Announcement

Collapse
No announcement yet.

Mouse Input investigation

Collapse
This is a sticky topic.
X
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • replied
    Originally posted by Stolid View Post
    I would definitely be up for some testing. An easy guide to do it offline at minimum would make it a lot more accessible.
    Agreed!

    It's quite strange to me that some people don't seem to notice it, I even sat down at a friend's rig who was certain there wasn't any lag at all and it had to be my system, but I experienced it just as much on his system as I did on mine. For me it's immediately obvious just standing still and moving the mouse, not only compared to games like Quake Live or CS:GO, but even to "busier" games like Overwatch or even Battlefield.

    Some objective data would probably help, especially in reproducing the problem. I bothered the Battle(non)sense guy, Chris, again and he said that he had an idea how to properly measure the lag between moving the mouse and it happening on screen, and he might give UT a look when he finishes his measuring-thingy.

    Leave a comment:


  • replied
    From what I understand maps have to be built with the option ticked so that the lighting is built differently. This would mean it isn't just a simple rendering option you set in your ini file of the public build, rather you have to create and provide those new map versions in some way. Whether that is sufficient for it to work with the public build I'm not entirely sure though.

    I would definitely be up for some testing. An easy guide to do it offline at minimum would make it a lot more accessible.

    Leave a comment:


  • replied
    Originally posted by frustzwerg View Post
    Is hosting of the server really necessary? I feel the input lag even in offline games, so it might be enough for you to modify UT's code and run it locally. I'm quite interested in this - how hard would it be for you to modify the code? If you could distribute the modified files somehow I'd of course be happy to test it.
    I need the server because ideally people should test how they feel using the new renderer and come to their own conclusions. Although there *seems* to be a reduction in input lag when I change the ShooterGame example game that comes with UE over the forward rendering, it could be placebo effect. While my aim and tracking are pretty good, I'm not as good as some people I know, and it might be good for who are really sensitive to input lag to weigh in on it. I can tell a difference between UE4 based games and Q3 based games, definitely. I also work in CryEngine for fun, and its input lag is even more perceptible than UE4. But again, it could be placebo effect on my part.


    However, getting UT4 to work is probably less than trivial. What I didn't realize when I wrote the first post is that UT4 is built on UE 4.12. The forward renderer is only stable in 4.14+. So, this would necessitate trying to migrate the project to the new version of the engine. I'm not a C++ coder (merely a hobbyist doing this for fun) so whether or not the project would migrate or not is beyond me at the moment. I can take a crack at it. I highly suspect the UT4 team have done some engine level modifications that will need to be merged with the 4.14 source, which is something that is beyond my skills.

    As far as changing any UE project over to the forward renderer when it's migrated, that is the trivial part. You merely need to open the project, open project settings, and check the forward rendering box. The editor just restarts and recompiles shaders. You'd then have a to cook the game.

    A way to test if there are any obvious improvements in the base UE4 code would be just to package the ShooterGame example and have people play it and see if there's a difference. However, that's not UT4, and there might be a lot of input tweaks happening that I'm not aware of (or completely different methods for controlling the character).

    Leave a comment:


  • replied
    Originally posted by Dashiva View Post
    I'm currently working on my own FPS, but I'm using Unreal as a base. The input lag was bugging me coming from quake3 style games. I am testing forward rendering right now, and it might improve things. If the UT4 code/assets are available for modding it would be trivial to enable forward rendering and test it. Would need someone to host the server but that's about it. Let me know if anyone is interested.
    Is hosting of the server really necessary? I feel the input lag even in offline games, so it might be enough for you to modify UT's code and run it locally. I'm quite interested in this - how hard would it be for you to modify the code? If you could distribute the modified files somehow I'd of course be happy to test it.

    Leave a comment:


  • replied
    Originally posted by Stolid View Post
    In the short term I'm personally most interested in the forward renderer that is being developed as it should give higher fps and lower input lag. Even though it is being developed for VR I hope the UT team puts some resources towards supporting it and soon making it possible to enable in UT as well.
    I'm currently working on my own FPS, but I'm using Unreal as a base. The input lag was bugging me coming from quake3 style games. I am testing forward rendering right now, and it might improve things. If the UT4 code/assets are available for modding it would be trivial to enable forward rendering and test it. Would need someone to host the server but that's about it. Let me know if anyone is interested.

    Leave a comment:


  • replied
    I have a feeling there is a very slight input delay on when you press fire and when the weapon actually fires.

    Playing with sniper (hitscan) i noticed that im on target (like bulls eye) i fire and you miss...

    I then started tracking/tracing my shots, like id fire but keep on holding crosshair over target for another second after i pressed fire and there was a 100% improvement.

