Anyone using a motion platform with VR?

[Einherjer]

This will kill ATW I think and it will need much configuration to get a smooth 1:1 tracking... don't believe it's worth it

[SkateZilla]

I Think if TrackIR is enabled and tracking that the Oculus Tracking is disabled.

[vJaBoG32]

Thank´s very much SkateZilla for this posts and this first steps...i also try to get this solved in the rift dev forum. And one thing about the Rift CV1 in emu. stereomode: In my case the 3D feeling in DCS (with stereo lua) dosn´t look that perfect, compared to the vanilla generated Rift 3D picture. I played a lot with the lua valures, but on of the 3D views (F2,F1 or F5) are not good, whatever.

[SkateZilla]

I'm not sure how other sims work or the Oculus API / Interface,

 

But if Oculus IMU Rotation Input is disabled, would the sim use the camera for tracking all 6 Movements, since the camera is there to sync the rotation axes because the IMU will drift occasionally?

 

if that's the case, then all that needs to be done is to ask ED for some type of Oculus VR Checkbox/command to disable the Oculus IMU.

Edited October 12, 2016 by SkateZilla

[dureiken]

Sorry but I didn't understand. Is rift usable for motion platform ? without trackir or freetrack ?

 

I'm doing blueprints for motion pltaform, I would like to know before buying everything :)

 

Thanks

[vJaBoG32]

I'm not sure how other sims work or the Oculus API / Interface,

 

But if Oculus IMU Rotation Input is disabled, would the sim use the camera for tracking all 6 Movements, since the camera is there to sync the rotation axes because the IMU will drift occasionally?

 

if that's the case, then all that needs to be done is to ask ED for some type of Oculus VR Checkbox/command to disable the Oculus IMU.

 

That´s would be i nice workaround!

Do you may know someone from ED Dev.? Elseway i´ve doubt to get even a response :music_whistling:

No response from occulus dev forum, yet.

[dureiken]

Sorry but I didn't understand. Is rift usable for motion platform ? without trackir or freetrack ?

 

I'm doing blueprints for motion pltaform, I would like to know before buying everything :)

 

Thanks

 

up ? :) thanks

[Frusheen]

up ? :) thanks

 

There is no easy answer to this. It depends on the platform and how tolerant you are of spurious headtracking movements. Seat movers doing fast shallow angle movements such as those used in car racing seem to be ok for some people from what I've read. Platforms which move to extreme angles show obvious tracking issues. In the extreme case where the platform rotates 360degrees then it won't be easily usable. The rift imu detects platform movement and translates that to head movement even tho the users head hasn't actually moved relative to the game view. This can result in nausea.

 

Edit: The easiest solution is to build a g-seat ;-)

Edited October 15, 2016 by Frusheen

[Einherjer]

Anyone using a motion platform with VR?

 

My platform is now ready for early tests with the rift

Atm the sensor is stationary which gives me this tracking issues , but it doesn't feel that strange but it can be better.

 

Some work to do

 

[dureiken]

awesome work :) what are your parameter ? linear speed, momentum, lever lenght ?

 

thanks

[Einherjer]

Thanks

It's in an early status.

 

I have 12cm lever length

 

155rpm @ 24V

 

So 24cm at 0.3 seconds

 

Seems quite fast, for flying (for driving it's to slow maybe)

The feedback line matches always the output line

 

Momentum is 37Nm per motor - but I can stand on the edge of the platform and it keeps it's center position with just 6A current (it's a wormdrive but can freewheeling so aktiv holding is required)

 

Motors are wheelchair motors 670w each

 

Max usage 36A ATM

 

Sabertooth 2x60 with arduino and smc3 software

 

So now BTT: lets see how tracking works

[dureiken]

155rpm => 0.193s to do half a round no ?

[Einherjer]

Oh you are right, but this are just theoretical calculations without load, but all in all I can't complain about the speed.

 

But I need to reinforce some structures

[MacThai_75]

Well, as I earlier indicated ... I ordered one of these Dynamic Kit Pro 2DoF motion cockpits for use with my CV1 for racing sims ... as this guy in the video is successfully doing, and hopefully for flight sims as well.

 

 

It came from Ukraine in two boxes.

One box was delivered last week after a loooong multi-week wait, but the other package is stuck in what I find by googling the tracking info that says "Your item has been processed through our facility in ISC NEW YORK NY(USPS)" really means that ... and I quote Harrison Ford from 6 Days and 7 Nights ...

 

 

 

:(

Edited October 20, 2016 by MacThai_75

[MacThai_75]

The A3 arrived today. It's 2:30am PST, and I just got a quick landing in. LOTS of thoughts even with a completely un-tuned profile. But the main one is...

