I always feel bad for companies trying to show off a higher resolution screen -- if my screen has a lower resolution, by definition it's kind of hard for my screen to accurately portray its quality; at best I can only see what my screen can deliver.
The photos in this article do the best job I've seen; they show close-up images of text at various DPIs, demonstrating the visibility of the pixels at lower DPIs. (It would have helped to have an image of the same text to view directly at normal size rather than zoomed in.)
Two answers, biggest is yield (number of 'good' screens made for a given run) and the second is bandwidth.
Assuming a 16:10 display a 4096 wide display would be 2560 lines 'tall' and have a total of 31,457,280 pixel elements (r,g,b). Each of those would include a transistor as well so another 31.5M transistors, coming out to driver logic at the edge of the screen. At 498 ppi that would be about an 8" x 5" screen.
To drive that screen at 60hz you would have to push out 31.5M * 60 bytes of pixel data (that is for 8 bits per color, higher fidelity 10 bits per color would be more. At 60hz that is a bandwidth of 1.8 GBps or 18 Gbit/sec. This is faster than the current DVI spec is rated for, although I believe dual-channel can get there.
The yield rate inversely increases the cost per good screen, so a yield rate of 50% doubles the cost, and a more likely yield of 25% quadruples the cost. This can be offset somewhat by bigger glass (more chances to succeed, semiconductor manufacturing is in many ways a game of statistics) but bigger glass may require different factory layouts (they might have to build a factory just for this type of display).
So Nvidia helped make some progress when they announced the Tegra-3 could drive 2560 x 1600 displays. That's more than halfway to your 4K display. And this will be the 'killer' display technology in the next gen 10" tablets.
If the tablets can keep up the ship rate then they will give manufacturers a better feeling about being able to sell enough even higher resolution displays to offset their costs and that is when we'll see something like this demo come out in a larger size.
30" @ 2560x1600 translates to 101 ppi [1]. Apple's Thunderbolt Display is 27" @ 2560x1440 (109 ppi) and costs the same. A 24" screen @ 1560x1600 would be 126 ppi, the same pixel density as the screen of a 13" MacBook Air. [2]
All of those are a far cry from Toshiba's prototype (495 ppi) or the screen Apple uses in its iPhone and iPod touch (330 ppi).
Although Sony has a thunderbolt display with the graphics card in the display not in the computer, so the long cable just sends higher level graphics commands and the bandwidth problem is over short proprietary stuff inside the display...
I thought 4K referred to the vertical resolution. If so, presuming a consistent aspect ratio, 2560x1600 is only 16% of the way to 4K in bandwidth usage.
That's what I'm wondering. I don't really care about these small displays with huge resolutions because I don't find a huge benefit beyond a slightly better picture. I remember going from 320x240 to 800x600 on a 15" and my mind was blown. That's because my icons went from taking 1/4 of the screen to actually being an 'icon'.
And it was $8,400 at release, and consumed 150w of power. It was also thick and heavy. Therefore, like I said, it would be very thick, heavy, expensive if produced even today.
Aside from the price and consumers only wanting the cheap crap there are no user interfaces that can handle it. No desktop GUI can scale well with high DPIs so you can only use it in niche applications anyway.
From John Siracusa's excellent OS X Lion review [1]:
"Resolution independence has been "coming soon to Mac OS X" since 2005. The dream of drawing the same interface elements at the same visible size but with more pixels was so close in 2007 that we could taste it. Then Snow Leopard arrived and the Mac's interface scalability features actually regressed. Depressing.
Meanwhile, Mac OS X's sibling operating system waltzed right into a high-resolution UI on its very first try. iOS's secret? Don't try to support arbitrary scale factors, just support one: double resolution. A 50x50-pixel square on a non-retina iPhone screen is exactly the same size as a 100x100-pixel square on a retina display. Graphics that have not been updated for the higher resolution are simply drawn with four-pixel squares in place of each low-resolution pixel. All dimensions are nice, even, integer multiples of each other. This is a perfect fit for physical screens which, of course, have an integer number of pixels. Fractional measurements necessarily require ugly compromises.
Lion has taken the hint from its younger brother. Arbitrary scalability is gone. In its place is a single check box to enable "HiDPI" display modes. (This option is still hidden away in the Quartz Debug application, so it's clearly not an end-user feature. But unlike all previous incarnations of resolution independence, this one actually works.)
After enabling HiDPI, new display modes will become available. In the screenshot above, the 720x450 mode is half native screen dimensions, and the 640x400 mode is half the (non-native) 1280x800 setting. After selecting a HiDPI mode, everything is drawn with twice as many pixels as its non-HiDPI equivalent. Here's a screenshot featuring TextEdit, our usual interface scalability workhorse.
It looks pretty good, right? The only flaws are the bitmap graphics that haven't been updated for HiDPI (look closely at the black triangles in the ruler). Unfortunately, there are a lot of these throughout the operating system and its bundled applications. But unlike in all years past, the framework is finally there for third-party developers and Apple itself to finally get their applications ready for a world in which 300-dpi desktop and laptop displays are more than just expensive curiosities."
First thought was "meh", requiring four times the pixels is quite steep. But after giving it some thought that might be as good as one can hope for and considering that the pixel-graphics of today can't really be made to look good scaled up 10-90% it probably is one of the better (the best?) solutions as well (forcing vector graphics is going to be a nightmare and it won't look good on standard-DPI-monitors anyway).
