Guardian Monkey
Member
I think its great that there are more companies entering the vr space. Will only help in the long run as far as I can see.
-
Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.
The High-end VR Discussion Thread (HTC Vive, Oculus Rift, Playstation VR)
- Thread starter Cartman86
- Start date
Fried Food
Banned
I've cancelled my Oculus CV1 preorder. The ship date had been changed to mid June. My Vive should arrive in May. The fact that my DK2 will continue to be supported, at least till the end of 2016, played a big part in this decision.
https://www.youtube.com/watch?v=SWJJY8cgvKo
Yep, OSVR is really committed. some might think it isn't because all the news headlines are Vive/Oculus.
serversurfer
Member
Remap locomotion to the hands? As I keep trying to explain to you, that doesnt require a remap. Weve been directing locomotion through the forelimbs for hundreds of millions of years. Its the recent trend of cutting the forelimbs completely out of the locomotion loop that requires the remap you so fear, and since its only been a few hundred thousand years, were actually still trying to get the hang of it.
Watch babies learn to crawl. Initially, all progress and guidance comes from the forelimbs. Eventually they realize they if they flail the hindlimbs around during locomotion, it significantly reduces friction and produces even more satisfying results from the forelimbs. It usually takes a while for them to get to the point where the hindlimbs can effectively push off and provide any real assistance to the locomotion being directed by the forelimbs, and longer still before they realize that with a lot of practice balancing, that extra push provided by the hindlimbs can often be sufficient for locomotion entirely on its own.
All of this typically comes with a fair bit of guidance and encouragement from the parents, and thats just us natural bipeds. Look at Bambi; even quadrupeds have a tough time persuading the hindlimbs to so much as cooperate.
The hindlimbs really arent as amazing as you think. They mostly popped out to keep the genitalia from dragging.
Thanks. <3 I really am trying to help.
No one at all, but again, the point is that the smaller markers mean you need a higher resolution camera to produce similar final results. Im not saying that no compensatory adjustments are possible; just that theyre needed due to the decision to use smaller, dimmer markers.
Sorry, the observers familiarity with Orion was assumed. The problem isnt with identifying the constellation once youve located the constituent stars, but rather in locating the constituent stars at all. The stars of Orion may be spread across the same 30º of sky as the blimp, but the blimp is a bright object which is 30º across, while Orion is in fact a collection of dim objects a mere 1º wide. Because the stars are so much smaller and dimmer, theres a much greater chance of a given star going undetected than there is with the large, bright blimp. When an unknown portion of our pattern of stars will be hidden from view, that further complicates confident identification.
If it worked like you think and the IR LEDs were always bright and clear with no false positives, just three markers arranged in a little triangle would be sufficient for providing them with a full position and orientation lock, just as the little triangle on the DS4 provides the same. As a matter of fact, it would look a lot like this:
If the markers were as easily spotted as you believe, then the three in the middle would be all that was required, but it actually works like Im trying to explain, and in any given frame, 0-3 of the markers that constitute that center triangle will actually be visible. While three markers give us all the information we need, if the camera is only able to detect one or two of them, thats decidedly less useful.
So the intermittent detectability inherent to small, dim markers is combated with redundancy; additional markers spaced over a larger area. Hence the four additional markers at the corners of the front face plate. Now even if 57% of our seven stars go undetected this frame, well still be able to draw a useful triangle out of whatever remains. If you remember that having more than 50% of a six-star constellation escape detection would be catastrophic, then you'll understand why we're using a seven-star cluster instead.
The need to space the markers over a larger area is less obvious on the face plate, where we have plenty of room to work with already, but it becomes more evident on the wands, where outriggers are required to achieve the needed marker spacing. Yes, the outriggers provides an occlusion-free view of the markers, but there are smaller packages which could achieve the same. HTCs donut wand-topper is an immediately apparent example. Perhaps a small pyramid or cone covered with an asymmetrical pattern of markers.
