Hacker News Reader

Comment by danpalmer | Hacker News Reader

I wonder how effective this VR actually is.

It seems that VR heavily depends on high quality screens, with very high resolution, and low latency and persistence. However this works with a screen that isn't optimised for those attributes.

$99 may be good, but it seems likely that it will be much worse than the Oculus Rift, I think is targeting ~$400-500, and could be far better. Is it worth it?

View on HN

I've tried most VR systems and I think the benefits of Gear FAR outweigh what you lose.

The biggest one is that it's wireless so you don't have to worry about the cord at all.

You do lose processing power, but quite frankly I think that is fine because it forces developers to really focus on making the UX/UI experience really tight.

From the Performance aspect the CV1 is much better and has better interactivity/presence abilities but if the goal is to get MM of people using VR, this is the way to do it. Everyone can thank cardboard for paving the way for this type of implementation.

Actually Palmer worked on a research project in university that was using 3d printed goggles and cheap off the shelf lenses which you could snap an iPhone into. See http://projects.ict.usc.edu/mxr/diy/vr2go/

I don't know if it was his idea or how involved he really was, just that he said he worked on it. The point is that cardboard was based on this.

Cardboard was based on the Oculus Rift DK1

That's $99 plus a $600+ smartphone. If you already have a Samsung phone that's great, if not... it's a very powerful Google Cardboard for $700+.

That's $99 plus a $600+ smartphone.

Poking around ebay I found lots of used Note 3's for less than $250.

FWIW, the Oculus DK2 uses literally exactly a Note 4 screen, display glass, Samsung logo, and all [1].

[1]: https://www.ifixit.com/Teardown/Oculus+Rift+Development+Kit+...

That is true, but the CV1 (consumer version) headset will have 2 separate screens that can be moved independently, which is something you can't get with a phone attached to a headset.

Not sure that feature will justify a $200+ price difference for most people. Plus, you will upgrade your phone every year or two anyway and get all the benefits (faster, better screen, improved features, etc) associated with that, while the CV1 will require a separate purchase to get new features.

"Plus, you will upgrade your phone every year or two"

Do normal people actually upgrade their phone yearly? I've kept my last several phones for about 3 years each, and I tend to think that'd be overkill for most folks (admittedly, I tend to buy higher end phones, with large storage, so they last a little longer; and I never buy on contract, so I'm paying full price, so I don't want to do it often).

Maybe not now; Apple is starting that "phone on subscription" program to replace your phone yearly, though.

I think the more typical behaviour is to buy on contract and thus upgrade on the one year or two year interval that coincides with the contract.

A quick Google suggests that even two years might be a bit long for Americans: http://www.phonearena.com/news/Americans-replace-their-cell-...

My family has been on contract for last 10 years and so has most people I met in America. The subsidy means flag ship phones at ~$200 with 2 year contract, or a cheaper phone for free. The other part is that AT&T and Verizon had much more reliable signal so it wasn't worth going to another provider.

The average lifetime of a smartphone increased last year from 18 months to slightly over 24. Analysts expect carrier and Apple leasing options to drive this back below 2 years.

That's actually really interesting.

If this is the case, then I'm not sure Oculus was worth the money Facebook paid for them - Samsung seems to own all the relevant IP and it's not clear what Oculus provides that couldn't be replicated by another company in short order. If Oculus Gear takes off as a product, it seems like Chinese manufacturers would start churning out knockoffs and commoditize the product space.

The screen is only a small piece of the puzzle. Most of the value of the Oculus products comes from the insane software optimizations that have been done at all levels of the stack to bring down the latency.

Oculus owns a lot of the IP, and more importantly know-how. Evidently the Gear VR was the result of extremely tight integration between Samsung and Oculus engineering, with Oculus providing the bulk of the VR-specific knowledge. Samsung provided screens and phones.

Late in the game, the Oculus engineers pushed for major changes to the architecture to allow the phone to communicate with the sensors in the headset in a very low latency way -- something nearly bare metal. Evidently it took a lot of convincing.

(A friend heard this from Carmack in person, for what it's worth.)

No, it uses a Note 3 screen. An executive from Samsung corrected me when I said Note 4.

It's significantly lower res than a Gear VR, and you can tell.

EDIT: your link confirms it's note 3, not note 4.

That it is. DK1 resolution was comparatively pretty awful. Even DK2 isn't great.

Crescent Bay and Gear VR felt fairly similar at GDC (rendering capabilities aside), so I'm really curious how much better the Rift CV1 is going to feel.

My time with Vive was very short, and felt so good I kinda forgot to look for flaws, but I hear CV1 is roughly the same (aka AWESOME!).

