Feel Three Motion Simulator update - launch incoming! - Page 2 — Oculus
New to the forums? Click here to read the "How To" Guide.

Developer? Click here to go to the Developer Forums.

Feel Three Motion Simulator update - launch incoming!

2

Comments

  • DaftnDirectDaftnDirect Posts: 4,613 Valuable Player
    kojack said:

    We haven't had confirmation that it's Vive only have we? They haven't replied to questions in this thread about whether they're able to get their software to communicate with Oculus' in a way that enables it to cancel out the chair's movement or whether they have some other method of achieving this but their website states 'every major headset supported'.

    The problem is people want to use motion rigs with games that don't support them (there's a big difference between games supported by a rig and games that support a rig). Motion cancelling by the game is easy. It's just a little quaternion/vector maths.

    Or the motion rig devs could do what my Oculus Injector project does: runtime hijacking. Shouldn't take more than a couple of days for a team with money. This can potentially cancel rig/headset motion in any Oculus based game without the game or oculus knowing. Although an online game might detect it as a cheat.


    Agreed. It's just that the devs have now stated on their website that the chair itself doesn't support the Rift so it seems they haven't figured out a solution at all, unless I'm misunderstanding.


    https://www.kickstarter.com/projects/feelthree/feel-three-virtual-reality-motion-simulator


    Under the section 'How it Works'

    'There is currently no acceptable motion cancellation for the Oculus rift. On OpenVR headsets such as the HTC Vive we have 100% functionality, but the rift currently doesn’t have a solution.'


    I'm just a little uncomfortable with the OPs initial proclamation type post in the Oculus forum without it including this statement, if that is indeed their position. Well, I'll leave it to you, nothing else really to add from me.

    Intel 5820K OC@4Ghz, Titan X (Maxwell), 16GB Corsair Vengeance DDR4, ASRock X99 Taichi, Samsung 500Gb 960 Evo M.2, Corsair H100i v2 Cooler, Inateck KTU3FR-4P USB 3 card, Windows 10 Pro v1809 (17763.348)
  • kojackkojack Posts: 5,149 Volunteer Moderator
    I was just playing around a bit with my Injector project (instead of doing actual work I was supposed to be doing).
    The following little script will use the right touch controller as a motion rig tracker, cancelling out it's rotation from the headset. Tested on my office chair (which can lean back and turn).
    (Not doing position cancellation yet, but that's not hard either)
    function HookGetTrackingState(time, latency)
    state = ovr_GetTrackingState(time, latency)
    state.HeadPose.ThePose.Orientation = state.HandPoses[1].ThePose.Orientation:conjugate() * state.HeadPose.ThePose.Orientation
    return state
    end
    On the monitor, it works correctly (I'm testing with Google Earth, that's my standard test program for this stuff), I can spin around and tilt but it looks like I'm staying still if my head matches the touch controller. Yay.

    However all is not well at the moment. Timewarp is getting confused.
    The problem is Time Warp knows the true headset orientation instead of the fake value I'm giving to Google Earth. Time warp also knows the headset orientation when tracking was last read. When submitting the frame to the runtime, earth says the fake value. Timewarp sees that and thinks "Wow, the current tracking is way off from when you rendered, I'd better rotate the view to compensate". So the view drifts off as a floating quad.
    Have you ever seen a game where a freeze from loading stuff makes the screen kind of detach from the headset and you see the edges? That's timewarp trying to fix things.

    Sadly timewarp is forced on, there's no way to stop it. But if I can hide the fake orientation from it, it should be ok. I've tried intercepting the submitframe call to modify it's layer data (the layer is where you tell the sdk which head orientation you rendered with), but it's still not working. I need to intercept a few more things to see what's going on.

    Interestingly, while looking at this I noticed the Oculus Debug Tool has a new field: Pose Injection. Turning it on makes the headset and touch controllers all move to a fixed position facing forwards with no tracking. I can't find any docs on what this is intended for (it looks like it's part of runtime 1.30, but the latest sdk is still 1.26).

    So that's a little hacking I did today. I REALLY need to get back to actual work for the next 2 days.


