Input lag AB test

[spacediver]

no luck with any of those suggestions - and i don't have a "preferred refresh" setting in my nvidia 3d settings. I recently rotated my spare FW900 into service and when I did so I think I screwed up something in the driver settings. CRU lists four monitors!

I'll try to figure out soon.

[flood]

fairly certain the custom res util would work. add a res for 960x600@160hz, change desktop resolution to that, and run the program

for me it worked to force 1280x720 to 120hz on my fw900 when i wanted to duplicate the image on it and my vg248qe

[spacediver]

I think I need to fix my driver issues first. I used the reset-all utility from CRU, now trying to install the FW900 driver, but having no luck. Keep getting this no matter what inf file I choose.

edit: fixed driver issue - thanks monitor asset manager! http://www.entechtaiwan.com/util/moninfo.shtm

@ 160hz I got 15/25 for the 10 ms test, so clearly at the time I took this test I was at chance level

[pisto]

Also there's an *average* delay due to the actual display refresh rate, which I would estimate to rougly half of the frame period. You would probably be able to minimize that if the visual feedback could span the whole height of the screen. Maybe instead of a pointer make a vertical line, that one drags horizontally?

yep, that's right, but I'm guessing some people will be moving the pointer in a circle or a figure-8. probably i'll add a toggle to switch between pointer, line, and both pointer+line

No, there must be something being draw on all the vertical span of the screen, because you can't know which scanline is being drawn when you update the mouse status. A figure of 8 won't help because at every single frame the pointer is a small figure.

[flood]

I know what you mean; what I was saying is that a straight line wouldnt provide as much visual feedback if you move the mouse in motions with vertical movement

though it's trivial draw both a line and a pointer, so probably that's our best bet.

[war1]

posting back with something interesting(well)

I was using newest beta driver 340 since last time, and although hard, 10ms was definitely noticeable enough for me to get it right.

NVIDIA DRIVER:
Now, i was playing happily, having convinced myself of input lag from driver being a placebo and went for the input lag test once again(dno why, impulse) and i was greeted with a black screen, computer not responding to anything, driver crash.

After a hit on the reset button, i was back in windows, and low and behold, i noticed mouse was lagged right away(this is important i believe, since driver lag might be a bug, only affecting some)

I tried input lag test again, but same crash as before, just a black screen. Ok, time to uninstall beta driver and install oldest driver i know to work flawlessly -> 175.

INPUT LAG:
with the old driver on the system, i tried lowest i could feel before, 10ms, and this time, i could not tell the difference. I know this driver to be without extra lag, so im fairly positive now that im not able to detect 10ms, but a chain of input lag, amounting to more. Example: old driver, total input lag 0, or 10, cant feel it. New driver, some amount from driver, or some amount + 10, resulting in > 10ms and that makes me pick up on the 10ms difference somehow.

I thought it was worth mentioning, but at the same time just adding to my insanity about it

[spacediver]

if this is true, then you're more sensitive to input lag with a pedestal value of above 0 ms (see my earlier post). If this is really what's going on, and your results are generalizable to the population, then this may indicate a dipper function.

[flood]

@war1

what numbers did you get with the old driver?

given this,

think I might have broken the test - I got 23/25 with 10 ms (@ 85 hz) and it wasn't due to me feeling anything

it's possible that the program is buggy and/or isn't running consistently from run to run... I don't know

if this is true, then you're more sensitive to input lag with a pedestal value of above 0 ms (see my earlier post). If this
is really what's going on, and your results are generalizable to the population, then this may indicate a dipper function.

could this be what is going on?

[spacediver]

yep that's exactly the sort of function that explains a dipper.

Here are two figures from a paper I published a couple years back.

The first illustrates what is sometimes called a neural transducer function. The idea is that it describes the neural response to a change in some dimension of a stimulus. Consider delta R as the amount of change in neural response that is required for a just noticeable difference. Because of the shape of the function, there is a sweet spot where we're very sensitive to changes in the stimulus (delta K in the figure).



If you were to derive thresholds from this transducer function, you'd get something like the following (ignore the predicted vs. actual thresholds - just pay attention to the shape of the function, and notice how it dips).

[war1]

@war1

what numbers did you get with the old driver?

given this,

think I might have broken the test - I got 23/25 with 10 ms (@ 85 hz) and it wasn't due to me feeling anything

it's possible that the program is buggy and/or isn't running consistently from run to run... I don't know

if this is true, then you're more sensitive to input lag with a pedestal value of above 0 ms (see my earlier post). If this
is really what's going on, and your results are generalizable to the population, then this may indicate a dipper function.

could this be what is going on?

its 16ms with old driver, refreshrate might mask lower on older driver, and thats why im only able to feel on the newer ones, which have lag on my system. It would be fun to try on 120hz CRT to see if it is refreshrate masking the input lag, or about one frame is what im able to detect, right now im on a 75hz lcd monitor.