I heard at a VR conference that the CV1 is likely to use the Note 4 screen, but that's hearsay. I had stronger confirmation that it will at least be the same resolution... the Note 4, S6, etc all are 2560 x 1440; we were hoping for 4k but it's not gonna happen this cycle. CV2, perhaps?

(And there are no guarantees, so this could be wrong... but probably not.)

Looks like it's actually slightly lower. According to Road to VR, it's going to be "2160×1200 at 90Hz split over dual displays" (http://www.roadtovr.com/oculus-rift-resolution-recommended-s...).

The Gear VR requires a phone with a 2560 * 1440 resolution. That's actually higher than Oculus DK2, which is only 1920 * 1080.

I have a Gear VR for Note 4 and it's pretty awesome (especially for 360° video)... But there's just no content yet.

I'm really looking forward to Land's End, a new Gear VR game from Monument Valley's makers: http://www.theverge.com/2015/9/21/9353235/lands-end-vr-game-...

Content is precisely the problem. Carmack even admits as much here:

http://uploadvr.com/john-carmack-talks-difficulty-mobile-pos...

That guy's pretty outspoken. (And yes, he's referring to live action content; I presume mobile VR will be tilted more toward video streaming etc than gaming, but I could be wrong.)

I'm not a sports person, but your "live action content" made me wonder how awesome it would be to be a sports fan who could view 'the game' from the sideline, or an aerial cam, and to be able to turn and zoom a camera or to switch cameras – as desired! Better yet, to be able 'be the athlete'. Cameras aren't there yet, but imagine being able to watch the scenery as an Olympic down-hill skier as flying downhill - AWESOME possibilities!

Cameras are here, the software is here too, I've actually toyed around with some implementations and they're definitely good enough.

The big problem that scares me is sports licensing and all the legal stuff attached.

VR will allow lots of cool stuff like that :) . One possibility I was thinking about was putting the cameras on the referees instead of the players themselves (for sports such as soccer).

But I think it'll happen as soon as there's a critical mass of people buying VR headsets.

Are there really cameras currently available that are light enough to provide 360° view without interfering with the athlete's performance?

No, unless the athlete is doing power lifting or something trivial.

It's possible an athlete already accustomed to encumbrance wouldn't mind more -- like a few cameras on a football helmet -- but keep in mind that even a GoPro is ~80g, and you need about 3-4 of those to get 360 (mono). Stereo is actually super important, and that requires a lot more cameras.

Personally, I put this in one of those 'future, maybe' categories.

Now, a VR view from courtside seats... well, that's literally already happening. Just look up NextVR, among others (that I'm probably not allowed to talk about).

2016 is the (first) year of VR, for real.

http://www.sporttechie.com/2015/09/24/portuguese-soccer-leag...

http://www.nextvr.com/

Also, OTOY is talking about live streaming. That said, I have spoken with more than one startup that wants to put cameras on athlete's heads/helmets. Dunno how practical it is, but I guess it'll happen eventually.

With VR, the question is (and has always been) "when". As in, "man, that's awesome, when will it actually happen?" Because everything is possible in VR, and that's part of the problem.

I've used the Cardboard on my S5 and it is pretty painfully low resolution (1080). 2560x1440 would be fantastic. 2000 vertical lines would be amazing.

Just because I've never done it and want to double check. How does one get content to the phone-VR system?

I currently develop my toy VR applications with Unity and only target PC/Mac. I'm assuming I can just publish to Android from Unity and be set (never done it, just know this exists)? That is I target Android and can use the phone-VR setup?

Pretty solid coincidence that I'm currently in the market for a new phone as well. I tihnk I'll go with this and then skip the regular Oculus and get a Morpheus instead since I own a PS4 but the only machine suitable for smooth VR is at work (and I surely want to play around at home, too) :D

The GearVR actually has a fair amount higher resolution (and arc-resolution) than the first-gen wired headsets. It also has a bit more SDE and lower FOV so there's a trade-off.

In terms of rendering pipeline, even w/ the latest Gameworks VR / Liquid VR drivers and WDDM 2.0, the GearVR may still beat out the PC's latency - Carmack's OC keynote from last year is worth watching for some of those details. Since the GearVR has both a direct kernel driver for the sensors and front-buffer access to the GPU, he quoted something eyepopping, like 4ms M2P latency.

IMO, the killer app for GearVR is watching movies on long trips, and the really compelling 360 photo/video content that's starting to get made. If you're getting an Android phone anyway, the extra $99 seems like a no-brainer and is a compelling reason to buy a Samsung device over the competition.

It is pretty effective. I think most people dismiss it as just a fancier google cardboard, which it isn't. With it's own IMU and impressive display the GearVR is a great entry level device.

Also, the Samsung S6 has a higher res screen than CV1 (the consumer version of the Rift coming out next year).

