Hands-on with HTC Vive and SteamVR

steamvr htc vive headset and controllers

Valve has partnered with HTC to bring their tech to market in the form of the Vive VR headset. Just so we’re clear, the Vive is the headset. There’s also the SteamVR controllers and ‘Lighthouse’ tracking stations. Together, all of these components form SteamVR.

The headset itself has a wide field of view that’s on par with Crescent Bay and Morpheus (for those of you who have the DK2 for reference, all three of the aforementioned have a noticeably larger field of view than the DK2, by about 10%).

The unit uses Fresnel lenses and dual displays that total 2,160 x 1,200 resolution (1,200 x 1,080 per eye), compared to Crescent Bay’s display, which I believe to be ~2,560 x 1,440x (1,280 x 1,440 per eye), though Oculus has yet to confirm. Crescent Bay also uses Fresnel lenses.

From my time in the Vive, I believe the sub-pixel structure is PenTile, just like Crescent Bay. From my time in each headset, I haven’t noticed an obvious difference in resolution, but of course every bit helps. For those curious about the screen-door-effect (grid like space between the pixels), it is significantly improved over the DK2. You can still see a fine grid if you look carefully, but your brain effectively stops noticing when you are concentrating on the world in front of you instead of the pixels. I didn’t notice a discernible difference in screen-door-effect between the two, though I wasn’t able to compare directly side-by-side.

The headtracking on the Vive, afforded by a combination of IMUs and the Lighthouse tracking system, feels every bit as accurate as Crescent Bay, but has the added benefit of tracking in a much larger space. During my time with Vive, which totalled about 45 minutes, I never felt a jump in headtracking (which could be caused by occlusion or pose-estimation errors).

_DSC0058
The Lighthouse base station that will be included as part of the initial SteamVR development kit that’s due this Spring.

This was with two Lighthouse base stations, which seems to be all that’s needed for a very reliable experience, though Valve says you could easily add additional units for more coverage (which might be a good idea if you want multiple people or more objects to be tracked in the space). We’ll be digging into Lighthouse in detail soon, but it should be noted that the base stations only need to be plugged into a power outlet—no USB cables running across the floor to a PC, or even a WiFi connection. This fact makes it much more practical to populate a space with the units.

SEE ALSO
Microsoft Reportedly Orders Samsung Micro OLEDs to Restart XR Hardware Ambitions

Constant Presence

Presence is the term used by the VR research community to describe a state of deep immersion—there is no longer the need for suspension of disbelief, on a subconscious level, your brain believes that the world around you is there.

See Also: Oculus Shares 5 Key Ingredients for Presence in Virtual Reality

Presence is something that can be felt in fleeting moments in other headsets. Incredibly low head tracking latency (and 6 degrees of freedom of movement) is vital to instilling the feeling of a solid world around you, which is essential to Presence. In headsets to date, like the Oculus Rift DK2, Morpheus, and Crescent Bay, I have felt moments of Presence, but they often vanish just as quickly as they come.

SteamVR creates a stage of constant Presence, and it maintains it for long periods of time, thanks, I believe, to the incredibly solid tracking on both head and hands, and also, crucially, to the huge space that’s navigable simply by walking. Navigation by other means (like a keyboard or mouse) causes the world to move around you without the cues of your body actually moving through space, and this quickly breaks the spell of Presence. The world around you feels all the more real as you move through it, just like you would in a real space, and your brain maps it as such.

Confidence in the space around you is key. If you don’t trust the system to prevent you from walking into a wall, you’ll always be worried about it—a constant reminder that you aren’t actually in the space that your eyes say you are.

SEE ALSO
Quest 3S Review – Value That Can't Be Beat, With the Same Rough Edges as Its Siblings

SteamVR builds this confidence quickly thanks to a combination of the solidity of their Lighthouse tracking system and Chaperone, which is simply what they call the system that makes a glowing grid pattern appear within the virtual world as you approach a real wall. After testing the boundaries of the space, and becoming confident that Chaperone is watching, you are free to believe that the virtual space you are walking around is real. I wasn’t in SteamVR for 5 minutes before I felt comfortable moving about without needing to distract myself with the location of real-world walls.

The only thing that kept me from losing myself completely in the virtual space was the cable that the headset is tethered to. Occasionally I could feel it on my leg or felt that I needed to step over it to avoid being tangled. The weight of the cables dangling from the controllers also felt a bit off, but I was told that the development kit controllers would be wireless. Unfortunately, I was also told that the headset would remain tethered. If I’m going to dedicate a room in my house to the system, I’ll definitely be investigating a ceiling-based cable management system, though that’s not to say that a wireless system wouldn’t be ideal.

