Headtracking software recommendations
-
EDtracker hm <6DOF is a big compromise. Apart from trackIR I would recommend either DeclanClip or DIY: cheap, effective and nowadays very snappy and smooth.
For DIY it seems declanclip offers the modded cameras but the PS3-eye mod is pretty easy to do if you have an old floppy disk lying around; plenty of guides on yt.
The clip you can also make yourself in give or take an hour. For the LED-mount, follow the serial approach listed here:
http://forum.free-track.net/index.php?showtopic=1856
The leds and resistors are quite specific but are easily ordered online; for the base I used a simple hook available in just about any hardware-store:
:rofl:
Cherry on the cake but not needed necessarily: transparent fimo-clay balls over the leds to offer a bigger target for the tracker-software.
Software-wise, go for opentrack. A trackIR unit is easy and good but if you want cheap, the whole setup above won’t cost you more than 20 bucks.
-
This smoothness thing you say is interesting. Is it lagging? What happens if you turn off all filtering in opentrack?
sh
I would say my Edtracker and Opentrack feels over-responsive. Don’t know how to articulate it. They are digital, I am analog. IRL, if I were to swing my helmet around to see my wingie, I would be quickly swiveling around but then smoothly slowing down as that jet came into peripheral vision, then smoothly stopping, finally focusing on that beautiful jet off my wing. Now though, I swing my head around, bam I am there, screeching to a halt on that aircraft off my wing, then focusing in on a beautiful jet. I realize, I couldn’t see the screen if I turned my head 90 degrees plus, I am just trying to communicate a subtle difference.
When I adjust for smoothness in the accela filter then I get an annoying delay at the start of the movement. (Could be I am annoyingly hypersensitive ). The Opentracker developer is working on this though and I check back for a stable release often.
-
Is anyone using freetrack?
I’m starting to make one, 3 Leds, see if I can get it mounted on my headset somehow.From the videos, the software seems pretty reliable (although I believe the webcam makes the difference). Anyway, if someone is using, is it accurate enough?
-
I found the facetracknoir setup quite easy, I also ordered a used & modified ps eye cam from Delan when I purchased the clip.
I find scrolling / head movement very smooth with no jitter whatsoever.
Hi Uwe,
I just downloaded facetracknoIR, I use glasses and the lamp is above and a bit behind me…
Let’s see how it works. -
I would say my Edtracker and Opentrack feels over-responsive. Don’t know how to articulate it. They are digital, I am analog. IRL, if I were to swing my helmet around to see my wingie, I would be quickly swiveling around but then smoothly slowing down as that jet came into peripheral vision, then smoothly stopping, finally focusing on that beautiful jet off my wing.
When I adjust for smoothness in the accela filter then I get an annoying delay at the start of the movement. (Could be I am annoyingly hypersensitive ). The Opentracker developer is working on this though and I check back for a stable release often.
You’re adjusting the wrong thing. leave smoothing on zero. Set sensitivity to 0.45. Set deadzone to 0.06 and leave it there. Adjust sensitivity for the right amount of jumpiness vs immediate response.
The sensitivity slider should be renamed to smoothing and smoothing itself renamed with a warning that it should stay zero for 95 percent of users.
Maybe smoothing should be removed entirely and added by a plugin.
-
TrackIR Track Clip (non-pro).
That was the one of the reasons why I went to TrackIR.It does occlude on extreme positions. Also changing seat position mildly makes it start occluding. Two companies are now dealing with a mini clip of my design. It’s impossible to occlude without getting beheaded, with a decent margin of error.
-
Is anyone using freetrack?
I’m starting to make one, 3 Leds, see if I can get it mounted on my headset somehow.From the videos, the software seems pretty reliable (although I believe the webcam makes the difference). Anyway, if someone is using, is it accurate enough?
See my post above, take a look at opentrack. It uses parts of the freetrack software and I can highly recommend it. The recording below uses a typical opentrack setup; quite smooth and accurate (I don’t use any sort of deadzone or something)
I guess this would probably best compare to a TIR 4 as my cam (ps3eye) will do 120fps in 320x240 but only 60fps in 640x480.
-
TrackIR Track Clip (non-pro).
That was the one of the reasons why I went to TrackIR.Are you saying the non-pro is better then the pro?
-
@TwanV:
EDtracker hm <6DOF is a big compromise. Apart from trackIR I would recommend either DeclanClip or DIY: cheap, effective and nowadays very snappy and smooth.
For DIY it seems declanclip offers the modded cameras but the PS3-eye mod is pretty easy to do if you have an old floppy disk lying around; plenty of guides on yt.
It’s repeating the bad design. The bottom “leg” is way too much forward. You can’t look up too much.
There’s a company about to sell an improved clip design internationally. Full disclosure: they funded PT support and provided hardware.
I would say my Edtracker and Opentrack feels over-responsive. Don’t know how to articulate it. They are digital, I am analog. IRL, if I were to swing my helmet around to see my wingie, I would be quickly swiveling around but then smoothly slowing down as that jet came into peripheral vision, then smoothly stopping, finally focusing on that beautiful jet off my wing. Now though, I swing my head around, bam I am there, screeching to a halt on that aircraft off my wing, then focusing in on a beautiful jet. I realize, I couldn’t see the screen if I turned my head 90 degrees plus, I am just trying to communicate a subtle difference.
