So when will high-resolution computer displays become available?

Jongseong's picture

When are we going to have access to computer displays with pixel densities even beginning to approach print, say 300 pixels per inch? Although mobile devices boast higher resolutions, I don't remember seeing significant improvement in the pixel densities of computer displays for several years.

I would dearly love to have even a tiny little 4-inch display as long as it has a high pixel density, that I can use as a secondary monitor to aid in the type design process. Is something like this possible with devices that are currently available on the market? When will such displays be available, or do manufacturers have no plans to improve the resolutions of displays for the average consumer?

Si_Daniels's picture

When the OS makers and applications running on those OS's truly support resolution independence. It's the software, not the hardware that's holding things back. I'd say 2015.

blank's picture

Sii’s prediction is probably right—if nothing else, by 2015 someone will sell some kind of hi-def screen to use for designing mobile apps and content. And I’m sure that we’ll be able to get a 300 DPI ebook by then.

In the meanwhile just keep using both sides of your test prints (I’m not kidding, on busy days I get a big pile of prints and use the other side.).

raph's picture

The frustrating thing is that prototypes were available quite a long time ago. In fact, the original IBM T220 came out in 2001. I've been predicting it would happen sooner, or at least wanting it, for a while.

If you want a small display, though, you can get a Nokia 810 (I have an 800, which is nice).

I do think it is inevitable at some point. The incremental cost is going down and down as compute power goes up, and the increase in quality would be dramatic. For a while I was expecting Apple to do it, as it would be a way to really differentiate their product line from what's available in the PC world, but currently their PC business model seems to be all about packaging completely commodity parts at premium prices.

The next challenge is to get Web graphics to work at higher resolutions. Bitmaps on the web are (more or less) locked at 96dpi. Of course, there's always SVG and so on, but at current trends it will be a while until most browsers actually support it.

Except in some applications, I actually suspect 200dpi is adequate, because you can make the antialiasing really good on displays. In print, you need higher resolution because antialiasing is either difficult or impossible (there's essentially no control over gamma curves). By my calculations, ordinary text at 200dpi has about 95% of the contrast as it would on an infinite-resolution display, even without doing stuff like ClearType. At some point, people may decide that going for that extra 5% is also worth it, but I'd be plenty happy just to get the 200dpi any time soon.

It's going to make a huge difference for type on the screen.

Jongseong's picture

Good point, Raph, about 200 ppi being adequate for displays. And thanks for confirming that the technology has already been around for a while, something that I suspected. The IBM T220 sounds like a monster!

So we've got resolution-dependent OSs and web graphics that lock us into the current resolutions and will make any substantial improvement in the density a bit of a hassle. Add the fact that manufacturers have been sitting on the technology for years, that the current trend in consumer technology has swung toward getting the same features cheaper rather than improving features and adding to the price, and I'm getting worried about the inertia. Reading discussions about the merits of different antialiasing technologies, I wonder whether such debates wouldn't be rendered obsolete if the resolution of displays simply improved. And of course, I would have less cause to worry about learning TrueType hinting...

Really, I think I'd be happy to get my hands on a 200 ppi monitor by 2015. I'll be following James's advice until then.

blank's picture

I’m getting worried about the inertia.

For a long time the RAM, CPU, and GPU power that new hardware gave up went to support bigger monitors and high refresh rates. In the last few years we’ve really reached the practical size limits for desktop monitors, and LCD adoption has locked refresh rates at 60Hz without complaints. Good 30" screens are affordable at consumer and professional price points, and most people don’t have the need or a desk big enough for anything larger. It won’t be long before the relationship of price to size or image quality is irrelevant to consumers in developed nations, so resolution will really be the best feature to sell displays with and inertia will be a non-issue.

Si_Daniels's picture

>Except in some applications, I actually suspect 200dpi is adequate, because you can make the antialiasing really good on displays.

Personal experience is that you I can still see improvements at 260ppi vs lower res. I know the organizers are planning some sessions around type on the screen for ATypI - hopefully they can organize a hands-on (or would that be "eyes-on") where type can be examined using a range of rendering options on different resolution displays.

Richard Fink's picture

I spent a long time today looking in a consumer electronics chain store.
Specifically, I took a very close look at the HP Touchsmart Laptop. The text was incredibly sharp and clear - haven't seen anything quite as good except on E-Paper.
Plus, the screen could re-orient itself from landscape to profile in all four directions instantaneously at the touch of a button.
Plus, it's a touchscreen "tablet", of course.
I was blown away by it, really.
All I could find out from the HP site is that it has an LED backlit LCD display.
I have a Dell Latitude notebook with a very good quality LCD screen with resolution set in Windows at 1280*800. It's quite sharp and clear.
The screen resolution on the HP Touchsmart was also set to 1280*800 but the text was much more print-like than on my Dell.
Any clues why that may be? Is it the LED backlighting or what?
(Price was $900, BTW)

kentlew's picture

> I would dearly love to have even a tiny little 4-inch display as long as it has a high pixel density, that I can use as a secondary monitor to aid in the type design process.