  • RiftFlyerRiftFlyer Posts: 112
    Art3mis
    Thanks so much for looking into this. A lot of sim enthusiasts with motion platforms would be very grateful if you find a solution. 
  • DaftnDirectDaftnDirect Posts: 4,613 Valuable Player
    @kojack I hope the OP is continuing to read this thread, I'd be very interested in what their thoughts are on your work. Plus if they find this useful, you need to make sure you get commission.... or recognition at least.
    Intel 5820K OC@4Ghz, Titan X (Maxwell), 16GB Corsair Vengeance DDR4, ASRock X99 Taichi, Samsung 500Gb 960 Evo M.2, Corsair H100i v2 Cooler, Inateck KTU3FR-4P USB 3 card, Windows 10 Pro v1809 (17763.348)
  • kojackkojack Posts: 5,149 Volunteer Moderator
    edited September 2018
    So... done it.
    :)

    I've now got the Oculus Tiny World demo working with the right touch cancelling my office chair rotation. Time warp has been tamed.
    The code is pretty horrible at the moment, hacks upon hacks upon hacks, in thousands of lines of C++, so you can write a 5 line lua script. :)  I mean, right now even the vr executable that you want to apply this to is hard coded.

    So it needs a LOT of work, but the test succeeded.

    Turns out a lot of the trouble I've had was because I was testing with Google Earth (it starts pretty fast, has touch controls, etc, seemed like a good test). It was built with an ancient oculus sdk, a bunch of data structures it uses are now different sizes (for example the current ovrLayerHeader is 128 bytes bigger than when Google Earth was built).


    Now I REALLY need to get back to real work. :)


  • brmmbrmmbrmmbrmm Posts: 7
    NerveGear
    kojack said:
    So... done it.
    :)

    I've now got the Oculus Tiny World demo working with the right touch cancelling my office chair rotation. Time warp has been tamed.
    The code is pretty horrible at the moment, hacks upon hacks upon hacks, in thousands of lines of C++, so you can write a 5 line lua script. :)  I mean, right now even the vr executable that you want to apply this to is hard coded.

    So it needs a LOT of work, but the test succeeded.

    Turns out a lot of the trouble I've had was because I was testing with Google Earth (it starts pretty fast, has touch controls, etc, seemed like a good test). It was built with an ancient oculus sdk, a bunch of data structures it uses are now different sizes (for example the current ovrLayerHeader is 128 bytes bigger than when Google Earth was built).


    Now I REALLY need to get back to real work. :)


    Such good news. There are so many guys on the xsimulator forums who'd love to get access to this solution .. including me :-) I'm tired of my head popping out the side window of my Cessna when I bank steeply.
  • traveltr0userstraveltr0users Posts: 21
    Brain Burst
    Thought you might like to see a quick video of the sim in action in Elite Dangerous....

    https://www.youtube.com/watch?v=ZETNFP0avwQ 

    Kickstarter is at 182% after 6 days now.



    Will try to find time to come back and answer some comments...
  • traveltr0userstraveltr0users Posts: 21
    Brain Burst
  • kevinw729kevinw729 Posts: 4,324 Valuable Player

    Any reason why he is not driving the experience, but sitting there?
    Even with no interaction (steering) there are a series of glitches from the visual / tracking performance seen in the video?
    P6ftmuw.jpg
    ** New Book **
    "The Out-of-Home Immersive Entertainment Frontier: Expanding Interactive Boundaries in Leisure Facilities"
    https://www.amazon.co.uk/Out-Home-Immersive-Entertainment-Frontier/dp/1472426959
  • Techy111Techy111 Posts: 5,559 Volunteer Moderator
    I think he is just watching a replay ???
    A PC with lots of gadgets inside and a thing to see in 3D that you put on your head.

  • kevinw729kevinw729 Posts: 4,324 Valuable Player
    Techy111 said:
    I think he is just watching a replay ???

    Yeah, I came to the same conclusion - but was surprised that a "replay" would include so many tracking and graphical anomalies.
    P6ftmuw.jpg
    ** New Book **
    "The Out-of-Home Immersive Entertainment Frontier: Expanding Interactive Boundaries in Leisure Facilities"
    https://www.amazon.co.uk/Out-Home-Immersive-Entertainment-Frontier/dp/1472426959
  • traveltr0userstraveltr0users Posts: 21
    Brain Burst
    The tracking isn't perfect, the cable arm the tracker is attached to needs to be a little stiffer, the wheels need some calibration and the motion cancellation software probably needs some filtering.. but hey, it's a prototype and again we wanted to get something out there to show it working even if it's a little rough and ready.

