auto s-3DTV "8-10 Years Away" From Taking Off,James Cameron

[Silversurfer]

While glasses-less 3DTV is already winging its way to Japan in the shape of Toshiba's two tiny TVs, James Cameron reckons it's "eight to ten years away" until it autostereoscopic 3DTV properly takes off. For now, we're stuck with glasses.

He spoke about 3DTV at the Blu-Con 2010 conference today, and shared some interesting details on the extended version of Avatar—apparently those extra 16 minutes ended up costing Fox $1 million per minute, which is the reason they chose to re-release it in cinemas, so as to try and recoup some of the costs.

Cameron has previously harped on about how the industry needs to shoot more in 3D, rather than applying it post-production like the studio of Clash of the Titans, and the latest Harry Potter movie did—or attempted, in the latter's case.

"You see another stumble with the most recent Harry Potter movie from the same studio making the same mistake - except really getting spanked for it now because they didn't get the film done.

They announced it in 3D - threw a bunch of money trying to convert it to 3D in post-production and it simply didn't work. They just didn't get it done."

Cameron's signed on to do Avatar 2 and 3, which he'll begin working on next year. While I do feel that particular story's been flogged to death, Avatar was probably the greatest example of 3D. While I'm not personally a fan of the format I do await the day when I don't have to squeeze a pair of 3D glasses onto the top of the glasses I already wear—nor fork out over $100 per pair of glasses. [BBC and TechRadar]

http://gizmodo.com/5680637/glasses+less ... Gizmodo%29" onclick="window.open(this.href);return false;

[Silversurfer]

as Nelly and Christina aguilera would say "in the song "Tilt Ya Head Back',' he's right you know, he's right". Curtis Mayfield Riff never sounded so good

[Fredz]

How can you say he's right, did he give any new information to bake his affirmation that autosteroscopic TVs are eight to ten years away ?

Autostereoscopic techniques have been invented more than 100 years ago (1896 for parallax barriers, 1908 for lenticular lenses) and have resurfaced each decade (Didik 3D-TV1 in 1978, Alioscopy in 1987, DTI-100M in 1990, Sharp LL-151 in 2004, Toshiba displays in 2010) without _any_ solution to the limited viewer postion. Why would it be different in 8-10 years ?

[Likay]

Fredz right. As is it's needed to circumvent physical laws or find a new brilliant never thought of way to do it. The latter is very far from being realized though.
For autostereoscopic displays the multiview seems the most plausible solution since it seems that 2d+depthmap is quite low hardwaredemanding after all. The other obstacles as low light, bad image quality/focus still remains. Maybe Cameron knows something we don't? Imo in this case i think he might be too optimistic though.

[cybereality]

I agree with Cameron. 10 years sounds about right. Although they have had auto-stereo in some form for years now, I think in the past few years there has been a lot more interest in the field. You can see with the Nintendo 3DS coming out next year, that we will finally have a mainstream product from a respected company utilizing this technology. I know Sharp had that auto-3D laptop a few years back and things like that, but it was a very niche market. Only a small group of people even knew it existed. So I think after the 3DS many more people will be exposed to auto-stereo and thus become fans of 3D. But it will probably take another 5 years or so to make a major breakthrough in this field and then a few more years to commercialize it. But it will happen. People will make it happen. I mean, just look at that 3D tube-looking display Sony was showing a few months ago. There are major companies with interests in perusing this.

[tritosine5G]

i don't care.

There are enough blokes who care about glasses ( and shining their own "ego", while they re at it ) way too much.

How about games + DLP + 150" + 2.39 : 1 cinemascope screen + 120fps

(DIY anamorphic Lens)

This is one thing 1080p is useful for. Its still overrated though.
http://seed-studios.com/undersiege_resmodes.jpg

[Fredz]

I agree with Cameron. 10 years sounds about right.

On which basis ? There've been no progress at all in autostereoscopy for the past 100 years, only adaptations to evolving display technologies (paper -> CRT -> LCD).

Although they have had auto-stereo in some form for years now, I think in the past few years there has been a lot more interest in the field.

Interest from the masses won't magically create a technological breakthrough. And interest in stereo have been even greater in the 50's, Russian even created autostereoscopic cinemas at this time. Without much success again...

There are major companies with interests in perusing this.

And ? Such companies are only interested in creating new markets, they've never invented the technology in the first place. Progress in this field has been made exclusively by academic research.

All the stereo 3D equipements we use now have been invented in research laboratories more than decades ago. A recent example being active stereoscopy used by the recent AUO 'deadzone-free' laptop whose technology have been used since 2000 in universities.

If we'd seen recent breakthroughs in scientific research concerning autostereoscopy I could've been optimistic, but unfortunately there's been nothing relevant in this field for the past few years.

[cybereality]

Well I will be honest, I have no facts to back up my claims. Just some hope and a dream. You gotta believe!

[PalmerTech]

I mostly agree with you Fredz, but you are wrong in one regard.

Autostereo is perfectly possible. It is perfectly possible with multiple viewing angles. In theory, it would work with hundreds of viewing angles!

The only problem is resolution. If in 10 years we have LCD panels with, say, 100x the pixels we do now... Well, then autostereo is perfectly viable. In theory, different views for different angles would work, too.