Valve’s SteamVR Controller

_DSC0055

Valve has taken the foundation of their Steam controller and adapted it for virtual reality. The units are lightweight and feature two buttons, a trigger, and a circular trackpad. The two buttons are on the left and right of the controller and can be activated by squeezing your hand around the controller. The trackpads have a real press to them (like a button), and also some very interesting haptics underneath which can create convincing clicks as you slide your finger about. The sensation of these clicks feels very localized, acting just below your finger no matter where it is on the trackpad.

Inside of the virtual world, Valve has demonstrated the ability to divide the trackpads up into virtual buttons. For instance, dividing the circle into 4 even parts, and then causing a haptic click as your finger moves from one region to the next. One demonstration in SteamVR showed the circle divided into many colored sections, with a white line extending from the center—like the hand of watch—to show your selection. Pressing down on the trackpad caused a balloon of a corresponding color to blow up from the top of the controller and float away. Changing the ‘button’ layout dynamically in this way means that the controllers are very flexible, and the ability to show the current selection eliminates the issue of players not being able to see real buttons on a controller.

SEE ALSO
Catch VR Games Showcase's 'XR Indies & Friends' Livestream @11AM ET Today Ahead of Meta Connect

Developers can represent the controller however they want in-game. I saw both the virtual replica of the controller (shown with virtual buttons, as above), and I’ve also seen the controllers represented as oversized cartoon hands. Developers are likely to come up with a bunch of interesting ways to show the in-game controllers; I’m already thinking of a realistic looking hand with a watch on the wrist to show your various buttons and selections.

In my experience, the controllers have been highly accurate and I’ve seen no erratic jumps (though I have heard reports of some from other users). Tied to the Lighthouse tracking system, which Valve says is accurate to less than a millimeter, they operate with absolute positional tracking. Touch the tops of the controllers together at the start of your session and note how their virtual counterparts touch together. At the end of your session, the controllers will touch together in exactly the same way with no drift. Absolute positional input is essential to a frustration-free VR experience and the SteamVR controllers are highly functional—hands down the best VR input system I’ve seen when considering precision, accuracy, and range.


There’s so much more to be said about SteamVR and the HTC Vive. Coming up, we have a technical analysis of the Lighthouse tracking system, as well as a breakdown of the various demos seen in Valve SteamVR demos.

1
2

This article may contain affiliate links. If you click an affiliate link and buy a product we may receive a small commission which helps support the publication. More information.


Ben is the world's most senior professional analyst solely dedicated to the XR industry, having founded Road to VR in 2011—a year before the Oculus Kickstarter sparked a resurgence that led to the modern XR landscape. He has authored more than 3,000 articles chronicling the evolution of the XR industry over more than a decade. With that unique perspective, Ben has been consistently recognized as one of the most influential voices in XR, giving keynotes and joining panel and podcast discussions at key industry events. He is a self-described "journalist and analyst, not evangelist."
  • seanlumly

    I would love to see optional external tracking for mobile VR. If the tracking is good enough, I would be that something like GearVR would be an incredibly compelling platform, with the benefit of being 100% un-tethered. Even a few built-in reflectors or LEDs on the GearVR and the ability to optionally use the Oculus camera would be a giant plus.

    Mobile VR still lacks the “90Hz” refresh required for presence, but I doubt this will be the case for long — Samsung is very aggressive about pushing ever higher specs, and mobiles can certainly handle this refresh with simple content and/or reprojection (so-called time-warping). Even Mediatek’s latest mobile SoC features a 120Hz display controller! Yes, fast VR-grade refresh is coming to mobile!

    I also have little doubt that GearVR will act as a tethered pass-through VR headmount in the future, leaving the heavy lifting to the CPU to render low-latency frames. Such a development would be another compelling reason to have optional external tracking.

    Still, I’m glad to see that Valve has been aggressively working on the problem of tracking and that they have largely solved it. I’m sure this isn’t the last we will see of interesting tracking solutions either. VR is just getting started!

  • Alkapwn

    That was a fantastic review Ben. I’m going to have to think of new questions for tomorrow’s AMA now. I also think this is the first time I’m hearing that Crescent Bay was using fresnal lenses. Do you think that they work better than the normal lenses?

    Also, having now tried the latest STEM system and Valve’s new motion controllers, which do you think is more accurate? And do you think that in a demo like the lightsaber one that Sixense is showing that there could be instances where there is occlusion of the controllers when swinging them around (for Valve’s)?

