VR Motion Cancellation - Time to test!

Latest at my end is the VIVE-Vive Pro is still working absolutely fine with the latest Steam VR updates.
if you want a good system that is reliable then for now the Vive is the only way to go.

The big question on my mind is the valve index now on preorder going to work with the Open VR emulator?
I suppose 460 quid is going to be the only way to find out,
Not unless a reviewer who has one can test it for us all.

I started with VR when oculus DK2 was out but since than I was completely out of the topic.

Back than, to obtain a proper response, the only thing one had to do, was to attach the front camera to the rig, and the headset 6-dof motion was acting locally not being affected by the motion of the rig.

I would have two questions:

1.)
Is someone aware of a headsets that doesn’t need the installation of fixed sensors everywhere in the room?
2.)
Do we have a solution on the market (including the upcoming products like Rift- S or Valve or others) which can be used without steam?

Thanks

Well, I pre-ordered the Index for June 28th so we’ll see.

Nice one, I hope it works for you, please let us know as soon as you can.

i am trying to find out if its possible to order with proxy services (i am in malaysia)
but its a pre order so i doubt it
probably need to wait it out some more for it to retail in online stores
there is no way anyone can prevent me to get my hands on this, its even better than pimax 5k performance wise
and it will be compatible with vrinput emulator plugin because its native steam/openvr, i mean no doubt that steam will make sure of that in foreseeable future

Looks like you and I had a very similar idea I think my approach is a little more involved though since it involved 3D Printing and pouring rubber into the part. I can't post a link as of right now because I don't meet the anti-spam post limit just yet, but everything is only Thingiverse, just look up "Vive Tracker Isolation Mount"

I don't suppose anyone is interested in testing this? My seats don't provide any properly rigid mounting spots (Corbeau Forza) so I can't rule out other sources of shaking.

A solution right under our noses guys!!!!

But please tell me If I have missed something.

I’ve been thinking about this a lot as I’ve invested so much time and dollars into my platform and I really need motion cancellation to work with VR.

I think that we can solve the problem 100%, AND with NO sensors being attached to the seat.

Here’s what I’m thinking:

All we need is absolute position data. XYZ position offset from zero origin AND XYZ angle. That way we get the position and orientation of the pilot/driver in 3D space.

Even better than using an IMU to estimate this which is prone to drift problems over time I figured that we already have the amazing HexPod app that plugs into SimTools already.

HexPod relies on the user inputting the exact dimensions of the users platform to work accurately, and so when the platform is in motion and being driven by the game we therefore already know it’s absolute position at any moment in time. (And in real world 1:1 scale measurements)

With this position information (available as a data stream from HexPod app/plugin), we can just feed it real time back into OpenVR as the offsets that it requires for our motion cancellation.

Please tell me I’m missing something and it’s really not this simple. After staring at HexPods 3D platform plugin for months, I’m pretty sure I’m on the right track. And I’m also sure that Ignacy (@hexpod) could get this data for OpenVR to use.

Also I’m assuming it would be trivial to have a variable to offset a position above the platform or CoG to estimate the position of the users head.

Cheers.

Matt

Some of us have been thinking of that too. I'm sure thats a good idea, even if motion couldn't be precize, because the rig never moves esactly the same way as the interface shows it. Because of actuator speed differences, the rig geometry or range and limit differences. But some millimeters don't cause a big issue, so I like that.
Especially if @pmvcda could make it with FlyPT interface

Would this work for any rig style because that would be swell! As far as accounting for what the rig is actually doing, you could still attach a controller/tracker and run a sequence of movements and different speeds and accelerations and record what it does and create a lookup table (assuming I know quite what the purpose of those are lol) Once you do this, you won't need the trackers anymore because it will know what to expect from the rigs movments. There could be some tuning involved but I think it could work.

I don’t know exactly how the modern headsets works.
If you don’t want to use the external position cameras, I guess (but I might be wrong) the headsets doesn’t know where it is located concerning the translational movements.

Following your assumptions, subtracting the angular positions could work but how about the translational movements ?
We also need to subtract the translations. Am I wrong? I don’t see how it could be done without an external reference device.

Concerning HeXpod, it just got an update changing from world coordinated movements to local coordinated.

You can try it :
http://hexpod.xyz/heXpod_2.6.zip

@hexpod All headsets behave differently but each has the same problem in common, that if the platform moves a lot, for example in a plane with the sway motion (my platform moves more than 30cm sway), then we feel like we are moving around (translating) in the cockpit. Imagine like if we got out of our seat and walked around the room during play in a race car game, we would walk outside of the car. This is what actually happens in the game with motion platforms and our head view falls outside the cockpit.


It’s these translational movements that we want to measure then subtract, so we can keep our bum constantly planted in our seat.


I believe OpenVR-InputEmulator already has the facility to readjust our position based on receiving offsets, but we are having trouble receiving those offsets for a number of reasons. The hand controllers are prone to vibration, the tracking stations themselves don’t work when attached to a rig.

Note that to a small but essential extent, we still need the external (or internal for new VR sets) position trackers to see us wobbling around in our seat due to G-Forces in both translational and angular deltas. It’s just the “synthetic” or artificial movements of the platform that are the movements that are unwanted and must be subtracted.

If you can output these translational measurements as a distance from the origin for us then I believe the OpenVR-InputEmulator developer can feed that data in for us and from x this problem. I’m happy to talk to the developer first if you can confirm that your end is possible to calculate. Basically to visualise this “offset” in HexPod we need a vertical stick locked to the middle of your 3D platform model, and of variable length based on users head height.

Let me know if that makes more sense.

I can take this chat offline and test with you if you can make that test above for me. (I sent u an email already yesterday)

I will test 2.6 in the morning when Australia wakes up. Is there a change log?

Seems Pimax is going open source, maybe that will open up some possibilities.

I think motion cancellation needs to work across all
Headset platforms.

So if I have the following:

Oculus Rift CV1
HTC Vive (with BaseStation 1.0)
HTC Vive Pro (with BaseStation 2.0)

Am I correct in interpreting that the HTC Vive Pro is going to be my best option for the motion cancellation? I have the HTC Trackers as well, so I can work it out to attach it to the rig.

Thanks,

Dan

I think so. My recollection is that getting cancellation working for the Vive was easy because the the tracker and headset are using the same tracking system -- so you don't have to deal with figuring out the offsets to get them aligned to each other. I'd assume the Vive Pro is the same.

hexpod 2.6 is local coordinated (translations dependent on rotations) in opposition to previous versions which are world coordinated.

I think there is no problem to output the platform positions through UDP or MMF.
Maybe MMF would be more suitable and easier to implement.