Sign in to follow this  

VR controllers to play Steam VR games on the smartphone

Recommended Posts

The biggest issue is that they only have one base station, so you'll have body occlusion issues with the motion controllers and the base station.

Yeah, no matter how good their product actually is, they've immediately gone into the "bullshit, do not trust" pile for referring to their capability as being "room scale" while only having a single base station. If you turn around and lean forwards, it will lose headset tracking, or if you put your body between the controllers and the base station, it will lose controller tracking.
That is not room scale VR; that's front-facing VR. There's nothing wrong with front-facing VR (that's Oculus Rift out of the box - until you purchase extra sensors - and it's great), but don't go lying about your capabilities to the consumer or try to muddy the waters about what these phrases mean.

 

I'd also be curious what the added visual latency is of this Riftcat->VRidge wireless video transfer actually is. My guess is at least an additional 60Hz frame of buffering, which can be the difference between giving people motion sickness or not.

 

[edit] and they mention that their tracking data is only 50Hz... which could be a problem when VR demands 60 to 120Hz rendering (90Hz is standard for SteamVR games). Also, can phones even display at 90Hz, or is mobile VR only 60Hz?

Edited by Hodgman

Share this post


Link to post
Share on other sites

Yeah, I was looking at the specs too and was a bit concerned. If the response latency is 20ms, but VR needs 10ms, then you're going to have misses on half your frames. It probably doesn't have much of an affect on the visual framerate, but I'm curious to see whether there would be any noticeable tracking performance problems within VR. Again, this is mobile VR though, and I don't have much experience with mobile VR other than disappointment and dismissal of the viability of the mobile platform for VR.

This product gives me the sense that it's an "As Seen on TV" knock off for the budget conscious consumer. That product category never goes well. It looks like they did a good job ripping off the vive grips. But... aside from that, this product looks like it isn't going to work. By that, i mean that the tech probably works, but as a product, it will fail in the market. My sense is that the beauty of mobile VR is that it's mobile. If you have to carry around a motion controller kit and set it up every time you want to use mobile VR, you're going to put limits on the mobility of it. Then you've got the price point of $89. That's a problem. It signals to me that the hardware was aiming for a rock bottom price and they cut corners to hit that price point (hence one base station). I'm almost certain that this product is being produced in a factory somewhere in China, so the actual per item manufacturing cost is probably about $15 per unit. Then you've got the freight shipping costs, which can be anywhere from $4k to $10k per shipping container, so the price point has to factor that in too. And then you gotta factor in the sales commission and retail cuts, so in the end that leaves a somewhat small profit margin. Why couldn't they just include a second base station and bump up the cost by $20? Then you could have non-occluded room scale VR, meet the marketing claims, and also keep customers happy because their VR experience is uninterrupted by tracking problems.

The other interesting claim is that they have 2mm tracking accuracy. With one base station. Out to 13 feet. I don't believe it. I need to see it. I'd go out to 13 feet and move the motion controller 2mm and see if it registers the movement. I doubt it would, I think the tracking accuracy drops significantly because you start running into tracking resolution problems the further you get from a base station. How much R&D did they spend on the base station resolution? With the $89 price point, I don't think it's a lot.

Overall, I get the feeling that this is "me too!" hardware which was rushed out the door. It doesn't do anything exciting to push anything innovative in the motion controller hardware category. If anything, it's inferior to what's currently available on the PC VR market. And, the trend in the PC VR motion controllers is to eliminate the motion controller as much as possible and increase immersion. That may be a few generations away in terms of hardware, but that is the direction its going. Well, at least mobile VR now has an option for tracked motion controllers, which is better than nothing at all...

Share this post


Link to post
Share on other sites

The biggest issue is that they only have one base station, so you'll have body occlusion issues with the motion controllers and the base station.

