Question iRacing Oculus Rift Tracking Motion Settings? (virtual cockpit shifts around)

I have been increasing my motion settings and I notice this annoying (and motion sickening) shift of the virtual cockpit. It is most pronounced during sharp movements like traction loss snap-back as the car regains grip. I attribute the cockpit shift to my head moving left & right quickly within the rig.

As I get more aggressive with the setting's and my driving, my virtual helmet will shift from the outside the driver window to the passenger seat and then return.

I have turned off the head wobble feature in app.ini.

The rift sensor is mounted to the rig just in front of me behind the steering wheel.

The head tracking seems accurate, but a little too good...any suggestions?

Does the Camera move with the rig ?

Yes it is.

UPDATE: I adjusted the SWAY AXIS in ENGINE up to 100 from 60, and lowered the SWAY in TUNER from 15 / -15 to 11 / -11. I made similar adjustments to the EXTRA1 (traction loss) and the aggressive snaping went away. The rig is much more in tune now and my head doesn't shift left and right as much. It is derivable now.

I bet the inertial sensors, which really do most of the tracking, were picking up the motion as you turned it up in your sim to where it got noticeable. The camera on the Rift, and the Lighthouse scanners for Vive, just correct the inertial sensors back to known positions.

Oliver Kreylos has an analysis (http://doc-ok.org/?p=1478#more-1478) where at least for Lighthouse, there are approximately three position updates by inertial sensors for every optical fix. I don't know if he has been able to dissect the Rift position update rates but would bet it's similar with more position information derived from inertial sensors than optical updates from the camera. That is the sticking point of mounting the camera on the rig since it provides a reference for the optical location but can't correct for inertia.

The Vive has at least on the surface a worse situation since the Lighthouses shouldn't be mounted on the platform due to the moving parts. The guy making the Feel Three (hope I got that right) mounted a new Lighthouse tracker module on it and said a simple difference got his head movements properly referenced to the platform but Valve/Steam doesn't support that yet in their driver and no idea of their roadmap for if or when that capability might come available.

So either way, unfortunately for now we are stuck with position and/or inertial errors.

The rift camera only updates tracking errors at 60hz so any large fast movement will result in tracking issues.

Thanks for the info.
The sad part about this is that I can't accurately tune... I'm stuck with soft motions that don't match how my race car feels in real life. The MX5 has a ton of grip with low power & weight which creates a fun & responsive driving experience in real life...but the sim cannot do this because I feel my head is moving too much causing the cockpit to shift

This is a problem with VR which I struggled with very early on when I considered building a motion rig. I abandoned my plans for a 6dof as a result. It was the reason why I decided to build a g-seat and not a moving platform.

Until the 1.12 Oculus update mine worked perfectly with the Sensor on the rig just natural movement. Being strapped in nice and tight also helps with less body movement.

I take it you have pointed any other sensors away from your HMD or put a cover over them so they cant see it ?

I've got the same camera shift flying Aerofly FS2 in Rift - spotted SilentChill's entry in Oculus Forum. Is mounting the sensor off-rig at all sensible - surely the Rift will perceive small movements of the platform as head movements?

Yeah it terrible with the sensor off the rig, I use to have big movement a long time ago when I didn't know any better but it spoils it so much now. The least movement you get the better and the only way is to mount the camera on the rig but that's gone down the shit pipe for me at the moment booohoooo. If I mount it off the rig I end up out the side windows and get my head sticking out the roof nearly

Well i've had my head completely through the roof with flight sim. VR is all about immersion and once you've had that you crave more, which should drive a lot more people to motion platforms than in the past. Hopefully they'll sort this soon.

I'm very new to motion platforms. I suspected they'd be a bit bleeding edge but hadn't realised how much.

Since Vive has the separate trackers available, we could have an easy way to correct for both artificial inertia and position contributions to head position if Valve would just write it into SteamVR. Even just strapping on a Vive wand and designating it as a motion reference. All they need to do is give a way to designate a tracker as a motion reference and then subtract it's motion from any motion reported by other devices.

The Feel Three guy did it himself in I think Unreal Engine and said it worked perfectly. Devs could do it themselves the same way but each app would need to be modified for cancellation. The best way is for people to add their voices over at the Steam Community forums in the VR feature request board.

If Valve adds the feature in the driver then no apps have to be modified. They would just work. But I'll settle for devs doing it but am afraid we are a small community. The Live For Speed guys are extremely responsive to user requests and they just might do it and it may even start a trend if we need to go the route of individual apps.

I don't know if this would be able to work for Rift users but if they use SteamVR I don't see why a Touch controller couldn't be designated as a motion reference the same way. Also, since Palmer is no longer at Oculus, maybe asking them directly again would help.

I have posted a couple of requests on the Steam community boards now but I think only the one I just did was truly in the proper place for a feature request. Thing is that as long as neither supports it, there isn't as much pressure on the other one to add it. I think we should request of both companies to do this as it should only speed up the process regardless of who breaks through first.

http://steamcommunity.com/app/250820/discussions/4/1326718197209837066/

You have an excellent point in that post, Noorbeast. (Or I believe you are making the point - didn't read the whole thread yet.) One I hadn't considered. The relative location of the tracker matters and it's more than just a straight subtraction. I think locating it near the headrest (as near the head as possible) would help minimize errors from the different motion axis locations. In a 2dof, the pivot point is low and beneath the seat so locating the tracker low wouldn't see the same motion as if it was mounted higher. For traction loss, mounting it too far back would accentuate that movement. For 6dof rigs, there really are no pivot axes and they are all virtual.

But for now at least, with no good way of cancelling motion, I think mounting a tracker as close as possible to head position but rigidly mounted to the frame would give a good approximation of the sim movement that the head experiences.

It would be cool if that dev can do this. It sounded like he is still working on it. I need to read the full thread but that's the impression I get.

I also sent an email to Scawen and the Live for Speed devs asking them about cancellation too. I completely agree that a proof of concept, no matter where it comes from, may be the spark that gets this moving.

It sure sounds like at least for now, without cancellation we are stuck with very limited platform movement.

Not sure if this has already been discussed but seems as though NxtLevel have come up with some sort of solution for their product?

http://www.nextlevelracing.com/motion-for-vr-introducing-vr-headway/

I would guess it is using the known dimensions of the rig and motion commands to know (calculate) how much the head position moves and uses that to compensate values for the game.

I wonder if you could use something like the FreeTrack protocol and an EDTracker to feed adjustment values to the game to compensate for the motion tracking. ie in essence reverse all the EDTracker values. I would assume doing this the camera would be better mounted off the rig? The EDTracker mounted somewhere near the head.