(the biggest drawback I guess is that it doesn't really do anything for people wanting a larger interface for readability reasons, having everything four times as large is a hard blow for the real estate when you really wanted 120% or something)
What the author constantly mixes up is that you need four times as many pixels, not twice. And thus we need ~400 DPI to have everything at the same size as today. But yeah, just hope that it becomes mainstream enough to be affordable, OLED to the rescue? Thanks to displayport the interconnect won't be a problem at least.
Very impressive technology, but I agree with the conclusion in the article: 6.1" seems like an odd niche. Too large to fit in most pockets, so it doesn't work for a smartphone, but a bit on the small side for a tablet.
The same 494.9 DPI would allow for a ~4.5" 1080p display, or ~3" 720p display, both of which seem far more practical for smartphones. For a tablet, the 2560x1600 10.1" screen from Samsung seems sensible, and personally I'd settle for a 10" 1080p display to reduce cost.
(All of this and still no 12" laptop displays with better than 1366x768 or so.)
I found myself wondering if it was for a vehicular application.
I've never had a car with a built-in LCD info/nav/etc system to play with them, but I can imagine 6" being a plausible size for some given other standard sizes. Because it's bundled as part of a much, much larger whole than with a phone it's likely able to support a potentially higher price point and it wouldn't surprise me at all to learn that a high-end car brand felt a 500DPI screen would be a good differentiating factor from competitor 100DPI screens. 'Read as easily as paper' is a compelling selling point.
I imagine that this display is merely a proof of concept. There aren't many electronic devices that use displays of that size, but perhaps 6.1" was the smallest that Toshiba could scale a 1600p display.
Or perhaps the electronics required to produce such dense pixels are still quite large, and whose rectangular dimensions wouldn't allow for displays of the same density without a wide bezel.
What's the proof of concept with high res screens, they've been around for some time, just not in any consumer products. I saw some 1280*1024 2" screens several years ago. From texas instruments, i think its main use was in projectors, i don't know if it could have been used for displays but from the demo it looked exatly as an lcd but semi-transparent.
6" sounds pretty optimal for e-books i would say. ipad is too big and smartphones are too small. This is also where there is good reason to have such high res.
I don't know, maybe the extra resolution is perceivable in the precise ends of strokes of ideographic Japanese, Chinese and related writing systems? If so, it is a huge market. (For the english alphabet, even old monitors seem fine to me.) Also, 500 ppi is just unbelievably awesome - useful or not.
> (For the english alphabet, even old monitors seem fine to me.)
Possibly. But operating systems have to go to great lengths to display text well, with subpixel antialiased rendering and all sorts of clever tricks to make the most out of the limited resolution. This often doesn't work well, leading to complaints like these:
If we get up to 500 ppi, that's well into inkjet-printer quality text without any antialiasing at all. Text would look rather nice. (Although I may be alone in peering at laser-printed text just to admire how smooth it is.)
I do agree that with a sufficiently high DPI, we can stop worrying so much about anti-aliasing and especially sub-pixel anti-aliasing
The articles you mention don't so much represent cases of "doesn't work well" as "no one solution satisfies everyone"; that problem does indeed go away when you have enough DPI.
(We still have to care about hinting, though, because among other things we want character boxes to overlap to avoid excessive whitespace in strings like "AV" or "Tr".)
As far as I know, the hinting instructions embedded in a font also includes information about how to do kerning for that font, as well as replacement glyphs for particular sets of adjacent characters (such as ff).
I thought the whole point of the "Retina display" is that it reaches the threshold of pixel density where you can't see pixels anymore (at that distance from the screen).
6.1" seems more the size of a small tablet than anything fit for a phone. Assuming people hold a tablet about the same distance away from their eyes as a phone, perhaps even farther, what's the point of having an even higher pixel density? Isn't this going to cause more design problems than it provides marketing benefits?
Higher resolutions will still be needed for implementing autostereoscopic 3D displays[1], where multiple angles of a scene are displayed at the same time.
Unlike 3D displays that use active glasses to time-division multiplex the viewing angles for each eye, autostereoscopic displays use parallax barriers or lenticular lenses to multiplex the images inside a single 2D image, reducing the effective resolution per angle/eye.
I wonder how these high-res tablets would affect HTML5 media queries. Right now, many of us do something like "screen and (max-device-width: 480px)". That's going to get rather awkward if pixel density varies a lot between devices. We won't be able to fool ourselves forever by implicitly multiplying that by 2 (or some other factor).
Perhaps we'll need to standardize on a unit that is independent of screen resolution. But it won't do to convert everything to millimeters, because different devices are also meant to be used from different distances. 2mm on a smartphone is OK. 2mm on a cinema display is virtually invisible.
A CSS px is not necessarily defined as a pixel -- The reference pixel defined in the CSS spec is the length crossed by a 0.0213 degree arc at the typical viewing distance (which works out as 96 DPI at arm's length).
This was solved five years ago: http://www.webkit.org/blog/55/high-dpi-web-sites/ Summary: px is a logical unit and 1 px = 2 (or more) pixels on high-DPI displays. Em is also a logical unit, although I think there are problems with em-based layout.
