Support NeoGAF

[Zalusithix]

SteamVR (also known as OpenVR) is the primary API by which VR games interface with VR headsets on Steam. While a diverse set of hardware can achieve interoperability with SteamVR, this thread is for the hardware that was built from the ground up for it using the Lighthouse tracking technology pioneered by Valve. It also covers the setup and software that directly interfaces with and alters SteamVR such as overlays and compatibility layers.

For general VR game / software discussion, please use the PCVR Game Discussion Thread. Only software that alters the base functionality of SteamVR is to be discussed in this thread.

The HTC Vive was the first official SteamVR headset released in April of 2016, and hardly needs an introduction at this point.

• Screen Type: OLED (Pentile)
• Resolution: 1080x1200 per eye
• Refresh rate: 90Hz
• Field of view: 110 degrees
• Lens type: Fresnel
• Physical IPD adjustment between 60 and 74mm
• Eye relief adjustment
• Connectivity at PC: 1x USB 3.0, 1x HDMI or DisplayPort
• Headset connectivity at link box: 1x USB, 1x HDMI, 1x Power
• Audio: 3.5mm headphone jack
• Built in microphone
• Front facing camera
• Tracking generation: 1.0

LG's headset is currently unreleased and scheduled to reach consumers sometime in 2017. It will usher in the first true competition within the SteamVR ecosystem and likely be the first commercial headset to use the new second generation tracking. Tracking aside, it is looking to be more of a half generational step with marginally increased specs over the Vive and a better designed mounting mechanism.

Unreleased product. Specifications are not final.

• Screen Type: OLED (Pentile)
• Resolution: 1440x1280 per eye
• Refresh rate: 90Hz
• Field of view: 110 degrees
• Lens type: Simple
• No physical IPD adjustment
• Eye relief adjustment
• Connectivity at PC: 1x USB 2.0 (3.0 recommended for camera use), 1x HDMI or DisplayPort
• Headset connectivity at link box: 1x USB 3.0 C
• Audio: 3.5mm headphone jack
• Built in microphone
• Front facing camera
• Rigid head mount with flip up visor
• Tracking generation: [Speculation] 2.0 [/Speculation]

The Deluxe Audio Strap for the Vive is an optional accessory which provides increased comfort and headset stability along with built-in adjustable headphones (removable). Unlike the original fabric strap, the DAS is a semi-rigid design with replaceable foam padding at the rear. Tension is achieved by spring loaded sides and an adjustable ratchet wheel on the back. The cable is rerouted to the side and to the back instead of over the head.

TPCast will be the first accessory to hit the market that gives users the freedom of being unchained from the PC via wireless transmission from the PC to the headset. It's based on the proprietary WirelessHD standard for HD video transmission over the 60GHz band. While this provides lots of bandwidth for minimal video compression, it is not capable of transmitting non-AV data over that signal. [Speculation] Given reports of audio compression, it might not even be sending audio over the 60GHz band to maximize bandwidth for video. [/Speculation]

The data that would normally be handled by the USB cable is therefore transmitted over a standard WiFi connection to a router (included). It currently does not send the mic data back to the PC properly, but that is supposed to be fixed for the US launch later this year. Camera data is also not sent, so the passthrough video will not be available in VR using the TPCast unit. There's been no word from TPCast as to this being a bug, so expect that to be the case at launch as well.

Also of note is that the TPCast unit it not forward compatible with future headsets. It can not handle resolution higher than the Vive. Additionally, 60GHz transmissions will not go through walls or objects in general. Even the human body can block the signal. Thus while the TPCast can provide wireless transmission, it can not act as a range extender to another room like you might consider with regular WiFi.

• Wireless technology: WirelessHD + WiFi
• Transmission bands: 60GHz + 2.4/5GHz
• Max resolution: 2k
• Latency: <2ms
• Battery capacity: 20100mAh (Anker PowerCore)
• Battery life: 5 hours
• Requires additional ethernet port on the PC or routing your internet connection through the supplied router. Advanced users may be able to configure an existing wireless router to provide the functionality.

The DisplayLink XR unit is Intel's reference platform for a wireless accessory, and not yet a prototype of hardware intended for commercial production. Another company will have to license the technology for it to see the light of day. That aside, it differs quite a bit from the TPCast unit despite running on the same 60GHz band. It's based on the WiGig standard, and is thus much closer to WiFi in nature. Because the underlying tech is focused more on general data, the DisplayLink XR unit supports transmission of the USB data (including camera) over the 60GHz band. The 60GHz band is subject to the same physical limitations of the TPCast unit however; the receiver and transmitter must be in the same room. Since all data will be broadcast over the 60GHz band and it supports higher resolutions on top of that, the DisplayLink device is obviously relying far more on advanced compression.

• Wireless technology: WiGig
• Transmission bands: 60GHz
• Max resolution: 4k + HDR
• Latency: 3-5ms
• Battery capacity: Unknown
• Battery life: 2 hours

The Vive's controllers are the culmination of Valve's initial testing of natural interaction within a VR environment. They have a clickable trackpad with an LRA haptic feedback motor attached - similar to the trackpads on the Steam Controller. Additionally they have a two stage analog trigger for the index finger and two face buttons to the top and bottom of the trackpad. The top button's function is defined by the VR software in use, while the bottom button is used to bring up the SteamVR menu. Finally there is a pair of grip buttons to either side of the wands that work in tandem as a singular input.