I watched their demo video here.  I will go to CES this year and may demo it myself and report it back lol

Share this post


Link to post
Share on other sites

Merry Christmas everyone! My name is Lisa Zhao and I'm the co-founder of LYRobotix, the company behind NOLO. I see that you guys have some questions and I would love to clear things up for you. 

 

[b]@[member='Robin90'][/b],  Thanks so much for initiating the conversation! You are very welcome to stop by our booth and try our live demo when you come to CES. If you want to make an appointment in advance to cut some waiting time, please email nolo@lyrobotix.com :)

 

[b]@[member='slayemin'][/b], [b]@[member='Hodgman'][/b], your feedback is very valuable to us and much appreciated. 

 

Firstly, what we provide is front facing VR based on your definition. From what we understood, "room-scale" doesn't necessarily mean 360 degrees. Since our system provides tracking for a bigger range compared to the PS VR and Oculus without the third camera, users could walk around more while playing. However, we understand that the word "room-scale" could be misleading, so point taken. We will be more specific when we describe NOLO in the future. 

 

Secondly, Mobile VR is more of a virtual reality experience using mobile phones. In my opinion, no matter how much mobility Mobile VR provides, VR in general is used at a set scenario instead of on the go. Since it's more of a entertainment system and less of a tool like a mobile phone, which AR probably is more suitable for. We believe that Mobile VR is very interesting with the potential to go mainstream due to its affordability and ease-of-use, however, it lacks the natural interactive system right now. NOLO is here to fill that blank without interfering with the features of affordability and ease-of-use. Therefore, we designed our product to only have one base station, so that users only need to put the base station on the table or a bookshelf and could start playing right away. On the other side, if users overwhelmingly demand two base stations and the full 360 degree at a higher cost, we would be happy to come up with a different edition in the near future. 

 

Thirdly, as for the latency problem. Our system provides 60Hz of positioning frequency. Mobile VR display is currently at 60Hz.  The visual latency of Riftcat->VRidge wireless video transfer varies based on internet connection, and we'll do more testing and come up with a range soon. We admit that our latency isn't as low as the PC VR systems like Vive or Oculus, but we do provide a decent experience for starters. Seeing is believing, we'll post more demo videos as updates and you all are very welcome to stop by our booth at CES :)

 

Lastly, for the price point, we've already gone through the trial production period and have a clear understanding of our BOM. Our PolarTraq technology inherently enables us to make a system that's more affordable than the ones existing on market, and we've calculated carefully before the pricing strategy is announced. As for the claim that "more affordable price = small R&D investment," We have invested a lot in R&D to have developed tracking technology, optimized the user experience and designing it with affordability in mind. 

 

Hope this clear things up a bit. Please keep the feedback and comments coming, as this is the best way for us to understand what you need and provide the best product for users.

Share this post


Link to post
Share on other sites

Lisa,

Thanks for taking the time to address some of our concerns. Much appreciated. I'm a VR developer (old game trailer), but not working in mobile VR. When I decided to start developing a VR game two years ago, I concluded that proper VR absolutely needs to integrate hands as seamlessly as possible. It's the most familiar way people interact with objects in reality, so if you can bring hands into virtual reality and let people interact with objects using their own hands, it would massively scale the sense of immersion and presence. I immediately ordered a Leap Motion device with my DK2 and went to work integrating both of them. It worked, but there were some serious field of view and performance limitations which made the solution difficult to design around. A few months later, Vive is announced with the motion controllers, so that was the moment that the leap motion became a stop gap solution until I could build support for motion controllers directly. The dream is to still have a players hands as the primary interaction system for VR.

 

So, the Oculus Touch became available to consumers on Dec 6th, 2016. The beautiful thing about the touch is that it has capacitive touch on several of its buttons, so you can detect when a finger is touching a button. Unfortunately, this only works for the pointer finger and thumb, so this severely limits the type and number of finger gestures which can be used. I think this was a missed opportunity on the part of Oculus.

