Dege wrote on 2021-03-20, 12:07:
Unfortunately no, dgVoodoo cannot affect mouse speed. But, in fact, it should be the same speed as with a pure 640x480 display because the physical cursor is still moving inside a 640x480 rectangle.
Do you get vsynced result with fake fullscreen and uncapped result with real fullscreen?
With both API (D3D11/12)?
Yeah, thing is mice back then had balls, compared to today's 20000 DPI (overcompensating much?), not an issue, just asking if it was a hidden feature.
I'm on W7, so I can test without desktop compositor (DWM/Aero) and it's forced vsync on/off in any windowed modes, but I can't test DX12.
Note that I'm not forcing vsync thru DGVoodoo or GPU driver.
I am also not artificially capping FPS.
This is what I noticed:
- the game has a bizarre case of FPS and/or framepacing
- when you move the mouse, the FPS will jump to literary 300-500FPS
- when you do not move and just let it play, it drops to 68 in fake fullscreen (with or without Aero), and 60 in real fullscreen
- it is ALWAYS smooth in fake fullscreen or pure windowed, even if the counter is 68 or 500FPS
- it is NEVER smooth in real fullscreen, even if the counter is 60 or 500FPS
- there is ALWAYS (all scenarios) a hiccup if you pan the game screen, even without mouse, using keyboard, it's barely noticeable, but in regular intervals > probably a frame being repeated?
There is just something about real fullscreen limiting the game to work at 60FPS (if you're not moving the mouse) compared to fake fullscreen with or w/o Aero, which is 68FPS (again, no mouse movement).
No Aero = no vsync
Aero = triple buffer vsync AFAIK, not sure which one, since people call a lot of stuff triple buffer
Fullscreen = either no vsync or double buffer or triple buffer, can't tell for signs since it's jerky anyway
Maybe thinking that vsync is the issue is the wrong direction, since I have no idea why mouse movement registers hundreds of FPS, that would imply no vsync is forced even in real fullscreen, yet it's 60FPS compared to fake fullscreen's 68FPS.
Also, forcing Vsync in DirectX in DGVoodoo in fake and real fullscreen results in stable 60FPS frametime (nice flat graph) BUT the game literary stops while you're moving the mouse.
All in all, it's perfectly playable with fake fullscreen, but I hoped as to why real fullscreen is so jerky wouldn't be such a bugger to figure out.