FPS limiters and Track-IR performance

I’ve been using this method for months....

144hz monitor but set it at 120hz through Windows.

In NVCP set the frame rate to lock at 60.

Set Vsync in NVCP at 1/2 refresh rate. (To match the monitor)

This gives me glass smooth Tracking with TrackIR.

The frames don’t drop below 60 no matter where I fly.

So this thread is really to discuss the best method of fps locking. In my tests, NVInspector "Limiter V2 - All Off" gets the best results and maintains the best lock on 60 fps. RTSS fluctuates a lot, especially under load in heavy missions. RTSS supposedly has lower input lag but no one has tested Limiter V2 so we don't know. And really, who cares? This isn't CS:GO we're playing here.

30mikemike, I think you mean NVInspector as NVCP doesn't have those options. I'll try the v-sync @ 1/2 refresh rate and report back. Thanks.

So this thread is really to discuss the best method of fps locking. In my tests, NVInspector "Limiter V2 - All Off" gets the best results and maintains the best lock on 60 fps. RTSS fluctuates a lot, especially under load in heavy missions. RTSS supposedly has lower input lag but no one has tested Limiter V2 so we don't know. And really, who cares? This isn't CS:GO we're playing here.

 

30mikemike, I think you mean NVInspector as NVCP doesn't have those options. I'll try the v-sync @ 1/2 refresh rate and report back. Thanks.

Sorry yes, Inspector, not Control Panel. There’s so many different options and what works for one might not for another but I thought I’d share my setup. See if it works for you mate...

We’re getting nowhere with this discussion. Read the article. It is perfectly clear what the advice is. It’s up to you if you want to accept it or not.

Maybe we should take the discussion over to guru3d forum and see what the folks over there say. I'm saying that capping my game at 60 fps while running at 144hz does not disengage G-Sync as you are claiming. You are absolutely right that capping at 58 fps on a 60hz panel is ideal for G-Sync.. No argument there. Is 58 fps ideal for TrackIR? Not in my experience.

Most people who have G-Sync have 100hz or higher monitors so advice that 58 fps is one size fits all setting is not good advice.

Maybe we should take the discussion over to guru3d forum and see what the folks over there say. I'm saying that capping my game at 60 fps while running at 144hz does not disengage G-Sync as you are claiming. You are absolutely right that capping at 58 fps on a 60hz panel is ideal for G-Sync.. No argument there. Is 58 fps ideal for TrackIR? Not in my experience.

 

Most people who have G-Sync have 100hz or higher monitors so advice that 58 fps is one size fits all setting is not good advice.

Good grief! I'm not suggesting 58fps is a one-size-fits-all. That was just relating to 60Hz monitors. The advice on Guru3D and blurbusters etc. is to use a cap that is about 2 below the refresh rate. So, in the case of 144Hz monitors, you would use 142.

I had a very long thread on this a few months back on Guru3D, and learned a lot from the experts there.

I really think this whole TrackIR is a red herring. If TrackIR had issues with variable frame rates, then this forum and others would be flooded with thousands of users complaining. As it is, we seem to be restricted to a small few.

As I said, the advice from those who know is out there, so it's purely up to you in you want to accept it or not. Whatever works best for you is the best solution for you!

a little bit off topic ... because i have a 60hz monitor, but ...

the trick "frame rate limiter mode V2 - force off" works very well for me

the anoying trackir stuttering has gone ... (most of the time :))

with vsync on i have a smooth DCS rendering and a smooth trackIR moving

thx a lot for this tipp :thumbup::thumbup::thumbup:

but i have to use most of the time ALT-Enter, but that is not new :)

Edit said ...

with more units and more traffic within the mission there are some breaks and stuttering available ... but i think it will be a little bit better than before ...

behaviour of TrackIR is a little bit anoying ...

Quote from Natural point:

"TrackIR is also a great option because it has a processor built in to the tracking unit, which means that all data processing and head movement translation calculating is done on the device itself: this means there is no performance hit on your system as this data is streamed into the TrackIR software."