    • Nick vB

      @Ben – “I’ll definitely be investigating a ceiling-based cable management system…”

      @Alkapwn – Well I finally got around to building that Slip-Ring prototype last year, I’ve necro-posted about it in your old MTBS3D thread: http://www.mtbs3d.com/phpbb/viewtopic.php?f=140&t=15933

      Just thought you might be interested…

  • mellott124

    Fresnel lenses. Yikes! How’s the eyestrain?

  • Sean Concannon OculusOptician

    Fresnel lenses are a poor choice for use in VR. Palmer said last year “fresnels cannot come close to matching the focal length/magnification of other optics tech”. http://www.reddit.com/r/oculus/comments/1jon7c/why_not_fresnel_lenses_for_the_rift/

    The HTC ReVive definitely uses fresnels based on the pictures I have seen, so I’m imagining observed artifacts, not just from the screen door but the actual lenses themselves. Oculus on the other hand is using proprietary lenses for Crescent Bay which reduce artifacting and the screen door effect significantly.

    • eidolon

      Oculus are strongly rumoured to be using fresnels in the Crescent Bay, despite Palmer’s earlier comments.

  • Curtrock

    Yesterday, I got out the old tape measure and pre-visualized what size a 15×15 room would look like. I have an unfinished bsmt, and I’ve been deciding how I would partition off the space. If the Valve HMD experience is as compelling as it appears to be, I will indeed dedicate a 15×15 room to VR. This will be my own personal “Holo-Deck”. It would be good if it could be scaleable, like optional 10×10, for those who have less available space. Great point, Ben….most families have already dedicated an entire room to their television. It’s already ingrained in our culture.

    • Dolfish

      Using an empty room is a good start to getting that authentic VR experience. For now this is the best and only way to achieve absolute presence, but it is also, unfortunately, limited to just moving through space on a 1:1 level and nothing else. Sitting on a virtual chair, leaning on a table, stepping on a ledge, all of these things remain impossible right now, and have yet to be addressed. For now the focus should be on pushing the resolution of these devices, refining head/positional tracking, wireless and portability, and the UX. Once these things have been worked on to death, then we can start thinking about how to tackle the issues aforementioned. Whatever the solution may be, you can bet on one thing: the solution will most likely involve a fusion of the neurological and computer sciences to further push the boundaries between the real and the virtual. Unless you have no problem purchasing a large mechanical platform that somehow morphs it’s structure to match what you’re seeing (much like MIT’s inFORM/Recompose project) than it obviously seems more likely the average consumer will prefer something a little more discreet.

    • Dolfish

      Using an empty room is a good start to getting that authentic VR experience. For now this is the best and only way to achieve absolute presence, but it is also, unfortunately, limited to just moving through space on a 1:1 level and nothing else. Sitting on a virtual chair, leaning on a table, stepping on a ledge, all of these things remain impossible right now, and have yet to be addressed. For now the focus should be on pushing the resolution of these devices, refining head/positional tracking, wireless and portability, and the UX. Once these things have been worked on to death, then we can start thinking about how to tackle the issues aforementioned. Whatever the solution may be, you can bet on one thing: the solution will most likely involve a fusion of the neurological and computer sciences to further push the boundaries between the real and the virtual. Unless you have no problem purchasing a large mechanical platform that somehow morphs it’s structure to match what you’re seeing (much like MIT’s inFORM/Recompose project) then it obviously seems more likely the average consumer will prefer something a little more discreet.

  • czero

    Ben, did You see fresnel lenses in Crescent Bay? As far As I saw pictures of Crescent Bay few months ago, I thought I saw regular lenses. In HTC Vive pictures I see definitely fresnel lenses. Did they change lenses in Crescent Bay?

    • Ben Lang

      I’m quite certain I made a mental note of interest when I saw Fresnel lenses with CB at GDC, but there is a chance I was seeing something else. It’s hard to tell because if the Fresnel groves are fine enough, they can be tough to see (and ideally they will be invisible when positioned correctly). I recently asked Palmer about his current stance on Fresnels, noting that it appeared that both Vive and CB used them, and he declined to comment, though I believe he would have corrected me on CB’s lenses if I was mistaken.

  • Jacob Pederson

    I had the idea of hanging some room dividing curtains up to prevent wall collisions, but looks like this won’t be necessary. With the amount of hype surrounding HTC Vive, I don’t think Oculus can afford to launch CV1 without controllers or hand tracking now. They already have Gear VR to target the casual market. Maybe they can team up with Sixense Stem?