[Fredz]

The problem is not much viewing angles, displays have been able to use 9 viewing angles for a long time, it's that when you are in a good viewing position you can't move your head much else you'll see inverted images.

This problem can be resolved for one (Varrier) or two (Dynallax) people using eye tracking but not for a greater number of people. And beside resolution loss there are still other problems to resolve like banding effects, fixed viewing distance or brightness and contrast loss. Without forgetting that like stereoscopic techniques, autostereoscopy cannot handle the vergence-accommodation conflict.

[PalmerTech]

You know, I pretty much agree with you. I will admit I was just looking for an argument.

Polarized contacts! Somebody make them!

[cybereality]

You guys are thinking too limited. I don't think auto-stereo will forever be based on parallax barrier principles, which inherently sacrifice resolution. For example, what if each pixel could emanate a different color based on the angle you look at it. If this were possible you could have a full resolution 3D display with say 180 different viewpoints. I don't see why this wouldn't be possible.

[Fredz]

Yes, having a pixel emanate a different color depending on the location of the viewer is the goal to attain, but none of the autostereoscopic techniques we know today are able to do this for more than a limited number of viewers (parallax barrier, lenticular lenses and autostereogram). Holography can do that but it's not an autostereoscopic technique nor a stereoscopic one, even if recent progresses in this field have been quite encouraging.

I don't say it's impossible, but if you do have even the slighest idea about how to do that, don't hesitate to keep us informed. That's not like if dozens and dozens of scientists had been trying to attain this goal in the past 172 years...

[Likay]

It's actually all about nature laws. I don't know if people are willing to sacrify 90% of resolution, quality and brightness for a limited autostereoscopic display compared to a 2d one...

[tritosine5G]

Hey , dont you remember those LCD premises over there at avsforums.

Maybe its true!

[skrubol]

Eventually display resolution is likely to become a moot point. Your eyes can only resolve so much. Why not use that extra resolution on autostereoscopic? 8-10 yrs seems like a reasonable time frame for the cost delta between 'good enough' resolution and 10x that to drop to very little. Eye tracking also should be good enough at that point to be able to show the right image to several people no matter where (to an extent) they are standing.

[yuriythebest]

Eventually display resolution is likely to become a moot point. Your eyes can only resolve so much. Why not use that extra resolution on autostereoscopic? 8-10 yrs seems like a reasonable time frame for the cost delta between 'good enough' resolution and 10x that to drop to very little. Eye tracking also should be good enough at that point to be able to show the right image to several people no matter where (to an extent) they are standing.

I think in 8-10 years people would have just barely all switched to stereoscopic 3d. Autostereoscopic displays will probably just begin to become available at reasonable prices

[BlackShark]

I hear the voices of the high-end home theatre folk wanting 4K cinema more than 3D, glasses or not.

Unless there is a major technology breakthrough, the day we'll have multi-view auto-stereoscopic 3D capable of outputting 1080p, we'll also have giant 4K3D displays with glasses for the same price.
That day, you'll be asking yourself :
"do I buy the big 4K3D TV with glasses or do I buy the small 2K (fullHD) glasses-less TV ?"
That's a tough choice, especially considering my recently acquired 106" screen which I absolutely love.

[Fredz]

And that same day we'll also probably have real-time holograms, so the choice will be quite simple...

[BlackShark]

Naah, we've been told those were at least 20 years away so it will take 10 more years

[cybereality]

Even if the real-deal free-floating holographic displays came out tomorrow, we have no infrastructure to produce or distribute content like that. That means all the existing catalog of 3D movies and games would be incompatible. I guess if we also had holographic cameras we could start filming some movies, but then thats another 1-2 years just to wait for the first holographic movie to come out. And how are they going to edit that movie? We would need all new computer software and techniques to do this. I cannot even imagine how you would render a game in full holography. That would take some serious horsepower. So I think we will be stuck with 2 view stereo 3D (or multi-view interpolated from 2 views) for the foreseeable future.

[PalmerTech]

Not to mention the fact that the field of depth can only go as far back as the holographic projection field. Suppose you have an absurdly large 10 foot by 10 foot projection area. How in the world would you deal with, say, mountains off in the distance? Or buildings a mile away? Besides all the problems Cybereality brings up, you still have that fundamental flaw of the medium.

[yuriythebest]

Not to mention the fact that the field of depth can only go as far back as the holographic projection field. Suppose you have an absurdly large 10 foot by 10 foot projection area. How in the world would you deal with, say, mountains off in the distance? Or buildings a mile away? Besides all the problems Cybereality brings up, you still have that fundamental flaw of the medium.

well, it could be cheated by offering objects near to the viewer in "full 3d" but as the scene gets further from the viewer it becomes flatter and flatter

[ERP]

The only vaguely promisin technology I've seen for auto S-3D recently is time multiplexed directional backlights.
Microsoft demonstrated this recently and I believe I read somewere that Toshiba has demonstrated something similar.
Still has a lot of issues, reduced brightness, limited number of users, reqires adequate head tracking/people changing places etc.
But since the technolocy was originally developed to reduce power consumption of display backlights, I could envisage some variation of it being practical in the next few years.