You can say that proves nothing, but this is one point why I am so septic.

But.

How is TrackIR software connected to the DCS software? TrackIR can deliver its data every 5ms. This is 0,005s, 60Hz is 0,0167s

TrackIR is almost 3x faster then the 60Hz monitor.

I still don't get the connection, then again many of you are rock steady on that there system IS influenced by TrackIR...

Ah and regarding the benefits of syncronising monitor and graphics card, what happens when you dont,

why tripple buffering might be a good idea from time to time:

The extra input lag, as also mentioned in that video, might be ok for Lara Croft ( as in that video ) but kills your formation flying, Air-Air-Refueling and hovering experience.

I'd defo keep that on "1" for a Sim like DCS, or Race games etc..

Quote from Natural point:

 

 

"TrackIR is also a great option because it has a processor built in to the tracking unit, which means that all data processing and head movement translation calculating is done on the device itself: this means there is no performance hit on your system as this data is streamed into the TrackIR software."

 

 

 

You can say that proves nothing, but this is one point why I am so septic.

But.

How is TrackIR software connected to the DCS software? TrackIR can deliver its data every 5ms. This is 0,005s, 60Hz is 0,0167s

TrackIR is almost 3x faster then the 60Hz monitor.

I still don't get the connection, then again many of you are rock steady on that there system IS influenced by TrackIR...

Yes, we are 100% sure that with TiR present we have stutter when looking L + R.

IF that is solely a TiR issue, software, hardware or the way it is implemented/injected into DCS, WHO KNOWS ??

didnt we discuss this for 3h last night and I was able to reproduce each and every claim I and the others made ? I admire your thrive to solve this tho you dont suffer from it with a 60Hz screen but I do not get that you do not accept our pure findings aka SYMPTOMS. We can argue about the root cause that we havent found yet, but we are passed the Symptom finding, that is really confirmed criss-cross from many.

Yes, we are 100% sure that with TiR present we have stutter when looking L + R.

 

IF that is solely a TiR issue, software, hardware or the way it is implemented/injected into DCS, WHO KNOWS ??

 

didnt we discuss this for 3h last night and I was able to reproduce each and every claim I and the others made ? I admire your thrive to solve this tho you dont suffer from it with a 60Hz screen but I do not get that you do not accept our pure findings aka SYMPTOMS. We can argue about the root cause that we havent found yet, but we are passed the Symptom finding, that is really confirmed criss-cross from many.

I'm not sure why some insist on acting like this issue doesn't exit. I've tested TrackIR on two different systems and in four different sims and they all do the same micro-stuttering unless locked at 60 fps.

I know that you and Goa tried to find a solution for this as well. There may not be one but I've opened another case with NaturalPoint to see what they say.

So this thread is really to discuss the best method of fps locking. In my tests, NVInspector "Limiter V2 - All Off" gets the best results and maintains the best lock on 60 fps. RTSS fluctuates a lot, especially under load in heavy missions. RTSS supposedly has lower input lag but no one has tested Limiter V2 so we don't know. And really, who cares? This isn't CS:GO we're playing here.

 

30mikemike, I think you mean NVInspector as NVCP doesn't have those options. I'll try the v-sync @ 1/2 refresh rate and report back. Thanks.

hi Bandit648, after some testings I have to agree with you that nvidia inspector fps limiter v2 all off seems to produce overall better perormance, a more rocksolid 60fps compared to rivaturner and also it seems lees input lag and most of all fps tends not to drop below 60 very often, nice discovery man:thumbup::thumbup::thumbup:

I'm not sure why some insist on acting like this issue doesn't exit. I've tested TrackIR on two different systems and in four different sims and they all do the same micro-stuttering unless locked at 60 fps.

 

I know that you and Goa tried to find a solution for this as well. There may not be one but I've opened another case with NaturalPoint to see what they say.

For the same reason as people with 9xx cards claim that VR is smooth and great quality. They aren't sensitive to it and can't see it/never seen any better therefore treat it as normal.