    Leave a comment:


  • replied
    how can i help giving u guys valuable information? i see some pc set-up posts but i cant imagine u can work with smt like that. i deffo am able to ask a whole lot of players what to do as long as we have a simple go-through

    Leave a comment:


  • replied
    Found this really good video testing vsync compared to gsync with mouse input lag.

    https://www.youtube.com/watch?v=F8bFWk61KWA

    Leave a comment:


  • replied
    Originally posted by burawura View Post
    Hardware: Intel i5 2500k @4GHz / 8GB DDR3 / EVGA GTX 970 / ASUS VG248QE @144Hz
    Drivers: Geforce v375.63 / Realtek ALC892 v6.0.1.6167 HD Audio
    Input: Corsair Strafe @1000Hz(w. updated CUE) / Zowie EC-1A @1000Hz+400dpi (both in USB3.0)
    Game: 1080p, windowed fullscreen, 100fps cap, no vsync, no smoothing, 111fov, 3.0 sens w. no accel, high settings, r.finishcurrentframe=1, r. oneframethreadlag=0
    Windows 10 6/11 no enhanced precision

    [ATTACH]31725[/ATTACH]

    I just switched from a Steelseries Sensei RAW mouse to a Zowie EC-1A and noticed a drastic reduction in the input delay in UT4.
    Estimated input delay (looking around, shooting the sniper rifle) with Sensei RAW: ~150ms (a noticeable pause)
    Estimated input delay (same activity, panning view & clicking) with EC-1A: ~50ms (almost instantaneous)
    Drastic and extremely noticeable improvement in the delay time. Let me know what else I can do to help.
    I didn't have any trouble with input delay using the Sensei RAW in any other games I play including CSGO, DOOM, Reflex, QL, Toxikk (not sure about Squad, only played like 10hrs)
    You should try running the game in true fullscreen rather than windowed fullscreen, pretty sure you have forced V-Sync if you use any kind of windowed mode.

    Leave a comment:


  • replied
    Very peculiar.

    That said, I don't think any significant percentage of people that has issues with the mouse feel has been having anywhere near that kind of input lag. I know many that tried it with a different mouse and other changes in hardware without any big changes in feel.
    Someone who does happen to have that mouse should definitely try to see if they can reproduce this though.

    Leave a comment:


  • replied
    i might have some input on the input delay issue :P

    Hardware: Intel i5 2500k @4GHz / 8GB DDR3 / EVGA GTX 970 / ASUS VG248QE @144Hz
    Drivers: Geforce v375.63 / Realtek ALC892 v6.0.1.6167 HD Audio
    Input: Corsair Strafe @1000Hz(w. updated CUE) / Zowie EC-1A @1000Hz+400dpi (both in USB3.0)
    Game: 1080p, windowed fullscreen, 100fps cap, no vsync, no smoothing, 111fov, 3.0 sens w. no accel, high settings, r.finishcurrentframe=1, r. oneframethreadlag=0
    Windows 10 6/11 no enhanced precision

    DxDiag.zip

    I just switched from a Steelseries Sensei RAW mouse to a Zowie EC-1A and noticed a drastic reduction in the input delay in UT4.
    Estimated input delay (looking around, shooting the sniper rifle) with Sensei RAW: ~150ms (a noticeable pause)
    Estimated input delay (same activity, panning view & clicking) with EC-1A: ~50ms (almost instantaneous)
    Drastic and extremely noticeable improvement in the delay time. Let me know what else I can do to help.
    I didn't have any trouble with input delay using the Sensei RAW in any other games I play including CSGO, DOOM, Reflex, QL, Toxikk (not sure about Squad, only played like 10hrs)
    Last edited by burawura; 10-25-2016, 04:35 PM.

    Leave a comment:


  • replied
    No, I'm just familiar with it.

    Leave a comment:


  • replied
    Originally posted by Stolid View Post
    Irrelevant.
    The way the engine works causes there to be significant and obvious input lag even under (near) optimal conditions. Under those circumstances it alone will already add ~30ms on top of whatever baseline of lag is caused by the hardware like mouse, screen etc.
    Oh, so you're one of the engine developers?

    Leave a comment:


  • replied
    Irrelevant.
    The way the engine works causes there to be significant and obvious input lag even under (near) optimal conditions. Under those circumstances it alone will already add ~30ms on top of whatever baseline of lag is caused by the hardware like mouse, screen etc.

    Leave a comment:


  • replied
    My mouse is a 500Hz mouse. I've heard 1000Hz over USB 3.0 (xHCI) can be an issue, mainly with mice, not with keyboards.

    Leave a comment:

Working...
X