Glasses-free 3D HDTV in five years?

[metalqueen]

Glasses-free 3D HDTV in five years?

3D HDTVs have just gone on sale, but they may be as obsolete as a gasoline-powered car in five years.

The powers that be are already working on glasses-less 3D, aka autostereoscopic 3D, with standards expected to be set in around three years with another two years or so for commercialization. So we could have 3D with no glasses by Christmas 2015? We're in!

Here's how it might work:

There are two distinct developing technologies for glassless 3DTV — the 3D video codec and the plasma/LCD display — that have to merge to create a new autostereoscopic 3D ecosystem.

Current 3D is encoded using an extension of MPEG-4 H.264 AVC called MVC (Multiview Video Codec). MVC fuses two nearly simultaneous HD "views," one for the left eye and one for the right. Overlaying these two synced "views" using active shutter glasses creates the perception of 3D depth.

Read the whole story here: http://dvice.com/archives/2010/03/glassless-3d-hd.php" onclick="window.open(this.href);return false;

[DmitryKo]

Dream on, glasses-based stereoscopic displays will be as obsolete in 5 years as gasoline-powered cars will be obsolete in 5 years (Hint: even hybrid gasoline-electric cars use gasoline, and expecting hydrogen-powered electric cars to totally replace gasoline ICE cars in just 5 years is wishful thinking at best).

The problem is not with video coding technology (BTW, it's Multiview Video Coding - is it that hard to remember?), it's with display technology that could offer high-quality autostereoscopic viewing and not involve some kind of limitation, such as ridiculous bandwidth and processing requirements, way too low resolution, or single-viewer only setup. Examples:
At CeBIT New Display Enables 3D without Special Glasses
Zecotek Demonstrates Auto-Stereoscopic Multiple Views
SeeFront 3D Display works without Glasses

[MadRussian]

Hmmmm...

I for one would be ecstatic to finally have a single-user auto-S3D system in my grasp. With head-tracking, the ability to look around objects becomes a reality. Bring it on!!!

Oh yeah, and make it semi-affordable. I would, no doubt, pay upwards of $1500-$2000 for 22-24" free-look no-zone auto-S3D if the display quality / drivers were there. Curious, how much would others be willing to pay for such a display?

btw - Wasn't the ELSA Ecomo 4D (the only other monitor I'm aware of that was similar in capability to this feat), around ₤ 20,000 back in the day? Maybe with all the momentum towards S3D currently, these enormous costs will drop and drop quickly. We can only hope!

[cybereality]

Serious, mainstream autostereo displays are at least 10 years away by my estimates. And even then it would require a massive change in the content infrastructure, it just doesn't seem realistic at all for the near future. I could see for select purposes (like medical/scientific imaging) maybe sooner, but not in the mainstream sense. These kind of articles really do not help 3D adoption at all.

[MadRussian]

@cybereality

Thanks for the interesting viewpoint. However, what do you mean exactly about a "massive change in the content infrastructure"? By my reckoning, regarding content for single-user auto-S3D displays, there's no extra bandwidth or anything like that required...

I think it all comes down to cost. If enough companies take on auto-S3D and get their product to market, that will drive the costs down. At some point, the niche users like us buy in. Then when the cost is driven below a certain threshold, everyone jumps on-board.

After reading again your last post cybereality, now I am even more curious. What would you personally pay for such a single-use auto-S3D 22-24" no-zone head-tracking display, provided the quality was there?

[DmitryKo]

I for one would be ecstatic to finally have a single-user auto-S3D system in my grasp. With head-tracking, the ability to look around objects becomes a reality.

That's until you realize that head tracking is supported by very few games, mostly flight sims, and that this custom display would probably cost in the tens thousands (dollars or euro, doesn't matter).

What would you personally pay for such a single-use auto-S3D 22-24" no-zone head-tracking display, provided the quality was there?

Nvidia 3D Vision can be had for $600 (120 Hz monitor + shutter glasses). I just can't see any of the above-mentioned autostereoscopic vendors coming down to this level in 10 or even more years. Prohibitively high costs are the reason why these displays are currently intended for medical instruments, not the consumer and PC gaming.

[MadRussian]

That's until you realize that head tracking is supported by very few games, mostly flight sims, and that this custom display would probably cost in the tens thousands (dollars or euro, doesn't matter).

"Game support for head tracking" and "auto-S3D display head tracking" are two completely different things. Game support for head tracking is intended to allow the user to look all around in the game, ala TrackIR (http://www.naturalpoint.com/trackir/). Very useful for flight simulators, etc.

In contrast, with the right drivers, auto-S3D display head tracking should allow the user to look around objects, within the constraints of the display specs (and not to mention provide the comfort/quality we've all been waiting for).

Nvidia 3D Vision can be had for $600 (120 Hz monitor + shutter glasses). I just can't see any of the above-mentioned autostereoscopic vendors coming down to this level in 10 or even more years. Prohibitively high costs are the reason why these displays are currently intended for medical instruments, not the consumer and PC gaming.