Now that Im thinking about it actively, it occurs to me this likely explains their camera placement recommendations as well. Because confident tracking requires a fairly large number of widely spaced markers, that means confidence will drop considerably when the face plate rotates out of view and youre left with only a few, closely spaced markers on the sides. By telling the user to face generally north with the cameras positioned to the northeast and northwest, that helps to ensure that at least one of the cameras always has a decent view of the face plate, even when youre looking away,
Thats too bad, but I suspect it will depend largely on the user in question. Looks exhausting either way. lol
It sounds like magnetic tracking is totally awesome at short ranges, and it falls over quickly after that. During that input panel I linked, the SixSense dude was suggesting hybrid tracking, where the difficult-to-occlude headset was tracked by conventional means and the often-self-occluding hands were tracked by a STEM base mounted in the headset itself.
Speaking of ideal tracking systems, Lighthouse is basically a laser-based GPS system, which is nice and all, but why not use radio beacons? Is it really that much harder to clock a radio wave, or is it just too dangerous to let us have such things?
If you could stop raging for just a moment, you may be able to see that I'm trying to explain that while motion not detected by the ears can cause sickness, that's not always the case. So if we look at the true causes of motion sickness a little more closely instead of simply dogmatically insisting that walking can be the only acceptable solution now and forever more, I bet we could come up with some solutions which are both comfortable and convenient, especially given the excellent input devices we have these days to communicate our intent more directly than we could in the past.
I've addressed this stuff as well, but I understand you're not particularly focused on the actual meat of the discussion at the moment.
----------
So I take it most of the room scale games use teleportation for abstraction, and that's why everyone assumes I'm opposed to teleportation, despite my having said several times that it's a solid abstraction and has my full support for use in VR? lol
That's good to hear. I was going to ask why it "wasn't getting much support" or whatever.
These words have nothing to do with your previous post. You said he has a fundamental misunderstanding of VR when he said "seen but not felt" motion is the #1 cause of nausea in VR.
What about his statement is a misunderstanding?
The side of the discussion coming from you has no meat because you've never tried any of this. It's all out the ass speculation and links to sony's youtube channel. Stop pretending that you know what you're talking about or that you're seriously interested in doing anything but convincing everyone that your (sony's) path is the correct one.
plagiarize
Banned
Tracking the motion controllers from the headset is a great idea, but I do still wonder about anything that uses a different tracking method to what the headset uses. Sony, HTC and Oculus haven't all hit upon the idea of tracking everything the same way by accident. It ensures that you get 1:1 tracking and also that everything is relatively tracked. When using my hydra it's obvious that this isn't happening.
Using one external tracking method is more of a matter of convenience than anything. People from valve have said that they envision the future of VR to have hybrid tracking systems involving cameras and lighthouse and other stuff depending on location and what kind of stuff needs to be tracked.
Just got my Vive today. Tracked controllers are awesome, but damn has it made me appreciate the build quality of the Rift that much more. Can't wait for the touch controllers to release.
Also for all that talk about FoV, I don't really notice a difference They both seem to have a binocular effect. Although I was sort of in a rush to check it out so maybe I didn't set it up right.
Also for all that talk about FoV, I don't really notice a difference They both seem to have a binocular effect. Although I was sort of in a rush to check it out so maybe I didn't set it up right.
serversurfer
Member
The same misunderstanding most people have upon hearing that statement; " then it follows that all motion not matched perfectly by the vestibular system must be avoided without exception." While many mistakenly believe the latter is implied by the former, this is not actually the case.
As I said, I listen to everyone, and yes that includes Sony and even MS. I listen to you too. <3 You should try listening sometime yourself; it's good stuff.
The wands would live in their own universe, but since that universe defines the headset as the origin, it should be trivial to translate position to the parent universe.
As for why nobody is doing it, I suspect it's the cost of marrying two systems on top of the fact that STEM isn't cheap on its own.
Croatoan
They/Them A-10 Warthog
Is anyone posting these walls of text actually a developer who is currently experimenting with VR locomotion?
I am a dev and in May i will be doing just that so it would be nice to know who is full of shit and who actually is worth listening to for my own education.
All of my research points to omni treadmills and redirected walking to be the only solutions to the locomotion problem if you want to stimulate human movement and remove all sickness.
Redirected walking will only be a thing at theme parks due to space requirements so that leaves us two options...
1)omni treadmills- this will always be the best solution for a multitude of reasons. Can even be used for fitness. We need treadmills that allow more range of movement though (crouching,jumping, ect).
2)traditional controls merged with room scale- Those that get sick will need to get use to it or be left behind. Basically small movements would be room scale but you would use a joystick for traversing terrain. Near lightfield displays will go a long way to limiting motion sickness with this method.
