Through the Looking Glass – holographic display hardware is great, but it's not enough Four years ago in a feature for The Register, I wrote about the latest technologies for three-dimensional photography and videography. At the time, the tech required an array of tens to hundreds of cameras, all pointed inward at a subject, gathering reams of two-dimensional data immediately uploaded to the cloud for hours of …
Indeed. Nobody has yet produced a fully effective static 3D display - not only the old artworks were questionable, the quality of the 3D rendering was as well. Personally I'll want to see this new system finding a sustainable market for simple printed 3D images before I'll get hot under the collar about e3D, let alone the groaningly cliche'd fourth dimension.
I suppose a decent lensing system would use little blobs rather than ridges and multiple spotted rather than striped images. Oh, and how many pixels to display all the different views per blob? I can see depth pixel resolution being a thing. But wait, what?! You mean we have to wear a special locator on our forehead so we look like a Brahmin, and the 2-px depth display is adjusted on-the-fly for just the one viewer; a film audience would each need their own screen?!
Not quite a tankful of fish, either way.
to Flatland. A charming book although the prose does somewhat date itself. I'd also recommend Flatterland which I believe was written early 2000s (I'd check but it's probably filed in my bookcase between "I know it's here somewhere" and "Ooh! Forgot I owned that").
As for 3D displays - I just don't see it being possible without becoming some uncanny valley/nausea inducing effect.
We have achieved videoconferencing for the masses. We can even use our smartphones for that now, something that the phone industry promised decades ago but failed to deliver.
The result ? Everybody hates it.
Now, we are witnessing a new gimmick : pics which have some depth to them. To make it work, you have to have bespoke hardware. That is very much going to limit the market. Yet another thing with batteries to follow.
It's an interesting idea, but I want it on my 26" widescreen.
Seems to me that 3-d displays are about as much use as smell-i-vision: nice idea until you realise that actual use of it is pretty awful.
3D headsets may be awesome especially since they spawned a whole sub-genre of "fail" videos on YT with people walking into things etc.
3D on flat screens, using some means of sorting the light between left and right eye, using colour ("it came fro outer space"), polariser goggles (cinemaxx etc), or lenticular film on screen (Nintendo 3DS, "looking glass") all suck and cause varying levels of nausea.
The fundamental disassociation of our visual systems's distance sensing through parallax (tick) and eye focus (non-tick) remains the same.
If i understand it correctly, an _actual_ hologram (*) doesn't have this issue though it has a ton of others, it's a static, analog, black-and-white photo after all.
(*) stop calling anything 3d a hologram btw., just sayin.
At Las Vegas several years ago my son and I saw a 3-D display for gambling (what else?). Sensors located where your head was and displayed appropriate images through a lenticular screen. It was all computer graphics. It looked pretty good though.
The system got really confused when both my son and I were within range of the sensors so single person use.
My phone doesn't do fancy depth information, so can somebody explain to me how that works?
You still need at least two separate shots from different angles to have an image of whatever is hidden behind an object in other angles, so how do those phones do that? Just having depth information for a single shot can't be enough by itself.
I mostly work with iOS but this is my field,
iPhones with dual cameras will use 2 cameras to get depth of field, single camera iPhones. won''t do it.
In addition phones that support FaceID have a depth camera on the front facing camera.
And now there's LiDAR enabled phones too.
Android has the depth API, which can work with a single camera, it just requires a bit of movement so it can compare multiple frames to derive some depth data, from our experiments though, it is nowhere near as accurate as the iPhone. Some android devices also have time of flight sensors, which really improve the results.
For a single image though, you are correct, they will not show information that's behind other things, in which case you can build a model using say a photogrammetry tool, or now, with phones with depth sensors, you can now build a model in app.
But I suspect, even from a single image, there are probably tools that can reconstruct missing background information with a bit of AI, but it's not my field.
If I understand correctly iPhones have the proper sensors to make an actual depth map and 3d image. Android just sort of approximates it for the purposes of blurring backgrounds and such so they can simulate DoF, but their "depth" map is more of a chroma-key matte of "this is foreground, this is background" than a proper 3d image. Not really useful for 3d images like this.
my brother has logitech webcam with "fake bokeh" software, it's hilarious: a houseplant in the background behind his head is interpreted like it's a fancy hat, so it's in sharp focus along with the face, in front of a blurred background. Oh the fun a simple mind like mine can derive from a zoom call...
"even from a single image, there are probably tools that can reconstruct missing background information with a bit of AI"
AI paint by numbers? Lovely. I'm sure the user's mummys and grans will tell them how much of an "artist" they are and stick it to the fridge with a gold-star magnet.
Hi jake. Using automated tools to separate subjects from the background of an old photograph - and then using other tools to fill in the 'missing' background - is often used in televisual history documentaries these days. The illusion of parallax twixt foreground and background when the virtual camera is slowly 'panned' really helps the old photographs 'come to life'.