Unfortunately the wands have a number of issues inherent to the way HTC is making them. The triggers are prone to squeaking and the trackpads are known to stop clicking over time due to the way the mechanism was designed. It's possible to fix both problems as an end user, but both require disassembling the controllers for any long lived solution. The grip buttons are also found to be uncomfortable to many for any prolonged use.

The Vive Trackers aren't traditional controllers so much as general tracking device that can act as an interface between a controller and SteamVR. By utilizing a standard tripod mount, a tracker can connect to third party devices. These devices can range from simple passive objects like a golf club, to an active devices like guns. For active devices a set of pogo pins on the bottom of the tracker or the USB port provides communication with the device to pass button and analog axis states back to the PC. Haptic feedback requests can also be forwarded to the device.

LG's controllers are an evolution of the Vive's. They maintain a similar form factor, but button locations and counts have changed. All face buttons have been moved to the top of the trackpad, and an additional application button has been added. This brings the face button count in line with the Oculus Touch controllers.

Valve's Knuckles prototype marks the first venture into a second generation control interface. Like the LG controllers, there are now two application buttons, but they've been moved to the sides of the trackpad . The system button sits below the trackpad, and the trackpad itself has been shrunk and made concave.

The real advancement however is in the way the controller is gripped. Unlike the wand designs, the Knuckles controller straps around the palm - keeping it in place regardless of whether the controller is actually being actively held. This is used with conjunction with capacitive sensors in the grip area, trigger, and touchpad to allow natural gripping and throwing actions. The capacitive sensors can also detect the extension of individual fingers as seen below.

Since the Knuckles are a complete superset of the original controllers, they'll work in any application without needing to be explicitly supported. In legacy compatibility mode, the capacitive grips can be configured to function as the old digital grip button or the analog trigger.

Tracking is where the magic occurs in VR. Without 6DoF tracking, VR would be little more than the head mounted displays of yore. Valve's lighthouse based tracking provides this core function for SteamVR hardware. Unlike the Oculus route of outside-in tracking using cameras to look at LEDs on the headset/controllers, SteamVR tracking is inside-out where the tracked objects look for light broadcast by external devices - the titular lighthouses.

The first consumer SteamVR tracking came in the form of two lighthouses with the Vive. Each lighthouse is comprised of two IR lasers and an IR LED sync array. Each laser shines into the open end of a cylinder. The cylinders are mounted perpendicular to each other and contain a mirror to redirect the laser 90 degrees to a lens in the cylinder wall. The opposite side of the cylinders are attached to specialized hard drive motors spinning at 3600rpm. The lenses in each cylinder fan out the laser beam into a structured line that runs parallel with the axis of rotation. The end result of this setup is two perpendicular laser lines being swept across the room at precise rates. Meanwhile the LED sync array floods the entire forward facing hemisphere with light at once.

The lasers and sync array are timed so that the LED sync array flashes, which is then followed by a laser sweep of one axis. Then another flash, and a laser sweep of the other axis. Both lighthouses will emit a sync flash simultaneously, but they will alternate between doing the pairs of laser sweeps. The gif below provides an idea of what would be seen in a room with a single lighthouse if it was greatly slowed and visible.

These light pulses and laser sweeps are picked up by a multitude of sensors embedded in every tracked object. The Vive headset alone has 32 photodiodes behind the shell.

By using the difference of the timing between the flash and sweep responses of the different sensors, the position of the object in relation to the lighthouse can be triangulated along with the rotation of the object at that location. This information is used as an absolute pose update (position + rotation) to correct drift from the IMUs which provide updates between each lighthouse sweep. It should be mentioned that while there are 120 sweeps per second, a full positional triangulation requires both the X and Y lighthouse sweeps, so there are only 60 full corrections per second.

This first generation of lighthouse based tracking does have a number of limitations.

• Multiple motors induce vibration which can cause minor jitter in tracking measurements which increases with distance from the lighthouses.
• Sync is handled by the LED flash array, so lighthouses must have a line of sight to each other or be physically connected.
• The distance that each lighthouse can cover is also limited by the distance that the flash can reliably be detected at, hindering the potential tracking range of the laser itself.
• No more than two lighthouses can be used for a single tracking volume.

All of these limitations will be addressed in the second generation of tracking.

The second generation of lighthouse based tracking is enabled in part due to the advancement of the actual sensors. While the first generation sensors were all discrete parts, the second generation makes use of dedicated ICs with improved capabilities, smaller size, reduced complexity, and lower total cost. The Triad TS4231 will make up the backbone of second generation sensors. Pictured below is the slightly older TS3633-CM1 prototyping module which can be compared to the sensors in the generation 1 section.

While the ICs reduce complexity on the sensor end, there's also been significant changes on the emitter end with the lighthouses themselves. The dual rotor design has been scrapped in favor of a single rotor containing two offset lenses with each fanning the laser at opposing angles.