Ive developed a wasd teleportation method in UE4 that has the freedom of wasd movement with the anti sickness of teleportation.
I am a dev and in May i will be doing just that so it would be nice to know who is full of shit and who actually is worth listening to for my own education.
All of my research points to omni treadmills and redirected walking to be the only solutions to the locomotion problem if you want to stimulate human movement and remove all sickness.
Redirected walking will only be a thing at theme parks due to space requirements so that leaves us two options...
1)omni treadmills- this will always be the best solution for a multitude of reasons. Can even be used for fitness. We need treadmills that allow more range of movement though (crouching,jumping, ect).
2)traditional controls merged with room scale- Those that get sick will need to get use to it or be left behind. Basically small movements would be room scale but you would use a joystick for traversing terrain. Near lightfield displays will go a long way to limiting motion sickness with this method.
Ive developed a wasd teleportation method in UE4 that has the freedom of wasd movement with the anti sickness of teleportation.
Not me, but it comes up a few times in this AMA I linked earlier - https://www.reddit.com/r/Vive/comme...developers_for_the_htc_vive_ask_us/?limit=500
Besides being a really weird way of interpreting his original statement, I think you will find few to no people who think in absolutes like that, let alone ''most'. It is obviously a combination of factors. How the artificial motion is implemented and how the individual player responds to it are the most important ones.
Lucky's Tale is free with every rift and has artificial locomotion. They wouldn't have shipped it if it was a vomit fest, but there are some people who say it makes them queasy when the camera moves on its own.
It is absolutely true that artificial locomotion is the main, and really only, cause of nausea in a correctly functioning VR system. The only way to guarantee a sickness free experience is to use 1:1 movement.
cleveridea
Member
I loved this description of the omni treadmill
http://www.gizmag.com/virtuix-omni-vr-treadmill-review-hands-on/41438/
http://www.gizmag.com/virtuix-omni-vr-treadmill-review-hands-on/41438/
plagiarize
Banned
Looks like Valve are working on official Hydra support. Sweetness.
https://github.com/ValveSoftware/driver_hydra
I'm going to try and build this and see if I can get it working. Hopefully this means I can actually throw the throwing knives.
https://github.com/ValveSoftware/driver_hydra
I'm going to try and build this and see if I can get it working. Hopefully this means I can actually throw the throwing knives.
OSVR Review on Gamespew
Full Review: http://www.gamespew.com/2016/04/razer-hdk-possible-low-budget-vr-solution/
Full Review: http://www.gamespew.com/2016/04/razer-hdk-possible-low-budget-vr-solution/
serversurfer
Member
All of the sources I've personally referenced are, so I would encourage you to check them out.
It's usually best to examine the claims being made and decide that for yourself, I'd say.
Can I ask why you've omitted or ruled out the third possibility of new abstractions which don't cause discomfort?
The comment you're focused on was brought up in defense of the commonly held belief that physical locomotion is the only acceptable solution.
Yea, I'm aware of that, which is why the solutions I've been proposing are careful to avoid such things.
Regarding the first statement, "correctly functioning" seems like a strange qualifier, since I would argue that if your method for abstracting locomotion causes discomfort, it's not "functioning correctly," in much the same way that a blurry display would fail to qualify. So you're basically stating that a shitty solution is no solution at all. I don't think anyone would argue that, but I don't understand why so many seem to have given up the search for non-shitty solutions when it's barely begun.
Regarding the second, that's not really true either. For example, sustained forward movement is comfortable if the velocity is kept constant and all acceleration occurs within about a third of a second. Similarly, comfortable, sustained forward travel can be achieved by simply pointing your wand to indicate the desired direction of travel, and can be adjusted as you walk.
Now that I have that image in my head...
But I was referring to this:
Moving on...
It sounds like you were when you said this:
Moving on...
Again though, you're speaking as if you know for a fact that Oculus haven't compensated for this by providing ample power to the IR diodes and/or a higher-sensitivity IR camera.
Multiple laser bars sweep across the room in different directions, and each sensor detects the laser pulses at slightly different times. I don't see any similarities to GPS.
You're not a developer yourself though, so a lot of the "if this then that" stuff you pepper your posts with will continue to be disputed.
