TrackIR5 Stutter at framerates other than 60 or 120 fps in DCS 2.8

[5ephir0th]

9 hours ago, hughlb said:

I have responded to their post, but it hasn't cleared moderation yet. Essentially though, minimizing didn't make any difference (or any of the other suggestions). The only strange thing was the 30 FPS of the TrackIR software, but it may be a red herring.

You are using an nVidia GPU, check the nVidia Control Panel, theres an option to limit the background apps FPS and its default to 30 fps, disable or increase the limit

[hughlb]

5 hours ago, 5ephir0th said:

You are using an nVidia GPU, check the nVidia Control Panel, theres an option to limit the background apps FPS and its default to 30 fps, disable or increase the limit

I thought you might be onto something there 5ephir0th. But no change. I tried both the global and program setting for Background Application Max Frame Rate, with combinations of 60 and 100 FPS, but no effect on stuttering and the TrackIR software itself still indicates 30 FPS.

[5ephir0th]

7 hours ago, hughlb said:

I thought you might be onto something there 5ephir0th. But no change. I tried both the global and program setting for Background Application Max Frame Rate, with combinations of 60 and 100 FPS, but no effect on stuttering and the TrackIR software itself still indicates 30 FPS.

Then, probably, it’s just that the refresh rate on the TrackIR software it’s just 30hz, it doesn’t have to affect to how it works, the TIR camera refresh is 120hz, at least on 5 version

The preferred target fps for not getting stutter is 30/40/60/80/120 fps

Edited August 28, 2023 by 5ephir0th

[hughlb]

FYI - Whilst I don't have a solution, after a lengthy troubleshooting period with the team at NaturalPoint, they have acknowledged and expanded on how the interaction between the game engine and polling rate of TrackIR can be a cause for this problem. But it also helps explain why some titles (MSFS, IL2 GB) don't present the same problem, under the same conditions. I feel there is hope that ED can solve or minimize the effect. As I have pointed out, prior to 2.8/2.9, for a long period with 2.7 and prior, this wasn't an issue for DCS, it appeared to be resolved. So it is possible to get TrackIR communicating smoothly with DCS, at variable refresh rates outside of 60 and 120 Hz.

I have attached the post provided by Jillian from NaturalPoint, with good information from a NaturalPoint software engineer. I have also pasted my response below, and you can also read the forum thread here - https://forums.naturalpoint.com/viewtopic.php?t=25141&start=10

"Hi Jillian, thanks again for your response and investigation.

Quote

It is really up to the developers and game engines to make sure that they retrieve the data at the right time and smooth out any stuttering that is a result of not being in sync.

This is really critical information, and something I feel developers, such as Eagle Dynamics, should be better alerted to by NaturalPoint. FYI, I have a lengthy troubleshooting thread on both the Eagle Dynamics forum, and through an open support ticket. Almost everything has been suggested with the exception of recognition that this may be an in-engine issue in how the tracking data is retrieved and, consequently, smoothed.

Quote

at some point windows full screen fixed this and made sure that the game was in sync and was receiving the right messages from TrackIR at the right times, and perhaps the game devs even programmed the game with the assumption this would not change.

Again, really critical information for developers. They may be implementing TrackIR support without fully understanding how to manage its data, and circumvent this stuttering issue.

Quote

any Microsoft products running in the background like Teams. He also mentioned that if your hard drive is near capacity, this could also contribute to the stuttering. He asked that I share this Reddit post with you to see if it helps:
https://www.reddit.com/r/hoggit/comment ... nd_system/

Again, no change after looking into these suggestions. I think we can, at this point, assume the issue is a TrackIR polling rate, frame rate interaction, and how that is handled by the game engine. The only "fix" to this issue we have come across, is switching game engines. MSFS can have a variable framerate, as can IL2 Great Battles, without causing this stuttering problem, or at least minimizes it significantly. However, DCS is severely hampered by stuttering at framerates outside of a fixed 120 or 60 fps. The fact also, that for several years, and prior to their current 2.8/2.9 versions, they had resolved this issue (perhaps inadvertently), again support that this isn't a user issue with hardware or software conflicts and settings.

