Bad input lag/desync issue: Testing with AW2521H 360Hz monitor w/ Nvidia latency module

[Boop]

I think you're supposed to use a dark background to get accurate results when using the latency analyzer. Could you try the same thing again with a dark background and maybe even darker targets so the analyzer only sees the flash?

This is how I did all my testing with Reflex Analyzer and had great results. Never tried a bright background like in the video.

[spoidah]

I think you're supposed to use a dark background to get accurate results when using the latency analyzer. Could you try the same thing again with a dark background and maybe even darker targets so the analyzer only sees the flash?

This is how I did all my testing with Reflex Analyzer and had great results. Never tried a bright background like in the video.

Look at the 2nd post, I uploaded the wrong video. I can't make the background darker than the 2nd video.

[nuggify]

I don't think so. I have this desync, bad hitreg problem also in LAN's, where PC's, monitors, network and of course electricity is different. I still have this lag. How this is even possible? I think the only logical reason is: Mice, keyboard, headset infected or damaged by EMI, RFI. Other thing could be that steam account causes this lag, but i have this lag in all fps shooters, not only in cs.
The only thing that helps a little bit with hitreg is old internet driver + make sure Realtek ethernet is unticked in MSI utility mode, and hyper threading disabled. Like I said, it helps just a little bit, but far away from real solution. So annoying :/

Not sure about your situation, but for mine I found locations where my exact hardware and peripherals work perfectly fine. This leads me to conclude nothing is damaged or wrong with the hardware. In my city I have tried 4 different homes and at every single one the problem is there (to different degrees, which also vary daily). At some places there are notably differing symptoms however in general they are the same. Blurred/degraded display (and graphical artifacts), dropped and delayed inputs, minor desync, poor audio quality, poor network performance, and random glitches/hangups in windows.

Therefore I have concluded that the problem here is area wide, (and somehow electrical in nature). I do not know how this is possible, I have tried a lot. Some of us have had some real luck with shielding the PC and cables. I have not tried an optimal version of this (server grade PC case with thick steel enclosure). Trying shielded cables actively made the symptoms worst for me. Current theories are that either somehow our PC is extremely sensitive to completely normal amounts of interference in our environment- or there is high RFI (very hard to properly measure) on the AC mains and it is radiating off all wiring throughout the home, and thus inducting onto the data and power cables to my PC. I have tried enough power filtering to know that it is not a solution. About $2k in power filtering for my PC. If there was a way power filtering could work it would have to be at the panel, and likely targeting RFI frequencies.

[spoidah]

I may have found a workaround maybe? I'm not gunna say it's a fix, because it's not quite a fix, just a big improvement, and maybe the result will be temporary (as have many tweaks and tests over the years), I'm hoping the effects don't degrade and go back to '''''normal''''', but whatever, at least I have video footage of it.

phpBB [video]

I know most of us have experience with overclocking and attempted to use lower vcore voltages and LLC's before, but I attempted to apply what I know about electronics and general electrical theory and put a bit of a spin on it (thanks to nug, unixko, numberfive and a bunch of the others in the discord, we were able to sort of educate ourselves and spread journal articles and training manuals around). I'm trying to avoid saturation of capacitors in the mobo (which filter unstable voltages to make things steady) by reducing power draw across the system. If capacitors aren't filtering the power properly, you can end up with harmonics inside your PC components if there's some external influence acting on it (e.g. current ripple, rfi, whatever) which increase heat, voltage draw, cause excessive EMI emission/radiation, and can make the system unstable, I wanna leave as much 'overhead' as possible for my voltages, to try and compensate for the issue.

... I'll just edit the first page so people see this faster.

[ashrr]

Could you try the same thing but this time offline? Something like this: (4:42)

phpBB [video]

If you really want to test to see if a specific change makes a difference with input lag you need to remove as many variables as possible. So test on the same map, offline, in a dark corner, looking at exact same spot, with the same fps so cap it to what you can hit with both 5GHz and 4.6GHz, and finally with the analyzer box in the same position. In the video you just posted the analyzer is in different places so this is probably affecting results too along with the lower overall framerate.

[spoidah]

Could you try the same thing but this time offline? Something like this:

If you really want to test to see if a specific change makes a difference with input lag you need to remove as many variables as possible. So test on the same map, offline, in a dark corner, looking at exact same spot, with the same fps so cap it to what you can hit with both 5GHz and 4.6GHz, and finally with the analyzer box in the same position. In the video you just posted the analyzer is in different places so this is probably affecting results too along with the lower overall framerate.

