Measuring input lag at 1000 fps

Discussion in 'General Discussion' started by SPASKIS, Dec 23, 2017.

  1. ebeninca

    ebeninca Registered

    Joined:
    Sep 7, 2016
    Messages:
    740
    Likes Received:
    526
    Single monitor here, gtx970
     
    juanchioooo likes this.
  2. juanchioooo

    juanchioooo Registered

    Joined:
    May 16, 2016
    Messages:
    2,242
    Likes Received:
    1,649
    gtx 1060 monitor 4ms , 31.5 "144hz, 100 fps stable with fx 8350, an old processor and 8 gb of ram , windows 7 . ....bad configuration ...
     
  3. patchedupdemon

    patchedupdemon Registered

    Joined:
    Jan 3, 2017
    Messages:
    1,602
    Likes Received:
    1,151
    Gtx 1060 here no problem running triples at 60 he with decent graphics.
    Rain will always be problematic though but still not impossible
     
    juanchioooo likes this.
  4. SPASKIS

    SPASKIS Registered

    Joined:
    Sep 7, 2011
    Messages:
    3,155
    Likes Received:
    1,426
    HDMI for the TV. The laptop monitor is integrated so I don't know. In both cases the result was basically the same. Maybe the integrated screen had a 90 ms input lag compared to the 100 ms for the TV.

    Have you tried to measure your input lag in a similar way to what I did? You might find that despite your 4ms monitor the overall lag is bigger that what you think as it happened in my case. I wouldn't have ever imagined it was so big as it resulted after measuring it.
     
    juanchioooo likes this.
  5. juanchioooo

    juanchioooo Registered

    Joined:
    May 16, 2016
    Messages:
    2,242
    Likes Received:
    1,649
    I replied to my partner that he had a bad configuration, for my opinion;):)
     
    patchedupdemon likes this.
  6. juanchioooo

    juanchioooo Registered

    Joined:
    May 16, 2016
    Messages:
    2,242
    Likes Received:
    1,649
    I can only tell you that on my monitor goes at the same time as my hands go, I do not notice any difference
    he hecho la prueba de mover el volante derecha izq a toda prisa y no hay delay ninguno, mis ojos no lo perciben , saludos
    pd.-the difference between a monitor and a TV is very noticeable, difference as night and day
    pd.- la diferencia entre monitor y tv es muy notable , como la noche y el dia
     
  7. SPASKIS

    SPASKIS Registered

    Joined:
    Sep 7, 2011
    Messages:
    3,155
    Likes Received:
    1,426
    I am sure that your input lag is great considering your monitor. For that reason it would be very interesting to measure it. Saying that the image follows your movements as you perceived with your naked eye is quite useless in order to provide a best class figure. Input lag is measured in ms.
     
  8. juanchioooo

    juanchioooo Registered

    Joined:
    May 16, 2016
    Messages:
    2,242
    Likes Received:
    1,649
    I do not know how I can measure it exactly, I leave you link of features
    http://www.benq.es/product/monitor/ex3200r/specifications/
     
  9. patchedupdemon

    patchedupdemon Registered

    Joined:
    Jan 3, 2017
    Messages:
    1,602
    Likes Received:
    1,151
    I’m not sure we can all get our hands on fancy high speed cameras.lol

    I’d love to test my monitors,1ms grey to grey,but I’d bet my last £ they are no where near that low
     
  10. Kaitsu_69

    Kaitsu_69 Registered

    Joined:
    Jan 13, 2012
    Messages:
    88
    Likes Received:
    8
    Favorite clothing at the home. :D
     
    SPASKIS likes this.
  11. juanchioooo

    juanchioooo Registered

    Joined:
    May 16, 2016
    Messages:
    2,242
    Likes Received:
    1,649
    I had said it above 4 ms. post 22
     
    SPASKIS likes this.
  12. SPASKIS

    SPASKIS Registered

    Joined:
    Sep 7, 2011
    Messages:
    3,155
    Likes Received:
    1,426
    It's not absolutely necessary to use a high speed camera to get a rough figure. For sure at 1000 fps you can easily measure up to ms precision. However from a qualitative point of view input lag can be measured with less precision.

    Many mobiles today have a low speed camera mode. In my mobile it runs at 120 fps which should be good enough to measure how many frames delay the display shows compared to the real input.