If you look out for this effect you may spot it.
I think the search for depth of field to create a three dimensional impression is ultimately something that won't get the uptake that researchers think it will, because it is just that, an impression, a cheat.
Also, because visual and auditory processing cue each other to some degree, visually 3 dimensional reproduction needs to work hand in hand with the audio.
Until 3D is actually an object I can walk around it will remain a gimmick for me.
"Until 3D is actually an object I can walk around it will remain a gimmick for me."
Liked, but really, even if it were, I argue that it would still be a gimmick.
For uses like movies, games, people are not interested in walking around a tiny object, they want a world around them, to be immersed in it.
For video calls? I do mine sitting on my couch in front of my TV: not useful for me. For people using a phone, are they really going to keep moving it from side to side to have a better look at the new hairdo of the caller? No, because if they do, the caller is going to complain that they can't see anything.
For portraits? The electronic kind never took off, people simply aren't going to replace their printed family pics on their walls by multiple expensive devices that require power.
There are surely some specific niche market, maybe tactical displays for the military, but I don't see a mass market use for those.
But sadly it isn't going to stop Marketing from selling it, existing or not ... nor is it going to stop suckers from buying into it.
The only remaining question is whether or not I want to ignore my scruples long enough to make bank off this latest mostly useless fad.
"For portraits? The electronic kind never took off, people simply aren't going to replace their printed family pics on their walls by multiple expensive devices that require power."
I wonder if the black/white rotating balls thing of e-ink could be repurposed in some way to use adjustable microlenses for "pixels" to give a sort of static 3D image display?
Stereoscopic TVs, mis-named 3D are not remotely 3D.
Lenticular displays are better than stereoscopic but lower resolution. The more angles of view the lower the resolution. There is no secret sauce. Compared with a spinning drum with an offset screen with the projected image changing as the drum rotates this is abysmal.
Also any flat display fails because the focusing of the eye can't change for the distance away of the object. A big clear spinning cylinder/drum with the offset ground glass screen has short comings (everything is transparent) compared to a real hologram.
It's a toy. It's not a proper 3D display and nothing remotely close to a holographic image!
You can do a lot of amazing 3D looking effects simply by doing head tracking
This YouTube video from 2007 shows it well - https://www.youtube.com/watch?v=Jd3-eiid-Uw (The 3D parts start at 2m30s in)
It did, well sterescopic like the display in this piece. The new 3ds even did it properly, by tracking your eyes and adjusting itself to keep you in the sweet spot.
Great feature, worked well on most games that used it, but it's just not enough of a wow to make people buy. By the end of it's life, the last pokemon game for example, they'd stopped bothering with the 3d render.
You need true depth of field for 3d display to be a system seller.
People don't want it. Maybe if you can do something like R2D2's Princess Leia message people will want it, but getting depth or "looking around the corner" on a flat display is something the general public doesn't want.
Some type of AR/VR where you move your phone/glasses/self to go "around" objects is where things are going, not yet another failed attempt at 3D TV.
Sorry, but this looks like the answer to the question of why 3d projections don't just look like real objects. I can't think of a use case beyond novelty until it improves. Any of the CAD modelling I would have guessed it'd be good for is ruled out by the poor resolution.
What is this 4th dimension he speaks of? Last I knew the "4th dimension" was time! Which we already have and do not need a display to experience it. Is he talking about some strange 5th dimension?
Anyway, I fail to see why I would need 3 dimensional display. To watch a movie? Nah! Play a game? cool, but we already have that for those who think this is important.
3 dimensions doesn't really give you much until you include all those other things we don't get from a screen. Touch, smell, taste, interaction! Now you bring all that together and you got a winner!
And when this happens, how do you know which of the competing technologies to invest your money in? Whichever one the porn industry standardizes on!
I am very familiar with 2.5 and 3D display technolgies because I have spent a lot of time working with 3 and 4D data. Affordable '3D' displays have limitations. However, there exists a class of display that can do what everyone dreams of. Problem - most mortals cannot afford them and may never get to see them 'in the flesh' so to speak. 2m diagonnal full colour UHD 'holographic' display (more reasonably call it a window to elsewhere) - yes it exists, right now, has done for a while.
Problem - fragile, incredibily low yield, some 'interesting' substances required in manufacture and a single display costs way north of a car park full of CT-5 Blackwings. Not for plebs, not even for Disney, for people who have a real need to do this now. The full colour is vertical, 'infinite' depth and for a static, changable scene. There is a reduced palette version that can do full motion 'table top' geat for urban ops. Who and where was under NDA.
Alternatives probably in someone else's lab. I hear promise of 'Light Field' display surfaces that trump 'Looking Glass'. We will see (sic). THe display is not the only problem - capturing and crunching the data and then moving it around is a biggy. If your broadband struggles with 4K VOD just wait till you learn what you'll need for live action holography.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