I've written a paper and will be 'publishing' it soon regarding VR locomotion.
Essentially it's an argument for RIPMotion (running in place motion) that I call CAOTS (Controller Assisted On the Spot) motion.
The short of it is... correctly implemented, it grants you all the advantages of an ODT system (fitness, vestibular noise, proprioceptive movement) and couples them with the ability to easily switch in and out of roomscale (and even combine them), without additional cost & space requirements of ODTs.
Also, will be working to get my specific ideas on implementation happening... even if I have to learn to code myself (not that hard... just requires a bit of time, but the resources are all there).
But RIPMotion (search for it) is already out there and partially demonstrates the idea. Protip if you use it: don't stick the controller in your pants, just hold it in your hands and use it as a pointer (the forward vector is the trigger pointing out).
plagiarize
Banned
Hydra support took another jump forwards.
https://github.com/betavr/steamvr_driver_hydra/wiki
Now the motion tracking works *much better*. You still can't throw the stick in The Lab or the knives in Budget Cuts however, but this gets much closer to proper support!
https://github.com/betavr/steamvr_driver_hydra/wiki
Now the motion tracking works *much better*. You still can't throw the stick in The Lab or the knives in Budget Cuts however, but this gets much closer to proper support!
The Vive was just released two weeks ago and we're already seeing official support of a third-party peripheral. This is fantastic.
I wonder how soon until we start seeing accessories with photosensors to work directly with Lighthouse.
Hydras is one of many controllers that OSVR will support
https://www.reddit.com/r/OSVR/comments/3wbt9p/hydras_seem_to_work_fine_with_osvr/
https://www.reddit.com/r/OSVR/comments/3wbt9p/hydras_seem_to_work_fine_with_osvr/
plagiarize
Banned
Well, I shouldn't overstate exactly what has happened. Valve had Hydra support in Steam VR at one point, in beta I guess. They turned over this source code to the community since they aren't going to be further developing it. That's still really cool.
There's a large issue right now with games that use the accelerometer data (there isn't any such data) for throwing objects, but the tracking is really good (it's as good as the Hydra gets, which isn't nearly as good as the Vive controllers or as good as touch looks). Hopefully someone can figure out how to generate the appropriate data... because this works really well for a lot of the Steam VR stuff I've tried out even without it.
It's mainly something I'm going to use to tide myself over until the touch release... but it sure does help being able to see some of these games first hand without having to rely on impressions. Even without being able to throw knives, Budget Cuts is something special. The modular controls and teleporter work brilliantly. The inventory system is great. It's fantastic stuff.
Hydra support isn't 1:1, but it's good enough that you can play a lot of stuff already... and like I said, it's going to help tide me over until the touch launch. Not that I can afford to really buy anything right now, but I'd much more readily buy Hover Junkers and Job Simulator now, for example.
It's great just being able to think about *the games*.
Your technique of microquoting makes it impossible to have the full context of what you're responding to. Even going back and looking at his original post I don't see anything that hints at even a minor misunderstanding of VR. If anyone is misunderstanding VR it's you. Anyone who has tried VR knows that what you think works in VR and what works when you actually try it in VR are totally different. Every developer says paper designs are useless and the only thing worth doing is prototyping. You can hypothesize all you want until october but it's not going to do anyone a bit of good.
By correctly functioning I mean the headset works and the game is running at 90fps. If your definition of correctly functioning requires no motion sickness then why are you interested in PSVR? Most of its games aren't going to use the stupid ratcheting motion that one sony guy seems to really like. Nobody is giving up on anything. We are saying YOU should give up your crusade until you've actually used a VR headset. You are not moving VR forward by pointing out obvious limitations in existing methods and championing someone else's idea that you haven't tried and isn't even being used in any games as far as I know.
I think anyone posting here knows what is meant by artificial locomotion causing sickness. We all know that acceleration is the main cause. "Seen but not felt" is the simplest way to describe it. There is nothing to feel when you are moving at a constant velocity so there is nothing for your brain to be confused about. That is why blink turns exist, that is why there is no-acceleration movement in joystick games.