 

... head tracking is a non issue. The Oculus camera is zip tied to the platform, and has quite a bit of play (Atomic is sending a proper fastener soon). Still not an issue. I haven't understood why this thread is concerned with this if the positional tracking camera is stationary relative to the head (ie, moves with the platform).

 

 

Everyone that I see that's actually 'using' one says there is no issue tracking with the rift on a motion rig, and that it works just like it's on a stationary rig ... and the different videos of people using a motion rig with VR while getting tossed about in a vehicle sim, looks to be doing just fine too.

I haven't come across any flight VR motion rigs viddys yet, so if you get a chance please do post up a youtube! :)

 

My DKP motion rig is hung up in the never-never land of US customs ... but if it ever gets out (and I can figure out the pitch roll as you go inverted thing) I'll be surprised if I see the tracking error that theory does seem to say it should see.

Possibly because the theory doesn't account for something in the rift that makes it work exactly as you and others have found ... by actually trying it. (???)

 

Really looking forward to your complete review, and certainly more on the roll (and I assume pitch) when rolling or pitching past inverted. :)

Edited October 23, 2016 by MacThai_75

[MacThai_75]

... Or maybe it's just cuz we all be genetically chicken disposed .... :)

 

 

If you look at the driver in the viddy a couple of posts above, that's using a CV1 and rig mounted camera ... you'll see at the first hard braking the rig pitches down and forward, but the viddy stays level ... as does his head.

This is repeated throughout the viddy.

 

I suspect that as long as a motion rig with VR does not roll or pitch beyond the travel of a rider's angular head movement (in pitch and/or roll) the VR tracking will work just fine, but if you have a rig that has motion beyond your ability to tilt your head to compensate (as we humans naturally do) I suspect that the video will tilt.

 

That 'might' explain why people are finding no issues with either an on rig, or off rig mounted camera as some have emphatically reported.

IOW: If you strap your head to the seat back you may find that the tilt is seen right away ... and full 6DOF might not be what you want in a VR motion rig.

 

Anyway ... once (read IF) my rig gets here I'll do some testing and post the results.

 

P.S. If you don't like/agree with this post and think it's just too far fetched ... you at least gotta love the chicken viddy. ;)

Strap a camera to a chickens head, add a perch to you dirt bike's handlebars ... and get you some really stable no multi-axis gimbal required GoPro viddys! :)

Edited October 23, 2016 by MacThai_75

[Einherjer]

Anyone using a motion platform with VR?

 

After testing a few days I can say, when you have little movement (around 5°) , camera of platform and no VR pilot in your virtual cockpit like it all is in racing games with a seat shaker, then you are absolutely fine .

 

But if the angle of the platform is about 15° you will notice the movement but it really gets problematic when you have a virtual pilot enabled because your viewpoint moves up and down and suddenly you look out of your chest...

 

When you mount your sensor on the platform, this problem vanishes, but then the discrepancies between headset and sensor becomes too big and the vision begins to jump to a new center position as soon as you cross a limit.

[MacThai_75]

^^^ That makes sense too.

And if you could tilt ... and stretch you neck more (for heave) like the chicken ... you may be able to compensate for more tilt in a motion rig before the rig's tilt exceeds your head's tilt ability, and things start to go Elvis. ;)

 

One thing that VR adds to the sim experience, is that it already simulates motion on a stationary rig, so setting a rig's angular travel limits with 'just' enough to add to the experience, but limited to whatever it takes to keep it working properly, might provide better overall results than trying to make a VR motion rig into a carnival ride. (???)

 

All of this is one reason I wanted to go cheap, and got the $1650 delivered DKP 2DoF rig.

Now ... if the second box (it was a two package shipment) will just get out of customs in NY and deliverd to me, I'll get to actually test my theory and see it it can stand up to the light of day.

If it's a bust for VR ... as many that I have a lot of respect for say it will, then I don't have thousands tied up in it.

 

If it does work as well as some have reported, then I'll get a more expensive rig ... but one that will have specs (and price) that match the needs, as opposed to "more is better" WRT things like maximum travel and etc.

 

As indicated, a VR motion rig with more DoF, travel ... especially heave travel, than I can stretch my neck or tilt my head, might be wasted money for VR use.

 

For just VR flight, the G-Seat as a commercialy available package might be a big seller! :)

But if you want a dual flight 'and' race rig, then a motion platform set with the appropriate amount of limited travel for each, saved in separate files ... might work well enough for both.

Edited October 23, 2016 by MacThai_75

[SkateZilla]

I'd limit the seat/platform to minor movements/vibrations, anything more than that would be visibly apparent.

[MacThai_75]

I'll be happy with the amount of pitch and roll that the guy in the 18+ minute race viddy in post #114 (that I posted above) is getting ... or possibly even less.

 

I don't think I'll need (or might even want) anything more than what he is successfully getting over the entire viddy.

