Quote: "If you get really low framrates in certain areas and looking into certain directions then the game engine is trying to render too many polygons from that view point. You can see how many polygons are being rendered at any point during the "test" by looking at the first number from the left that is displayed on the bottem of the screen."
Correct - this can be because FPSC sometimes appears to calculate polygon counts incorrectly and is one of the reasons that slowdowns occur when they should not.
There is no such thing as a leak in FPSC levels - though this effect could be described as such as the effect is similar to bsp engines where leaks do occur - in the same fashion as with bsp leaks the problem can be cured where such slowdowns occur I have found by looking to see if you can ascertain why and what mught cause a "leak" to occur and encompassing the offending area with a surronding completely closed and sealed outer room structure.
Some things which I have found and others have reported that cause slowdowns (reduced fps) are default ducts and corridors and also walls and such like inserted as static entities (FPSC reads through them) and not as segments. If you have any of these in slow areas - enclose them in an outer envelop of standard rooms segments and although technically you will have more walls to calculate the fps should go back to a stable 30/32 fps. It does with me.
The excessive polygon counts are seen by FPSC it seems when polygons even around corners and way out of sight are calculated by FPSC tracing and calculating polygons it sees as in view - via any unblocked or sealed route. This means that its particularly difficult to achieve a workaround in outdoor levels where FPSC could concieveably calculate every polygon in the level from any offending position.
FPSC V1 as with EA is still cabable of calculating widely varying and erratic polygon counts - i.e. from one part of a level x ploys and from another where you may think there is much less in view (and there is) a much greater number of polys.
Who knows exactly why this occurs/I doubt anyone knows the answer.
In indoor areas you can certainly have what could be described as very large open rooms with much content and action going on as long as you follow the above guidlines and close off any apparent leaks that crop up. In outdoor areas the answer seems to be much trial and error to get stability of fps - whether or not you have a great deal in view is not always the defining cause or answer due to the nature of the way FPSC seemingly erractically calculates what actually can be seen. Certainly it does not always see what you as the player sees in its polygon calculations so you have to go with the numbers it returns and try to workaround problem areas.
Everyone is aware that FPSC could do with some help in the area of fps and the erratic poly counts that can crop up that drag the game seemingly unecessarily in certain locations can be a level killer. In outdoor levels designing them out can be successful - occasionally - it wont work so you are stuck with no option but to try removing content, but be warned these areas can move around as you remove and build elsewhere so you could end up going around in circles, rather frustratingly.
An example of one form of this erratic behaviour is a City level I have where when looking directly at a single wall close up with nothing else in view at one side of the level I have unacceptably slow fps (14fps) when I cant see anything else - just the bricks in front of me. On a rooftop with most of the level and polys in view from above I have perfectly acceptable frame rates - If I add a couple of roof segments to a flat topped building in view the fps drops dramatically by around 9 fps and thats a lot when an engine has nominal fps at the best of times.
In outdoor levels the further you move from the centre of the world towards the extremities the greater the chance of high poly counts and slowdowns.
Thats what I have found anyway - hope that helps someone a little.
I would very much agree with Daniel Silverman whos opinion I respect very much.
