Blog

An overview to current HMDs

Miguel Villar February 6, 2014 Virtual Reality 2 Comments

A few weeks ago we released a post where we showed real people testing the Oculus Rift (from now on the “OR”) in the University of Alicante, close to Lemon Team’s headquarters.

Before the testing I introduced all the assistants with a brief explanation about what VR could mean for the future of the entertainment, including other medical/military/productivity/healthcare applications.

And after investing a lot of time with the OR and other similar devices like the Sony’s HMZ-T3W (“T3”), I though it was time to share my experience with all of you who are not so much involved with this kind of technology. In this post I’ll be somehow reviewing current Head-Mounted-Displays (“HMD”) only, this is, I will not comment on other devices also critical for VR experiences such as motion sensors, feedback sensors, etc.

Maybe I’ll cover those too in a second post, specially focusing on the Razer Hydra, Sixense, PrioVR, Virtuix Omni and ARAIG, once I have a chance to try them out. Note that each one of these addresses some of the issues caused by VR self-nature.

Before I start shooting technical details, it must be said that VR IS NOT FOR EVERYONE. And following the same concept, VR IS NOT FOR EVERY TYPE OF GAME either. It’s just a really cool piece of tech (both hardware & software side of things) not applicable to everything nor to everyone. It can produce unprecedented sensations ever experienced by the consumer but at the same time it can result very annoying, troublesome, inconvenient… And the worst part is that most players feel a different experience, mostly on games that try to re-create real-life conditions. This is because each person has a different perception of the reality (different height, different walk/run speed, different inter-pupillary distance, different color perception, etc).

And we all know VR is not something new. The race started back in the eighties and continued over the nineties (Nintendo was there with its Virtual Boy) but the technology hasn’t evolved enough to go mainstream yet.

So what do we need to enjoy a great VR experience?

The problem with the answer is that most people would provide a different affirmation although most of them would be correct or partially correct. Since any of us haven’t yet experimented a truly VR experience is very difficult to tell. One way to achieve it could be “a way, by technological means, to pass sensorial information to our eyes, ears, nose, fingers, skin, muscles, etc, in order to simulate a virtual world that can be felt as close as it was real”. Another theory, closer to Matrix’s scenario, would be to “simply pass information to our brain in form of micro-electrical signals that we can naturally process”. Even if both systems use completely different approaches (sending information external vs internal means), there’s still one question waiting to be reveal: “how much information and what level of quality (that at the end, the quantity of information goes in pair with the quality of it) do we need to actually induce the subject into another world without he/she getting notice?” Be aware that I’m not referring to 3D rendering quality. For instance, being able to render New York in stereoscopic 3D at >=60hz with realistic materials and global illumination is out of the question here. Game designers might want a realistic representation of their worlds while others not (it depends on the type of game). Audio quality is out of the question as well, mainly because we have already been able to reproduce hi-def + true positional audio nowadays.

Despite most people tend to think that we would need huge amounts of information to really cheat our brains, the truth is that we only need a decent set of information being provided externally (no need to connect cables into our brains by now :-D), and that’s because our brain is very good at filling the “gaps” of missing information. It is something it does naturally without us being aware, and does it better the more we focus into the game/experience. So the truth is that, for the majority of situations, we only need a device that can provide decent image & audio quality and that is able to track our body movement/rotation with high precision. This device is what we call the HMD and is the angular stone of any kind of VR experience. Unfortunately, the technological barriers behind doing a decent HMD have been hard to solve, until now…

So right now we have 3 kind of HMDs, each one with its pros and cons:

Definitively the InfinitEye represents the pinnacle of the field of view :-). By using two separate screens with high quality Fresnel lenses they achieve 210º FOV. In case there’s a butterfly fluttering by your ear you would be able to see/notice it! For scary moments in horror games it would be great! Unfortunately, it is too much of a stone on your head… too much wide, too much heavy… Difficult to move your head fast without noting its weight… Although I haven’t tested this device on my own, personally I don’t see this form factor to be the winner, at least in the sort term.

The Oculus Rift sits between the InfinitEye and the T3. It produces a decent FOV of 110º, which is what most humans can focus on (in normal conditions the sharpest part of our cortex-generated image is ~90º, up from there the image starts getting blurry). Despite being smaller than the InfinitEye, it still feels big but not heavy, and I must admit is quite comfortable. However, picture quality (“PQ”) on the current version of the development kit is clearly under the expectations. It only delivers 640×800 per eye, which is totally insufficient. For you to have an idea, the T3 features 1280×720 per eye in a considerable smaller FOV, and still doesn’t look perfect. So this means you can clearly see the pixel matrix on the Oculus, and not even a gorgeous 2560×1600 5-7” flat/curved panel would fix the problem. It is now clear that for the OR to get mainstream ~4K is the minimum resolution it should run (~1920×1600 per eye). Also, another thing to get notice is that IPS (the LCD technology behind the display) is not suitable for HMDs. Yes, it can reproduce accurate plus natural colors in wider viewing angles, but in general this kind of technology refreshes very low and do not provide deep blacks, which means that high contrast in dark scenes is really poor. When moving your head fast, you can clearly see the “ghosting” effect since the pixels do not really refresh at 60hz, but at much lower rate. Gladly the Oculus Team has mounted an OLED panel for the new prototype (shown at CES past January), which should look much better and luckily will be used in the final/commercial version. Another issue with the OR is lens aberration. Color & blur artifacts (saturated red and yellow and focus problems) appear in the corners of the image, which is not pleasant. The head-tracking though, works flawlessly with almost no noticeable latency issues! All in all, the OR is a good prove of concept for a technology that should be “ready” in 2015 or 2016. If they manage to mount a CPU+GPU combo with its own operative system and dedicated game/movie store (making it an actual “game console”, not just a dependent peripheral) the future of the OR could be brilliant 🙂

