[Max] Several questions concerning modeling edicate
Hey guys! I've been struggling with some of the concepts I've learned about games. I figured it would be a good idea to ask them all at once. So here it goes:
1) Whats a decent poly/triangle count for a high and low model? I know this depends on the model but I guess I just don't understand how much a regular PC can handle, or a phone for that matter. What about a multiplayer game?
2) I'm attempting to develop a game that will have many models on screen simultaneously, I don't know how to budget my tris
3) Do you make the high poly model first? I've heard some controversy over this matter.
4) What's the difference between an AO and a normal map? I know they both bounce light on the model to define depth but how are they different? What other maps do you guys use?
5) Should every model have an AO and a normal map? A high and low poly? What would be an exception?
6) Whats the most optimal units to work in? Or what do you prefer for that matter, and why?
Here's a model of a table I made:
7) Should the legs even be connected to the bottom of the table? That just creates more triangles - I don't see how it would be an issue if they were not connected
8 )How would I go about making my higher poly model, or would this be considered the low poly? I've read some of EQ's tutorials but since these aren't cylinders I expect that the 90 degree edges are not a problem?
Here's another screenshot from a different angle:
Any other critique concerning my model would also be greatly appreciated.
Thanks for putting up with my novice questions
and if you guys just want to drop me links to threads I should read that's fine too!
1) Whats a decent poly/triangle count for a high and low model? I know this depends on the model but I guess I just don't understand how much a regular PC can handle, or a phone for that matter. What about a multiplayer game?
2) I'm attempting to develop a game that will have many models on screen simultaneously, I don't know how to budget my tris
3) Do you make the high poly model first? I've heard some controversy over this matter.
4) What's the difference between an AO and a normal map? I know they both bounce light on the model to define depth but how are they different? What other maps do you guys use?
5) Should every model have an AO and a normal map? A high and low poly? What would be an exception?
6) Whats the most optimal units to work in? Or what do you prefer for that matter, and why?
Here's a model of a table I made:
7) Should the legs even be connected to the bottom of the table? That just creates more triangles - I don't see how it would be an issue if they were not connected
8 )How would I go about making my higher poly model, or would this be considered the low poly? I've read some of EQ's tutorials but since these aren't cylinders I expect that the 90 degree edges are not a problem?
Here's another screenshot from a different angle:
Any other critique concerning my model would also be greatly appreciated.
Thanks for putting up with my novice questions
and if you guys just want to drop me links to threads I should read that's fine too!
Replies
Ambient occlusion = some "generic shading" which looks kinda like an overcast day which helps make models look more solid by adding shadows in the places we expect them to be generally.
I'd not bother joining the legs to the bottom; most of the time that table's gonna be viewed from above, and the triangle will likely be backface culled as a result so the slight fillrate cost of not having holes in the bottom is negated anyway.
Games don't work in any specific units; there are of course exceptions like simulators, but generally stuff is sized so it plays well and doesn't look wrong. As long as your table is the right size for your chairs, which are the right size for the characters, it'll be good.
You could optimize this a little more but this is generally about low poly. Look up subdivision modelling - you make a model specifically designed to smooth nicely using turbosmooth by adding triple edge loops to support the shape.
Thanks, this really clears up many of my questions.
-Separate the top of the table from the legs and get rid of unnecessary construction edges..your table only really needs 10 faces if you could fake the top with a normal map (you could even bake a normal map with the current table top's geometry and just use that on a 10 sided table top)
-You could remove the edge loop that runs down the center of the legs and you would regain a good bit with minimal loss to the contour of the shape
-Unless this table is going to bounce & bend around like a cartoon, there is no need to have the model as one piece, you can just combine separate objects into one and that will save you some geometry.
Which brings me to say one last thing..the reason you want to model in quads is for triangulation and deformation purposes...whenever you hit triangulate the operation goes smooth because it is just cutting each 4 sided polygon in half so there is a flat surface to stretch the Normal,AO,Diffuse..ect...across...when you model with triangles up front you are making that process much more complicated mathematically for ray tracing and interpolations or deformation....when something is animated the edge flow is really important. Until you really understand why you are doing some things with the geometry....try to make everything with 4 sided faces and you will be on the right track.
-Keep It Up!-