I have the key to One Eyed Willy

[Wasp]

QUOTE (KGJV @ Apr 13 2016, 01:33 PM) ...I haven't done any tests, I'm just looking at code...

Have you found anything that may be responsible for this?

I've done some tests and can say that the response time from keyboard input to ship movement has almost doubled. The input seems to be inconsistent and ships jumping around suggests that there is a problem with frame management?.

Any ideas?

[KGJV]

Very busy here and I was waiting for more "narrowing" from you

I'll have more time at the end of new week, i'll look at frame management differences between dx7 & dx9.

[Wasp]

QUOTE (KGJV @ May 10 2016, 02:26 PM) Very busy here and I was waiting for more "narrowing" from you

I'll have more time at the end of new week, i'll look at frame management differences between dx7 & dx9.

I think I COULD narrow it within the Dx9 builds however, there are no builds to test (that I can find). Wasn't there a trail of builds getting Dx9 working before doofus introduced it?

It seems logical that the problem would not be in one Dx9 build vs another...but instead, in all of the Dx9 project. I've been reading that this occurs when constants are turned into floating values or when clocks go from fixed steps to variable and when frames are improperly managed or from huge amounts of lag causing what is being displayed being out of sync with inputs....

Seriously, game input / response and synchronization are so critical to the very basic functions of Allegiance, I'm shocked that the Dx9 version was ever released with this apparent flaw.

Thanks again for looking into this, I'm guessing you're probably the only one left who can fix this.

[Wasp]

Kage,

Was the time step changed in R5? Was it a fixed time step in R4? Is the time step now variable?

Gaffer on Games

Not being a coder I can only speculate and present my findings...

1. The Dx9 version brought on this bug. It appears in the build that Doofus submitted. Joysticks lost high precision accuracy and feel laggy. Mice feel laggy compared to R4.

2. The Dx9 version also brought micro stuttering which can be seen when observing miners close up and there are tons of videos showing this behavior in ship movements.

3. The Dx7 version that BT produced did not solve the problem however, that must be key to solving this. We know it's not Directx causing this.

Wouldn't there be a way to measure R4's loop, timing / precision and compare?

[KGJV]

QUOTE (Wasp @ Jun 22 2016, 05:00 PM) Was the time step changed in R5? Was it a fixed time step in R4? Is the time step now variable?

I don't know. I'll have to look into this.

QUOTE (Wasp @ Jun 22 2016, 05:00 PM) Wouldn't there be a way to measure R4's loop, timing / precision and compare?

probably, check the performance counters (default alt-f iirc) if isn't in there , it should be easy to add infos to the perf counters.
you can also output the perf counters to the log (search for "m_bFPS" in the code but i don't think there a cmd line option for it so you'll have to force it).

[Wasp]

QUOTE (KGJV @ Jun 26 2016, 02:02 PM) I don't know. I'll have to look into this.



probably, check the performance counters (default alt-f iirc) if isn't in there , it should be easy to add infos to the perf counters.
you can also output the perf counters to the log (search for "m_bFPS" in the code but i don't think there a cmd line option for it so you'll have to force it).

I've been using alt f to look for fluctuation or differences between r4 and r5. I can say that with vsync forced off in nvidia control panel, the fps in r4 engine (dx7) stays within 128fps (lobby) to 512fps at launch and then drops to 341'ish a few seconds later and remains there the whole game. If I unlock vsync in R7 (dx9) the fps goes off the charts (over a thousand) and varies by hundreds during the game without ever stabilizing.

I've tried everything under the sun in every combination to see if managing the fps would resolve the over sensitivity issue, with no success. I can get wonkyness (lag) out but no matter what I do, the x and y axis seem "off" and the controllers are over reacting with almost no fine precision when making minute adjustments.

One more thing, it appears that the x axis moves quicker when moving the joystick out of the deadzone as compared to the y axis. It feels as though the x axis was stretched. Don't know if that means anything but since there was so much focus on widescreen and additional changes regarding video modes, I thought it may be related.

[JimmyNighthawk]

*hach*

[Wasp]

Kage or any Coder...

Is there a way to measure input lag? The difference between the Pre-doofus (DX7) and Post-doofus (DX9) is incredible.

I've read This repeatedly and even though I can't understand the specifics, the problem sounds identical to ours.

[Wasp]

nm

[Wasp]

Alright, I'll quit.

Clearly a PROBLEM! that is of no interest.