The issue is definitely there. I capped my FPS at 120 (via autoexec.cfg) and live with stutters below that, thankfully on my system it rarely goes below 100fps. I have tried everything before, nothing really works. Natural Point needs to address this, they have created TrackIR when 60fp@1080 was top of the game, well, it isn't the case for a long time now. I doubt they will do anything though, we may see useable VR headsets before that ;)

I can't play DCS at 60fps in 2D (don't like how it feels), so I lowered my settings despite having up to date hardware to keep FPS high.

hi Bandit648, after some testings I have to agree with you that nvidia inspector fps limiter v2 all off seems to produce overall better perormance, a more rocksolid 60fps compared to rivaturner and also it seems lees input lag and most of all fps tends not to drop below 60 very often, nice discovery man:thumbup::thumbup::thumbup:

Glad it works for you! I'm thinking we should ask the ED dev's to fix the in-game limiter so we can all just use that. Technically the in-game limiter should offer the best performance but the DCS one is broken.

For the same reason as people with 9xx cards claim that VR is smooth and great quality. They aren't sensitive to it and can't see it/never seen any better therefore treat it as normal.

 

The issue is definitely there. I capped my FPS at 120 (via autoexec.cfg) and live with stutters below that, thankfully on my system it rarely goes below 100fps. I have tried everything before, nothing really works. Natural Point needs to address this, they have created TrackIR when 60fp@1080 was top of the game, well, it isn't the case for a long time now. I doubt they will do anything though, we may see useable VR headsets before that ;)

 

I can't play DCS at 60fps in 2D (don't like how it feels), so I lowered my settings despite having up to date hardware to keep FPS high.

Yes exactly.. this used to be not so much of an issue since everyone was gaming on 1080p @ 60fps with v-sync on. Now many more people own high refresh-rate displays with VRR in the mix as well.

Glad it works for you! I'm thinking we should ask the ED dev's to fix the in-game limiter so we can all just use that. Technically the in-game limiter should offer the best performance but the DCS one is broken.

+1

Also I dont understand why for someone ingame fps limiter seems to work but for others don't...

ED : yes pls take a look at max_fps limit which is not properly working, thank you :)

Hey guys,

Check out the latest version of RivaTuner Statistics Server that comes bundled with Afterburner. I own EVGA cards and still use Afterburner/Rivatuner in lieu of Precision XOC.

The developer has incorporated a frametime limiter. It is actually a minimum limiter as opposed to a maximum limiter. Averages for frames per second can mask microstutter. It is measured in microseconds so you need to move the decimal right 3 places from 16.6 milliseconds to 16666 microseconds assuming your target is a smooth 60 frames per second.

1 second per 60 frames, 1/60 = .01666666666... seconds per frame = 16.6 milliseconds = 16666 microseconds

I can't speak to any overhead or induced lag but seems to be working well on my end as the OSD frametime graph will show even the slightest hiccup.

You need to click on framerate on the control panel to change it to frametime limiter. 0 is unlimited.

Hope this helps

I tested several techniques limiters FPS and for DCS and the best is based on the legacy #232 from ReaINC.

https://forums.guru3d.com/threads/the-truth-about-pre-rendering-0.365860/page-12#post-5380262

Settings according to the largest delay on the weekend I'll try.

Hey guys,

 

Check out the latest version of RivaTuner Statistics Server that comes bundled with Afterburner. I own EVGA cards and still use Afterburner/Rivatuner in lieu of Precision XOC.

 

The developer has incorporated a frametime limiter. It is actually a minimum limiter as opposed to a maximum limiter. Averages for frames per second can mask microstutter. It is measured in microseconds so you need to move the decimal right 3 places from 16.6 milliseconds to 16666 microseconds assuming your target is a smooth 60 frames per second.

 

1 second per 60 frames, 1/60 = .01666666666... seconds per frame = 16.6 milliseconds = 16666 microseconds

 

I can't speak to any overhead or induced lag but seems to be working well on my end as the OSD frametime graph will show even the slightest hiccup.