    I just demo'd to a group of people and they were extremely impressed...
  • kevinw729kevinw729 Posts: 4,324 Valuable Player
    Thanks for coming and commenting here @traveltr0users - any chance we can discuss with you your plans for VR arcades, I have a number of clients that have questions on the viability of the system, and the demos are obviously aimed at consumer users at this moment.

    P6ftmuw.jpg
    ** New Book **
    "The Out-of-Home Immersive Entertainment Frontier: Expanding Interactive Boundaries in Leisure Facilities"
    https://www.amazon.co.uk/Out-Home-Immersive-Entertainment-Frontier/dp/1472426959
  • traveltr0userstraveltr0users Posts: 21
    Brain Burst
    Demo with me actually driving and fixed tracking...

    https://www.youtube.com/watch?v=8z6OFH8b5Gc
  • DaftnDirectDaftnDirect Posts: 4,613 Valuable Player

    @traveltr0users ;can you please clarify progress with regard to using the motion simulator with the Rift?


    We've made a lot of comments in this thread regarding motion cancellation and how it may or may not be achieved with Oculus' hardware/drivers. This is the Oculus forum after all.

    Intel 5820K OC@4Ghz, Titan X (Maxwell), 16GB Corsair Vengeance DDR4, ASRock X99 Taichi, Samsung 500Gb 960 Evo M.2, Corsair H100i v2 Cooler, Inateck KTU3FR-4P USB 3 card, Windows 10 Pro v1809 (17763.348)
  • kevinw729kevinw729 Posts: 4,324 Valuable Player

    @traveltr0users ;can you please clarify progress with regard to using the motion simulator with the Rift?


    We have been talking about this on the VR arcade dev forum on FB - it seems that OVR have been lacklustre in returning any of the questions that the teams sent them regarding making this compatible with the CV1 - it was felt best to focus on getting the system out to the investors rather than wait on OVR dragging their feet. 

    I think OVR did not see any incentive to support the devs as they had already got it working well with HTC and that is a kinda of red-rag to OVR management at the moment.
    P6ftmuw.jpg
    ** New Book **
    "The Out-of-Home Immersive Entertainment Frontier: Expanding Interactive Boundaries in Leisure Facilities"
    https://www.amazon.co.uk/Out-Home-Immersive-Entertainment-Frontier/dp/1472426959
  • DaftnDirectDaftnDirect Posts: 4,613 Valuable Player

    kevinw729 said:

    @traveltr0users ;can you please clarify progress with regard to using the motion simulator with the Rift?


    We have been talking about this on the VR arcade dev forum on FB - it seems that OVR have been lacklustre in returning any of the questions that the teams sent them regarding making this compatible with the CV1 - it was felt best to focus on getting the system out to the investors rather than wait on OVR dragging their feet. 

    I think OVR did not see any incentive to support the devs as they had already got it working well with HTC and that is a kinda of red-rag to OVR management at the moment.


    I understand the problem as it's been discussed here and in an associated thread.


    A possible solution has been offered by kojack, albeit a with further development needed, if indeed it's able to become a working solution at all. Regardless of where the problem lies, I think it's reasonable to ask the dev to confirm the problem is being worked on or that something like kojack's solution is a dead end or whatever.


    Otherwise this seems increasingly like advertising to the wrong people i.e. to Oculus users, hardware that is not Oculus compatible.

    Intel 5820K OC@4Ghz, Titan X (Maxwell), 16GB Corsair Vengeance DDR4, ASRock X99 Taichi, Samsung 500Gb 960 Evo M.2, Corsair H100i v2 Cooler, Inateck KTU3FR-4P USB 3 card, Windows 10 Pro v1809 (17763.348)
  • traveltr0userstraveltr0users Posts: 21
    Brain Burst

    Otherwise this seems increasingly like advertising to the wrong people i.e. to Oculus users, hardware that is not Oculus compatible.
    It is Oculus compatible, this is the Oculus developer forum and this is where I should be talking about it. If you read my Kickstarter section on Motion Cancellation you'll see I have 6 solutions to it not working with any application in the rift. We even have a developer program so developers, Oculus developers, can add support and get a free development sim.

