Choppy Trackir View with Vsync off on 60hz monitor

[Hiob]

I once also believed that trackir-stuttering was related to monitor (or gpu) refresh rate, but - for me at least - that was debunked several times. I don't have proof - other than anecdotal experience, but by now I'm strongly convinced, that it has more to do with resource management in windows, problems with the usb-interface and alike. 

I would advice to pass on any kind of v-sync IF possible (prefer g-sync or freesync if possible), disable the hotplug function, and check every windows (and mobo-) setting that affects the usb interface. If possible, don't use an usb-hub (though, atm I'm using an unpowered usb 2.0 hub without problems), try to change the usb interface (use another port on your mobo from 3.X to 2.0 or vice versa). Deactivate the power-savings options for usb in windows.... and so on

Note that most of those actions were aimed at a then occuring disconnect problem I had with trackIR, but along the ride I got rid of the microstuttering and it never returned since. Even! if I drop below 60 fps (which has trackIR related downsides OTHER than microstuttering).

[SharpeXB]

30 minutes ago, Boosterdog said:

Link?

https://www.naturalpoint.com/trackir/trackir5/

[Boosterdog]

Just now, SharpeXB said:

https://www.naturalpoint.com/trackir/trackir5/

Thats the main product page.  link to the referenced thread? 

[SharpeXB]

18 minutes ago, Boosterdog said:

Thats the main product page.  link to the referenced thread? 

The product page answers the Hz question easily enough. 

I don’t believe the TrackIR sample rate has anything to do with stuttering. It doesn’t need to be synced to the screen refresh rate anymore than your mouse does. But you could ask NaturalPoint. 
 

Regarding Vsync

https://forums.naturalpoint.com/viewtopic.php?f=60&t=13056&p=59296&hilit=Vsync#p59296

Edited November 29, 2021 by SharpeXB

[Boosterdog]

4 minutes ago, SharpeXB said:

The product page answers the question easily enough. 

I don’t believe the TrackIR sample rate has anything to do with stuttering. It doesn’t need to be synced to the screen refresh rate anymore than your mouse does. But you could ask NaturalPoint. 

You quoted a topic. The product page supports nothing. 

29 minutes ago, Hiob said:

I once also believed that trackir-stuttering was related to monitor (or gpu) refresh rate, but - for me at least - that was debunked several times. I don't have proof - other than anecdotal experience, but by now I'm strongly convinced, that it has more to do with resource management in windows, problems with the usb-interface and alike. 

I would advice to pass on any kind of v-sync IF possible (prefer g-sync or freesync if possible), disable the hotplug function, and check every windows (and mobo-) setting that affects the usb interface. If possible, don't use an usb-hub (though, atm I'm using an unpowered usb 2.0 hub without problems), try to change the usb interface (use another port on your mobo from 3.X to 2.0 or vice versa). Deactivate the power-savings options for usb in windows.... and so on

Note that most of those actions were aimed at a then occuring disconnect problem I had with trackIR, but along the ride I got rid of the microstuttering and it never returned since. Even! if I drop below 60 fps (which has trackIR related downsides OTHER than microstuttering).

My experence differs from yours (thats not to say I think you are wrong mearly that our experience is not the same). Disabling hotplugging, checking all usb related functions, usb hub, powered usb hub, USB 3 or 2 direct make no difference for me.  I think ive tried them all over the past 15 years, often more tha once!:-)

This is my understanding based on a non gsync monitor. My own expereince of trying out many things leads me back to it every time.

TIR polls at 120 fps and has an update rate of around 9ms. As such it points your camera to somewhere about 120 times a second. Since the camera is acting independently of the monitor or your setings in game, to be smooth that needs to coincide with a fully rendered frame at roughly the same location it is pointing the camera at during that refresh cycle. A lower "harmonic" of the 120 (60) is acceptable but anything much other is problematic.  If the frame cannot be rendered fully in harmony with the latest polled TIR position, it will result in a noticable stutter which may be more or less noticable dependent upon the subsequent ability of the gpu to render the following frames/s and, most importantly the range of head movement accross those frames itself (ie it will be fine if looking at a fixed point but not if looking left to right)

Since you cannot alter what the TIR camera is doing and that its doing it independently of anything, one needs to adjust the external perameters to best match it. In this case doing all you can to ensure an even frame rate that is matched to the polling rate or a harmonic thereof for the camera. At 60hz this is what I have personally found vsync does best provided I can maintain 60 fps minmum. Without vsync, and given my fps would vary in the main between 70-90 fps, the resultant images would not only be unevenly paced but would no longer match the positional updates from the TIR camera inducing the jitters. 

I would be delighted to be proven wrong and pursuaded otherwise as TIR is a bugbear of mine. 