Furthermore, the LED sync array has been removed; the lighthouses and devices now sync on the actual laser sweep instead of a dedicated sync flash. The two lighthouse limitation has also been removed and tracking volumes will support an arbitrary number of lighthouses.

These changes, unfortunately, make for some incompatibility with the first generation. While second generation devices will function properly with first generation lighthouses, the opposite is not true. First generation devices such as the Vive will not be able to operate in a tracked volume provided by second generation lighthouses. Second generation lighthouses are scheduled to enter mass production for consumers in Q4 2017.

Continued in next post.

 

[Zalusithix]

The play space in SteamVR is representative of the area you have for playing VR unobstructed. A common misconception is that because the lighthouse based tracking provides the capability for large tracked volumes, that a large area is required to use the product. In reality, SteamVR supports everything from seated play, to standing 360, to full roomscale. The first step in achieving a functional play space is mounting the lighthouses.

First generation lighthouses have a 120 degree field of view and are best mounted above the user looking slight down at opposite corners of the tracked space. This provides 360 degree tracking while providing the highest amount of occlusion resistance. As mentioned in the tracking section of this OT, lighthouses need general line of sight to each other, so if something is blocking that, they'll need to be connected by the aux cable and channels changed from B and C to A and B.

If mounting them overhead isn't an option, they'll still work, but the setup will lose a measure of occlusion resistance. Additionally, if the VR system will only be used for seated experiences such as simulators, then both lighthouses could technically be mounted to either side of the cockpit/desk area facing the user. That said, such a setup should be limited to cases where opposing corners is impossible or highly impractical as 360 tracking capabilities would be lost.

Once the lighthouses are set up, the next step would be to run the actual SteamVR setup and choose the type of play space you're trying to create.

Chaperone is the system within SteamVR that keeps users aware of the boundaries of their real life environment while they're in VR. It does this by projecting virtual walls as the user approaches previously defined boundaries.

In the case that a roomscale play space was selected during the SteamVR room setup, the user will be asked to define their safe area. This is best accomplished by using the "advanced" setup option. This mode effectively creates a point on the floor at the controller's location every time the trigger is pressed. Each point is then connected by a straight line segment. By placing points at every wall corner and around objects, a low poly outline of the safe area is created and each of the line segments becomes a chaperone wall.

Once the outline is set, the SteamVR room setup will best fit a rectangular play space within the outlined area. The dimensions of this rectangle are what is reported to applications that make use of the size of the user's play space. The play space itself is independent of the chaperone boundaries though. In the case of a non-rectangular chaperone area, so long as the game doesn't model the virtual environment to fit the play space, the user can make full use of the entire bounded area.

Overlays are a means by which applications can be interacted with from within VR, but not be the primary running VR application. They're similar in concept to traditional desktop gaming overlays, but properly integrated into SteamVR.

OpenVR Adavnced Settings adds another icon to the SteamVR menu and exposes many settings that otherwise would be relegated to the desktop based settings window. In addition to exposing many of the built in options, it also adds many more features that otherwise don't exist. These features include being able to save and load setups complete with different chaperone settings, modifying the chaperone interaction distances, adding a center marker to the play space, removing chaperone, manually moving the virtual world relative to the play space, and much more. OVRAS is a must-have for anybody that uses SteamVR.

OVRdrop is an overlay that brings a desktop window or monitor into a virtual pane within the currently running VR application. The pane can be interacted with, and can be set to either remain at a static location with the VR environment or attach itself to the viewport or user's controllers.

TurnSignal is an overlay to help users combat cable twist during gameplay. While OVRAS has a page that will tell the user how many times they've turned, it requires bringing up the OVRAS overlay every time to check - something that's not always viable in the middle of gameplay. TurnSignal, in comparison, overlays a simple starburst pattern on the floor. As the user turns around, the lines warp into a spiral corresponding with the direction turned. The more the user turns, the more intense the spiral pattern becomes. As the user turns back in the other direction, the spiral gradually straightens back out to the original starburst. This way the user can figure out roughly how many times they've turned from the starting point with a simple glance down - not needing to bring up a menu and take themselves out of the game.

VR isn't perfect on the PC right now. There's two competing store fronts and the Oculus platform does not support SteamVR headsets. Given the similarity between the VR systems, however, it is possible to bridge the divide using third party compatibility layers that translate between the VR APIs.

Revive is the compatibility layer created by CrossVR to enable the Vive (and by extension SteamVR devices) to work with Oculus games. It includes an overlay to launch any games purchased on Oculus' storefront directly from SteamVR. Compatibility with Oculus games is not 100% however. Before purchasing an Oculus game expecting it to work, please reference the compatibility list.

What are the minimum specs for VR?

This varies depending on the game, but in general any gaming rig with a GTX 970 equivalent or better should be able to run most games. When in doubt, run the SteamVR Performance Test.

How should I mount the lighthouses if I can't put holes in the wall?

