windows color profile for lowest input lag

[Chief Blur Buster]

Nobody cares[1] about your little ego[2], and how you're trying to prove that "you know better"[3], and the other party has just fallen victim to placebo...

Strike 1,2,3*
*by itself, Strike 3 isn't really a strike, but becomes a strike when supplemented by mood-damage in strike 1, 2 and 5.

You're not helping anyone.[4]

Strike 4*
*by itself, this isn't really a strike, but becomes a strike when supplemented by mood-damage in strike 1, 2 and 5.

I was defending your position, dummy[5].

Strike 5

This is the best, most open-to-controversial-topics forum that I've ever been on!
I love it.

I appreciate the compliments (and other productive contributions!) but we have a "Be Nice To New Users" rule that all forum members must follow.

Coincidentially, this rule also helps us maintain our "most open-to-controversal topics" while being scientifically reputable -- this enforced forum rule prevent attacks on fully meritworthy topics relevant to the refresh rate race.

While I can relate to frustrations by users, there are published rules that we all humbly have to enforce to make sure this is one of the better on the Internet in these spheres. Next strike by you could result in some temporary administrative action (even if the other party is wrong and you are right) -- as part of our well-known "Be Nice To New Users" rule. It is appreciated if you could avoid making other people feel bad on our forum.

Thank you...

[Chief Blur Buster]

And what is that supposed to achieve ???

Easy, it's a legitimate question. To foster good discussion, we need to be patient even if some questions are redundant. "Be polite" is also a published forum rule here. I rarely have to remind people about forum rules thankfully (it happens only approximately once a month!), but let's get back and start off on our good feet, shall we...

We thrive on scientific basis such as photographs and measurements. Sometimes this requires us to annoyingly redo tests, so evidence is successfully recordable. We can help you with creating a test that will allow creation of recordable proof that we are extremely interested in!

Your topic is meritworthy.

Pretty sure ICC profile input lag is a myth.

While I know that ICC profiles can be lagless and not a problem at all almost all the time (even lagless >99% of the time), I'm not 100% convinced that they are guaranteed always 100% lagless on all configs.

Sometimes ICC is transmitted to the hardware with certain graphics drivers on certain graphics chips -- and sometimes some firmwares do it laglessly, and other firmwares does it laggy. Normally, ICC is completely done only on GPU, but there exists some DDC/CI commands (VESA Monitor Command Control Set Version 2.2) that may transmit various ICC setting changes, and some graphics drivers may transmit certain picture settings via DDC commands to the monitor. At that point, certain picture settings becomes monitor processing responsibility. And if one of the DDC commands (transmitted from PC to the monitor) hits one model's laggier-than-expected monitor-processing firmware feature unexpectedly, and.... boom, question becomes legitimate!

Just like monitor scaling is often hit-miss, sometimes lagless, and sometimes laggy. I know GPU scaling can be virtually lagless, but it wasn't always in previous GPUs. And I've seen GPU scaling less laggy than monitor scaling. And I've seen monitor scaling less laggy than GPU scaling. But you've noticed, nowadays, I recommend sticking to GPU scaling for simplicity because it's consistent behaviours on modern GPUs)

I think it needs further study. As part of our Blur Busters philosophy, we've found unexpected things matter in the refresh rate race to retina refresh rates. Whether it is "0.5ms-vs-1.0ms GtG is human visible in some situations", or that it is "1000Hz-is-worth-it". Some topics are quite out-there, but we keep an open mind.

I deem this topic meritworthy of further study, though our mental battle is currently on other battlefronts in the refresh rate race. If other users have the equipment to accurately test this, I'd love to hear!

[KNUTZ]

If you remove icc color profile, you can notice you get a lot of less input lag. However it is still unclear if some windows color profile isn't selected. As in advanced tab, you can't choose none, but you have to select something. I am not expert on windows color profile and i don't know how it affects input lag. I wasn't able to google anything. Is there any way to measure it, it may be difficult, if there is less difference than 1 ms, you can't even test that with 1k camera... I ask, because you could make custom profile with intent to lower input lag even further. Anyone have any idea how windows color profiles work ???

So which of the default profile you think is the fastest? The srgb or the scrgb? And which one was the default one?

[Chief Blur Buster]

I think that direct-ICC-induced latency is so rare that it's not one of the topics usually worth worrying about.

However, there is another common latency issue: Different latencies of different pixel colors, as known from GtG heatmapping.

Also, I have thought about this further in a scientific tick-tock in my grey matter.

