View Issue Details
ID | Project | Category | View Status | Date Submitted | Last Update |
---|---|---|---|---|---|
0000563 | FSSCP | graphics | public | 2005-09-23 11:06 | 2008-02-22 07:37 |
Reporter | SadisticSid | Assigned To | taylor | ||
Priority | normal | Severity | minor | Reproducibility | always |
Status | resolved | Resolution | fixed | ||
Fixed in Version | 3.6.10 | ||||
Summary | 0000563: FOV becomes warped for the death screen | ||||
Description | It's been the case for probably more than a year now, but when the player dies the FOV for the death screen changes (looks like a widescreen image crushed on to a 4:3 screen). This causes some strange visual artifacts like ships suddenly disappearing off the screen prematurely - I'd assume that the game thinks the ship shouldn't be in the FOV and decides not to try rendering it at all. The same also goes for weapon effects disappearing before they reach the edge of the screen. Ships are also warped in various dimensions depending on what angle you view them from. | ||||
Tags | No tags attached. | ||||
|
Is this with OGL and/or D3D? I know it's a problem with OGL but just haven't done anything about it yet. If it doesn't happen with D3D then I'll just take this one, otherwise I'd want to coordinate a fix with Sticks. |
|
I can confirm weapon effects disappear before they reach the edge of the viewable area in D3D, and background bitmaps become stretched out of their original scale too when the camera focuses on a ship going by them. The effects are clearly seen if you bump up time compression. |
2005-10-08 16:41
|
|
|
This FOV problem is also noticeable when using the set-cutscene-bars SEXP. Try out the mission, as the cutscene bars fade in, the fov changes. Once the camera turns back around to face the ship, make it fly straight up, and you'll see it disappear before reaching the edge of the screen. The same goes for all directions. 3.6.7 Linux Build. edited on: 10-08-05 12:43 |
|
Anyone? Bueller? |
|
Haven't had another chance to look at this. I probably won't have time for quite a while either so if someone else wants it then feel free. |
|
I'm almost certain it's a side effect of gr_set_clip, but I don't know what causes it. |
|
At least on the OGL side what I think it's when set_clip is called (looked fine to me), the viewport is changed (looked fine to me), the frustrum is changed (properly as far as I can tell), but it could be the modelview/projection matrix getting messed up somehow. I have never had time to look any further than that, though I never ruled out some clip/viewport issue either. Don't know if that's the D3D problem as well since I'm not sure how the code works on that side. I think it's just some math issue or perhaps call order. |
|
This has something to do with 2d drawing. I can pretty much fix this one problem, but not without breaking something else (mainly the HUD targetbox). Still a work in progress though, haven't given up yet. |
|
Ok, 2D related problem is fixed now, as is a projection matrix fubar that I only just noticed. The death stuff still isn't 100%, maybe something wrong with the frustum, I guess. I'll keep working on it, but at least it's been identified as two different problems rather than just the one. The camtest mission works properly now though. |
|
This is partially fixed in CVS now. The distortion is fixed, but the object culling is apparently unrelated to it. I'm looking into the culling problem, then this will /finally/ be fixed. :) |
|
Ok, found the second half of the problem. It's the game's own object culling which makes things not draw. The game math is based on it's own view setup, not what the OpenGL view/proj matrices are set to. This is why things get culled, OpenGL has a wider/larger view than the game math is using. Outside of making an API specific culling function, I have no idea how else to fix this. For now I'll just go with an OpenGL culling function (when I get it working properly) and have it used instead of the game culling math when OpenGL mode is used. |
|
The second half of this should be fixed thanks to Murleen's patch. :) Fixered. |
Date Modified | Username | Field | Change |
---|---|---|---|
2005-09-23 11:06 | SadisticSid | New Issue | |
2005-09-23 11:30 | taylor | Note Added: 0003457 | |
2005-09-24 15:17 | SadisticSid | Note Added: 0003461 | |
2005-10-08 16:41 | Fractux | File Added: CamTest.fs2 | |
2005-10-08 16:41 | Fractux | Note Added: 0003504 | |
2005-10-08 16:43 | Fractux | Note Edited: 0003504 | |
2005-11-29 23:34 | phreak | Note Added: 0003950 | |
2005-11-30 04:53 | taylor | Note Added: 0003958 | |
2005-12-14 00:45 | WMCoolmon | Note Added: 0003988 | |
2005-12-14 01:54 | taylor | Note Added: 0003989 | |
2006-05-22 18:22 | taylor | Status | new => assigned |
2006-05-22 18:22 | taylor | Assigned To | => taylor |
2006-09-09 02:57 | taylor | Note Added: 0006565 | |
2006-09-21 11:36 | taylor | Note Added: 0006660 | |
2006-09-24 12:21 | taylor | Note Added: 0006701 | |
2006-10-14 20:15 | taylor | Note Added: 0006890 | |
2007-02-27 16:38 | taylor | Relationship added | has duplicate 0001282 |
2008-02-22 07:36 | taylor | Relationship added | has duplicate 0001563 |
2008-02-22 07:37 | taylor | Status | assigned => resolved |
2008-02-22 07:37 | taylor | Fixed in Version | => 3.6.10 |
2008-02-22 07:37 | taylor | Resolution | open => fixed |
2008-02-22 07:37 | taylor | Note Added: 0008891 |