Lighthouses come with a standard 1/4-20 threaded mounting hole that can be used with many tripods and related gear. While tripods consume lots of space, using a floor to ceiling construction support pole or lighting pole in combination with a pole clamp will allow lighthouses to be mounted virtually anywhere without holes

What are the minimum dimensions required for full roomscale?

Roomscale requires 1.5 x 2 meters (5 x 6.5 feet).

What are the maximum dimensions for roomscale?

There's no real maximum, but there are practical limitations. Lighthouses aren't recommended to be mounted more than 5 meters (16 feet) apart from each other. By using the sync cable this distance can be extended further, but the sync flash still has to reach the headset and controllers. Tracking accuracy in general will also decrease the further you are from the lighthouses making spaces greatly exceeding the recommendation less than ideal.

If I'm blind in one eye, can I enjoy VR?

Admittedly half the display would be rendered redundant, but a person with one eye can appreciate VR as much as anybody else relative to their own experiences in real life. Depth perception with a single eye in VR would be accomplished the same way the person does in real life aside from eye accomodation at shorter focal distances.

Can I use the headsets with glasses?

The Vive allows eye relief adjustment and is compatible with a large number of frames. The LG headset also supports eye relief adjustments, so it should end up similar in that regard.

Will I even need glasses?

This depends on your vision. In the case of the Vive, if you're able to focus clearly around a meter or so away, you'll probably be fine without them. If you're at the point where you need to wear glasses to read text on a monitor, however, you'll probably need your glasses. Focal distance for the upcoming LG unit is unknown.

Are there any options to not have to wear glasses if we need them?

Aside from contacts, there's a couple providers of prescription lenses that clip to the Vive's existing lenses. VR Lens Lab is one, and VR Lens EU is another. If going with the Lens Lab option, the RABS option is highly recommended to reduce the distortion that the normal lenses they sell introduce to the image.

Can I increase the breakout box distance from the PC?

Yes. In addition to a longer HDMI cable, you'll need a USB extension cable, or one of the few male/male type A cables out there. Beyond ~15 feet you'll need active cables. Use this resource on Reddit to get started. Extremely long runs are theoretically possible using ethernet extenders, but the chances for things to go wrong increases with every extra device in the chain.

What about extensions from the breakout box to the headset?

In the case of the Vive, the USB and HDMI advice are the same as above, but you'll also need a 1.3 x 3.5mm power cable extension as well. For the upcoming LG headset, extension would be provided by a USB 3.1 type C extension cable, but those are rare, relatively short, and obviously untested right now.

What about extending the power cables for the lighthouses?

First generation light houses for the Vive will need a 2.1 x 5.5mm DC power cable extension.

What is reprojection, and what's the difference between all the different modes in the settings?

An entire thread could be devoted to the details behind reprojection. In short, Reprojection is a method by which an artificial frame is created from a past rendered frame so that the headset can display a rotation corrected frame when a true frame hasn't rendered in time for the headset's refresh.

The different modes alter the way by which these artificial frames are created. In interleaved mode, when a frame is late, the rendering is reduced to half the headset refresh rate. After every rendered frame presented to the headset, an artificial frame is created by warping the previous frame to match where the headset is currently looking. This continues until the framerate recovers. In asynchronous reprojection mode, the rendering continues as fast as possible at all times, and in the event the frame will be late, the rendering is preempted and an artificial frame is derived from the last render.

Always on reprojection alters the way that the timing for CPU -> GPU command submission works. It allows more time for the rendering application to submit GPU commands, but because of this incurs more latency on the rendered image relative to headset position. Due to the added latency, all frames are reprojected before presentation on the headset. This costs the mode on positional accuracy, but can actually increase rotational accuracy. The downside is this mode seems to ignore other modes of reprojection and can result in frame drops that otherwise wouldn't exist if the headroom it provides is used up.

Force interleaved projection forces the application to always render at half the headset refresh rate, and should never be used.

OK, what about supersampling? What is that and what should I set it to?

Supersampling is the process of rendering above the native resolution of the headset's displays and then downsampling the render to the display's native resolution. This results in an image with less aliasing / shimmering, and more detail in general. On the down side, it greatly impacts rendering performance.

The default render target for applications is already by default set to 1.6x the native resolution, so any change to the supersampling setting is actually in addition to that. Currently the scale functions as a liner representation of absolute pixel count. A setting of 2.0 for supersampling will render exactly 2 times the number of pixels as a setting of 1.0, rather than multiplying the x and y resolution values by 2x.

The ideal supersampling value will differ from game to game, and from GPU to GPU. There is no "best" setting. At best another user that has experimented with the same GPU might be able to give advice for a certain game, but it remains something that the individual user should experiment with themselves.

As large as this thread is, it is merely a drop in the ocean of the available information out there. This section contains links to other sources for further reading and / or discussion on various related subjects.

• HTC Vive launch thread
• The high-end VR discussion thread
• Valve Knuckles discussion thread

• Doc-Ok's Lighthouse tracking article
• Doc-Ok's article on VR headset optical properties
• Alan Yates presentation on the creation of SteamVR tracking
• Video of Valve's past VR prototypes
• Gabe Newell & others talking about VR
• Alex Vlachos' presentation on advanced VR rendering (GDC 2015)
• Alex Vlachos' presentation on VR rendering performance (GDC 2016)
• Discussion on Lighthouse based tracking (gen 1&2) with Doc-Ok and Alan Yates on Reddit (2.0 specific post)

 

[Aztechnology]

Awesome OT, thanks for taking the time to do it, we really needed it.