Since monitors have different lag for different pixel colors (as GtG latency heatmapping shows, like those done by TFTCentral and ApertureGrille, etc), it would so scientifically surmise -- in a Sherlock Holmes style deduction -- that ICC changes could shift specific games' main colors into faster color territory or laggier color territory -- and would thus hugely vary depending on the panel and panel type (IPS vs TN vs VA) as certain colors are faster pixel response than other colors. And that it varies on a monitor-by-monitor basis.

Thus, I believe there is an additional potential cause-effect study. Latency of different colors may need to be considered. ICC calibration can also move a game's dominant colors into laggier territory (example: slow-responding VA greys).

Also, numeric latency stopwatching is often arbitrary and nonstandard (e.g. latency benchmarking ending at GtG2% versus GtG10% versus GtG50% versus GtG90% versus GtG100%), and different websites stopwatch input latency very differently. GtG10% is still human visible (it's like RGB(25,25,25) in the pixel journey from RGB(0,0,0) gradually fading/transitioning to RGB(255,255,255). I know at least one website stopwatch lag til GtG2% and a different website benchmarks till GtG100%. And lag of VSYNC ON versus VSYNC OFF have extremely different mechanics that makes numbers impossible to directly compare to each other across different websites.

TL;DR: There are also situations where ICC calibration is lagless, but a specific panel might have laggier pixel colors than others. Example is laggy dark colors of VA panels. A darker ICC calibration may thus increase VA lag, and a brighter ICC calibration may thus decrease VA lag. Recalibration moves the colors around. This would not be ICC latency but a side effect of ICC calibration moving common colors to faster colors in the GtG heatmap.

GtG heatmaps are often totally different between different monitors, and different games have different dominant colors that affects different parts of the GtG heatmap. So there's no one-size-fits-all advice except create a custom calibration that feels better for you, such as spiking the gamma extremely bright -- sacrificing your colors to improve latency.

It's quite the rabbit hole to fall into.

[1000WATT]

What are your Input Lag AB results?

At 16ms I got 21/25 without ICC and 20/25 with ICC, so I'd say input lag is identical. Maybe Windows is screwing with you.

Suppose the total system delay in this test is 10ms. Minimum 10ms, maximum 26ms.
Assume icc profile added 5ms.
16ms 21/21 (10-26)
16ms 20/25 (15-31)
In either case, there was a delay of 16ms between a and b.
Most likely I did not understand what exactly this test does

[hleV]

What are your Input Lag AB results?

At 16ms I got 21/25 without ICC and 20/25 with ICC, so I'd say input lag is identical. Maybe Windows is screwing with you.

Suppose the total system delay in this test is 10ms. Minimum 10ms, maximum 26ms.
Assume icc profile added 5ms.
16ms 21/21 (10-26)
16ms 20/25 (15-31)
In either case, there was a delay of 16ms between a and b.
Most likely I did not understand what exactly this test does

Your calculations are correct, however with added input lag the test results should improve somewhat. It's gonna be hard to notice a 1ms vs 6ms difference, however if 5ms is added on top of 16ms lag, you should have an easier time noticing that it's lagging. So e.g. 1ms vs 21ms should mostly feel the same as 6ms vs 21ms (the laggy side becomes more noticeable while non-laggy side not so much), but 6ms vs 21ms should be more noticeable than 1ms vs 16ms. Basically it's not about the ms difference between non-lag and lag but at what point the feeling of lag becomes more noticeable.

If my ICC added [an impactful amount of] input lag, I should be able to feel the laggy side better in Input Lag AB.

[1000WATT]

Your calculations are correct, however with added input lag the test results should improve somewhat. It's gonna be hard to notice a 1ms vs 6ms difference, however if 5ms is added on top of 16ms lag, you should have an easier time noticing that it's lagging. So e.g. 1ms vs 21ms should mostly feel the same as 6ms vs 21ms (the laggy side becomes more noticeable while non-laggy side not so much), but 6ms vs 21ms should be more noticeable than 1ms vs 16ms. Basically it's not about the ms difference between non-lag and lag but at what point the feeling of lag becomes more noticeable.

If my ICC added [an impactful amount of] input lag, I should be able to feel the laggy side better in Input Lag AB.

I understood your logic.
If it worked that way, then you can easily type 25/25 at a 300ms b 316ms.
Although it seems to me it will be even more difficult than 1-16.

[empleat]

I think that direct-ICC-induced latency is so rare that it's not one of the topics usually worth worrying about.

However, there is another common latency issue: Different latencies of different pixel colors, as known from GtG heatmapping.

Also, I have thought about this further in a scientific tick-tock in my grey matter.