[Hiob]

I was just reading through the naturalpoint forum and did some googling out of interest. No general solution, but an idea.

I've read something about usb polling rate of windows, but I can't figure out, what it is supposed to be. Some suggest 125 Hz (which would heavily conflict with the 120 Hz polling rate of trackIR), other 1000 Hz.

Does anyone know how to test this and more important how to change that? Note that Im not talking about the polling rate of any device, but the usb-port itself.

[Hiob]

2 minutes ago, Boosterdog said:

You quoted a topic. The product page supports nothing. 

My experence differs from yours (thats not to say I think you are wrong mearly that our experience is not the same). Disabling hotplugging, checking all usb related functions, usb hub, powered usb hub, USB 3 or 2 direct make no difference for me.  I think ive tried them all over the past 15 years, often more tha once!:-)

This is my understanding based on a non gsync monitor. My own expereince of trying out many things leads me back to it every time.

TIR polls at 120 fps and has an update rate of around 9ms. As such it points your camera to somewhere about 120 times a second. Since the camera is acting independently of the monitor or your setings in game, to be smooth that needs to coincide with a fully rendered frame at roughly the same location it is pointing the camera at during that refresh cycle. A lower "harmonic" of the 120 (60) is acceptable but anything much other is problematic.  If the frame cannot be rendered fully in harmony with the latest polled TIR position, it will result in a noticable stutter which may be more or less noticable dependent upon the subsequent ability of the gpu to render the following frames/s and, most importantly the range of head movement accross those frames itself (ie it will be fine if looking at a fixed point but not if looking left to right)

Since you cannot alter what the TIR camera is doing and that its doing it independently of anything, one needs to adjust the external perameters to best match it. In this case doing all you can to ensure an even frame rate that is matched to the polling rate or a harmonic thereof for the camera. At 60hz this is what I have personally found vsync does best provided I can maintain 60 fps minmum. Without vsync, and given my fps would vary in the main between 70-90 fps, the resultant images would not only be unevenly paced but would no longer match the positional updates from the TIR camera inducing the jitters. 

I would be delighted to be proven wrong and pursuaded otherwise as TIR is a bugbear of mine. 

I think, we both agree, that we don't have the definite solution or answer.

My reasoning against your suggestion is, that any device (input-) has a polling rate which will be, more often than not, out of sync with the refresh rate of the gpu. Also it should affect any game in the same way. Further - depending on the speed of your head movement (aka the amount of pixels the trackIR camera counts which every update), you will be put of sync with everything. (Sorry if I can't explain my thought here in a more understandable way). In short, if desync would create stuttering "per se" it would be a much more prominent problem overall. Last point (very technical) - electronic clock generators have a very wide spread deviation. Not two of the same kind give the exact same speed. That is why in very precise time depending applications, the devices are calibrated and fit with a correction value. So, being "in sync" is a very very theoretical thing.

[Boosterdog]

7 minutes ago, Hiob said:

I was just reading through the naturalpoint forum and did some googling out of interest. No general solution, but an idea.

I've read something about usb polling rate of windows, but I can't figure out, what it is supposed to be. Some suggest 125 Hz (which would heavily conflict with the 120 Hz polling rate of trackIR), other 1000 Hz.

Does anyone know how to test this and more important how to change that? Note that Im not talking about the polling rate of any device, but the usb-port itself.

All i ever found was the average run of the mill mouse USB  polling rate is 125hz. Seems that can be set higher on gaming mice which would indicate that windows isnt too fussy until 1000hz then it goes batshit crazy. All topics point to shooters and reducing lag. 

I took this as being a way to get the view to a point in a twitch game without lag. Refresh rate still matters but in competative shooters most will forsake everything for fps and reduced lag. 

[Lange_666]

My experience:

If FPS isn't exactly on the TrackIR (and with Track IR i mean the Natural Point stuff, not some 3th party DIY stuff that's also sometimes called TrackIR) pull rate of 120Hz or a division of it (60Hz) i "always" experience stutter on fast moving terrain.
Over the years i used my TrackIR on various GPU/monitor combinations and the result is "always" the same, stutter when FPS and TrackIR pull rate don't match. 
The closer the FPS comes to the TrackIR pull rate the less noticable it is, the further it drifts away, the bigger the stutter gets.
It's OK and almost not noticable with terrain movement looking straight ahais but when looking sideways it's always there.
G-Sync is not a solution to this type of stutter.
For the moment i'm running a 144Hz monitor but i set it to 120Hz and then use VSync + FPS cap at 60 (i can't maintain 120 FPS with my setup) to make TrackIR real smooth. It's the "only" solution i found so far that works all the time to prevent TrackIR stuttering.

[Boosterdog]

22 minutes ago, Hiob said:

I think, we both agree, that we don't have the definite solution or answer.