Yeah, I can make one of those, but I was considering just making a new Kovaak vid, since inputs on that game were much, much worse, and doing a side-by-side comparison between 5GHz and 4.5GHz (I dropped from 4.6 to 4.5 btw, CPU didn't like 1.15V at that frequency). The change doesn't look quite as dramatic on CS:GO, but I assure you it is, I just don't know whether I should keep dropping the vcore and frequency lower and lower. Maybe I'll be able to reach voltage nirvana.

[ashrr]

Could you try the same thing but this time offline? Something like this:

If you really want to test to see if a specific change makes a difference with input lag you need to remove as many variables as possible. So test on the same map, offline, in a dark corner, looking at exact same spot, with the same fps so cap it to what you can hit with both 5GHz and 4.6GHz, and finally with the analyzer box in the same position. In the video you just posted the analyzer is in different places so this is probably affecting results too along with the lower overall framerate.

Yeah, I can make one of those, but I was considering just making a new Kovaak vid, since inputs on that game were much, much worse, and doing a side-by-side comparison between 5GHz and 4.5GHz (I dropped from 4.6 to 4.5 btw, CPU didn't like 1.15V at that frequency). The change doesn't look quite as dramatic on CS:GO, but I assure you it is, I just don't know whether I should keep dropping the vcore and frequency lower and lower. Maybe I'll be able to reach voltage nirvana.

The game doesn't really matter but CS:GO would be better since you can compare your results with FR33THY's and we're only looking at the difference in input lag. If you do want to use Kovaak's try in sandbox mode where you can change the colour of the floor and walls to solid black.

[empleat]

Here we go folks, did some input lag testing with MSI afterburner and RTSS for your viewing pleasure.

How do you test input lag with Afterburner and RTSS??? Google found nothing.

Only method I ever heard about - is to use 1000fps camera and led plugged into a USB, so you can see when you move your mouse, or click - not sure if click latency is same every time!

And even that doesn't 100% test input lag, what if you have multiple bottlenecks under 1ms, you would have to address multiple of them. And input lag will differ every time slightly, because various reasons. Still it can help for testing major input lag. I do not have it, because it is expensive and I Am very sensitive to input lag, so I can usually tell, if there is a difference.

PS: Must be annoying issue! I heard UPS can help with EMI, because it will smooth out current. But don't know how much and it is expensive! Also someone said this can be used to test your current from outlet: https://www.amazon.com/gp/product/B000RUL2UU/

[f1ndus]

hello boys, after few days testing and taking informations i am sure PC is not affected (maybe in some cases)
- change place with only PC and strager monitor = everything FINE.
- change place with only MONITOR = PROBLEM, stutters, unsmooth , input lag

After 1 day reading information aboute behavior monitor to electricity power.

Monitors are really prone to unbalanced power electric voltages.
ASUS tech support confirmed - Monitor can be damaged by inconsistent voltages in house from outlet.
Online UPS with AVR and pure sine wave can help but does not have to.

Probably YOURS monitors are already damaged and nothing will help.

+ new monitor directly plug to online UPS or some of type stabilizer power.

[spoidah]

The game doesn't really matter but CS:GO would be better since you can compare your results with FR33THY's and we're only looking at the difference in input lag. If you do want to use Kovaak's try in sandbox mode where you can change the colour of the floor and walls to solid black.

Done.

phpBB [video]

I looked at sandbox mode, but I just didn't experience the laggy stuttery behaviour on those community maps. I think it's specific to the 'training mode'. Anyway, to compensate for this, I picked a weapon with a much brighter muzzle flash.
One or two of the high values on the 5GHz side was me missing the 'reset' part of the screen that spawns a new bot, and the game-mode requires you to spawn a new bot in order to fire, so that value isn't valid. I also switched back from 4.5GHz to 5GHz in order to test this, but the problem seems to get to peak awful after keeping it at the higher frequency for a long while, so while it was a little unstable feeling compared to the 4.5GHz attempt, it still felt more stable than the previous video I did the first time (the 2nd video I posted on the first page of this thread)

I tried doing a fr33thy style CS video, but decided against it. The main problem is that I don't get stutters randomly, it's in response to enemy players, when they peek me, and when I try to shoot at them, shooting at a wall in a dark part of a map doesn't display the problem, I even tried to gather footage on a deathmatch server, shooting in CT spawn just like fr33thy did, it didn't show anything unusual.