Phone screens on top-end phones are high enough quality to manage (and are higher resolution than most PC monitors anyway). They also have the added benefit of being able to render graphics internally as opposed to being attached to a PC by a cord. Yeah, the graphics may not be quite as good, but the gap in perceived graphics quality between a $1000 PC and a flagship cell phone is rapidly closing.

I think the Rift will be the choice of the "PC Master Race", but a solution like this is much more mass-market friendly. There's no reason a headset like this should cost more than $30 once mass-produced - and that's low enough to be an impulse purchase for a lot of people.

If VR is going to matter for movies and tv it's only going to be on devices like this that everyone can participate on economically. Streaming game services will also close the gap between this and an expensive dedicated gaming rig per headset.

Streaming games have far too much latency for VR applications. Latency requirements for VR are much lower than console games because too much latency on head tracking can cause nausea and breaks the "immersion" effect.

But yeah, I agree that when VR blows up big, it'll be something like this rather than the PC-based Oculus. That solution has always felt clunky and more like a development kit than a consumer product.

> perceived graphics quality between a $1000 PC and a flagship cell phone is rapidly closing

Is this true? Can an iPhone 6S theoretically run CS:GO or Crysis (an eight year old PC game)?

Perhaps not the actual CS:GO binary - but one built on Metal could certainly come close to CS:GO on lower settings.

Wow, that's awesome! My 3-year-old video card uses 160 watts of power and roars like a jet engine running CS:GO at 1080p.

It blows my mind that we've come so far so fast.

The trend lines showed a year ago that a high end smartphone in about 3 years from now (mid-way through the PS4's lifetime) will have higher paper specs than a PS4 (paper specs, not necessarily as well-utilized).

An Nvidia Shield tablet is on paper roughly equivalent to an XBox360. The Shield set top box is double that. A few more doublings and the prediction will play out.

3 years from now, the PS4 will be 5 years old and much cheaper than a high-end smartphone.

yes, but a high end smartphone will always be cheaper than a high end smartphone and a PS4.

the point isn't that phones are cheaper than other things, it's that you need one anyways.

You may need a smartphone, though that's debatable, but you don't need a high-end expensive one. And those wont't have paper specs anywhere near PS4-level anytime soon.

>higher paper specs than a PS4 That's not very difficult. My midrange graphics card beat it one year before the PS4 was even released.

Does your midrange graphics card run on a battery that fits in your pocket? That's the hard part.

The GPU in a Shield tablet uses literally the same tech as a 980. But, the best that tech can do on <10 Watts is to keep up with an Xbox360.

Sure, but there is no way the thermals will be there.

If you take a look at games like Deus Ex the Fall which was also ported to PC (with mobile assets even though they could have just taken the high res assets from Human Revolution) you'll see that it's not even close.

You may not notice it on a tiny screen but on a monitor or a HMD it's absolutely terrible. It looks like a game from 8 years ago.

The key overlooked point for VR - a point that Carmack had enough pull to persuade Samsung to address - is latency. http://oculusrift-blog.com/john-carmacks-message-of-latency/... The system must render "motion-to-photons" the image in less than 20ms to avoid discernible (& disorienting) lag. That's fast enough that the graphics system doesn't have time to buffer the image (which would cut available render time by a large percentage); Carmack was able to convince Samsung of the importance of "unbuffered display" in a market where otherwise nobody cares about such minute timings.

I had the chance to use the first generation of this setup together with a variety of other VR headsets a few weeks ago at a VR film festival. It was fun and worked pretty well, but non of the screen that I got to try had the needed pixel density. I was able to see the pixels on all of them, because the screen is so close to your eyes.

The screen differences aren't that far off between Gear and Rift CV1, but the biggest difference (aside from processing power) is positional tracking. You only get directional tracking of the user's head on the Gear, but games/experiences can be designed around current limitations.

The other frustrating limitation to the current generation of headsets is field of view, but FOV is fairly similar between the two as well.

IMO the value of each of these platforms is going to come down to content.

The screen is literally the same as what you get in a DK2. Oculus couldn't get any manufacturer to make them a screen that qualified all their requirements. Then they discovered that if they pry out the screen from a Galaxy Note, it is nearly the best they can hope for. So every DK2 right now just uses a screen that they literally tore out of a normal Galaxy Note.

I owned the Note 4 Innovator Edition. Its better than you'd think. Very low latency, excellent tracking, solid design. But a noticeable Screen Door Effect is a downside, more so than resolution IMO. You may not get full presence, but undoubtably its worth a mere $99.

Screens on these phones are pretty advanced. Even the Oculus Rift is just using screens designed for phones. From what John Carmack has said in the past, mobile platforms are easier to control for latency and lag than PCs.

it's effective enough for the wow effect, and that's pretty awesome because that's how VR might become mainstream.

Reply on news.ycombinator.com