Reply 20 of 31, by robertmo
- Rank
- l33t++
i think hal is talking about dosbox version linked to here:
game speed is not constant
i think hal is talking about dosbox version linked to here:
game speed is not constant
The game uses a 250hz rate for IRQ0 (mode 3), and the interrupt is like its heartbeat. Too slow or too fast, and the effect on the game is obvious: music and animations speed up or slow down in a linear fashion if the rate is changed. Point being that any results seen in a virtualizer where the game is obviously too fast or too slow are not terribly useful as an example of correct behavior.
As a test, I had DOSBox use a 2x timer rate (i.e. 500hz), and the 3D flight is generally not harmed by the faster rate, but animations and music are too fast. The interesting result from this test, though, is that the platforms open fire at a range that seems normal when DOSBox is running at max cycles, supporting the idea that the problem is related to timing in some way; although it's not at all conclusive that there is a problem with DOSBox.
Slightly off topic, but I suggest avoiding any branding of games as "important" or "significant". Even if it is generally considered to be true, it is nonetheless subjective, and should be irrelevant as criteria for which games receive attention for issues. An issue in one game could very well apply to others.
h-a-l-9000: OK will try the build when I've got some time. I remember reading the "boom" thread, wouldn't have guessed that this build could be a solution/test.
robertmo: thanks for the link (;
ripsaw8080: I was trying to apply a bit of logic based on observations, but as I've said I can't really tell what's going on "inside" the emulation. The point is more or less moot anyway as I didn't manage to run the game at "normal" speed (yet?). The BOCHS system of setting emulation speed is horribly complicated, at least for what I'm trying to do. I understand what you're saying about "important" games (I've used the ""'s for a reason) but I still think there's games that are more known, famous, liked or whatever you want to call it than others. For instance, questions about certain games pop up very often on VOGONS, others seldom or not at all. But, of course, all bugs should receive attention if possible.
No new word on folks' experiences with Tie fighter Collector's Cd-Rom and CPU cycles right? Still seems tragic one must choose
between totally fluid gameplay and an incorrectly emulated program, or an oft chuggy yet correctly emulated program...
Yeah can't even express how tragic this is.
Probably of no interest, but since my original post I've upgraded to a Sandy Bridge 2500k @ 3.8ghz, 4 gigs RAM, and Win 7 - still using the latest dosbox .074 release to both install and run the game, - I'm seeing exactly the same behavior as observed with my opteron, the cycle count used to summon the effects this thread highlights are unchanged between the processors. I've observed some other interesting things related to rendering, - @ 60,000 cycles I can stop a mid-LOD platform from firing its turrets simply by shutting off my CMD and thus eliminating the extra rendered ship it displays...
Max LOD platform will fire regardless, as will platform with laser/missile impact bitmaps on it or death-roll explosions...
It really seems that all non-fighter objects of small and midsize are affected. The largest ships seem to behave normally when onscreen at any LOD but don't fire much when off-screen...
Yeah can't even express how tragic this is.
Uncalled for. Don't you have some other place to act out your bad moods?
Anyway, TomJeffersonJones: did you by any chance test the game with the DOSBox build mentioned by h-a-l-9000? I have to admit I forgot about this thread altogether, and don't have the game installed atm. You would be the perfect candidate for testing, because you seem to have a pretty exact idea about what the error/bug/whatever is looking like (; .
Don't you have some other place to act out your bad moods?
What bad mood?
He posted his non-contributing bump, so did i.
wrote:Don't you have some other place to act out your bad moods?
What bad mood?
He posted his non-contributing bump, so did i.
Yes I have no solution, but the issue is unresolved and there may be some users who have the program running perfectly, and are unaware of issues observed by other players.
Some new ideas or standards for comparison are always welcome, and your sarcasm has at the very least taught me to avoid hyperbole when reopening threads.
And Addict, i have tried the mem-delay .dll's and .exe via h-a-l with no success, but will keep experimenting. 60,000 cycles with some chop during explosions is far from unplayable, but still seems curious.
I looked at the example of the first platform in TIE Fighter Bomber Mission 2 with core=dynamic cycles=max, and saw that it usually did not shoot until some time after I arrived at its location, as long as I didn't shoot it first. After trying a few things, I had some success in getting it to open fire as I approached.
The point where the platform should start shooting is not clear to me, so to those who can judge correct behavior, please try running the attached TSR program before running the game, and report what you think. Note that it will only be effective with 640x480 graphics in flight mode, and is probably more of a workaround than a fix.
So... any patch for this today ?
Not for you
Thread closed.