 

You need to click on framerate on the control panel to change it to frametime limiter. 0 is unlimited.

Hope this helps

This has been in RTSS for some time now. Personally, I don't see that it is any different to limiting framerate in RTSS.

This has been in RTSS for some time now. Personally, I don't see that it is any different to limiting framerate in RTSS.

You have to look at the graphs in Afterburner or the OSD. 1 second is a long time in computer time. A 60 fps average will mask all kinds of fluctuations that you will see with your eye but will still show as a solid 60 on an OSD or in a graph depending on the polling rate you have set. Frame time is a much more accurate measure of performance and smoothness that your eyes can perceive. I doubt many people can perceive any difference in milliseconds. You can also do fractional framerates out to several decimals but that isn't any more accurate than frametime.

If you don't notice a difference it is perhaps because your rig can hold a solid 60 fps? Others may find this information useful as this has only been a feature since late October and most people don't actually read the changelogs and it wasn't mentioned in the thread.

You have to look at the graphs in Afterburner or the OSD. 1 second is a long time in computer time. A 60 fps average will mask all kinds of fluctuations that you will see with your eye but will still show as a solid 60 on an OSD or in a graph depending on the polling rate you have set. Frame time is a much more accurate measure of performance and smoothness that your eyes can perceive. I doubt many people can perceive any difference in milliseconds. You can also do fractional framerates out to several decimals but that isn't any more accurate than frametime.

Yes, I agree. However, I think RTSS does the same job with frame rate or frame time being limited. See

Also, is this not what Gsync is for?

Gsync is for variable frame rate where the output of your graphics card matches the refresh rate of your monitor and works very well when you are not using TrackIR. Unfortunately, as the OP pointed out it doesn't work with TrackIR. You have to lock your frame rate at 60 or 120 with some method. Which method will work best for you I can't say, it's nice to have options. You'd have to do some very precise testing to be sure.

I've only recently realized that for a lot of the people complaining about stuttering on these boards it's most likely from this issue. I thought it might have been power management settings, unstable OC, background processes, or general user error based on people expecting PCs and software to be plug and play when they inherently aren't. I've been playing at 60Hz with Vsync on this whole time. TrackIR does not work with Freesync, Gsync, or any method of VRR that I'm aware of. If the general player is not aware of this they will waste many hours on unnecessary trouble shooting and the experience won't be smooth. FYI.

Gsync is for variable frame rate where the output of your graphics card matches the refresh rate of your monitor and works very well when you are not using TrackIR. Unfortunately, as the OP pointed out it doesn't work with TrackIR. You have to lock your frame rate at 60 or 120 with some method. Which method will work best for you I can't say, it's nice to have options. You'd have to do some very precise testing to be sure.

 

I've only recently realized that for a lot of the people complaining about stuttering on these boards it's most likely from this issue. I thought it might have been power management settings, unstable OC, background processes, or general user error based on people expecting PCs and software to be plug and play when they inherently aren't. I've been playing at 60Hz with Vsync on this whole time. TrackIR does not work with Freesync, Gsync, or any method of VRR that I'm aware of. If the general player is not aware of this they will waste many hours on unnecessary trouble shooting and the experience won't be smooth. FYI.

I’m afraid I’ll have to respectfully disagree with a lot of what you said, especially with regard to TrackIR, but I’m not going over it all again as it has already been stated in this thread.

I’ll leave you to it.

Sigh...smh.

G sync/ Freesync/ VRR does not work with Track IR. What about that are you not understanding? You have to lock your frame rate at 60 or the experience will not be smooth. If you are playing with a variable frame rate monitor TrackIR will not match the frame rate. I can't explain it any more clearer than that. If you can get TrackIR working with some kind of variable frame rate let us know, but you're wasting your time and you will have stuttering. Telling people they won't isn't going to change that.

Sigh...smh.