In the mean time, why not just take your current secondary monitor, place it 2.5x further away and set the type 2.5x larger? ;-)

Si_Daniels's picture

Or get a 2" monitor and place it twice as close and make the type 2 times smaller? ;-)

kentlew's picture

I was being half-serious. Si, your suggestion won't give the effect of more pixel density.

Si_Daniels's picture

True - I was not being at all serious.

Kevin Larson's picture

Kent is right about distance being an important variable. By my back-of-the-envelope calculations, we can't detect additional resolution beyond 300 dpi from a distance of 50cm. The problem is that when we want to evaluate text quality, we typically get closer to the screen. But any screen becomes a high density screen from a far enough distance.

I'm pessimistic that screen resolutions are going to get drastically better by 2015. LCD pixels are inefficient at letting the light from the backlight through, and smaller pixels are much less efficient. Backlights are already the single biggest power users on laptops. Many consumers are asking computer manufacturers for longer battery life and fewer are asking for higher pixel density. I think we're going to see battery life increase to a full day's use first.

Jongseong's picture

In the mean time, why not just take your current secondary monitor, place it 2.5x further away and set the type 2.5x larger? ;-)

That's how I would set up a secondary monitor if I had one, in the absence of a higher-density monitor, of course. I think I've seen similar set-ups used by type designers (maybe it was a Typophile thread). But I'd rather evaluate type from not much more than typical reading distance, I guess.

Quincunx's picture

Well, it's already going the right way, with most current lcd's being between 120 and 150 dpi (like Apple Cinema Displays). If they keep up this rate of development it probably won't take long to hit 300dpi. And yes, of course you would need a resolution indepentant OS to really take advantage of those screens.

dberlow's picture

>with most current lcd’s being between 120 and 150 dpi (like Apple Cinema Displays)

(max res H = 1920 / inches H = 22.57) The Apple Cinema Display is 88 dpi, unless counting subpixels ;)...

The iPhone is 160, the Palm Pre is 180, (your average TV is 80).

Linear thinking says just one thing, displays will be at a million dpi in a few decades. lol.

>When will such displays be available [] to improve the resolutions of displays for the average consumer?

Real thinking, on the other hand, says "n e v e r." I will gladly stake $1,000 on 2015 having no impact on this topic.

There is little or no market for improved print wysiwyg in page or letter design.

OS don't scale their desktop, screwing developers, and OS don't care about reading on these pathetic resolutions, screwing users, teachers, type developers and publishers.

Since we can't completely hint in this particular hell, there is little to do but design robust contours (from a distance if ness.).

Any questions? :)

Cheers!

Richard Fink's picture

>The iPhone is 160 (ppi)

The text on an iPhone is terrifically sharp and clear.

Quincunx's picture

> (max res H = 1920 / inches H = 22.57) The Apple Cinema Display is 88 dpi, unless counting subpixels ;)...

Yes, I was a bit too optimistic. The 24" LED ACD is 94dpi. The 30" ACD is 100 dpi.

I've always used sqrt(HorizontalRes^2+VerticalRes^2)/ScreenDiagonal.
e.g. for the 30" ACD: sqrt(2560^2+1600^2)/30 = 100.629132

Erik Fleischer's picture

It’s the software, not the hardware that’s holding things back.

Why is that so? I would have thought that if the processing power required for rendering were available, it would be trivial to develop software -- including patches for existing software -- to take advantage of higher resolution displays.

By the way, I'm not expressing a different opinion; I really am ignorant and would appreciate it if anyone could develop this topic a little further. :-)

Jongseong's picture

Thanks for all the insightful comments!

Erik, I think one important issue is that current OSs are resolution-dependent, as the windows, menus, and icons of the user interface are bitmap-based.

If you double the resolution in ppi, the windows, menus, and icons become a quarter of the original size.

So it's not simply a case of pumping more processing power. The interface has to be re-designed to be scalable, so that an increase in resolution doesn't lead to tiny unreadable menus. Replacing the bitmap-based OS interfaces with vector-based ones probably won't be a trivial matter. And it's not just OS interfaces. Think about what web pages would look like shrunk dramatically in size.

Kevin Larson's picture

There are some interesting articles about the challenges of providing a high dpi OS at the Engineering Windows 7 blog. Application compatibility is the big problem.

http://blogs.msdn.com/e7/archive/2008/09/13/follow-up-on-high-dpi-resolu...
http://blogs.msdn.com/e7/archive/2008/09/16/more-follow-up-to-discussion...

Richard Fink's picture

We are going through a transition right now. I recently bought, and had to return, a 24" LCD display because no PC I currently own is capable of driving it at it's native resolution.
It worked great as a TV working off my digital video camera, but I couldn't justify holding onto it for just that.

bowerbird's picture

jongseong said:
> And it’s not just OS interfaces.
> Think about what web pages
> would look like shrunk
> dramatically in size.

web-pages _should_ be able to
rescale to any size/resolution...

except that our c.s.s. overlords
have sold us a bill of goods, and
their stuff fails as often as it works.

moreover, they don't seem to know
that it needs to be fixed, or which
parts are broken, or how to fix it...

-bowerbird

Richard Fink's picture

IE8, FF3, Chrome, Safari, are all resolution independent.
Zoom, zoom, zoom.