My reasoning against your suggestion is, that any device (input-) has a polling rate which will be, more often than not, out of sync with the refresh rate of the gpu. Also it should affect any game in the same way. Further - depending on the speed of your head movement (aka the amount of pixels the trackIR camera counts which every update), you will be put of sync with everything. (Sorry if I can't explain my thought here in a more understandable way). In short, if desync would create stuttering "per se" it would be a much more prominent problem overall. Last point (very technical) - electronic clock generators have a very wide spread deviation. Not two of the same kind give the exact same speed. That is why in very precise time depending applications, the devices are calibrated and fit with a correction value. So, being "in sync" is a very very theoretical thing.

Got most of that all though you went a bit "Brian Cox" in places for my tiny brain 

Id say back that the TIR captures a point in time every 8.3ms on a 3D grid which it passes to the game. I dont think TIR suffers from having to think about rendering scenes. It just crunches the positional data and makes it someone elses problem.   Getting the image in game to roughly match that works well enough for our brains if the image can be redered in time or in a time that is a specific  harmonizastion of the time ( half rate. This is good enough for our brains to be fooled at the rate of display (60 or 120 hz). It would be harder to fool at lower rates.  Sync is the wrong term I guess. Small ballpark is perhaps a better one.  As long as its there or threabouts its sufficient to allow the motion to be perceived as being smooth. 

Every game has different demands but by enabling a frametime limiter such as vsync you do actually affect all games equally dont you? 

Anyhow - when do we get onto religion, football and politics??

 

Edited November 29, 2021 by Boosterdog

[SharpeXB]

43 minutes ago, Boosterdog said:

The product page supports nothing.

The product page says the device works at 120hz, not 60-120hz

[Hiob]

 

13 minutes ago, Boosterdog said:

anyhow - when do we get onto religion, football and politics??

As long, as we stay civilized and humble - there is no harm in discussing this matter, is it?

And I'm honestly very interested in this issue. I mean part of the problem is most likely, that there are different perception of stuttering. E.g. I can't confirm the experiences of Lange_666 with my own, but that doesn't mean, that I don't believe him. I'm only 100% sure that the stuttering I experienced in the beginning (a very prominent and very regular jittering of the picture when doing long continuous head movements), is gone for good. And I'm playing well  below 100 FPS most of the time.

Edited November 29, 2021 by Hiob

[SharpeXB]

47 minutes ago, Boosterdog said:

I would be delighted to be proven wrong and pursuaded otherwise as TIR is a bugbear of mine.

It’s possible that you’ve got this all wrong about how the device works. It’s no different than pointing with a mouse to free-look around the game. It doesn’t need to sync to your GPU or anything. Stutters happen to be more noticeable when looking around with it but the source of that isn’t the TrackIR

I would suggest asking this to NaturalPoint before confusing anyone here further with this. 

[Hiob]

4 minutes ago, SharpeXB said:

It’s possible that you’ve got this all wrong about how the device works. It’s no different than pointing with a mouse to free-look around the game. It doesn’t need to sync to your GPU or anything. Stutters happen to be more noticeable when looking around with it but the source of that isn’t the TrackIR

I would suggest asking this to NaturalPoint before confusing anyone here further with this. 

I'm with you - mostly.

But the difference in polling rates between a mouse and trackIR is a magnitude. (Given the USB port isn't the bottleneck) Just saying. (which is - I'm absolutely aware, countering my own arguments...)

Edited November 29, 2021 by Hiob

[SharpeXB]

22 minutes ago, Boosterdog said:

Id say back that the TIR captures a point in time every 8.3ms on a 3D grid which it passes to the game. I dont think TIR suffers from having to think about rendering scenes. It just crunches the positional data and makes it someone elses problem.

You’re overthinking this. TrackIR doesn’t render frames. It’s basically a mouse pointer strapped to your head. Although using 6DOF makes any stuttering more perceivable, the device itself is not causing that. 

6 minutes ago, Hiob said:

But the difference in polling rates between a mouse and trackIR is a magnitude.

TrackIR polling at 120hz or your mouse polling at 125-1000hz makes no difference to your GPU or monitor. They don’t have to be synced with it. 

[Hiob]

11 minutes ago, SharpeXB said:

It’s possible that you’ve got this all wrong about how the device works. It’s no different than pointing with a mouse to free-look around the game.

You're absolutely right in saying that trackIR - in principle - is nothing else than a mouse (with a 3rd dimension). However, to make it comparable, you would need to reduce the mouse refresh rate 120 Hz. Would be interesting if looking around with the mouse then would give a similar result.

[Boosterdog]

14 minutes ago, SharpeXB said:

You’re overthinking this. TrackIR doesn’t render frames. It’s basically a mouse pointer strapped to your head. Although using 6DOF makes any stuttering more perceivable, the device itself is not causing that. 