Zalusithix
Member
AFAIK no game directly accesses any of the IMUs. OpenVR abstracts hardware access away from the devs. Everything I've seen (in my cursory looking into things) falls back to the GetControllerStateWithPose or GetDeviceToAbsoluteTrackingPose functions which return a TrackedDevicePose struct. That contains the device position, velocity and angular velocity. How these things are calculated is a black box, but none of them would need accelerometer data.
plagiarize
Banned
Well, it's weird. I can throw stuff in Surgery Simulator, but cannot in The Lab or Budget Cuts. Both on the OSVR workaround for the Hydras that first came out, and on the officially released driver work from Valve... so I'm sure both of those games are looking at the same data to calculate the velocity of a thrown object.
Guardian Monkey
Member
plagiarize
Banned
I definitely need to get my sensor up higher for motion controller based games. In Brookhaven, when I aim at a zombie over in the direction of my sensor, the sensor can't see the headset, and in cloudlands mini golf, if I am looking down (as you are want to do in such things) with my back to the sensor it can't see the headset either.
The headset tracking with one camera has been pretty solid aside from those scenarios though. Tracking points on the back of the headset are probably a good idea for everyone. I know some people said it wasn't fair to knock the Vive for being more complicated to set up, because you could just as easily set that up with one sensor for seated play... but then you wouldn't be getting as good tracking as you do on the Rift.
Not saying that to knock the Vive, just saying that tracking points on the back of the head are a good idea all around.
Motion controllers with cables are definitely not ideal. I've done a decent job of not getting tangled so far, but I failed at that on Cloudlands and got the cables all twisted up.
The headset tracking with one camera has been pretty solid aside from those scenarios though. Tracking points on the back of the headset are probably a good idea for everyone. I know some people said it wasn't fair to knock the Vive for being more complicated to set up, because you could just as easily set that up with one sensor for seated play... but then you wouldn't be getting as good tracking as you do on the Rift.
Not saying that to knock the Vive, just saying that tracking points on the back of the head are a good idea all around.
Motion controllers with cables are definitely not ideal. I've done a decent job of not getting tangled so far, but I failed at that on Cloudlands and got the cables all twisted up.
Zalusithix
Member
Yeah, I'm not sure what's causing that. My guess would be some sort of issue between the Hydra/Sixense driver and OpenVR where the incoming data just isn't precise/consistent enough for OpenVR's calculations. Why it would affect only certain games could come down to whatever algorithm they're working with to compute the throw. I know Budget Cuts for sure isn't just taking a raw single data point. Pure conjecture though given I'll never see the code of any of the programs, and I haven't even looked at Sixense SDK yet which is a part of Hydra driver.
Yeah, he botched the quote origin, but it's pretty obvious who it is supposed to be.
Edit:
Well for seated play, one lighthouse would be enough for most situations. If you're going to be facing 360 degrees away from the forward position, you're probably no longer in seated play. If you're not in seated play, you're likely using tracked controllers and you want both lighthouses. If you have both lighthouses, you don't need the rear tracking 99.9% of the time. The .1% of the time would be a rather awkward situation with you crawling with your head looking between your legs while you're perfectly in line with the lighthouses.
XiaNaphryz
LATIN, MATRIPEDICABUS, DO YOU SPEAK IT
What's the issue? Oculus funded the studio to make them an exclusive title. Almost sounds like port whining/platform warrior bullshit to me. Can you elaborate?
I already did in a post following that one.
GuitarAtomik
Member
This Razer Hydra stuff and the new Oculus exclusive from Insomniac got me thinking. What are the chances Vive will get Touch support? They'll be sold separately and come with a camera so I imagine you'd have everything you need (unless it requires two cameras). Having another set of motion controllers with 90% of the same functionality doesn't sound awesome but I'd imagine most of the Oculus exclusives will take advantage of that 10% which will make it a pain even if you can get the games to work on Vive. Though the out of the box occlusion issue would suck, it would be nice to have the option with Vive games even.
XiaNaphryz
LATIN, MATRIPEDICABUS, DO YOU SPEAK IT
It's been happening outside the console space forever. Games that only work on certain versions of Windows or higher because of a DirectX feature, games that used to only have 3D accelerator support for specific proprietary APIs, etc. etc. Complaining about it contributes nothing regardless of whether it's in the console space or elsewhere and threads end up flooded with those kinds of posts, hence the rules against it.
I'd be surprised if it didn't happen at this rate. edit: apparently more complicated than I thought!