When I adjust for smoothness in the accela filter then I get an annoying delay at the start of the movement. (Could be I am annoyingly hypersensitive ). The Opentracker developer is working on this though and I check back for a stable release often.
Immediate response is doable with proper rotation sensitivity and a small enough deadzone (don’t exceed 0.06). I go about 0.35 to .45 rotation sensitivity with deadzone to 0.06. If you get a response to steady movement in the form “jump -> stop -> jump -> stop -> done” I’d like to hear about it.
Also see the test build changelog <https://github.com/opentrack/opentrack/issues/688> before the 2.4.0 release. I have an EXPERIMENTAL change for the PT tracker that needs more exposure and feedback. See <http://opentrack.misaki.pl>. It points to dropbox but the domain is under personal ownership if this needs to change. I’ll be releasing test builds with the experimental change(s) shortly.
I should create a wiki page for opentrack PT settings.
-
Hi sthalik, it’s sanpats in the GitHub. I am very interest in this mini clip of your design. When will it be available?
-
It’s to be shipped by TrackHat in the EU and MoveHead in Russia.
Despite normal open spirit, let the companies start shipping before I provide any details on differences.
-
It’s to be shipped by TrackHat in the EU and MoveHead in Russia.
Despite normal open spirit, let the companies start shipping before I provide any details on differences.
Trackhat already ship it. Its called Trackhat Clip Plus Mini.
-
sthalik, February 8th, 2018, Post #18.
You’re adjusting the wrong thing. leave smoothing on zero. Set sensitivity to 0.45. Set deadzone to 0.06 and leave it there. Adjust sensitivity for the right amount of jumpiness vs immediate response.
The sensitivity slider should be renamed to smoothing and smoothing itself renamed with a warning that it should stay zero for 95 percent of users.
Maybe smoothing should be removed entirely and added by a plugin.sthalik, February 15th, 2018, Post #22
Immediate response is doable with proper rotation sensitivity and a small enough deadzone (don’t exceed 0.06). I go about 0.35 to .45 rotation sensitivity with deadzone to 0.06. If you get a response to steady movement in the form “jump -> stop -> jump -> stop -> done” I’d like to hear about it.Thanks for the tips. I did have smoothing in the Accela filter on zero. But I am trying the other suggestions as a baseline for more experimentation this weekend.
I don’t think its the software but that I am turning my head 45 degrees (so that I can still see the screen) and my view on screen is 160 degrees back (so I can see over my shoulder in the sim). It is a short head movement of my head versus a long movement of the view in the sim.
The other issue might be the athleticism of my neck muscles Maybe I need to go to the gym more!
-
Thanks for the tips. I did have smoothing in the Accela filter on zero. But I am trying the other suggestions as a baseline for more experimentation this weekend.
I don’t think its the software but that I am turning my head 45 degrees (so that I can still see the screen) and my view on screen is 160 degrees back (so I can see over my shoulder in the sim). It is a short head movement of my head versus a long movement of the view in the sim.
The other issue might be the athleticism of my neck muscles Maybe I need to go to the gym more!
I think you need to play with curves.
-
Thanks for the tips. I did have smoothing in the Accela filter on zero. But I am trying the other suggestions as a baseline for more experimentation this weekend.
I don’t think its the software but that I am turning my head 45 degrees (so that I can still see the screen) and my view on screen is 160 degrees back (so I can see over my shoulder in the sim). It is a short head movement of my head versus a long movement of the view in the sim.
The other issue might be the athleticism of my neck muscles Maybe I need to go to the gym more!
Hey,
Your muscles are fine. It only a matter of configuring the software and getting used to it a bit.
Regarding mappings:
First off, use the mapping screen. Left mouse button adds/drags, right removes. You have exact coordinate tooltips. The “max input” selector allows you to make more precise adjustments.
I heavily don’t recommend using “deadzones” in the mapping window. If you try to solve one problem with “deadzones” then you have two problems.
The max rotation angle you want for yaw/pitch depends on how far you sit from the screen. So it indirectly depends on screen size.
With 40 deg. I’d only be able to see with one eye, the other occluded by the root of my nose! And even then, very uncomfortably due to eyes turned sideways up to 11. I just tested and at 80 cm it’s comfortable up to 23 cm. YMMV.
Regarding filtering:
In just-released opentrack 2.3.10, the “smoothing” and “nonlinearity” sliders got euthanized because they frequently got misinterpreted. They were also pretty useless.
Just to be very clear, the old “smoothing” slider never helped with noise and wasn’t even meant to.
Hi sthalik, it’s sanpats in the GitHub. I am very interest in this mini clip of your design. When will it be available?
by = 75, bz = 35, ty = 40, tz = 18;
This is the design we arrived at with Alex Orohovatski from MoveHead. He made an adjustable engineering clip for testing various configurations. You can notice that this a common clip with “Z” values cut down in half.
-
I’m gonna do the FIMO trick so to go from IR reflectors to IR LEDS.
I use simple IR Leds from old TV remote controllers and no resistor. First try was not good…
I’m bored to get the known leds with the flat head… So I’ll experiment on this.So I got the 2€ FIMO thing.
if you want some info look here:
http://forum.free-track.net/index.php?showtopic=2378Has anyone here done it already?
Any impressions?