Hey Elliott, just noticed your reply, good to see you found MTBS at last! It’s a bit quiet around here these days...
Shame you didn’t drop by 2 years ago, would have liked to get your take on my crazy GameTrak idea!
There must have been quite a few GameTrak owners on here, I did a few tracking experiments with it myself
but I ended-up using a modified retractable cable for my slip-ring setup:
I finally got around to building it, about this time last year actually, but didn't get time to post a proper DIY guide so
I’ve just chucked some photos of the build up on imgur for anyone who's still interested: http://imgur.com/a/BWXPf
Your Roto VR system looks great! Good luck with the KickStarter, nice to see you’re getting some press now, I’m plugging it whenever I can for you.
Good call moving the PC off the base, hard disks don’t like pulling too many G’s, and I prefer my water-cooling loop neither shaken nor stirred!
I really hope you get some traction on the Oculus SDK integration stuff, to compensate for the camera rotation issue etc, I’m not a software guy but it should be possible to hack a shim in there somehow if it comes to it. Maybe OSVR or Sixense’s new SDK could present a possible solution at some point, it will be interesting to see how all the input API stuff pans-out in the long run.
Personally I’ve been hooked on standing VR and fully-decoupled controls for a while now, after trying Teddy’s excellent Hydra-deck demos I could never go back to “gun-face” VR. In fact I couldn’t really play much else on DK1, even cock-pit stuff without positional tracking was a problem after 20 minutes, and I’m still waiting to miraculously develop immunity to stick-yaw turning, I'm not holding my breath. I ended-up modding HydraCoverShooter to work with some old UDK maps and just spent a few days exploring them! Even when calibrated properly, Hydra positional-tracking obviously isn’t a patch on DK2, but I still think the intuitive control scheme (360 degree, fully de-coupled) makes those early HydraDeck demos some of the most immersive publicly available VR demos - to this day.
I still remember the first time I got the calibration dialled-in just right, I had been playing HydraCoverShooter for about 10 minutes when I accidently dropped my Hydra controller, I looked down to see it’s virtual representation lying on the ground behind me, and (without even thinking about) just crouched down, picked it up, then turned to aim at something and carried on playing. A few seconds later it struck me – I was actually *in* VR… I know I experienced my first fleeting moments of presence in that demo, so I just want to say thank you Teddy0k – I’ll never forget it. HL2VR used to be amazing to, but sadly its Hydra torso-tracking mode has been broken for a long time now, hopefully Nathan will get around to fixing that with the STEM update (was nagging them about it on Reddit but didn't get very far).
- It’s a real shame they've pulled the plug on UDK…
It would be nice to try these demos in the DK2, any chance of releasing that internal build some day?
After the recent UE4 announcement I'm not sure what harm Epic thinks it could do now??)
I think most people have been really under-estimating the importance of proprioception in general, just look at the response to the standing Oculus Connect and (Re)Vive Lighthouse demos, this aspect of presence seems to have come as a revelation to many. Even when seated, adding inertial cues into the mix will be really compelling for certain content.
- How are you getting on with the Roto’s new slip-ring design? The quotes I was getting for custom wired units had a few too many zeros on the end, so I just built one myself for less than £50! There are still some cheap Moog capsules left on eBay if you’re quick, but the Chinese clones should work ok at low rpm. It would be interesting to compare notes on the electrical performance; I pushed mine to 1920x1200@75Hz without any visible pixel artefacts. I’ve not actually tested it with DK2 yet, been really busy lately so lent it to a friend and he’s taking his time getting it back to me. But I’ve already made up some break-out adapters to send power, USB & sync down standard HDMI cable, and I’m hoping voltage-drop won’t be as much of an issue for DK2. I added a simple EEPROM circuit to spoof some signals on the PC side of the slip-ring, I was expecting detection issues anyway and needed another pair free because I’d originally planned to try and send the two Hydra controllers signals over the slip-ring to, but I chickened-out of that in the end, didn’t want to risk it after the Hydra supply dried-up and the price gouging started, might still try it some day though.