Even though I ended up with an Oculus rift, I'm always just blown away by the Vive, especially the tracking solution. It's incredible and I hope it becomes the standard going forward.

 

[Zalusithix]

Awesome OT, thanks for taking the time to do it, we really needed it.

Even though I ended up with an Oculus rift, I'm always just blown away by the Vive, especially the tracking solution. It's incredible and I hope it becomes the standard going forward.

Ugh, between this mammoth thread and all the work on the PCVR thread, I'm wiped lol. At least now we can start anew with future proofing for what's coming later this year.

As for getting a Rift, I honestly can't blame anybody that took advantage of Oculus' current offer. Hopefully hardware on the SteamVR side sees a drop when we get more competition from within.

 

[Matthew23]

Yes nice job on this. A new HMD OT for every product on the platform would have been a mess.

I also hopped on the OR Amazon deal but will be closely following how the Knuckles and LG HMD turn out. If they end up as good as they are looking I won't be able to resist if/once they hit that $300-400 sweet spot.

 

[VlaudTheImpaler]

Amazing job. Subbed.

This will be my new home for the foreseeable future.

Will be sticking with Vive it seems since I think exclusivity is harmful to VR and HTC is playing it buddy, buddy with Valve who is so open like PC and VR should be.

I'd totally switch HMD manufacturers though, for the first company that is as open comes up with a way to sell their products just like PC hardware where it's designed to be as interchangeable and customizable as possible.

 

[staticneuron]

Ugh, between this mammoth thread and all the work on the PCVR thread, I'm wiped lol. At least now we can start anew with future proofing for what's coming later this year.

As for getting a Rift, I honestly can't blame anybody that took advantage of Oculus' current offer. Hopefully hardware on the SteamVR side sees a drop when we get more competition from within.

Amazing thread. Def look forward to future of Steam VR. Subbed as well.

 

[ISee]

Very good OP. Just an info because this is about open VR.
PSVR also works with steamVR thanks to clever people working on it and the open nature of PCs/Steam. You can even use it with PSmove controllers (two ps3 cameras and a blue tooth adapter needed) or use any other motion tracking device. The big downside is that PSVR only runs at 60Hz currently instead of it's native 120Hz. Which is a huge problem, but the TRINUSPSVR devs are apperently working on a solution for that.


Also get bridge crew. It is such an amazing MP and star trek experience!

 

[Iced_Eagle]

I'd totally switch HMD manufacturers though, for the first company that is as open comes up with a way to sell their products just like PC hardware where it's designed to be as interchangeable and customizable as possible.

I think it's on the horizon. However, I think it will really get started once both HMD's and controllers are fully inside-out tracking (aka, no lighthouses at all) and truly independent of each other. That way, every piece can be self-sufficient, upgradeable whenever you'd like (without needing to worry about tracking tech or the like) and you'd just need a way to have them wirelessly communicate to your PC through some sort of receiver.

 

[Compsiox]

Great thread. I can't wait for the Valve games in October (at least one).

 

[aaaaa0]

Great thread, subscribed.

 

[Wollan]

Thorough update on Tracking tech 2.0 and current state of wireless with details new to me. Great OT.

 

[innervision961]

So let me get this straight, Gen 2 lighthouses will not track the Vive? So this means a new headset inbound, or a revision?

 

[Auto-Reply]

So let me get this straight, Gen 2 lighthouses will not track the Vive? So this means a new headset inbound, or a revision?

I remember an interview with some dude from HTC earlier this month, saying that the new sonsor tech has to enter mass production first, then they will build new products upon that tech, these sensors are supposedly going into mass production at the end of this year.

I assume that means it can take a while until you see new stuff from HTC.

-

Sweviver also tested the Vive n Chill add-on https://www.youtube.com/watch?v=wte9po0_FEM, I'm at work and can't see the video so I don't know what's the verdict here.

 

[Rygar 8 Bit]

Great thread. I can't wait for the Valve games in October (at least one).

for a game with a full budget kinda crazy we havent heard or seen what the game thats coming out this year is looks like they are going with announce and release same day strat

 

[Samaritan]

Mmmm, gotta love that new thread smell!

So I've just experienced an odd, new problem that I've discovered my Vive was causing, and was hoping someone might be able to shed some light on how to solve it.

My system is connected to my monitor via DisplayPort, and my Vive is connected via HDMI. Until the other night, when I tinkered with my system and repositioned some fans, my system has booted up no problem. But upon plugging everything back in and powering on my machine, my monitor was suddenly no longer receiving a signal until it booted into the Windows environment.

Upon disconnecting my Vive's HDMI cable, I discovered that my PC was attempting to display my UEFI and boot screen on my Vive and not my monitor. This has never happened before, despite always keeping my Vive connected to my PC.

