[KARMA]

1: performance hits on background images.

I tryed the same 1024x1024 image in 2 versions: 8bit pcx and jpg full color (they were IIRC of the same order of size in kb)
While the 8bit pcx was working fine, the jpg caused a major performance hit (btw you know my system is crap..).
I tryed the same jpg image on a skybox (more than 500 triangles):
The skybox was mapped with two textures, one was the same background image I used above, another one (jpg 1024)was tiled 5 times over the rest of the skybox model.
The performance hit was IIRC half of the loss with the background image.
Since it was the same JPG image, maybe there is something specifically wrong with background images when using jpg images? I remember many reported some performance losses with some versions of lightspeed's planet, it may be the same problem?

Second:
I had a weird bug with skybox in 1_20, which I never had before.
When I move (even linear movement) the skybox "shakes", which caused me a terrible headache:).


request:
would it be possible to set the minimum % of stars generated by FS2 as 0% (no stars) instead of 10%?. There are issues with some kinds of background images (stars in the images stay still, stars generated by the game have motion blur)

[Lightspeed]

Originally posted by KARMA
Since it was the same JPG image, maybe there is something specifically wrong with background images when using jpg images? I remember many reported some performance losses with some versions of lightspeed's planet, it may be the same problem?

no.

no.

The shaking skybox thingy is the remnant of the 'skyboxes vanishing' bug - it is known and will be sorted out eventually.

[Bobboau]

and eventualy means before 3.6

[KARMA]

btw I don't see anything strange with an 8bit pcx being rendered 4+ times faster than the true color version jpg,  althought the jpg is smaller in kb size.

I find instead at least pretty curious that the same jpg image on a skybox is rendered 2+ times faster than the SAME jpg as background image.

I thought it was going to happen the opposite: the skybox has a second jpg image of about same size/res, and many polys to be rendered. And I don't have an HTL hardware, so the skybox could be even faster.
The only counterpart is that the performance loss of the skybox is always present , the performance loss of the background image is present only when you face the image.

BTW not being a coder it sound pretty strange to me this performance difference between background images and skyboxes when using the SAME images...

[J3Vr6]

I had a problem with 1_20...  Whenever a ship warps out, it shows up on the other side of the warp hole before it disapears.  So the warphole opens and the ship starts to go thru it but comes out the other side.  And once the whole ship is on the other side of the warphole will it disapear.

I don't know if this a known bug or not so I won't mantis it til someone says so.

[KARMA]

the phantom tie!:lol:


when zooming out it seem like if it intersect the background sphere. Is this normal?

[kasperl]

looks like z-buffer jokes, dunno for sure. is your skybox large and far away enough? OGL/DX? what are your specs? any additional flags?

just a list of standard issue questions, it might be usefull for a coder to have the anwsewrs to those.

[Bobboau]

1_20 I think still had the z buffer problem, it's been fixed, though skyboxes them selvs still jump around quite a bit

[KARMA]

mmmm I really thought that with the skyboxes it doesn't matter the size....ops:p

[Bobboau]

it doesn't, when it's working corectly