I will update my ticket with Eagle Dynamics and make these suggestions, but it would be useful if NaturalPoint could reach out and alert ED to the fact this issue can be both triggered, but also resolved by the developer. They might then look at how they can smooth TrackIR input data, perhaps even review previous versions where the issue wasn't apparent."

Edited October 27, 2023 by hughlb

[hughlb]

We have established that 'mouse look' or panning the camera using the mouse is smooth, whilst looking around with TrackIR is juddering. However, when I change my mouse poling rate from 500 to 125 Hz, I notice the mouse panning also starts to begins to stutter, not as badly but noticeably less smooth than 500 Hz or 1000 Hz. I haven't noticed this in other games.

[SharpeXB]

As much as I’ve tried I haven’t been able to duplicate or notice this behavior. Do you find this to be very pronounced and jarring or something you really need to look for? I actually notice the difference between 60Hz and 100+Hz more so than anything to do with the view panning. 

[hughlb]

On 11/2/2023 at 12:56 AM, SharpeXB said:

As much as I’ve tried I haven’t been able to duplicate or notice this behavior. Do you find this to be very pronounced and jarring or something you really need to look for? I actually notice the difference between 60Hz and 100+Hz more so than anything to do with the view panning. 

It is really pronounced. The easiest way to describe it is this - it looks like the frame rate is 20-30 FPS. For example, if my frame rate and monitor is set at 100 FPS and 100 Hz, respectively, but I look around with TrackIR, then the frame rate resembles 20-30 FPS. Keep in mind, the frame rate is not 20-30 FPS, it is 100 FPS. If I hold my head still, and instead use the mouse to pan the camera around, it looks like 100 FPS.

My original video demonstrates that every five frames or thereabouts, we essentially get an anomalous position update for TrackIR. In that there is a 'jump' or 'spike' in in-game head position. It is subtle. There are more or less 25 of these erroneous frames every second, if the frame rate is 100 FPS. This presents as a rhythmic judder.

I think people often misdiagnose the issue because they look for a stutter. The terminology isn't great at defining the appearance of the problem. A stutter can be rythmic, but the intervals can be seconds or more, or a stutter might not be rythmic at all, and instead be a hitch or pause, or sequence of pauses. This is a predictable, repeatable problem that occurs all the time, 25 times a second, when using TrackIR in DCS.

For anyone following - below is a collection of different forum threads, over several years, detailing this specific problem, and how locking frame rate to 60 or 120 FPS resolves the issue. The question however, is why did TrackIR in DCS once work perfectly (2.7) at variable high frame rates, and why don't I currently have this issue in MSFS or IL2? It is possible to resolve this problem, and it appears to be game engine related, and how TrackIR input data is resolved by the engine. Also, notice how many threads relate to DCS. I rarely see this mentioned in MSFS, for example. 

2016 - https://forums.naturalpoint.com/viewtopic.php?t=13991
2018 - https://forum.dcs.world/topic/186271-track-ir-stutter-fix-petition-at-natural-points-forum/
2018 - https://forums.blurbusters.com/viewtopic.php?t=4242
2018 - https://forum.dcs.world/topic/172576-25-trackir-stuttering/
2018 - https://forum.il2sturmovik.com/topic/40171-stuttering-while-turning-my-head-in-the-cockpit-solved-to-95/
2021 - https://forum.dcs.world/topic/293359-does-ir-tracking-have-issues-stuttering-above-60fps/
2021 - https://forum.dcs.world/topic/288806-stutters-flickery-motion-with-new-144hz-freesync-monitor-and-track-ir-running/
2021 - https://www.reddit.com/r/hoggit/comments/smlqdj/alot_of_people_say_they_have_to_disable_gsync/
2022 - https://forum.dcs.world/topic/293359-does-ir-tracking-have-issues-stuttering-above-60fps/
2022 - https://www.avsim.com/forums/topic/619132-track-ir-stutters-with-msfs/
 

Edited November 6, 2023 by hughlb

[SharpeXB]

44 minutes ago, hughlb said:

is really pronounced. The easiest way to describe it is this - it looks like the frame rate is 20-30 FPS.