Well I guess we'll see. $600 for the mainstream is not unreasonable, but I think we, the niche users, will go quite a bit higher for a quality auto-stereo product.

To me, there's BA, and AA, before and after Avatar. Suddenly there are a lot of minds / companies working to produce glasses-free 3D, and there's certainly an enormous goldmine in profit to be made for whoever gets the comfort/quality to market first, at a reasonable price. Somehow based on the frenzy we've been seeing develop, I think we'll have these displays sooner, rather than later.

[DmitryKo]

"Game support for head tracking" and "auto-S3D display head tracking" are two completely different things.

Not really. Both require in-game support and every head tracking device or motion tracking device is just an input controller which can be used in different input modes, such as character movement, turning your head around, or look-around with head movement. It's just the matter how do you program the game to respond, and both PS3 motion controller and Project Natal can certainly be used in many different ways.

Unfortunately look-around (moving your head to the side of the screen) would be very hard to implement in drivers, since current games do not support fully independent head movement - most games do allow look-up/look-down, but to look left or right, you move the body of your character.

Most games cull in-game objects that do not fit in the expected view frustum, so the drivers would not even receive 3D information needed for rendering these previously invisible objects.

Also, changing the view frustrum in drivers can lead to weird rendering artifacts, since in-game graphics depend on correct viewing angles which are passed to graphic shaders for calculating various effects, so there should be some highly intellectual (or even manual per-game) shader rewriting like the one they use in NVidia stereo drivers and probably iZ3D drivers as well.

there's certainly an enormous goldmine in profit to be made for whoever gets the comfort/quality to market first, at a reasonable price

So far I haven't seen autostereoscopic technology that does not suffer from some weird limitation.

People seem to expect that "glasses-free" stereo 3D in Full-HD quality for multiple users will be available "soon"; this is not going to happen, just like hydrogen-powered cars won't be a commodity for the next 30 or 50 years. Just look how much time it took for complex display techologies like take up to a "sub-$1000 consumer device" level (plasma panels - 40-45 years, CRT - 35-40 years, active-matrix LCD - 30-35 years, DMD - 15-20 years, SED and FED - 15-25 years in research, commercialization still uncertain).

I remember reading in a 1989 issue of a scientific journal that the US military had pruchased a large quantity of 20" or so monochrome flat-panel display in bright orange, based on plasma panel technology, for use in Pentagon situation rooms; I recall they cost about $50 000 or more per panel. I think autostereoscopic displays are currently have just recently achieved a similar stage. It would take far more than 5 years for them to become a commodity.

[cybereality]

Unfortunately look-around (moving your head to the side of the screen) would be very hard to implement in drivers, since current games do not support fully independent head movement - most games do allow look-up/look-down, but to look left or right, you move the body of your character.

Actually the DDD TriDef drivers already support this. However, since they use a webcam for the tracking, it is not very accurate and thus not fun to use. But it does work from a purely technical standpoint.

[Dom]

Hi I kind of agree but mainly the entertainment electronics field and market need the money and captial from their sales of now and soon technology like the new 3d tv's and especially the 300 dollar glasses to be able to afford to make the new autostereoscopic tv's for 1200 dollars. They also need to be able to have the panels made with their own patents so other companies can't claim their patent cheque when duties arrive on licensed patents. Probably texas intsruments will make their dlp autostereoscopic projector sometime when a science breakthrough is engineered, mostly just software and 3d pixels layered. Go T.I.

[DmitryKo]

Actually the DDD TriDef drivers already support this

Even the developer admits it doesn't work as good as expected and the users point out multiple gameplay issues which you can't really work around without in-game support.
http://www.ddd.com/forum/viewtopic.php?f=2&t=167" onclick="window.open(this.href);return false;

They also need to be able to have the panels made with their own patents so other companies can't claim their patent cheque when duties arrive on licensed patents

The problem is not with licensing fees, these costs are negligible for mass production. It's about lack of autostereoscopic technology that is really ready for mass market.

For example, mechanically shifting lenticular lens arrays may sound good in theory, but in practice, what about reliability and cost of repair? I wouldn't like it if my multi-thousand monitor fails when someone accidentally touched the screen with a finger. Etc.

[craylon]

I think predictions in a 5-10year window are pretty unreliable anyway since there are rivaling techniques developing towards mass market at the same time.

From the back of my head I would think multi-user autostereoscopic 3D would compete with
a) holographic technology
b) augmented reality glasses
within a 10year time frame.

At the moment we should be happy that s3d is going that strong and if s3d becomes the standard in content production/consumption within the next 3 years it will also bring money to research and development to these other area pushing for improvement. We also don't know if people would just get used to wearing glasses after a year or two when it becomes obvious that almost all content (sports, entertainment...) is vastly enriched from the s3d experience.

My personal favorite prediction btw. would be that in 10 years we all wear augmented reality glasses i.e. lightweight, mobile huds in form of glasses that combine stereoscopic presentation with augmented reality. I believe that being forced to gather around a device like a screen to play games or view content could be thing of the past someday.