 

G sync/ Freesync/ VRR does not work with Track IR. What about that are you not understanding? You have to lock your frame rate at 60 or the experience will not be smooth. If you are playing with a variable frame rate monitor TrackIR will not match the frame rate. I can't explain it any more clearer than that. If you can get TrackIR working with some kind of variable frame rate let us know, but you're wasting your time and you will have stuttering. Telling people they won't isn't going to change that.

exactly,

trackir operates under OPENGL , the game and gync operates in direct3d and this is the reason of stuttering when using trackir withouth fps cap.

The only solution which I along with other have discovered long time ago ( there are several threads in this regard from at least 3 years) is to cap fps at 60 or 120.

The problem is quite simple :

- stable 60 fps trackir and the game is smooth

-stable 120 fps or higher trackir and game is smooth

ANY VALUE OF FPS BETWEEN 60 AND 120 , meaning fluctuation of fps rtange from 60 to 120 will generate JITTERY and stuttery movement when paning around.

If the game runs higher than 120 fps, ie. fluctuation from 120 up to inifnity , this will not generate jittery and will be smooth anyway.

Trackir software its 3 years OLD, 3!!!!!

I can gurantee you all I run all kind of tests:

- change 3 cpu and 3 motherboards : z170/z270/z370 + i5 6600k/i7 7700k / i7 8700 k

- changed 4 gpu : 1060/1070/1080/1080ti

- changed 3 hd : normal mechanicl hd 7200 rpm / samsung evo 850 / samsung nvme 970 evo

- changed : 3 psu from 3 different manfactures

- formatted windows 10 at least 50 times in 3 years : tried all the builds of 10 starting from 1511 up to now 1807

- reinstalled dcs at least 50 times as well

run all kind of bios settings, usb hubs, usb pci-e cards with their own chipset etc etc etc

-tweeaked windows so many times from telemetry options to cpu afifnity to deep windows and bios settings you dont even imagine.

Trackir , in any game is only smooth at perfect 60 fps, 120 fps or higher.

This is not a theory, this is a PURE FACT.

exactly,

trackir operates under OPENGL , the game and gync operates in direct3d and this is the reason of stuttering when using trackir withouth fps cap.

 

 

The only solution which I along with other have discovered long time ago ( there are several threads in this regard from at least 3 years) is to cap fps at 60 or 120.

 

 

The problem is quite simple :

- stable 60 fps trackir and the game is smooth

-stable 120 fps or higher trackir and game is smooth

 

 

ANY VALUE OF FPS BETWEEN 60 AND 120 , meaning fluctuation of fps rtange from 60 to 120 will generate JITTERY and stuttery movement when paning around.

 

 

If the game runs higher than 120 fps, ie. fluctuation from 120 up to inifnity , this will not generate jittery and will be smooth anyway.

 

 

Trackir software its 3 years OLD, 3!!!!!

 

 

 

I can gurantee you all I run all kind of tests:

 

 

- change 3 cpu and 3 motherboards : z170/z270/z370 + i5 6600k/i7 7700k / i7 8700 k

- changed 4 gpu : 1060/1070/1080/1080ti

 

- changed 3 hd : normal mechanicl hd 7200 rpm / samsung evo 850 / samsung nvme 970 evo

 

 

- changed : 3 psu from 3 different manfactures

 

 

- formatted windows 10 at least 50 times in 3 years : tried all the builds of 10 starting from 1511 up to now 1807

- reinstalled dcs at least 50 times as well

 

 

run all kind of bios settings, usb hubs, usb pci-e cards with their own chipset etc etc etc

-tweeaked windows so many times from telemetry options to cpu afifnity to deep windows and bios settings you dont even imagine.

 

 

 

 

Trackir , in any game is only smooth at perfect 60 fps, 120 fps or higher.

 

 

This is not a theory, this is a PURE FACT.

Very well said, no argument to this as it is indeed fact.

I have my FPS set at 60, however; what should I be doing with my GSYNC and screen refresh rate of 144Hz? Do those even matter as long as my FPS target is set to 60?