I definitely don’t see anything like that in DCS nor do other games seem smoother. 

[SkateZilla]

this seems more of a System Software/Hardware issue, with USB Polling and latency causing the FPS Stutters you're seeing..

60 FPS is 1:2 pulldown from the TrackIR's sampling rate, w/ Native being 120Hz, which is why those FPS look the best,

If TrackIR is Polling at 120Hz and your FPS is some random low number like 37, then there's going to be latency introduced due to the pulldown and sync, 
the same happens in VR, TrackIR has yet to update their software to be asynchronous to the GPU Rendering rate. NaturalPoint (is it OptiTrack?) has been pretty much sitting comfortable with trackIR4 and 5 and hasnt done anything to update them to work with today's GPUs.

Edited November 6, 2023 by SkateZilla

[Lange_666]

2 hours ago, hughlb said:

The question however, is why did TrackIR in DCS once work perfectly (2.7) at variable high frame rates, and why don't I currently have this issue in MSFS or IL2?

Because it never did, not in 2.7 nor in any other version. Same goes for MSFS and IL-2, they suffer from the same problem.
The issue has been around since TrackIR exist, the only solution is to sync your frames with the TrackIR pull rate (or a division of it).

[Hiob]

1 hour ago, Lange_666 said:

Because it never did, not in 2.7 nor in any other version. Same goes for MSFS and IL-2, they suffer from the same problem.
The issue has been around since TrackIR exist, the only solution is to sync your frames with the TrackIR pull rate (or a division of it).

Sorry, that is just not true. I had this problem once but no more for a very long time now. My FPS (using G-sync and fast V-sync) are all over the place and there is no problem whatsoever with TrackIR. 

Nobody denies that the problem exist for some people, and maybe the synced fps at 60 or 120 does alleviate the symptoms. But if it was the root cause, everybody would have that problem and that is simply not the case.

[Lange_666]

Yeah but then you are using "new" technologies like G-Sync "and" fast V-Sync to "solve the problem which initiates in TrackIR in the first place.
Those who don't have a G-Sync are still left with the problem so it still exist as to today.

And it IS the root cause and i'm 200% sure if i come and watch your setup in a non G-Sync environment i still would see the problem.

BTW: i still prefer a fixed FPS V-Synced FPS over any fluctuating FPS anytime. It's just A LOT SMOOTHER then fluctuating FPS all over the place.

Edited November 6, 2023 by Lange_666

[SharpeXB]

I always found Nvidia Fast Vsync to work very well in DCS before I had G-Sync. 

[Hiob]

30 minutes ago, SharpeXB said:

I always found Nvidia Fast Vsync to work very well in DCS before I had G-Sync. 

Yeah, sometimes I use it, sometimes I don’t. I‘m not really sure how much it helps with G-sync activated. My TV is only 120 Hz G-sync compatible (so no active G-sync chipset). Sometimes I think I get a tiny bit of screen tearing and I activate it and then I forget about it and after the next driver update I‘m on default application controlled. Which is set to off in DCS.
 

So no, I‘m not sure how much V-sync is helping here. G-Sync is a very helpful feature though. Wether it is responsible for alleviating the infamous TrackIR-stuttering, I don’t know. But that is easily tested.