    We also have a couple of developers already on the program of very popular and appropriate titles who we will be announcing soon.

    Don't confuse the fact that just because we can support pretty much any game in SteamVR that it doesn't work in rift, it could work with a injection wrapper, OpenXR support, Oculus support, Unity/Unreal support, or developer support. The first four fix the issue for all games or developers do it themselves.

    Full native support is an afternoons work for 2 Oculus engineers, they've probably already done it, so I will be back at OC5 asking for the fourth time that they implement it. If you guys did the same they might start to listen.

  • snowdogsnowdog Posts: 6,392 Valuable Player
    I thought it should be easy enough to fix. Are you mounting the sensor to the motion simulator? That's how the Yaw VR motion simulator guys have done it, but I think they're working on alternative methods too.
    "This you have to understand. There's only one way to hurt a man who's lost everything. Give him back something broken."

    Thomas Covenant, Unbeliever
  • DaftnDirectDaftnDirect Posts: 4,613 Valuable Player
    edited September 2018

    Otherwise this seems increasingly like advertising to the wrong people i.e. to Oculus users, hardware that is not Oculus compatible.
    It is Oculus compatible, this is the Oculus developer forum and this is where I should be talking about it. If you read my Kickstarter section on Motion Cancellation you'll see I have 6 solutions to it not working with any application in the rift. We even have a developer program so developers, Oculus developers, can add support and get a free development sim.

    We also have a couple of developers already on the program of very popular and appropriate titles who we will be announcing soon.

    Don't confuse the fact that just because we can support pretty much any game in SteamVR that it doesn't work in rift, it could work with a injection wrapper, OpenXR support, Oculus support, Unity/Unreal support, or developer support. The first four fix the issue for all games or developers do it themselves.

    Full native support is an afternoons work for 2 Oculus engineers, they've probably already done it, so I will be back at OC5 asking for the fourth time that they implement it. If you guys did the same they might start to listen.


    I'm not saying that it shouldn't be talked about in this forum, I'm saying that this forum is definitely the place to talk. I'm simply suggesting that clarification is needed as to whether it's compatible with the Rift.


    You're saying it's compatible which is great but you're also saying it doesn't work with the Rift until Oculus deliver the solution. I would say that that's the point at which it becomes compatible.


    If you're also saying that some sort of injector is not an option (as suggested by kojack above) in the event that Oculus aren't forthcoming with a solution, then we'd appreciate confirmation on that also.


    I'm just approaching this from the point of view of a prospective end user who very interested in this great looking product but I just think this thread needed that clarification from yourself. Which we now have... I think.

    Intel 5820K OC@4Ghz, Titan X (Maxwell), 16GB Corsair Vengeance DDR4, ASRock X99 Taichi, Samsung 500Gb 960 Evo M.2, Corsair H100i v2 Cooler, Inateck KTU3FR-4P USB 3 card, Windows 10 Pro v1809 (17763.348)
  • traveltr0userstraveltr0users Posts: 21
    Brain Burst
    you're also saying it doesn't work with the Rift until Oculus deliver the solution. 

    Thats isn't what I said. Oculus can solve this across the board if they choose, in a few hours.

    If a developer adds their own motion cancellation code it works, and a few will. Getting Unity and Epic on board would also make this trivial for developers, just tick a box. OpenXR is also a pretty good option (if implementation by committee is 'good').... and there is also hooking and injecting (and yes, the wheels are in motion).

    At every Connect, every Oculus employee or management member sends me to someone else and so on and so on. I guess I'll have to try to get Carmack on the record in two weeks saying that they should add it.....
  • DaftnDirectDaftnDirect Posts: 4,613 Valuable Player
    edited September 2018

    You said Oculus could have it working across the board which I took as meaning Oculus delivering the solution.


    So ok, we have clarification that any developer can solve the problem for their game, game engine developers can solve the problem and you can solve the problem and have the wheels in motion. Or Oculus can solve the problem across the board.


    Great, everyone can solve the problem. It just needs to be... erm… solved.


