View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0001789 | FSSCP | graphics | public | 2008-10-07 06:03 | 2008-12-19 06:14 |
| Reporter | Mongoose | Assigned To | |||
| Priority | normal | Severity | minor | Reproducibility | always |
| Status | closed | Resolution | not fixable | ||
| Product Version | 3.6.9 | ||||
| Summary | 0001789: Certain particle effects and animated textures become garbled | ||||
| Description | Certain in-game graphical elements, including warp animations, damage particles, thrusters, and certain beam weapons, are showing up as distorted or garbled in-mission. In the case of the thrusters (particularly afterburners) and particles, the normal texture seems to be replaced by whitish squares; the warp animations turn into a flickering rainbow mess. The issue isn't present when I first start a mission, but they "kick in" at some point during one; the framerate slows down somewhat, my hard drive starts reading/writing more than usual, and the glitches appear. Sometimes it'll take a few missions played straight through to be triggered; at other times, it'll happen in the first mission I play. Quitting the game and re-loading it seems to reset the problem. I've attached a few screenshots of the various manifestations of this, although I don't have a shot of an affected beam; at least one of the Shivan beams displayed garbled animations down its length while firing. | ||||
| Additional Information | Running on ATI Radeon X300 64MB card with the latest Catalyst drivers, although the same problems were occurring even before I updated the drivers. I've witnessed the issue on several of the nightly builds going back to Sept. 15 or so, including the most recently-released build. It does not occur under 3.6.9 final. I'm running the 3.6.10 beta VPs. | ||||
| Tags | No tags attached. | ||||
|
2008-10-07 06:03
|
|
|
2008-10-07 06:03
|
|
|
2008-10-07 06:04
|
|
|
2008-10-07 06:04
|
|
|
|
Just wanted to add that this issue seems to usually manifest itself in missions with a large number of ships on-screen, particularly when there's a sizable amount of weapons fire. I can go through several smaller missions in a row without it occurring, but as soon as I hit something more substantial (think "Number of the Beast" from Derelict), it'll show up. It's usually accompanied by a noticeable reduction in framerate, and I've been noticing audio crackling occurring that doesn't take place in normal situations. I'd love to be able to track this down, if someone could give me some sort of idea about how to possibly do so. |
|
|
It is probably not related, but for the slim chance, I'll add this note anyway. I had the same problem with the new Maxim impact effect. I was able to fix it by reducing the fps in the EFF file to 30 instead of 60. |
|
|
While this is up here, the same issue seems to be cropping up with static textures as well...or, to be more specific, at least the shinemap portion of them. I've attached a shot of a rather fetching absurdly-colored Orion; I also managed to snag a shot of a beam cannon firing. As I said before, I never experienced any of this while running 3.6.9. I'll try fooling around with a few of the more recent nightly builds, just in case anything changed; I've been playing on the build from October 6th during all of this. |
|
2008-10-24 04:33
|
|
|
2008-10-24 04:34
|
|
|
|
@DaBrain: Can you attach the new impact effect to me to test with (or just PM me with a link)? |
|
|
That's a fairly low amount of video card memory to be running the MediaVPs. I'm not saying it should do that, but my bet would be that it's related to that. Especially considering it happens when there are many items being paged in for use at one time. It could also be a hardware issue, such as a faulty memory chip near the end that only gets used when the card is 'filled up'. Personally, if I had to guesstimate system requirements for the MediaVPs, I don't know if a 64mb card would cut it nowadays. I'm not an expert here though, and if it was working prior to a certain date just fine then I imagine it's at least partially a code issue. Have you re-tested older builds to make sure they still work fine? |
|
|
I've been running the MediaVPs on this card for a few years now without major problems, other than the occasional stuttering slowdown during intense missions. In fact, most of the recent 3.6.10 builds have given me a marked improvement in performance, other than the bug at hand. I'm going to run some of the worst-offending missions under 3.6.9 final, just to see if anything happens, but I've never experienced anything wrong in the past while running that build. That's not to say that it isn't a hardware issue that serendipitously decided to manifest itself now, of course. |
|
|
Yeah, try making sure that you can't reproduce those still on 3.6.9 with the 3.6.8 mediavps, and also perhaps try some of the older 3.6.10 testing builds. Some trunk builds that were pre-shaders maybe, around May or June, and some just after July 12. Most of the links in the recent builds forum should still be working. |
|
|