 

He's using the same rig I have coming (a Dynamic Kit Pro) with the same SimTools for it, along with SimTools game specific plugins.

 

I guess the real question is why does it work for him, and others here that have tried it with a few different rigs, control SW, and DCS, Vs why it didn't work for you?

 

Possibly it was the specific configuration you used?

What motion rig did you use, with what control program, and how did you mount your camera, etc.?

Maybe there's a hint to be found in that information as to why it works well for some, then only so-so for some, and apparently not at all for others. (???)

 

If people that have posted it being unsuccessful could post what rig/SW combo ... it might help to know ... if only to avoid that type of rig, and conversely for people that got it to work, and to what degree (no pun intended).

 

Or ... if the people that have never actually experienced one, but know technically that it wont work, might explain in lay persons terms 'specifically' why the pitching and rolling in the race viddy above works, but if it was a flight sim, it would not work the same ... even with it limited to the same degree of roll and pitch he is experiencing in the viddy.

 

His straight and level view through the left side of the windshield is the same at the end of viddy as it is at the beginning when he starts the race, and it remains the same through the race except for times that he turns his head ... but it returns back to the same view when he looks straight again, as it also basically does when he holds his head level during curves, acceleration and braking, although the rig is pitching and rolling all about.

 

So ignoring flight, how specifically does the CV1 work just for only 'that' racing roll and pitch?

Then ... what specifically makes that different from a flight sim that moves roll and pitch to the same amount of angular travel?

 

Ignoring the going vertical issue (which the rig below seems to handle nicely) ... if a rig like the one below (which BTW is using DCS) and looks to have about the same amount of limited travel as the race rig above (but moving much smoother) ... why would it not work as well as the race rig if the pilot was using A CV1?

 

Not meaning to be argumentative, but I'm having a difficult time getting my mind around why there is any difference at all, when the position in space of the HMD is via the camera, and the roll and pitch of the HMD are via the HMD's own internal gyro/accelerometer/magetometer feeding it's IMU with it's roll and pitch's axial rotation (and the acceleration of same) data, as well as possibly actual magnetic heading data ... the same way when used in either racing or flight applications.

How the heck does the camera and/or HMD generated signals know you rolling and pitching an aircraft Vs rolling and pitching a vehicle? :dunno:

 

For now I'm stickin' with my chicken ... theory. ;)

Edited October 24, 2016 by MacThai_75

[Einherjer]

What you see in the video of post #114 is a good tracking because he stays in the limits of the oculus system- if he would move further, the tracking will go wild.

That's absolutely the right approach... for driving

There you need short , fast and hard Feedbacks to make you feel you rock over a street

 

On flying, this short hard and fast movement doesn't feel that good to me (but this is a very personal feel)

It makes flying to feel like a roller coaster- so for me a little slower setting with longer ways feels more natural, Especially to feel your attitude in bad weather, head down time or formation.

But then the tracking issue begins to kick in.

[MacThai_75]

I agree ... but I'm not sure what you mean by "limits of the Oculus system".

It is full 6DoF (both rotational and positional tracking) with unlimited angular travel on any axis, so I'm not really sure what that means unless you mean going outside the range of the camera ... which I suppose could happen if it was mounted off rig ... especially if the line of sight was blocked by the rig as it moved.

But other than that ... :dunno:

 

IMHO: The limits are there because of any given person's ability to how well they can match up to the (impossible) full 6DoF HMD on a given motion rig, with their head's travel in angular movement.

And I believe that's the real and only issue.

If you could swivel your head in a full sphere that would match the CV1, there would be no problem at all.

 

But ... we obviously can't do that, and different people will have different experiences (as has been reported by different motion rig/VR users).

So the difference between a positive and negative experience with tracking would appear to be based on how far a rider can swivel their heads, and if they have matched the rigs motion travel to those limits, which are probably much easier on pitch than roll for most.

 

As far as the speed of the movement ... with a rig mounted camera, as long the acceleration does not exceed the ability of the camera to track the HMD's spacial movement (where it is located in space) nor the the HMD's mounted sensors that track spherical rotation, then all is good.

I would be willing to bet that the speed needed to out pace either the positional and/or rotational tracking is well beyond the average human and commercially available motion rig. ;)

On an off rig mounted system .. it should still all work fine, but will require the rider to have even more head tilt ability and/or more reduced rig motion.

 

My theory as to why it even works at all is because we humans have a sense of balance via our inner ears, and we can see a target/place in space with our eyes.

 

As the rig pitches down the HMD image will also pitch down, and if we held our head stiff we would see towards our knees in the HMD.

But we automatically try to keep looking at the same place in space, so at the same time as the rig moves down we compensate by tilting our head up at the same rate and the HMD image remains level ... until such time that we can no longer match the angular travel of the rig with our head tilt.