Valve is currently working on their next prototype for the vive motion controllers. I got to see it myself at Steam Dev Days and ask a few questions to the hardware developer. The most exciting part of their controller is that they have capacitive touch for most of the fingers (maybe all of them?). This is a step closer to the per-finger VR interaction which would be the holy grail for VR.

So, the place where Leap Motion fails is that their tracking volume for the hands is too small. They recently announced that they're coming out with a device which features 180 degrees of tracking, but that is still insufficient. They are the definitive leaders for optical per-finger tracking, so they have the best chance / tech for determining hand positions. However, the 180 degree hand tracking is not good enough. Imagine you're holding a baseball in VR and you want to throw it. You pull your hand behind your head and then throw your hand forward while releasing your fist. As soon as you pull your hand behind your head as is natural in baseball, the hand is no longer tracked. And now we don't know whether the player opened their fist or closed their fist until the hand re-enters the tracking volume. What happens if the hand which re-entered the tracking volume is open? Then it means that at some point between when the hand exited the volume and re-entered the volume, the player opened their hand and we were supposed to release the held object. Since we couldn't track that event, the hand didn't properly release the held object and we break VR immersion. However, leap motion is worth paying very close attention to because their tech and hardware is probably some of the most advanced stuff for tracking analog finger movements on a per bone basis with sub millimeter accuracy. I think the future of VR hand tracking will need to integrate the Leap Motion tech.

Back to the Valve prototype: Their hardware is currently an 80% solution. But, it's a prototype so it's not the final product. The beautiful thing about their hardware is that it uses a hardware attachment which is strapped to the hand and the hardware has the tracking sensors which can read the laser light pulses from the vive base stations, giving the hands 360 degrees of tracking and handling obstruction fairly well. Where they fall short is with the per-finger tracking. Since they only support capacitive touch, you only get a binary touching / not touching response, similar to what you'd get from the oculus touch currently. But, if they support all five fingers, then its a little bit more robust than the Touch motion controller which only supports two. However, it does limit the number and types of finger interactions.

The ideal motion controller would be a combination of the current vive prototype, with capacitive touch on each finger and 360 degree tracking, but would also have the leap motion hardware tech embedded into the HMD on three faces: left face, front face, right face. So, you'd be able to get per-finger optical tracking when your hands are fully extended out on the left and right (as well as elbow positioning!), and when the hands move out of the optical tracking frustums, you'd still have the capacitive touch and hand tracking as a fall back solution. In order to save on bandwidth and CPU, you could toggle the active state of the three face cameras depending on whether the IR camera could actually track the hands from that camera. And since you still have a hand grip attached to your hand, you can also receive button input for some game specific inputs (such as toggling artificial locomotion). In my mind, this would be "perfect" for VR and would really advance the industry. Lastly, a final enhancement could be to place a bottom facing camera to look downwards at the players feet and knees and try to assess their positions. If the players feet can be observed or derived at all times, then the locomotion problem for VR could effectively be solved: Just walk in place to move. If we know the wrist, elbow, knee and foot positions, we can do some amazing IK work for full body avatars to create one of the most amazing, immersive VR experiences possible. If this can be done, then "motion controllers" can be devolved into just tracked wrist bracelets for the non-optical tracking fallback. I know a lot of contemporary VR game developers aren't going to like it when they don't have buttons on their motion controllers anymore, but It's going to take a bit of a design paradigm shift to adapt to the better, more immersive form of user input collection. A really powerful enhancement could be to use a form of foveated camera capture for the tracked hand positions to eliminate false positives, interference and reduce CPU costs -- you already know where the hands are supposed to be because the wrists are tracked, so the camera only needs to process the optical data around that region instead of capturing everything and trying to find the hands and their orientation.

I don't know which hardware manufacturer would ultimately end up coming up with this ideal VR input solution, but I won't hesitate to pitch it to anyone who will listen.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

Sign in to follow this