Page 1 of 2

Scanline sync - Input Lag tests

Posted: 16 Jul 2020, 12:09
by Defox
I tried searching but never found any "Scanline Sync" tests compared to "No Sync". This feature is awesome, but I'm worried about how much input lag it adds. I would like to use this feature in competitive shooters where low latency is very important and would not want this feature to degrade my performance. In my case, I have a 144Hz monitor.

There are tests anywhere with comparison of input lag "Scanline Sync"? Thank you in advance!

Re: Scanline sync - Input Lag tests

Posted: 16 Jul 2020, 13:06
by jorimt
Defox wrote:
16 Jul 2020, 12:09
This feature is awesome, but I'm worried about how much input lag it adds.
Scanline Sync is V-SYNC off, but with the tearline manually steered off-screen. So it's effectively the same thing as limiting no sync to your max refresh rate with an FPS limit, i.e. 144 FPS @144Hz, 240 FPS @240Hz, etc, but without the visible tearline when properly configured.

No sync would only have notably lower input lag over Scanline Sync if your system wasn't GPU-limited AND it could reach framerates far above the refresh rate (3x or more). And this input lag decrease wouldn't be due to any decrease in actual frame buffering, but directly due to the reduction in frametimes from the higher sustained FPS (and thus more possible tearlines inside a single scanout).

As for input lag tests, I've already done similar:
https://blurbusters.com/gsync/gsync101- ... ettings/9/

Take the second scenario listed in the below image (ignoring the 142 vs. 144 FPS cap; not enough to make a difference with no sync), and then look at the remaining scenarios below it:

Image

That's your effective Scanline Sync vs. uncapped (non-GPU-bound) no sync input lag difference @144Hz. The higher the framerate above the refresh rate = the more tearlines in a single scanout = the lower possible input lag due to said tearlines per single scanout.

Depends on if your system can reach and sustain such high framerates without maxing out the GPU (at which point the pre-rendered frames queue begins to increase, introducing more buffers, aka lag).

Any tear-free method is always going to limit your display to delivering a single complete frame per scanout cycle for the given max refresh rate. There's no way around that if you want no tearing at currently achievable refresh rates.

Re: Scanline sync - Input Lag tests

Posted: 17 Jul 2020, 11:57
by Defox
Thanks for the detailed answer!

If I understood correctly, there is practically no difference in Input Lag between No Sync and Scanline Sync. Therefore, when Scanline Sync is on, the more I have a GPU power reserve, the less input lag will be (which is similar to the additional number of frames with No Sync).

I have already tried this function and set the Scanline Sync to a negative value, hiding the tear line above the top of the montior screen, which is recommended for lower latency. And I can say that I really didn't notice any additional input lag! It looks amazing with motion blur reduction, perfectly flat picture without stutters and tearing. Now I don't want to turn off scanline sync ever!

Re: Scanline sync - Input Lag tests

Posted: 17 Jul 2020, 12:11
by Defox
P.S. I read your post again and finally understood. There will be no difference in the scenario when 144 fps with No Sync and Scanline Sync. But with fps above 144, the option with No Sync will slightly reduce the latency (judging by the tests, by another 4 ms). I can say that it is absolutely not worth it. A crystal clear picture will provide more benefits than an additional 4 ms of latency reduction.

At one time I was disappointed that motion blur reduction cannot be used with G-Sync. And using a regular VSync adds too much input lag. Finally, there is a really great solution to this problem.

Perhaps someday I hope we will see the scanline sync function in the official nvidia drivers. Or the monitors can somehow support G-Sync simultaneously with motion blur reduction.

Re: Scanline sync - Input Lag tests

Posted: 17 Jul 2020, 13:24
by jorimt
Defox wrote:
17 Jul 2020, 11:57
I have already tried this function and set the Scanline Sync to a negative value, hiding the tear line above the top of the montior screen, which is recommended for lower latency. And I can say that I really didn't notice any additional input lag! It looks amazing with motion blur reduction, perfectly flat picture without stutters and tearing. Now I don't want to turn off scanline sync ever!
Yeah, it's a great option if your system can sustain the framerate above the refresh rate in the given game at all times (and you can keep your GPU usage below max simultaneously), especially for low-lag, tear-free strobing.
Defox wrote:
17 Jul 2020, 12:11
Or the monitors can somehow support G-Sync simultaneously with motion blur reduction.
There is ELMB Sync (FreeSync + Strobing), but it ironically doesn't handle the variable framerate part very well currently. No doubt it could improve in the future though and maybe even ultimately support (hardware and/or G-SYNC Compatible) G-SYNC as well.

Re: Scanline sync - Input Lag tests

Posted: 03 Sep 2020, 03:42
by alexander1986
Defox wrote:
16 Jul 2020, 12:09
I tried searching but never found any "Scanline Sync" tests compared to "No Sync". This feature is awesome, but I'm worried about how much input lag it adds. I would like to use this feature in competitive shooters where low latency is very important and would not want this feature to degrade my performance. In my case, I have a 144Hz monitor.