 

My rig isn't here yet so I can't test this theory, but anyone with a motion rig and CV1 can.

Simply adjust the rig's travel to optimum results for no error tracking while wearing the HMD.

Then try it again with the same settings but with HMD alone strapped to the seat so it can't move.

 

I believe the tracking will then suffer when it didn't before, because the HMD will then obviously be doing exactly as what others in this thread say it will do, because they may not have taken into consideration the ability of the rider using his head movement to naturally ... and even unconsciously keep the image in the HMD pointing where he/she wants to look by simply tilting their head and compensating for the rig's movement.

 

But once the rig moves beyond the ability of the rider to keep up with the rig's angular travel by tilting, rotating and stretching their head/neck (assuming a 6Dof rig) then things will go bad.

 

That's my theory as to what's happening, and the explanation for the mixed reviews on the different forums that discuss VR and motion rigs, and I think it's good that I only ordered an inexpensive 2 DoF rig, and plan to stack a G-Seat on it.

 

Minimize rig movement to that which will comfortably match my head tilt ability, and since it has no heave axis, there will be no need to stretch my neck (thankfully! ;) ) but the advantages of the G-Seat will continue to fill in for the limited rig travel, as well as the fake the other axis ... by its continued synthesis of motion on all axis, combining for a full and better 6DoF experience.

 

The rig plus a G-seat, instead of just a G-Seat is 'only' because I just can't get past my intuitive feeling that a 2DoF motion rig will add significantly more enhancement to a G-Seat experience, than what the additional $1650 price of the motion rig is.

 

The only other adder I might try is the SimX yaw system which is basically just a very low profile 'T' shaped slider system that a motion rig sits on to add yaw with not a lot of actual angular movement.

Edited October 24, 2016 by MacThai_75

[Einherjer]

By the limits of the oculus tracking system I meant the that the calculation of the movement by the sensors of the headset and the camera system goes off because there are some not expected behaviors when the camera is set on the platform.

 

If you pitch up , your IMU in the headset says "looking up" but your camera says " no movement "

This will result in some kind of recalibration process which will recenter your view in an strange position.

Hard to describe, just test it when your rig arrives.

 

About compensation of the movement by your head, you are right- you will take much of that for real, it doesn't feel that bad as long as it not gets too far off (10° maybe)

This joyrider/seat mover only moves 3 or 4° and only shakes around a lot - this is absolutely enough for driving and feels good

 

A g seat will be an good enhancement for a 2dof platform (like mine too) and it's quite cheap and easy to build, can be done for 100€

 

6dof would absolutely be great but not without a software solution for the tracking system

[MacThai_75]

Yes, I think the a G-seat mounted on a 2DoF motion platform with the possible addition of a SimX $1300 yaw system would be cost effective and offer an ideal "dollar per immersive experience performance".

 

As you may know @Frusheen is doing an amazing job on a G-Seat for helicopters.

His work is excellent and inspirational, as well as providing some very good lessons/tips, so I'm hoping for more posts from him as he proceeds with it. :thumbup:

You can see it here:

https://forums.eagle.ru/showthread.php?p=2733045#post2733045

 

I'm also hoping that the rest of my rig gets out of customs and delivered soon ... or if ever!! :cry:

 

In the meantime ... I would like to mention that I had emailed Oculus about things that can confuse the CV1's positional tacking recently, and posted the results of that contact earlier in this thread.

 

When you say "the camera mounted on the rig will sometimes cause the tracking to go wild" has me thinking about what Oculus said in their reply to me.

And I'm wondering if when the camera being mounted on the rig and moving when it should normally be stationary, is picking up stray IR reflections as it moves about when the rig pitches and rolls, that it might be causing it to interpret them as pilot/HMD positional movement in the cockpit, and changing the POV.

 

So ... could you do me a favor and place a toilet paper roll' tube, or some other type of paper etc. tube over your camera, so as to act like a camera sun shield ... just a tube that blocks out light from outside of the intended target area ... in our case just the cockpit.

 

 

It might take some experimentation with the length of the light shield tube to get the camera to 'just' see the HMD's IR leds, and no other extraneous IR reflections.

 

Also, although more troublesome, if you could move your rig to another area, or maybe just rotate it a bit in case something behind or off to the sides is reflecting IR ... just to see if the symptoms change at all.

Obviously any mirrors or other reflective objects in the close proximity to the camera should probably be avoided.

 

It might just be a wild goose chase, but like I said (and posted earlier in this thread) Oculus 'did' caution about reflected IR as a cause for poor positional tracking.

So it might be worth the trouble to experiment a little. ;)

Edited October 25, 2016 by MacThai_75

[Einherjer]

Ok thanks, that was what i thought first and shutting off all lights and removed reflective surfaces. It only was a short try when i was building my platform, I will try it again soon