If anyone else has any history of this and may know a way to prioritize display devices in my boot/UEFI environment, that'd be very helpful. It's a minor annoyance in the grand scheme of things, but considering this wasn't happening until an hour ago, it's probably more annoying than it should be.

GPU: Gigabyte GTX 980Ti G1
Motherboard: Gigabyte GA-Z170XP-SLI
Monitor: Acer XB240H

 

[GuitarAtomik]

So let me get this straight, Gen 2 lighthouses will not track the Vive? So this means a new headset inbound, or a revision?

Gen 2 basestations don't work with the Vive, correct. I would assume a gen 2 Vive would use them, but I don't think that means it's imminent. The prevailing assumption is that the new LG headset will use the gen 2 basestations.

 

[Veritigo_X]

Nice OP! Subbed.

 

[NuMiQ]

Hot damn.. What a thread.
Great job Zalu, 2 awesome VR threads in a single day.
Colour me impressed

 

[Kaako]

Great OT. Quite informative!

 

[capylikesgames]

The Knuckles controller looks super intuitive for movement within VR space. I'd love to try them out one day. Also, does anybody know if they pulled out the dual stage triggers on them? I saw an unboxing a little while back, and the triggers seemed to lack a click at the end. A bit of a shame if so; Valve was smart to adopt that concept in the first place.

 

[Tain]

good new threads, getting me even more pumped for the future

 

[ArtHands]

LG seems like they only want to release a headset and nothing more. Hopefully they will pour in money to help improve the VR ecosystem as well like HTC did with viveport and various investments.

 

[Zalusithix]

So let me get this straight, Gen 2 lighthouses will not track the Vive? So this means a new headset inbound, or a revision?

Well, the LG unit is inbound this year unless something changed, and the expected release date coincides with the general availability of the second gen lighthouses. The correlation between those two things is enough for me to expect that it will be based on second gen tracking.

The Vive may be iterated on to use the new Triad modules (and thus support 2.0 tracking), but I have my doubts that such a unit would officially support it or use the updated lighthouses. It makes sense to use the new tech from a cost reduction standpoint, but it creates user confusion. They'd have to stock two different lighthouses on the accessory page, and deal with people ordering the wrong one. They'd also have to field tons of questions on why a second hand Vive headset/controller doesn't work in an existing setup.

Clear marketing would help to some degree, but you're in a situation where you need to make it obvious that it isn't the original Vive, but also isn't a Vive 2. You effectively have something like the "New Nintendo 3DS" situation - except applied to two different pieces, along with the introduction of a third piece that will be used with the Vive 2. It's a bit of a marketing mess to say the least.

Ideally it would also look noticeably different, but then you're redesigning everything for a mid-gen incremental update. An update that probably won't last much more than a year before it's superseded by the Vive 2. Unless they plan to keep selling that hypothetical update after they launch the Vive 2 as a lower tier product, I'm not sure how much sense it makes to devote that much effort into it. Especially since the existing product will support all new accessories as is.

LG seems like they only want to release a headset and nothing more. Hopefully they will pour in money to help improve the VR ecosystem as well like HTC did with viveport and various investments.

Frankly I've heard of way too many conflicts with VivePort screwing up the installation of the Vive and SteamVR functionality. To make matters worse, the one major feature of Viveport, the subscription, is plagued with outdated or "lite" versions of games that you'd otherwise get on Steam. While I understand HTC's desire to have their own software distribution center to make money from, from a user's perspective, it's not worth it.

 

[Blam]

That's a very very nice OT good job OP.

 

[Deleted member 22576]

Subbed! Hey guys!

 

[TheMadMan007]

Subbed. Amazing job on the thread by the way!

 

[Zalusithix]

There was a recent conversation about Lighthouse tracking on Reddit with Doc-Ok and Alan Yates chiming in. While there was a lot of good information in general there, more importantly, it included information about the second generation tracking.

Yep, that's where we are headed. 2.0 bases are simpler and cheaper, they have no blinker and only one motor and laser. The newer sensors can multiplex more bases in the same volume concurrently. The upper limit is now the computational and communication bandwidth available in the receiver, and the power/cost/complexity trade-offs of doing more crunch locally vrs shipping more data and doing the crunch on more capable hardware.

Lighthouse 2.0 will likely spin a little slower but will be multiplexed with spin rate and modulation codes. By removing time division multiplexing you get more data per second per base, so spinning slower doesn't cost you, any latency cost is already mitigated by the IMU. The slower spin means the bases use less power and can run quieter, cooler and longer from batteries. It does cost a little in stored energy for jitter minimisation, but improvements in the control system mean the new bases are actually even lower jitter than previous designs despite less mechanical Q. Additionally the optical scanner subsystem now has less uncorrelated noise sources which makes the effect on range noise lower and in general the per-beam updates more consistent with each other which reduces low frequency noise (which is more challenging to reject).

I want to get to a place where bases are so affordable you just dot them around and can get single point tracking everywhere in your volume. The prototypes of the new bases are the best bases I have ever seen. They are sub 100 microradians RMS (absolute, even better relative/locally) over the entire ~108x160 degree FOV. They are also largely software defined, so they will be even more upgradeable, supporting almost any kind of digital modulation scheme.

