Flux poly limits

Second Chance (or anyone else who puts models into the game):

Have you done any experiments on models with high poly counts?

What's the highest poly count you've placed in the game without experiencing a slowdown?

No, I haven't.

The highest poly models I have at the moment are the B-Wing at 2,808 polys, followed by the A-Wing at 2,120 polys and the Star Destroyer at 1,914 polys. No problems have been reported so far, even in multi-player.

Actually, a long time ago, I tried to get a custom Klingon Academy model into EoC. It met with limited success. But I believe now that the fault was mostly my inexperience at the time. But it did work, and that model was several thousand polys altogether. With individual parts easily exceeding 1000 polys.

I think you're barking up the wrong tree for poly limits. Since each mesh is handled and displayed seperately (even when only a part of a larger model), the only poly limits you're likely to come across are those of an individual video card and not from the game engine. A good example would be someone's comp that can handle a battle with 20 - 30 ships on-screen at once, versus a comp that starts to choke at over 10. Maybe the second guy is running out of vid memory, or maybe his slowdown is because there are too many AIs at once and he's got no processor left.

I think that as long as you keep all your meshes under around 1000 polys, there's no real point to worrying about an individual model's total limit. Because a ship that runs smooth as silk on one machine may not even load on someone else's.

btw - Is this for something specific, or just a general question?

mailto:second_chance@cox.net
The Ultimate Guide To Modding: I-War 2 - Edge Of Chaos (on hold during SW MP mod)
cartoons.sev.com.au/index.php?catid=4
.

I think you're barking up the wrong tree for poly limits. Since each mesh is handled and displayed seperately (even when only a part of a larger model), the only poly limits you're likely to come across are those of an individual video card and not from the game engine.

Yes. But your answer of around 2000 polys for the more detailed models is pretty much what I was after. I did not phrase my question well.

btw - Is this for something specific, or just a general question?

Both actually. I was considering building a more-detailed version of the Dreadnaught corvette. One where comsec and pasties detach when the main body is destroyed.

But, it's just a consideration. I'll have to investigate the scripting that would be needed to make it work properly.

Thanks!

Sure, glad to help.

btw - The B-Wing is actually closer to 3,000 polys, so I think you've got some good room in going for detail.

mailto:second_chance@cox.net
The Ultimate Guide To Modding: I-War 2 - Edge Of Chaos (on hold during SW MP mod)
cartoons.sev.com.au/index.php?catid=4
.

Yes. But your answer of around 2000 polys for the more detailed models is pretty much what I was after.

Here is an idea if you guys want to gather more detailed statistics: Between all of us we have greatly varying hardware. A "known" set of models could be used (or created) with a varity of polys. They could be simple spheres broken into ever finer detail. From 500 to X thousand polys.

Given almost no internal systems, they could then be placed into the game. Everyone could report back thier system specs and the frame rate when they view X number of type X model. At that point you would know just how much FPS damage your new model is going to cause to each segment of the population.

I would be glad to run some tests. Since I only have one game computer, and it is pretty fast, it would be helpful to have some idea of just how much stuff can be tossed around stations before people are playing at a crawl. I don't want to block people out because they have other life priorities than hardware upgrades.

Glow effects and transparency would also be interesting to understand the FPS effects. Hardware does not seem to scale evenly with respect to effects. Sometimes it seems to bound ahead - other times, not at all.


-Gtrout

Lots of variables. Texture resolution would effect the frame rates... also graphics-quality settings (lighting quality, etc.) the user currently has active would make a dent in the framerates.

But, I'm willing to try if others are. I'll make the spheres. I run a 1.1 GHz, 512 Meg Ram, with NVidia GeForce MX/MX 400 (64 Mb video ram I believe). Given the age of my system, I'd expect it to represent the lower-end of the scale.