[SharpeXB]

49 minutes ago, Hiob said:

Yeah, sometimes I use it, sometimes I don’t. I‘m not really sure how much it helps with G-sync activated. My TV is only 120 Hz G-sync compatible (so no active G-sync chipset). Sometimes I think I get a tiny bit of screen tearing and I activate it and then I forget about it and after the next driver update I‘m on default application controlled. Which is set to off in DCS.
 

So no, I‘m not sure how much V-sync is helping here. G-Sync is a very helpful feature though. Wether it is responsible for alleviating the infamous TrackIR-stuttering, I don’t know. But that is easily tested.

I don’t know if Fast Vsync would do anything with G-Sync. I understand G-Sync is supposed to function with Vsync set to On. Vsync will act as a frame rate limiter so you shouldn’t get any tearing. 

[hughlb]

11 hours ago, Lange_666 said:

Because it never did, not in 2.7 nor in any other version. Same goes for MSFS and IL-2, they suffer from the same problem.
The issue has been around since TrackIR exist, the only solution is to sync your frames with the TrackIR pull rate (or a division of it).

You are not correct. It did work, in fact it worked well prior to 2.7 as well, for several years. I first reported on this issue in DCS back in 2016. It was a problem then, and then at some point ceased being one. I can only speak from my own experience, it may have remained an issue for others.

As for MSFS and IL-2. I can jump into either of those right now, and it is a night and day difference. The issue simply doesn't exist. With MSFS, I need to enable both vsync and gsync, and with both I cap my framerate at 100 FPS. But I don't need 60 or 120 FPS to get it smooth, it is smooth at variable framerates.

This nullifies the idea it is a poling problem, or hardware issue, per se. If it was, shouldn't I have this problem in all applications that use TrackIR? I don't, therefore a more probable cause is how the game engine interprets TrackIR data.

To quote NaturalPoint themselves from the other week "TrackIR does attempt some smoothing, but again, only reports that same data 120 times a second. If the game is out of sync, you will see periods of time where the game and TrackIR line up, and periods of time where it does not. This results in multiple frames in a row receiving the same data, and some frames even being lost. The jumping/acceleration the user observes is likely the game engine not polling at the right rate with his configuration. This is especially a problem if the game is depending on the framerate to poll TrackIR information. It should be its own process independent of framerate. I also see games often do their own smoothing for other input sources so that this is less noticeable if anything goes out of sync."

https://forums.naturalpoint.com/viewtopic.php?t=25141&sid=937eb55a2ec8cea58a870dd11272d602&start=10

Not to repeat myself from my original post - but I have tested this across different systems, seven years apart in terms of hardware, different OS, and even a different TrackIR units. The issue is identical on both systems. I have even moved house several times. Everything susggests this is game engine related, and perhaps more broadly, including interaction with GPU drivers, and how they work with TrackIR position data.

It is a solvable problem, at least in as much as it can be minimized or controlled to the point of it not being noticeable. And I think, based on the feedback from NatrualPoint, that this needs to be mitigated by the game developer. TrackIR spits out its data, the game engine uses that to address camera position.

Edited November 7, 2023 by hughlb

[Dr Zaius]

This may be opening another can of worms but I find running DCS in 'true' full screen mode (check various posts on this subjext as just ticking the full screen box in the game options is more like boarderless windowed mode) smoothes out TrackIR massively, granted I've a system that can run at 120Mhz but the difference is still really noticiable, worth giving it try perhaps? 

[hughlb]

On 11/15/2023 at 9:23 PM, Dr Zaius said:

This may be opening another can of worms but I find running DCS in 'true' full screen mode (check various posts on this subjext as just ticking the full screen box in the game options is more like boarderless windowed mode) smoothes out TrackIR massively, granted I've a system that can run at 120Mhz but the difference is still really noticiable, worth giving it try perhaps? 

We tried this method, and in fact tried using a technique to activate exclusive fullscreen in the single threaded open beta. There was a time before 2.8 where alt+entering was possible, and improved the smoothness of the image. Sadly, this didn't affect the TrackIR issue.