There are tests anywhere with comparison of input lag "Scanline Sync"? Thank you in advance!

jorimt wrote:
17 Jul 2020, 13:24
...

hm, looking at ashuns button-to-pixel lag tests over at aperturegrille, it does seem scanline sync adds a frame or so (?) of delay vs no scanline sync at the same framerate :

Image

(taken from his input lag / framelimiter test here: https://www.youtube.com/watch?v=8ZRuFaFZh5M )


although this is with gsync enabled, not sure how that affects the result..

from personal subjective testing in games like fortnite and fifa, playing with 120 hz / 120 fps in-game limit vsync off, vs 120 hz / 120 fps limit via scanline sync ON, and unlimited fps ingame, there is a clear difference in input delay to me at least, scanline sync feeling "heavier" in terms of mouse responsiveness, same thing in fifa, I can turn around and dribble faster and easier without scanline sync as compared to with it..


anyway it would be interesting to see the real numbers without any gsync or freesync involved as well.

Re: Scanline sync - Input Lag tests

Posted: 03 Sep 2020, 07:20
by RealNC
Scanline sync uses the RTSS frame limiter, so you need to compare against that to find if it adds any latency on top of just the frame limiter. So first test RTSS cap at FPS=Hz, then test scanline sync and compare.

The RTSS limiter having the "natural" latency of the cap (like 16.7ms at 60Hz) is known.

Re: Scanline sync - Input Lag tests

Posted: 13 Dec 2020, 22:29
by Caronizeeee
alexander1986 wrote:
03 Sep 2020, 03:42
Defox wrote:
16 Jul 2020, 12:09
I tried searching but never found any "Scanline Sync" tests compared to "No Sync". This feature is awesome, but I'm worried about how much input lag it adds. I would like to use this feature in competitive shooters where low latency is very important and would not want this feature to degrade my performance. In my case, I have a 144Hz monitor.

There are tests anywhere with comparison of input lag "Scanline Sync"? Thank you in advance!

jorimt wrote:
17 Jul 2020, 13:24
...

hm, looking at ashuns button-to-pixel lag tests over at aperturegrille, it does seem scanline sync adds a frame or so (?) of delay vs no scanline sync at the same framerate :

Image

(taken from his input lag / framelimiter test here: https://www.youtube.com/watch?v=8ZRuFaFZh5M )


although this is with gsync enabled, not sure how that affects the result..

from personal subjective testing in games like fortnite and fifa, playing with 120 hz / 120 fps in-game limit vsync off, vs 120 hz / 120 fps limit via scanline sync ON, and unlimited fps ingame, there is a clear difference in input delay to me at least, scanline sync feeling "heavier" in terms of mouse responsiveness, same thing in fifa, I can turn around and dribble faster and easier without scanline sync as compared to with it..


anyway it would be interesting to see the real numbers without any gsync or freesync involved as well.
Hello friend, I'm having problems with FIFA 21.

I'm getting 350fps + in-game on a 240hz monitor. I happen to be having micro stutterings, the screen is not as fluid as it was before I changed my processor yesterday. I traded a Ryzen 7 3700x for a Ryzen 7 5800x. My video card is an RTX 2070 and I have 2x8gb 3800mhz 1: 1. I tried to limit my fps by NVCP but I felt a ton of input lag like you said, I can't dribble right anymore, the players seem heavy. Do you have any solution? Thanks

Re: Scanline sync - Input Lag tests

Posted: 13 Dec 2020, 23:40
by RealNC
Caronizeeee wrote:
13 Dec 2020, 22:29
Hello friend, I'm having problems with FIFA 21.

I'm getting 350fps + in-game on a 240hz monitor. I happen to be having micro stutterings, the screen is not as fluid as it was before I changed my processor yesterday. I traded a Ryzen 7 3700x for a Ryzen 7 5800x. My video card is an RTX 2070 and I have 2x8gb 3800mhz 1: 1. I tried to limit my fps by NVCP but I felt a ton of input lag like you said, I can't dribble right anymore, the players seem heavy. Do you have any solution? Thanks
Also try an RTSS cap. If you use g-sync, cap to 237, then to 230 and see if the somewhat lower cap gives better latency.

In any event, it will take a while to retrain your muscle memory. The latency difference isn't much, you just need to get used to that difference.

Re: Scanline sync - Input Lag tests

Posted: 13 Dec 2020, 23:47
by Caronizeeee
RealNC wrote:
13 Dec 2020, 23:40
Caronizeeee wrote:
13 Dec 2020, 22:29
Hello friend, I'm having problems with FIFA 21.

I'm getting 350fps + in-game on a 240hz monitor. I happen to be having micro stutterings, the screen is not as fluid as it was before I changed my processor yesterday. I traded a Ryzen 7 3700x for a Ryzen 7 5800x. My video card is an RTX 2070 and I have 2x8gb 3800mhz 1: 1. I tried to limit my fps by NVCP but I felt a ton of input lag like you said, I can't dribble right anymore, the players seem heavy. Do you have any solution? Thanks
Also try an RTSS cap. If you use g-sync, cap to 237, then to 230 and see if the somewhat lower cap gives better latency.

In any event, it will take a while to retrain your muscle memory. The latency difference isn't much, you just need to get used to that difference.
Thank you very much, I will test the RTSS. Should I just limit the FPS on it or do I have any specific settings to do with it? In NVCP I leave vsync on? Latency mode? It is difficult to understand how a stronger processor can make the game worse by not having a limiter within the game, sad.

Another doubt, I never used RTSS. I can only limit the FPS of Fifa in it and not the other games right? And it will only work if the application is open, correct?

Where can I download only RTSS, without afterburner, in a more updated version?

Thanks!!