Since monitors have different lag for different pixel colors (as GtG latency heatmapping shows, like those done by TFTCentral and ApertureGrille, etc), it would so scientifically surmise -- in a Sherlock Holmes style deduction -- that ICC changes could shift specific games' main colors into faster color territory or laggier color territory -- and would thus hugely vary depending on the panel and panel type (IPS vs TN vs VA) as certain colors are faster pixel response than other colors. And that it varies on a monitor-by-monitor basis.

Thus, I believe there is an additional potential cause-effect study. Latency of different colors may need to be considered. ICC calibration can also move a game's dominant colors into laggier territory (example: slow-responding VA greys).

Also, numeric latency stopwatching is often arbitrary and nonstandard (e.g. latency benchmarking ending at GtG2% versus GtG10% versus GtG50% versus GtG90% versus GtG100%), and different websites stopwatch input latency very differently. GtG10% is still human visible (it's like RGB(25,25,25) in the pixel journey from RGB(0,0,0) gradually fading/transitioning to RGB(255,255,255). I know at least one website stopwatch lag til GtG2% and a different website benchmarks till GtG100%. And lag of VSYNC ON versus VSYNC OFF have extremely different mechanics that makes numbers impossible to directly compare to each other across different websites.

TL;DR: There are also situations where ICC calibration is lagless, but a specific panel might have laggier pixel colors than others. Example is laggy dark colors of VA panels. A darker ICC calibration may thus increase VA lag, and a brighter ICC calibration may thus decrease VA lag. Recalibration moves the colors around. This would not be ICC latency but a side effect of ICC calibration moving common colors to faster colors in the GtG heatmap.

GtG heatmaps are often totally different between different monitors, and different games have different dominant colors that affects different parts of the GtG heatmap. So there's no one-size-fits-all advice except create a custom calibration that feels better for you, such as spiking the gamma extremely bright -- sacrificing your colors to improve latency.

It's quite the rabbit hole to fall into.

Wow that's a great post. Problem: even if you had 1000fps camera. It can add less than 1 ms input lag per frame, but it can scale over time. Like 100 us here and there. What if it adds like 100 us to each frame, but than it would become greater value over time. And after 1 second, it can exceed 1 ms of total input lag, as it scaled! I don't know how exactly rendering on monitor works and whether, or not: there would be discernible difference - even on1000fps camera... I usually go by feel, as i can tell even 6 ms difference in input lag! Tested myself with input lag AB test from this forum... I will download more profiles and test them out probably, but there are so many of them! I tried couple and nothing was better, than what i was using...

On both ASUS VG248QE and AOC G2590: SCRGB virtual profile has lowest latency, but SRGB virtual profile is more consistent. Using no profile was fine on my old monitor (ASUS), but on my new - sRGB IEC61966-2.1 was set automatically in Windows and feels best. I also turn off DDCI, it lowers input lag at lot. On my old monitor i did also set sharpness to 0%, as it added input lag as well.
And scalling on monitor definitely has lower input lag, on both: my old and new PC! And on 2 monitors, but ASUS vg248QE, didn't habe scalling on monitor, which was shame, it was one of best gaming monitors ever!

It would be interesting, if someone did ultimate low input lag color profile: for best gaming monitors! I don't know how much manufacturers care about this, but not every default color profile for a certain monitor feels best. I also i tried even play black and white, but i didn't notice difference, i would even play 16 bit, if it was possible. It used to be possible on Windows 7 and it reduced input lag!

Hmm i tried couple color profiles of low input lag monitors and it wasn't better than default profile in windows it was set to. I am not sure, if this was default windows profile, or Windows selected best for my monitor. Because i couldn't find specific color profile for my monitor. And i thought default is ICC not IEC. Anyway there are to many of them to try them one by one and by some chance one may be better.

Btw i love how some people randomly jump into discussion and tell you these things are myths and don't do anything and yet they start swear to you for no reason. So if you think, it doesn't matter. Why you are bothered so much by fact, that someone else think it does? It is like that on every forum. If you discuss something about input lag. Random people jump in and start to flame you LOL. These are probably people which think - eye can't see more than 24 fps...

[vurt]

I also turn off DDCI, it lowers input lag at lot. On my old monitor i did also set sharpness to 0%, as it added input lag as well.

Is "DDCI" a setting in the monitor or is it a Windows setting?

Interesting topic btw...

[Chief Blur Buster]

VESA DDC/CI is a monitor setting. But keeping it enabled doesn't add input lag on most modern monitors -- some models of monitors may have had issues with input lag for that, however.