    I'm all for putting pressure on Oculus if that's where the pressure needs to be applied, but until today you hadn't stated in this thread that there was a problem so it was left to us to speculate and ask the question, which we did.... and now we know.


    Why is this conversation so difficult? we're on the same side.


    Edit: Oculus, sort this out!

    Intel 5820K OC@4Ghz, Titan X (Maxwell), 16GB Corsair Vengeance DDR4, ASRock X99 Taichi, Samsung 500Gb 960 Evo M.2, Corsair H100i v2 Cooler, Inateck KTU3FR-4P USB 3 card, Windows 10 Pro v1809 (17763.348)
  • kevinw729kevinw729 Posts: 4,324 Valuable Player
    At every Connect, every Oculus employee or management member sends me to someone else and so on and so on. I guess I'll have to try to get Carmack on the record in two weeks saying that they should add it.....

    Ouch!





    P6ftmuw.jpg
    ** New Book **
    "The Out-of-Home Immersive Entertainment Frontier: Expanding Interactive Boundaries in Leisure Facilities"
    https://www.amazon.co.uk/Out-Home-Immersive-Entertainment-Frontier/dp/1472426959
  • traveltr0userstraveltr0users Posts: 21
    Brain Burst
    Yeah, I have a whole section on this topic in the Kickstarter, 500 words... I didn't want to have to write it all again... :p 
  • traveltr0userstraveltr0users Posts: 21
    Brain Burst
    snowdog said:
    I thought it should be easy enough to fix. Are you mounting the sensor to the motion simulator? That's how the Yaw VR motion simulator guys have done it, but I think they're working on alternative methods too.
    That doesn't work.... it takes 10 seconds to test it... unless you don't turn on the Yaw axis.... which would be pretty ironic :p 
  • DaftnDirectDaftnDirect Posts: 4,613 Valuable Player
    Yeah, I have a whole section on this topic in the Kickstarter, 500 words... I didn't want to have to write it all again... :p 

    Well we asked the question a while back... just a quick reply and a link to that page would have been good... computers are great like that ;)
    Intel 5820K OC@4Ghz, Titan X (Maxwell), 16GB Corsair Vengeance DDR4, ASRock X99 Taichi, Samsung 500Gb 960 Evo M.2, Corsair H100i v2 Cooler, Inateck KTU3FR-4P USB 3 card, Windows 10 Pro v1809 (17763.348)
  • snowdogsnowdog Posts: 6,392 Valuable Player
    snowdog said:
    I thought it should be easy enough to fix. Are you mounting the sensor to the motion simulator? That's how the Yaw VR motion simulator guys have done it, but I think they're working on alternative methods too.
    That doesn't work.... it takes 10 seconds to test it... unless you don't turn on the Yaw axis.... which would be pretty ironic :p 

    Well the Yaw VR guys managed to do this ages ago:





    You can see yaw, roll and pitch axes in operation in both of these videos.
    "This you have to understand. There's only one way to hurt a man who's lost everything. Give him back something broken."

    Thomas Covenant, Unbeliever
  • traveltr0userstraveltr0users Posts: 21
    Brain Burst
    snowdog said:
    Well the Yaw VR guys managed to do this ages ago:
    You should know better than to quote old videos from February :p 

    Pretty sure that was before Oculus updated home and broke the implementation that was working. I tried it a few weeks ago.... pick up your sensor, point it at your head and turn around.... and then tell me it works perfectly :p

    Actually, look at the video, the sim is moving and he's holding his head level. There is about 10 degrees of pitch, roll and yaw so it's easy. Would you like me to fake a similar video too where my sim barely moves?



    this video shows the jump as the home tracking readjusts at around 52s and 27s....

    Shocking deception really.... maybe we should have done the same...
  • lluisgl7lluisgl7 Posts: 1
    NerveGear
    snowdog said:

    Well the Yaw VR guys managed to do this ages ago:





    You can see yaw, roll and pitch axes in operation in both of these videos.

    Fake videos, the joystick is not even plugged in..
  • RedRizlaRedRizla Posts: 5,833 Valuable Player
    edited September 2018
    @lluisgl7 - That Joystick could be plugged into an extension lead because you cannot see the end of the Usb. But it does look a bit fake though given the guys not a good actor.
Sign In or Register to comment.