The touch receivers are in the Rift headset. And you'd need two cameras if you want to turn around.
plagiarize
Banned
Eh, I had a DK2 and I'd regularly lose tracking when playing racing games or space sims where you'll often want to look directly behind you. For a few extra sensors, I think it'd be a worthy addition to the next Vive hardware refresh.
Like how the Vive controllers connect to the headset, the Touch controllers connect with the Rift for wireless communication for buttons and the sensors inside the controller. It would require some craziness to pull off.
Technical limitations are not exclusivity deals though. And this is a new market with its own challenges already that doesn't need to be segmented even more because of these deals. Discussions on this topic are already present in this thread which covers VR in general. Notice that I didn't go in the specific thread about the game and complained about it. I posted it here, because it's one of the topics that has been discussed here, as I said. Maybe the language was not the best, so I took care of that now. But I didn't think that exclusivity deals are excluded from discussion altogether on GAF, since we had even threads dedicated to some of them.
Zalusithix
Member
I supposed if you really whip your head around, it's possible. Not that it really matters since the whole "single lighthouse on the desk" was more of a hypothetical thought exercise than any realistic solution to begin with. Besides, I'd like to see the logic boards and connectors moved to the back of the head anyhow in the next generation for weight distribution purposes, so a few extra sensors on that rigid section couldn't hurt anything.
XiaNaphryz
LATIN, MATRIPEDICABUS, DO YOU SPEAK IT
Discussion is fine, yes. Random one liner "more exclusivity bullshit" reeks of drive-by shitposting. Thanks for going forward with your edit.
GuitarAtomik
Member
Ah didn't know that. Probably out of the question then.
Croatoan
They/Them A-10 Warthog
Custom abstractions (like walking by moving your hands or whatever) are great for games based around them. But if you game is based around a dude (or dudette) walking around an open world then you need to be...uh...walking.
Think of a VR Skyrim for example. You need to be able to freely walk around a giant world in a way that is immersive (doesn't break immersion, isn't noticeable), immediate (happens instantly), dynamic (can move any direction), smooth (no pauses, stutters, or breaks in movement), and additive (means that the upperbody can do things (aim a gun, swing a sword, ect) separate from the lowerbody (locomotion).
Using those five criteria there are only two locomotion methods available for this gametype...
Absolute Best: Treadmill/redirected
Gets the job done: traditional joystick + roomscale
Using the arm pulling method doesn't work as you break the additive criteria
Using the rotating world (ratchet?) method doesn't work because you have to pause the game to rotate the world with the player which breaks both immersion and smooth.
Now if your game was based around being a wolf or four legged creature you could use the arm pulling method.
Or if your game was based around being spiderman you can shoot webs and swing.
The only games I can see using the ratchet (rotating world) method are slow adventure games with no combat. And even then they will suffer from the NASCAR effect.
I guess what I am saying is that locomotion can be anything, but it HAS to make sense with your game. Shoehorning something in that isn't natural wont work.
Zalusithix
Member
Actually in regards to the controllers connecting to the Vive headset, are we actually certain of that? Has anybody shielded the breakout box to make sure they're not connecting to that instead? Given the prototypes connected to a USB receiver on the PC, it's entirely possible the headset has no wireless chip in it. We need a teardown of the Vive already damnit.
They connect to the break out box. You can connect vive controllers without the headset being plugged in
I read a review pretty much saying the omni direcitonal treadmill available right now is pretty bad anyway since you cant walk around naturally like you would in IR and instead have to slide your feet.
Then you have to factor in the insane price for those things as well, which makes me think that an omni directional treadmill will never be a truly viable final solution to locomotion in VR.
Zalusithix
Member
I mused about it in the Vive thread, but for an omnidirectional treadmill that doesn't cost tens of thousands of dollars and take up dedicated floor space, you could theoretically invert the concept. With powered shoes (think a smaller, lighter, tamer version of this) that are positionally tracked, you could negate forward movement in a traditional roomscale environment. It would allow normal foot motions and allow a certain amount of inertia on starting and stopping to sell the illusion more.
It's not foolproof by any stretch of the imagination, but it seems like a viable solution within its limitations if you can get the shoes light enough.
It's not foolproof by any stretch of the imagination, but it seems like a viable solution within its limitations if you can get the shoes light enough.