Sorry your browser is not supported!

I know it might seem like an odd question. but in a setting such as a single urban street, would invisable segments lets say blocking off the top of the fenceline plus an invisable segmented ceiling stop the lag that usually caused by open areas?

if not what are some suggested methods?

Sometimes it does, but it's not significant, I think...

An interesting question, so I did a small experiment. Nothing conclusive and it would have to be taken much further but interesting none the less.

I made a long, straight, double-wide hall. On one side of the hall I lined up 10 characters (Medic) with original scripts and in dynamic mode. I then place the player at one end of the hall.

In each scenario, I paced the hall from one end to the other multiple times and averaged the FPS.

Results:

1.Medics on one side of the hall, me on the other: FPS=29
2.Invisible wall the length of the hall between the medics and me: FPS=28
3.Full wall the length of the hall between the medics and me: FPS=31
4.Full wall with windows between the medics and me: FPS=23
5.Full wall with doors between the medics and me: FPS=25

Thoughts:

A lot of this is already known but I found some things to be interesting.

1.Just the result of the number of characters in a level. The fact that they were all running the medic script didn't seem to affect FPS, but only the sheer number of characters.
2.It may be that it takes more processing to render through an invisible wall than no wall at all.
3.No surprise here. Entities not rendered through solid walls.
4.Rendering thrugh windows definitely take more processing.
5.What I found (again, we already knew but I never saw it before) was that it didn't matter if the doors were closed or open, it still took the same processing and reduced FPS.

Again, this is small and non conclusive but verifies a lot of the recommendations we hear concerning level design. Don't use long halls, but make turns to keep distant rendering to a minimum. Windows and doors greatly reduce FPS.

The question at hand however remains. This tiny test showed little to no change between no walls and invisible walls. It would take more testing on a larger scale.

Best.

xplosys, thanks for this test. It does prove a lot of things

Doors, windows (or for that matter an invisible wall) do have to render everything on the other side of the wall. So obviously the frame-rates remain low.
What puzzles me is the difference in framerates : 23,25 and 28 (for windows, door and invisible wall respectively) Since they all make the game render everything on the other side, why this difference? Also, if anything, the order should be the opposite since more stuff can be seen through an invisible wall than through a door. And windows are even smaller openings. funny.

Also, all these have worse framerates than now having them at all. So I guess its better to have no walls rather than have walls with openings in them.

Thanks again for this xplosys!

Characters are still the main issue in serious lag in outdoor areas and you cant have a shooter without them. We need DarkAI to remove that influence then we can concentrate on the separate Serious Lagg Issue without having to account for characters influence.

As with most serious issues inside FPSC TGC will be the best people to help due to the complexity of FPSC and the BSP compile process which determines the poly counts from any given position and view and the resultant fps returns. As FPSC is very inconsistent and erratic in this it will be very difficult for users to successfully find a solution. Unfortunately I dont believe the Serious Lagg Issue is likely to be tackled any time soon if ever by TGC.

When DarkAI is a reality it may be easier for us to try at that time to find a solution at least that may provide some help and I am sure some users will eventually discover workarounmds that will at least go some way to partially improve things somewhat. I dont see any way myself at the moment but will have a go when I get around to it after DarkAI integration.

Who knows too if current development updates may help - I guess its always worth testing for this issue and any improvement as we go if anyone has the time.