I've included a link to the topic in the link section of the main post.

 

[ArtHands]

Well, the LG unit is inbound this year unless something changed, and the expected release date coincides with the general availability of the second gen lighthouses. The correlation between those two things is enough for me to expect that it will be based on second gen tracking.

The Vive may be iterated on to use the new Triad modules (and thus support 2.0 tracking), but I have my doubts that such a unit would officially support it or use the updated lighthouses. It makes sense to use the new tech from a cost reduction standpoint, but it creates user confusion. They'd have to stock two different lighthouses on the accessory page, and deal with people ordering the wrong one. They'd also have to field tons of questions on why a second hand Vive headset/controller doesn't work in an existing setup.

Clear marketing would help to some degree, but you're in a situation where you need to make it obvious that it isn't the original Vive, but also isn't a Vive 2. You effectively have something like the "New Nintendo 3DS" situation - except applied to two different pieces, along with the introduction of a third piece that will be used with the Vive 2. It's a bit of a marketing mess to say the least.

Ideally it would also look noticeably different, but then you're redesigning everything for a mid-gen incremental update. An update that probably won't last much more than a year before it's superseded by the Vive 2. Unless they plan to keep selling that hypothetical update after they launch the Vive 2 as a lower tier product, I'm not sure how much sense it makes to devote that much effort into it. Especially since the existing product will support all new accessories as is.


Frankly I've heard of way too many conflicts with VivePort screwing up the installation of the Vive and SteamVR functionality. To make matters worse, the one major feature of Viveport, the subscription, is plagued with outdated or "lite" versions of games that you'd otherwise get on Steam. While I understand HTC's desire to have their own software distribution center to make money from, from a user's perspective, it's not worth it.

I'm also talking about initiatives like ViveX Accelerator which invest and mentor VR studios , or investing in VR arcades.

Viveport is shit, but at least they are trying. Viveport plays a part in the infrastructure for VR arcade business. The arcade operators get to offer a library at an affordable price, VR devs get an additional channel to expose their games and earn passive money, while general public a way to try out VR.

Not saying that LG should offer their own app store too, but I hope they will do something beyond just releasing a headset. Maybe fund some titles , sponsor VR events and the like.

 

[Zalusithix]

I'm also talking about initiatives like ViveX Accelerator which invest and mentor VR studios , or investing in VR arcades.

It'd be nice, but realistically you can't expect every hardware manufacturer to do investments like that. Manufacturing the hardware is already a risk at this early stage in VR's life cycle. Doing stuff like the accelerator program is a gamble on top of that. I'll be happy just having more hardware options. So long as they're not half-assed products at least.

Unrelated to that, a new potentially useful overlay has been developed. TurnSignal places a design on the ground that warps as you turn so you can tell at all times how much you've rotated from the neutral position and in which direction. OVRAS has a turn counter, but that's something you have to bring up. This is something that would be available to see at all times (while still being unobtrusive), and thus less cumbersome to keep track of.

It's still in the early stages, and being actively worked on. I haven't tried it out myself yet, but I plan on giving it a whirl tomorrow along with updating the overlays section of the main post..

 

[Teknopathetic]

I actually really like that. I've been using the rotation counter in OVRAS a ton in Echo Arena because of how easy it is to end up turning around countless times. I'm actually surprised at how often I've actually spun in that game without realizing it. Definitely going to give that a spin.

 

[Raticus79]

Oh yeah, I guess too many turns in the same direction would eventually cause the cable to twist up. I haven't run into problems with that yet, but I can see how it would be useful.

 

[PeterVenkman]

So I'm having issues with a stuttery headset.

Problem is, even in the steam home, when moving my headset the image jumps or lags behind every few seconds. In games like Echo Arena the stuttering is intensified.

It isn't dropped frames. I've tried various 2.0 and 3.0 USB ports. I've repositioned base stations, counted LED lights, disabled the camera, disabled display mirroring, cleaned smudges off the stations, restored nvidia control panel settings, rotated the base stations - I'm at a total loss. I can't trigger the stuttering in any particulr way - just moving around and eventually it hitches.

Any clues here? Hardware is a 980ti and i7 4790k

 

[Veritigo_X]

So I'm having issues with a stuttery headset.

Problem is, even in the steam home, when moving my headset the image jumps or lags behind every few seconds. In games like Echo Arena the stuttering is intensified.

It isn't dropped frames. I've tried various 2.0 and 3.0 USB ports. I've repositioned base stations, counted LED lights, disabled the camera, disabled display mirroring, cleaned smudges off the stations, restored nvidia control panel settings, rotated the base stations - I'm at a total loss. I can't trigger the stuttering in any particulr way - just moving around and eventually it hitches.

Any clues here? Hardware is a 980ti and i7 4790k

I had the same issue and had to roll back the latest Nvidia drivers. If you have 384.76, roll it back to 382.53 and it'll fix it.

 

[PeterVenkman]

I had the same issue and had to roll back the latest Nvidia drivers. If you have 384.76, roll it back to 382.53 and it'll fix it.