If it wasn’t for the sync signal I could have used a cheaper 12 track slip-ring, if you are keeping the camera on the Roto’s base you could probably get away with that, by EDID cloning and tying all the shields to a common ground etc. When I get time I’m going to make up some DisplayPort adapters to see how far I can push the bandwidth on my prototype. The HDMI repeater isn’t actually needed for most short cables, and to be any use on the other side of the slip-ring it would need some manual re-tuning via it’s SMbus, not played with any of that yet though, something like the RedMere chip would probably be better for CV1 anyway. Not sure if Oculus will use DisplayPort or HDMI 2.0 for CV1, dual-link DVI seems very unlikely, but who knows, they could go with ThunderBolt, SuperMHL or something really weird I’ve not thought of?
http://www.feelthree.com/2015/03/ultra- ... layport-2/
Longer term we might end up having to use something like a 12G-SDI coax or optical hybrid slip-ring to get the bandwidth, but that would be expensive and signal conversion might add some latency. If nothing turns-up on the wireless front maybe these magic light-bulbs could be the solution one day!
http://www.tweaktown.com/news/43630/res ... index.html
Wireless is not really a pre-requisite for standing VR anyway, back-top PCs are one solution, but those retractable cables actually work pretty well over a small area, just don’t stray too far off your “magic carpet”. Because it’s a ribbon cable it actually helps drive the slip-ring around as it retracts, but they are not very robust and I think they’ll struggle electrically with CV1 (very thin cores, no proper shielding etc). There are better retractable cables out there but a custom design would probably be needed, short coiled “phone cord” type HDMI cables would work ok for ODTs but they stretch easily and are a bugger to untangle.
I found that you really need a thicker cable to transfer enough torque to get the slip-ring moving, very thin cables just want to loop-up like you see in all the Omni / Virtualizer videos. I’ve been thinking about a way to use a longer very thin cable by integrating the retractor spool & slip-ring together, putting the slip-ring at the centre of a lightly-sprung spool and mounting it to the ceiling as one unit, turned on it’s side and maybe angled down towards the player on some sort of rotating gimbal mount. Unfortunately that would be the easy part…
The cable to body interface remains the biggest problem, as Palmer suggested I normally fixed the DK1 box to a belt-clip on my back, this worked ok but you really need it higher-up, walking backwards is unnatural and turning around quickly onto the extended cable was still an issue sometimes. But you don’t want the cable going directly onto the HMD either, the weight / tension tries to pull it off - and no one wants to wear a chin strap! For safe head-clearance you really need at least a foot of rigid stand-off fixed high up on your back, centred between the shoulders probably makes the most sense. I looked at camera poll-mounts
but they are really cumbersome, you probably want something like this
or a H.A.N.S harness
but with a much taller "collar" structure to run the cable down onto (lightly tensioned to take up any slack). It's obviously not ideal but it solves most of the safety and comfort issues without adding too much complexity. I suppose you could try and integrate it into a tracker-mount harness of some kind, but it’s not exactly practical, let alone productizable.
Sorry for the brain-dump but I’ve been meaning to post about this for a while now. It’s great to see someone else working on the locomotion problem has actually gone and built something! I've always loved a good contraption (apparently that word has a negative connotation!?) - didn't John Carmack tweet something recently about "a prototype being worth a thousand meetings…” ? I have to agree, I think your final design is looking great.
Hopefully once the various tracking solutions finally start to roll-out we’ll see more demos with fully de-coupled control schemes. Then you just need to find your killer-app for the motorised view-point stuff, a great turret-shooter maybe or some cool 360 video experiences should do the trick, I guess you've already reached out to the big flight-sim / space-shooter developers? (DCS, Elite: Dangerous, Star Citizen, EVE: Valkyrie, DarkField VR and so on)
Valve’s Lighthouse tracking seems to have got people thinking about standing VR again, maybe that could actually be a thing, but I guess we knew that all along really…
I say damn the lawyers Oculus – you’ve got to stand up for your rights sometimes!
After all, accidents will happen, even whilst sitting down: http://www.theregister.co.uk/2015/03/05 ... o_trouble/
Be careful with those trade-show demos Elliott!
Will keep an eye on http://www.rotovr.com
Best of luck with the KickStarter launch - “build it and they will come...”
https://www.kickstarter.com/projects/ro ... next-level