(Rasterizer)
So, since both come as a bundle, your total consumption will be the sum of both.
It’s true that low poly meshes with more textures will give better results, but the old advice is: Use more small textures than few bigger ones. Balance the polycount and the textures resolution according to your game(max. triangles and textures to be rendered at any time).
Don’t forget that a game is not only about graphics. The physics will kick in as well, so your polycount must be not only for textures, but for animation and physics as well.
Well, this is a thing you need to study and test it for your case, in anycase.
Thanks a lot guys! I appreciate it.