On the other side, we have the T3 as a different kind of HMD. By default it does not provide head-tracking (you can easily install one though!) and it only delivers 45º FOV, which is perfect for watching movies but lacks peripherical vision to feel immersed in a video game. If you ask how this looks, I can sure tell you this is like watching a 1080p ~90” display at 2m distance (maybe it doesn’t look that sharp). Sony claims the T3 simulates a ~750” display at ~20m distance. I wouldn’t say this is tricky marketing stuff but what I can tell you is that it really feels like a cinema screen (perhaps a bit smaller than you might expect). When I use it in the sofa, close to my 50” TV (just at 2m distance) I can raise my head and be able to see both screens at the same time. The T3 looks almost twice the size of my Pany 50”, so you can get an idea of the actual size… As in the OR, lens aberration is also present, being almost impossible to see all picture focused despite the smaller FOV. I must say that even with its stylish design and relatively low weight it doesn’t feel comfortable after hours of use (I prefer the OR design to this respect). However, Sony provides integrated audio-jack with support for Dolby 7.1 where the OR doesn’t and the wifi model (T3W) allows you to connect to your PC or console wireless, which is great!. Last, but not least, one of the coolest features of the T3 is the ability to adjust the position of the lenses to match your IPD. This could seem unnecessary but trust me, is critical for any HMD out there. For further details on this Sony device, feel free to read my personal review on the AVS forums. Just be aware, it’s a long read 😛

There’s a similar device regarding the T3-specs which is called Silicon ST1080. I’ve only had one opportunity to test it and I must say pixel matrix in this device is totally invisible. It features 1920×1080 per eye, delivering truly crystal clear, retina images (it’s almost 3000 dpi!). Despite its super high-res, it still suffers from lens aberration, who causes blur on all borders of the screen, specially on the top and bottom sides. You might be thinking how the hell Silicon achieved this DPI, and the answer remains on the core of its proprietary technology. They do not manufacture standard LCD or OLED screens, but a proprietary version of LCoS instead. LCoS is very similar to LCD but it is built on silicon, which allows to reduce pixel size (and pixel gap) and related circuitry by a lot. At the same time they use micro-mirror-matrix technology to refresh pixel color/lighting at higher refresh rates (240hz = 80hz per color channel), which is incredible compared to the current standards. I remember also how light it is (~180g) compared to the OR or T3. The down side though, as it happens with LCD technology in general, is poor contrast and poor blacks, which is essential for an HMD 🙁

There’s also a new prototype coming to crowd-funding from Avegant called Glyph. It features integrated headphones (Stereo, 20hz-20000hz, 115dB) and 9 DOF headtracking, plus a micro-mirror-array display of 1280×720@120hz per eye.

There are other devices similar to T3, ST1080 or Glyph, but they have even smaller FOVs (30-40º, simulating 40-50” screen sizes at 2m distance), so no worth mentioning them…

Once briefly introduced each of these devices, I would like to focus my review in the OR and the T3, putting both into a comparison. You might think the OR and T3 has nothing to do one with each other, and in part that’s true (45º vs 110º FOV is a long way to go), but after playing Doom 3 BFG edition on both I can share a couple of thoughts.

The good thing of Doom 3 BFG is that it supports both Oculus Rift as well as standard stereoscopic 3D in TVs, PC monitors and HMDs such us the T3 or the ST1080, so you can easily compare your experience across devices. And although I’ve loved playing Doom3 on the OR I must say I prefer the T3 experience after all. Even if my OR was 4K resolution and Doom3 had better support for VR, I would still prefer my T3 over the OR for “standard” gaming. Don’t get me wrong, the Doom3’s VR-mode is spectacular and really provides an immersive experience, but what you should ask yourself is whether it’s really comfortable for long gaming sessions. For me it’s not. The fact that you’re actually immersed into the world but cannot move around with your own legs ends up with many annoying moments of frustration and sickness. Believe it or not, you really need some sort of Virtuix Omni for playing FPS on the OR, otherwise you’re doomed to static games (where you only have to rotate your head) or cabinet-based games such as driving or spaceships simulators (where you can’t actually move your legs and your body). The T3, as well as the ST1080 or the future Glyph, despite its low FOV, suits best for long gaming sessions using a gamepad just laying on your sofa. It feels like playing in the cinema with nothing catching your attention on the screen surroundings, thus avoiding distractions and getting more focused into the game, while limiting sickness thanks to a not-so-heavy FOV. Also, no need to fatigue your neck trying to aim at anything every time (but you have the option to do so installing a head tracking device), then better for the long term comfort. Also have in mind that since current HMDs are not able to track your eye-balls there’s no way you can auto-focus wherever you want. Auto-focus and depth-of-field is always based in the center of the screen, not in your real focus, who might become very annoying when FOV is larger than ~45º.

So you see, VR is great in theory but it is not perfect technology in practice unless you go along with other peripherals such as the Virtuix Omni or the PrioVR, all being fully supported by the software/game itself.

That’s it for now. I’ll keep posting my thoughts on VR and HMDs as long as I continue testing different games and technical demos. In addition, as we keep developing for the OR, I’ll eventually post some useful information about development challenges, technical constraints, game design considerations and so on!