@BB

I saw this on the side of building. You're a poet. Is it yours?

"The CSS gurus have failed us, I fear
But they've promised to fix it with standards next year
I follow their rules with the sites I construct
So now, in the meantime, I guess I'm just f-cked
"

(Sorry about the dash. Takes away the punch. These forums and their dang rules!)

So, what platform other than standard browsers would you suggest?

rich

bowerbird's picture

richard said:
> IE8, FF3, Chrome, Safari, are all resolution independent.
> Zoom, zoom, zoom.

sorry. maybe it's just me, but zooming seems stupid to me...

i mean, it's fine as a momentary thing when one particular item
seems undersized and needs to be magnified to be perceived,
but since my view-port is usually maximized top-to-bottom
(and, less often, left-to-right as well), zooming means that
_something_ is run off the screen, so it has to be momentary.

consequently, i almost never zoom, even when i know i can...
and since i use it very rarely, i eventually forget it altogether.

meanwhile, the resizing command-keys are in my muscle-memory.

> I saw this on the side of building. You’re a poet. Is it yours?

well, i like the sentiment, and it does rhyme...
but that meter is far too confining to be mine.

> what platform other than standard browsers would you suggest?

i have no suggestions. i suffer. i just don't suffer in silence.

-bowerbird

trevhutch's picture

Would the next logical step to be to have monitors that output double the ppi (192ppi), and an OS that presents its applications where 1 pixel becomes 4, with the option to use the increased detail. In other words, bitmapped graphics like a photo on a web page would appear the same size, but vector based text could be much sharper.

dberlow's picture

Trevhutch: I wish it was that easy, but not all resolutions are equal now, so doubling them is not as easy as 1 pixel becomes 4. 88 gets to 176 dpi, and 96 gets to 192...and all the graphics on the planet would shrink, (millions needing to be re-rendered)... wouldn't they?

Cheers!

Don McCahill's picture

> and all the graphics on the planet would shrink, (millions needing to be re-rendered)... wouldn’t they?

Not necessarily. They will just be smaller within a larger frame. Their clarity will remain consistent, as their pixel dimensions will be maintained.

Notice that as the frame has increased from the original 640x480 the actual size of characters on the screen has not changed all that much. I note that this site is one of many on the web that no longer use the full width of the screen, as was the norm in 1995.

The change will be that designers will start to use the bigger frame to create new larger graphics. Background images may have to be enlarged, but many elements will remain useful at their current size.

k.l.'s picture

Though about high-resolution displays and resolution-independent UI, the discussion* suffers from the pixels thrown from all sides.
As with other things too, there would need to be a break with the past. With res-independent UI there are only two factors to speak about:
-- resolution-independent UI with elements measured in mm, inch, etc.**
-- scaling of this UI including the mouse arrow, FLS5 nodes, etc.
The hardware driver would tell OS and apps about its size (mm, inch) and resolution (dpmm, dpi).

How to deal with old pixel-based apps is another question. Maybe present them in special red-flagged windows ("Legacy! Update or forget!") that scale content or not, depending on user preferences.

* Here, MS blog, Bill Hill's blog. Resolution-independence per se does not look like a big issue. Not for type designers who are used to think in abstract units (per em). Many contributions try hard to defend the status quo by complicating matters, mixing res-independent and pixel-based approaches, presenting calculations from pixels to display size to resolution and back to pixels.
Not sure about the UI in total but this how-to for making icons for OLPC's Sugar shows that it is possible.

** Ideally, this was something as abstract as fonts' UPM, yet I think it is fine to starting from a real-world size that at the same time indicates what the designer considered ideal at normal viewing distance. The user would still be free to scale the result.

Richard Fink's picture

Re: "Real-world size"

From a web development perspective, knowing the screen resolution - which you can query via JavaScript - gives you a very incomplete picture of what's physically being displayed to the user.
Example: A resolution of 1280*900 can be showing on a 15" wide-screen display or a 20" wide-screen display. And there's no way to know the difference. The 1280*900 number is being supplied by the OS, not the display.
What is the "real" size of the text? Only the user knows.
The plug'n'play spec was designed at least in part to have devices "announce" themselves with basic info about their capabilities.
I do not understand why, after all this time, displays do not report their actual physical size.
A puzzlement.

k.l.'s picture

Re: "Real-world size ..."

Don't take "real-world" too seriously, the second part of the sentence was the important one:
"... that at the same time indicates what the designer considered ideal at normal viewing distance."

Of course one needs to start somewhere. Design is making decisions. Unlike in the world of material things (books have "fixed" sizes for centuries, including margins and type!), the digital world allows users to scale both UI and content according to their individual preferences (that was point two in my list).

And there's nothing that would hold display device drivers back from specifying both size and resolution. (And please forget about actual pixel count.) With a new generation of OSes and software this could be a requirement. If or why it is not done today, does not really matter.

Miguel Sousa's picture

Windows Vista seems to handle pixel densities up to 144dpi well enough: http://www.istartedsomething.com/20061211/vista-dpi-scaling/

Syndicate content Syndicate content