    You just need someone that can record you and your monitor while you move the wheel to both sides as fast as possible. The delay is better measured when the wheel carries the highest speed which is in the central position. In movie maker or other video editing programs you can later go frame by frame.
    Screenshot_20171227-200458.png
     
    juanchioooo likes this.
  13. juanchioooo

    juanchioooo Registered

    Joined:
    May 16, 2016
    Messages:
    2,242
    Likes Received:
    1,649
    @SPASKIS I already lost my partner ... if I tell you that my eye sees no difference. in the tv that used before it was clearly noticed the enormous difference...but I think the features of the monitor says it very well ;):)
     
  14. SPASKIS

    SPASKIS Registered

    Joined:
    Sep 7, 2011
    Messages:
    3,155
    Likes Received:
    1,426
    I believe you juanchio. I just was very curious to have the global figure. As you know you have to add the amount of time the game needs to generate the frame. If you are running at 120 fps it means that the game is generating one frame every 8 ms. For your monitor that would mean a global of 12 ms delay which is really good. A 120 fps recording would barely show a one frame delay. However it is always good to check if there is some extra delay adding to the theoretical calculations. Mabe some other top monitor user will try to measure it.
     
    juanchioooo likes this.
  15. juanchioooo

    juanchioooo Registered

    Joined:
    May 16, 2016
    Messages:
    2,242
    Likes Received:
    1,649
    very important screen refreshment 144 hertz
     
    SPASKIS likes this.
  16. Lazza

    Lazza Registered

    Joined:
    Oct 5, 2010
    Messages:
    12,382
    Likes Received:
    6,600
    Everything adds up, but a 144Hz monitor gains at best 10ms over a 60Hz one (assuming no prerendered frames). Add another 10ms for slow vs fast response rate, and that's still only 20ms - a small part of the ~100ms lag measured.

    The practical effects (driving speed/consistency) of even 100ms aren't known. Again more lag can't help, but is 100ms making more than a negligible difference? Would 50ms? Without a lot of proper testing we can't really say.

    Still interesting to see though.
     
  17. juanchioooo

    juanchioooo Registered

    Joined:
    May 16, 2016
    Messages:
    2,242
    Likes Received:
    1,649
    I just know that when I look at the screen the pilot of it rotates at the same time as me, I do not see delay
     
  18. Nygaard

    Nygaard Registered

    Joined:
    Feb 1, 2017
    Messages:
    29
    Likes Received:
    8
    i did the same test with a 120 fps mobile phone in iracing, and i did count 6 frames of input lag. 6 frames / 120frames/s or 50 ms seconds of input lag. It is on a cheap 60 hz pc-monitor.

    It would be funny to se similar tests on different setups. i doubt a 144hz highend Screen will change it that much. But it would be fun to se how much difference it is.

    To relate this to something relevant, your reaction time is about 200ms. So input lag is in my case 25% of my reaction time.

    having an input lag of 50 % of your reaction, you should have an excuse of bad driving.. ;-)

    remember going from from a 40 inch lcd Screen to 22 inch triple made a difference when driving on the limit..
     
  19. SPASKIS

    SPASKIS Registered

    Joined:
    Sep 7, 2011
    Messages:
    3,155
    Likes Received:
    1,426
    As a rule of thumb multiplying delay by speed will give you an idea of how late are you perceiving something happening. 100ms at 200 kph is 5,5 m. For the best achievable values it seems that this could be reduced to less than 1 m. Considering this you have to react quite ahead before where your input will actually take place which I understand makes it more difficult. This is regarding visual input. For FFB input hopefully the delay is much smaller so I am probably able to perceive tires losing grip prior to noticing it visually.

    Overall it is difficult to say how much it affects driving as you say. Testing would be required and the results might be very different depending on the user and FFB configuration. However, I think that it is clear that nothing good can be obtained from a high input lag.

    With my new PC I am not sure how much I will improve. Hopefully I will also be able to run my DK2. I would love to be able to measure input lag with it. I read in the past that it has a very good response.
     
  20. Lazza

    Lazza Registered

    Joined:
    Oct 5, 2010
    Messages:
    12,382
    Likes Received:
    6,600
    You don't see delay, but we know there isn't zero delay. This doesn't mean the delay isn't very small, and perhaps small enough to be able to disregard it (as something affecting your driving), but at the same time someone else that can't see any delay might have more or less than you.

    What I'm saying is, Spaskis is measuring the delay. Your non-measured observations, though probably correctly indicating less delay than him, lack precision.
     
    patchedupdemon and juanchioooo like this.

Share This Page