Jesus man - I didnt say it rendered frames - In fact I said the opposite. Read the words. 

And at NO POINT have i said the device itself is causing the stuttering. Ive said, several times now, that its the gpus inability to render a frame prescribed by the TIR within a refesh cycle that causes the issue.  Ive tried to explain everything as well as I can. Certainly Ive done more than point to a product page which says absolutley nothing about the issue.  

23 minutes ago, SharpeXB said:

I would suggest asking this to NaturalPoint before confusing anyone here further with this. 

Seems your the only one confused by this. 

[Bossco82]

Running at 60hz is not a requirement for Track IR to function. IE, you move your head and the first person camera view in the moves around. In that sense whoever says that is correct.

However, running at an fps IE 60hz which is inline with the polling rate of the Track IR at 120hz, it can be divided. This helps to reduce the effect of stutter seen when panning the head movement and looking around.  So V-sync is it is not a requirement to "run" Track IR with a game. It simply increases the chances of seeing a lot less stutter or tearing.

I have a freesync monitor so I can lock my fps between 55-75hz and DCS will look smooth as long as I dont use Track IR. As soon as a I start looking around with Track IR at an fps other than 60hz there is a visible tear line and micro stutters. Running your game in sync with the Track IR's polling rate is simply good practise to ensure "butter" smooth performance when playing something like DCS on a flat panel with Track IR. Its not that V-sync is essential, V-sync is the setting that keeps everything at 60fps/hz if you have a 60hz monitor which is then inline with the Track IR's polling rate.

Something else to take note of. Some users dont notice this and others do notice this. I use a 32" monitor and I do notice it, its obvious. If I used a smaller monitor say 24", and I have checked this. It is not nearly as noticeable I have to look for it but its still there.

Its only my opinion but whatever you have if your running a Track IR with any monitor maintaining a constant 60fps is a sure way of getting the smoothest experience.

[SharpeXB]

32 minutes ago, Boosterdog said:

its the gpus inability to render a frame prescribed by the TIR within a refesh cycle that causes the issue

But that has nothing to do with TIR, the same stutter could happen when using a mouse. I think I see your point if what you mean is that TrackIR would be like setting your mouse down to 125hz. Could that be perceived as stutter? Maybe. I get down to <60 FPS in DCS frequently and don’t perceive it as stutter with Fast Vsync enabled. 

44 minutes ago, Hiob said:

you would need to reduce the mouse refresh rate 120 Hz. Would be interesting if looking around with the mouse then would give a similar result

Maybe

Edited November 29, 2021 by SharpeXB

[Hiob]

See, I completely agree with you, that the stuttering is not due to the polling rate of trackir - because it doesn't fit my own experience/observation.

However, I have no hard evidence nor sufficiant explanation for it and IF stuttering is a result of a asynchronism or disharmony, it would very much make a huge difference if the polling rate is 120 or 1000Hz.

And at least for now, you only have given a strong opinion but no explanation or evidence for it. Which I would be very interested in...

Edited November 29, 2021 by Hiob

[SharpeXB]

2 minutes ago, Hiob said:

However, I have no hard evidence nor sufficiant explanation for it and IF stuttering is a result of a asynchronism or disharmony, it would very much make a huge difference if the polling rate is 120 or 1000Hz.

If mice could cause stutter by being out of sync with a monitor then for certain Nvidia would be selling G-sync mice! $$$

[Hiob]

4 minutes ago, SharpeXB said:

If mice could cause stutter by being out of sync with a monitor then for certain Nvidia would be selling G-sync mice! $$$

Well, I happen to sit in front of my office pc. (60Hz widescreen, logitech mx wireless mouse)

My mouse cursor is stuttering like crap when I drag it from left to right and back.... just saying.

[SharpeXB]

1 minute ago, Hiob said:

Well, I happen to sit in front of my office pc. (60Hz widescreen, logitech mx wireless mouse)

My mouse cursor is stuttering like crap when I drag it from left to right and back.... just saying.

If it’s a wireless mouse, make sure the USB receiver is very close to it. I’ve found that putting this on the back of the PC under a table is too far away. The gaming move come with extension cables. 

[Hiob]

1 minute ago, SharpeXB said:

If it’s a wireless mouse, make sure the USB receiver is very close to it. I’ve found that putting this on the back of the PC under a table is too far away. The gaming move come with extension cables. 

Plugged in a cabled mx cherry - no difference whatsoever.

Just give in and admit that you don't have a clue, like all of us....

[SharpeXB]

17 minutes ago, Hiob said:

Plugged in a cabled mx cherry - no difference whatsoever.

Just give in and admit that you don't have a clue, like all of us....

When I have that problem with a mouse I just buy a new one. I’ve had some odd stutter problems from them too.