[b252-256] mouse behaviour and fps
elodea
Editlodea Join Date: 2009-06-20 Member: 67877Members, Reinforced - Shadow
Editlodea Join Date: 2009-06-20 Member: 67877Members, Reinforced - Shadow
Since the linux patch, mouse control seems to have become much more sensitive to client rate. When it drops, the mouse gets
- decreased sensitivity
- higher input delay. I'm pretty certain there's input delay all across the scale of client tick rates, just it gets less noticeable the higher the tick rate
- a feeling akin to increasing ammounts of mouse smoothing. Maybe this is just a perceived effect due to suddenly decreased sensitivity, I have no idea. All i know is that something that isn't lower sens or higher input delay is going on as well that doesn't feel right at all.
As the client tick increases, mouse behaviour seems to approach true raw input, no delay. It's kinda hard to show on a fraps/dxtory or whatever because there's no easy way to standardize a mouse movement or show a mouse feeling. Also, using programs to artificially cap fps/client rate does not reproduce, so it seems to be tied to some underlying congestion?
For me, the effect is quite apparent just going from a small range of ~130 to ~70 client ticks in any normal round. On just a private lan server, you can noticeably feel the effect going from the max 200 ticks to ~100 ticks from viewing different sections of a map. So even a normal 100 'fps' in your average game does not seem to give a true mouse behaviour.
Historically even before the linux patch, sensitivity has always felt like it changed as client rate (fps) changed, but never really to such a significant degree that you could feel it and be 100% sure it was happening.
Is NS2 still using raw windows mouse input? It seems like first you need to get used to how the mouse behaves specifically within ns2, then you need to get used to the micro sensitivity and input delay fluctuations that happen as client tick changes in any gameplay.
Additional information that might be helpful.
-For me, it happens both on dx9 and dx11. Although less on dx9, it's still a significant problem.
-I've done mf's raw input/accel switch fix many multiple times, so it's not related to that bug i'm certain.
-Aero is off
-Happens also on 30 tick stable servers
- decreased sensitivity
- higher input delay. I'm pretty certain there's input delay all across the scale of client tick rates, just it gets less noticeable the higher the tick rate
- a feeling akin to increasing ammounts of mouse smoothing. Maybe this is just a perceived effect due to suddenly decreased sensitivity, I have no idea. All i know is that something that isn't lower sens or higher input delay is going on as well that doesn't feel right at all.
As the client tick increases, mouse behaviour seems to approach true raw input, no delay. It's kinda hard to show on a fraps/dxtory or whatever because there's no easy way to standardize a mouse movement or show a mouse feeling. Also, using programs to artificially cap fps/client rate does not reproduce, so it seems to be tied to some underlying congestion?
For me, the effect is quite apparent just going from a small range of ~130 to ~70 client ticks in any normal round. On just a private lan server, you can noticeably feel the effect going from the max 200 ticks to ~100 ticks from viewing different sections of a map. So even a normal 100 'fps' in your average game does not seem to give a true mouse behaviour.
Historically even before the linux patch, sensitivity has always felt like it changed as client rate (fps) changed, but never really to such a significant degree that you could feel it and be 100% sure it was happening.
Is NS2 still using raw windows mouse input? It seems like first you need to get used to how the mouse behaves specifically within ns2, then you need to get used to the micro sensitivity and input delay fluctuations that happen as client tick changes in any gameplay.
Additional information that might be helpful.
-For me, it happens both on dx9 and dx11. Although less on dx9, it's still a significant problem.
-I've done mf's raw input/accel switch fix many multiple times, so it's not related to that bug i'm certain.
-Aero is off
-Happens also on 30 tick stable servers
Comments
P.S. Yeah, I noticed FPS screws up the mouse "feel" (for lack of a better word for it). I wasn't monitoring server rate. Dx11 "felt better" on 254, but after 255 dx9 seems to be about the same - probably due to the fullscreen exclusive fix.
I hope these issues do get fixed.
That's the spirit, comrade! I am nominating you to the "Mild Enthusiasm Award 2013" with the right to hang a special commemorative plaque on the wall of your office. Make sure you hang the plaque by Friday when the inspection is scheduled, or it will be you who will hang there.
Glory to the TSF!