Wow night and day difference. Thanks a ton - I was neck I deep in third party jitter tests and htc customer service.

 

[Dreams-Visions]

I'm glad this thread got bumped. Prompted me to check the store where I found the deluxe head strap is available again. Ordered.

 

[PeterVenkman]

I do still get stuttering on any quality settings while in the Echo Arena lobby. Could be ReVive hiccups or internet issues I guess, as the tutorial is super smooth.

Edit: starting to really think it's ReVive as any time I join a matchmaking game, the whole thing crashes.

Double edit: I am a fool - new ReVive dropped half a day ago that says to disable async reprojection for issues in who Arena, AND looks like they implemented trigger grip. Still crashes on matchmaking

 

[Teknopathetic]

I haven't had any crashes. I don't even have reprojection disabled, I do have super-sampling turned off. Generally pretty smooth.

 

[PeterVenkman]

I haven't had any crashes. I don't even have reprojection disabled, I do have super-sampling turned off. Generally pretty smooth.

Hm. Well I've spent a good hour trying to get into a matchmaking game and all I get are crashes, even after fresh installs of both software (Echo Arena and ReVive). It happens right after "joining match" announces in the lobby. I feel like it's gotta be something obvious on my end.

I'll try reinstalling oculus software next, but I dunno.

 

[Durante]

Excellent thread. I approve of moving from a Vive-specific thread to a general SteamVR thread -- hopefully we'll have a larger selection of native SteamVR HMDs soon.

I don't have much time to post right now but I'll keep up with this one.

 

[Deleted member 22576]

I'm regretting uninstalling Oculus/Revive after Superhot came to steam. Now I want to set it all back up to play echo arena!

 

[Durante]

I just fully read the OP and have two remarks:

• It's a very minor thing, but the Vive actually doesn't need USB3, it works with USB2
• I had not known about TurnSignal before. Looks extremely useful and the visualization approach is very clever.

Oh, and to reiterate: fantastic job on the thread!

 

[robotrock]

Great thread. I can't wait for the Valve games in October (at least one).

Where is the October date coming from?

 

[Zalusithix]

I just fully read the OP and have two remarks:

• It's a very minor thing, but the Vive actually doesn't need USB3, it works with USB2
• I had not known about TurnSignal before. Looks extremely useful and the visualization approach is very clever.

Oh, and to reiterate: fantastic job on the thread!

Yeah, the USB 3.0 bit was more of a recommendation for usage with the front facing camera. I've clarified that now. It's not even that it needs 3.0, but the chance of a USB 3.0 bus being saturated by other devices is far lower, so the tracking is less likely to hitch in less than optimal setups (real world situations).

As for TurnSignal, it's not surprising you didn't know about it; the overlay has only been around for a couple of days. It's certainly a simple yet elegant solution to the rotation problem. In retrospect it seems like something so obvious, but the use of a visual feedback system like that never occurred to me before.

 

[Nzyme32]

Amazing job with the thread. This is going to be a fun upcoming year for this thread

 

[Auto-Reply]

https://uploadvr.com/htc-vive-gm-happy-sales-met-targets/

Doesn't sound like HTC is going down with the price of the Vive anytime soon, maybe it's too early to see any effects of the Oculus price reduction and when they come into effect (or if they come into effect) then that will change.

Maybe they will react to the LG headset once that is available (and not more pricey than the Vive)

 

[Alexander DeLarge]

Where is the October date coming from?

Valve said they'd show new games at the next Steam Dev Days last year.... Or at least hinted at them saying people who complain about Valve not being a game development company are going to be very happy around this time next year (October).

 

[MaDKaT]

I had the same issue and had to roll back the latest Nvidia drivers. If you have 384.76, roll it back to 382.53 and it'll fix it.

Exactly what I was looking for. I have been trying to resolve a minor but consistent jutter that is amplified in Revive.

*Update
Certainly improved but still some issues in lone echo. Was scanning the revive issue thread and seems I just need to go back a bit further. *.05 was recommended more than once.

 

[SimplexPL]

I have 382.33, there is no judder in lone echo on revive (at least inside the station).

 

[GuitarAtomik]

So I finally got around to picking up one of these magnetic gun stocks and gave it a try today in Pavlov. Man does it make a difference having steady aim lol. I've never been that great at the game and usually end up in the middle to bottom in a match, but I was top of the team in three straight games using this today. It definitely has it's awkward points though. It takes a while to get used to the magnet placement without being able to see them and they tend to move around a bit since the magnets are so strong which exacerbates the issue. Trackers on the magnet parts of the stock could be a solution to that since you could see the points in game. The movement might be remedied by tightening the screws on the parts though. Also doing anything with your primary hand besides holding the gun takes a lot of getting used to so switching to grenade or knife or sidearm is more awkward. I think the best course of action is probably to hold your off hand in place (ie in magnet) because trying to find the right orientation of the stock after you've completely let it go is a herculean feat, but I haven't really tried that strategy yet. Definitely had moments of occlusion too though they were brief. So yeah, definitely not the perfect solution but, again, overall it improved my performance. Still, I wonder if there's some kind of better out of the left field solution to the issue the gun stock is trying to solve.