Okay, so after doing more repetitive testing than I ever want to do again, I came up with some interesting results. I tested all of these build/VP combinations on the Derelict mission "A Paper Wall," since it managed to trigger the bug consistently every time and doesn't modify any of the relevant game elements. The following builds all fully triggered the bug, including both corrupted beam and shinemap textures, as soon as the Nyarlathotep (a Lucifer) jumped in near the end of the mission: --Chief's July 23 3.6.10 build w/3.6.10 beta VPs --Oct. 6 build w/3.6.10 beta VPs --Oct. 24 build w/3.6.10 beta VPs The following builds triggered the bug only on shinemaps, not on beam textures, when the Nyarlathotep jumped in: --3.6.9 final w/3.6.10 beta VPs --Swifty's May 28 3.6.9 build w/3.6.10 beta VPs --WMCoolmon's June 29 3.6.10 build w/3.6.10 beta VPs The following builds showed no evidence of the bug, but did display massive framerate drops when the Nyarlathotep jumped in: --3.6.9 final w/3.6.8 zeta VPs --Swifty's May 28 3.6.9 build w/3.6.8 zeta VPs I'm really not sure what to take out of all of this. There's clearly something going on in each circumstance, but the exact symptoms differ depending on which combination is in use. I can't see the Lucifer model itself as being the problem, since I was able to look at it just fine in three or four other Derelict missions without any framerate drops, including another one where it jumps in mid-mission (albeit without glowmaps activated on that one occasion). The slowdowns on this particular mission seem to be more severe than most others I've experienced, no matter what particular build I was using. If it does come down to potentially being a hardware issue on my end, is there any sort of memory test utility out there that would allow me to determine it for sure? |
|
|
Very useful. I have one more build you may want to try. http://www.freespacefaq.com/Misc-Downloads/Builds/Old/StandaloneFixes-1.7z It's one of the last builds before the shaders commit, and it's recommended as being pretty stable. For the other 3.6.10 builds you tested, you may want to also try using the Disable GLSL option (-no_glsl). Those tests should help narrow some stuff down even more. Also, if you could post an fs2_open.log from a couple of them and mark them as such, might also be useful. |
|
|
I'm extremely sorry that I haven't provided any additional feedback for this. I haven't yet run the full gamut of tests you prescribed, but I did perform a limited test using the Dec. 7 nightly build and the new MediaVP release. I ran the same mission with GLSL both enabled and disabled, and the bug manifested itself in both instances. I've attached the fs2_open.log generated from the debug build; I ran this instance without disabling GLSL. I can perform a more thorough testing run on several of the builds I tried before if you deem it necessary. |
|
2008-12-10 04:00
|
|
|
|
You might want to also draw attention to this on the MediaVP release thread. |
|
|
At the suggestion of some of the chat denizens, I tried removing MV_Advanced and replaying the problem mission with the 12/17 nightly build. The results were rather telling, as I didn't get the stutter-step slowdown that had occurred on every other play-through, and none of the graphical effects were glitched. I'd guess that the problem lies with my video card: either there's some sort of physical problem with it, or it's simply telling me that it can no longer handle the super-detailed graphics in the latest release. If there's any way of definitively proving/disproving the former case, I'd be happy to try it. |
|
|
I think we're going to have to chalk this up as 64MB cards just can't handle Advanced anymore. The SCP really has made leaps forward in terms of graphics capabilities and it's going to put more pressure on those older cards with all those features turned on. So I think this is safe to close. |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2008-10-07 06:03 | Mongoose | New Issue | |
| 2008-10-07 06:03 | Mongoose | File Added: screen0002.png | |
| 2008-10-07 06:03 | Mongoose | File Added: screen0003.png | |
| 2008-10-07 06:04 | Mongoose | File Added: screen0004.png | |
| 2008-10-07 06:04 | Mongoose | File Added: screen0006.png | |
| 2008-10-20 07:42 | Mongoose | Note Added: 0010066 | |
| 2008-10-23 23:43 | Da_Brain | Note Added: 0010108 | |
| 2008-10-24 04:32 | Mongoose | Note Added: 0010109 | |
| 2008-10-24 04:33 | Mongoose | File Added: screen0008.png | |
| 2008-10-24 04:34 | Mongoose | File Added: screen0011.png | |
| 2008-10-24 04:44 | taylor | Note Added: 0010110 | |
| 2008-10-24 15:10 | chief1983 | Note Added: 0010113 | |
| 2008-10-24 20:43 | Mongoose | Note Added: 0010116 | |
| 2008-10-24 22:05 | chief1983 | Note Added: 0010118 | |
| 2008-10-25 04:56 | Mongoose | Note Added: 0010119 | |
| 2008-10-25 05:33 | chief1983 | Note Added: 0010120 | |
| 2008-12-10 03:59 | Mongoose | Note Added: 0010390 | |
| 2008-12-10 04:00 | Mongoose | File Added: fs2_open_20081207.log | |
| 2008-12-10 05:02 | chief1983 | Note Added: 0010391 | |
| 2008-12-19 02:15 | Mongoose | Note Added: 0010434 | |
| 2008-12-19 06:14 | chief1983 | Note Added: 0010435 | |
| 2008-12-19 06:14 | chief1983 | Status | new => closed |
| 2008-12-19 06:14 | chief1983 | Resolution | open => not fixable |
