Vsync off vs Scanline Sync, what is the difference in input delay? with frametime comparison

Discussion in 'Rivatuner Statistics Server (RTSS) Forum' started by alexander1986, Oct 22, 2019.

  1. alexander1986

    alexander1986 Member Guru

    Messages:
    192
    Likes Received:
    17
    GPU:
    RTX 2060
    Hello, I have been looking for real numbers/hard data on input delay testing between vsync off and scanline sync enabled, but I can't seem to find any!



    I'm using all lowest graphics settings at 1080p on a quite strong PC (9700k / rtx 2060 / 2x8gb 3200mhz cl15 ram) and my gpu load never exceeds 30-50% no matter what the ingame scenario is, so for me in this case scanline sync is a great option as it eliminates any tearing while not affecting input delay much, it is really hard for me to say with 100% certainty if it adds any input delay or if mouse feels much heavier than just vsync off / 120 fps limit ingame, which is a good problem I guess, but I would really like to know how much, if any input delay it really adds?



    from my testing in Fortnite with my 120 hz monitor, using vsync off / 120 fps limit set in the ingame options compared to vsync off / scanline sync enabled is VERY similar in terms of how light/responsive mouse feels, only big difference is of course with scanline sync there is zero tearing and the picture is butter smooth, but not only this, when using scanline sync and looking at frametime overlay graph, the frametime will be a perfect flat line / glassfloor locked to 120 ms / 8.3 ms in any and all scenarios when using scanline sync,


    however, when not using scanline sync and just limiting ingame fps to 120 with ingame limiter with vsync disabled, the frametime will hover around 8.5-9.0 ms and frametime graph is not completely flat, especially with many players in my vicinity the graph will have slight peaks and valleys, compared to scanline sync enabled where the frametime graph will be completely flat, 100% of the time, no matter what is going on ingame,



    picture example of frametime comparison :


    vsync off, 120 fps limit ingame:


    [​IMG]



    vsync off, scanline sync enabled:


    [​IMG]



    As we can see on the bottom picture with scanline sync enabled, the frametime is perfect 120 fps/8.3 ms with completely flat line, this is the case for me on my hardware/settings no matter what is going on in game and how many players are near by and so on, and since my GPU load never exceeds 30-50% with my specs and settings, I assume it is working really well,


    I would still very much like to know 2 things!

    1. What is the real numbers of added input delay when using scanline sync in my scenario, if any at all?

    2. Why is the frametime graph not flat/perfect when using the ingame fps limiter set to 120, compared to the perfect frametime we see when using scanline sync? is it an issue with engine / game it self perhaps?



    In both examples, the mouse/responsiveness is very light and feels great, and using scanline sync eliminates tearing so i'm currently leaning towards using that and just be happy instead of not using it, but sometimes I get feeling that mouse after all is slightly heavier when using scanline sync instead of regular vsync off / 120 fps limit ingame, i'm a competitive player so for me input delay is priority #1, but if I can get away with using scanline sync and have zero tearing while not adding any input delay then that would of course be great,


    Hope my post makes any sense and would greatly appreciate any knowledge and input on this, especially hard numbers / data on added input delay with scanline sync would be really nice, if it exists, and the frametime difference in the pictures would be nice to understand also why that exists,



    if it is okay I want to tag/mention Unwinder and couple users here that I have experience with from before that are super knowledgeable, that maybe can help me with my questions, but anyone is free to comment and answer of course, @Unwinder @RealNC @Mda400


    Thank you.
     
  2. Andy_K

    Andy_K Master Guru

    Messages:
    435
    Likes Received:
    41
    GPU:
    MSI GTX 960 OC
    I do not have any experience with scanline sync, my rig is not powerful enough to get it working.

    But I read this here and it may help you:
    You may want to read along the thread to get further information.
     
    alexander1986 likes this.
  3. alexander1986

    alexander1986 Member Guru

    Messages:
    192
    Likes Received:
    17
    GPU:
    RTX 2060
    thanks for helping, I already know that with the tearline though, I've set the tearline to the bottom of my screen (setting scanline sync to "1000" in RTSS on my 24" monitor puts it at the bottom),

    I am still looking for hard data/real numbers on actual measured input latency when using scanline sync vs vsync off at the same FPS (120 fps/120hz for example in my case)

    searched a lot for this, only found some suspect numbers by some guy measuring with his high-speed cellphone camera or something like that, not good enough IMO :p


    cheers though!
     
  4. RealNC

    RealNC Ancient Guru

    Messages:
    3,106
    Likes Received:
    1,337
    GPU:
    EVGA GTX 980 Ti FTW
    Does it matter? You use ssync to get rid of tearing at a lower latency cost than normal vsync. I don't think there's any proper latency tests of ssync, but in my experience it's the same as g-sync + normal RTSS cap. There are tests for that setup so you could look those up.
     
    alexander1986 likes this.

  5. alexander1986

    alexander1986 Member Guru

    Messages:
    192
    Likes Received:
    17
    GPU:
    RTX 2060
    :p in the end I guess it does not matter that much, just interesting to know! still appreciate the answer, cheers !
     

Share This Page