Font hinting on iPad?

Does iPad make any use of font hinting, if available, please? I mean not just safari browser, but the device itself, e.g. in purchased apps.

As far as I know, the CoreText rendering engine on Mac OS X completely ignores font hinting and rather employs subpixel antialiasing (source: http://www.typotheque.com/articles/hinting).

However, iOS devices do not use subpixel antialiasing (probably due to portrait/landscape orientation switching) and do only standard "grayscale" antialiasing, which obviously leads to less precise results. So I am wondering whether or not do they make any use of font hinting (to compensate this)?

The reason why I'm asking is we want to embed Neue Helvetica in an iPad app we are developing now (the Neue Helvetica included in iOS does not have Latin Extended-A charset, unfortunately). And we are wondering whether we should invest in the hinted version offered by Linotype, labelled as XSF (more about it here http://www.linotype.com/2049-16432/xsf.html)

Any ideas, please?

Thank you very much.

clauses's picture

Apple devices does not use any hinting at all. Early iPod's use a bitmapped pixel font. iOS and OS X uses font metrics (stated or derived – I don't know) to do some gridfitting in the y-dimension.

blank's picture

Does this means that screen fonts need to be designed for three Mac rendering environments in addition to all of the Windows environments? Cheap 300 DPI displays definitely can’t come quickly enough.

Richard Fink's picture

Hinting does not effect rendering on Apple devices.
Hinting is done "on the fly" in a sense.

If you're developing strictly for the iPad, no, it doesn't make sense to pay more for a font hinted for Windows. But if you do opt for the hinted font, it won't have any adverse effect. (At least I haven't noticed anything. And I do a lot of looking at hinted TrueType fonts on Mac, iPad, and iPhone.)

John Hudson's picture

Hinting is done "on the fly" in a sense.

I think that is a misleading and potentially confusing statement. What Apple do is apply some vertical gridfitting to outlines as part of their rasterisation process, and then fuzz the whole lot with antialiasing. You don't need any hinting for this.

Tom Gewecke's picture

@josefrichter

On my iPad, the app TypeFaces seems to indicate that the font Helvetica Neue does in fact include Latin Extended A (U+0100–017F). Are you sure that is incorrect?

Chris Dean's picture

Post 320 PPI at a viewing distance of ~30 cm hinting doesn’t matter anymore. Finally.

Curcio, C.A., Sloan, K.R., Kalina, R.E. & Hendrickson, A.E. (1990). Human photoreceptor topography. Journal of Comparative Neurology, 292. 497-523.

http://prometheus.med.utah.edu/~bwjones/2010/06/apple-retina-display/

Si_Daniels's picture

>Post 320 PPI at a viewing distance of ~30 cm hinting doesn’t matter anymore.

...even if this is true...

>Finally.

...A prediction as to when this will be the norm? My guess is July 17th 2022.

Chris Dean's picture

I’ll bet you $1 that all Mac products will have that resolution in <3 years.

And the competition will be <2 years to catch up.

We’ll finalize the terms through email, but I’m pretty confidant I'll walk away with a buck (CAD).

It's past my bedtime. Sleep well (:

Si_Daniels's picture

>It's past my bedtime. Sleep well

I think you're already dreaming ;-)

Rob O. Font's picture

>Mac OS X completely ignores font hinting
This is not entirely tr
>Apple devices does not use any hinting at all.
This is not entirely tru
>Hinting does not effect rendering on Apple devices.
This is not entirely true.

But... there is no need to pay extra for hinting if one wishes to publish on the iPad. In addition, impishly I add, there is no need to freak over the size of the fonts or screens, or resolution, as the IPad allows design based on "Actual Size". And all the browsers on the iPad behave the same, so a whole multitude of other type harpies that plague, e.g. Windows and the web, are additionally avoided. Also, the iOS uses the broadest type standard, ttf fonts, so other format complexities can be trashed.

Just what i always wanted, as exciting as the early days of PostScript -- publishing that's about design again, not bugs, # mark tricks or endlessly nodding standard issue gents without a clue. Heave that huge hammer home honey! ;)

k.l.'s picture

iPad2 still with same resolution as iPad? :(

apetickler's picture

And the competition will be <2 years to catch up.

I'll die if it takes two years. Back in the 90s, I expected us to all be using 300ppi displays by now.

John Hudson's picture

Karsten: iPad2 still with same resolution as iPad?

No, no, Karsten, that's a good thing. Didn't you read what David wrote? Now publishers and designers can pretend there is only one device in the world and one fixed resolution for which one can design based on Actual Size. It's as if the past two decades and never happened, and the challenges of scalable dynamic display have been met by, um, not scaling or being dynamic. :)

k.l.'s picture

And give Apple a chance to come up with "expert" tricks to address (work around?) resolution-independence. Like their explicit viewport parameters, when viewport is implicit in CSS and size/orientation-dependent layout adjustments can be done via media queries too. Or their very "special" implementation of CSS's position:fixed.

Rob O. Font's picture

Hmmm. First, KL, do you have a problem with the iPads current resolution?

Ummm, JH, if the challenges of which you speak, had been met, as opposed to more or less faked, we would not be having this conversation.:)

Rmmm, My point was not that the iPad, in it's entirety, is the only way to go. Do you have one by the way, John, Simon?

The point was that key typographic ingredients are present, like browser irrelevance as far as displaying content and presentation, designer awareness of size and space, fonts that always match the device, and predictable scaleability.

In the meantime, no, the fonts used on iPads don't need hints... But if you want to, you can grid-fit a font for any size below around "12 pt" and improve it. To drive the nail through the tissues, in the standard world of web publishing, one is compelled to try and make one font that can do 5 things, usually doing the web text thing unwell, (aliased, aa greyscale, aa CT, aa CTDW and Mac). In this non standard platform, one has two ways of doing the web text thing, one way well, one way perfect.

Standard or not, if you are a typophile you gotta love that.

k.l.'s picture

Hmmm. First, KL, do you have a problem with the iPads current resolution?

Yes. After having seen the iPhone 4, every other display makes me feel like going back to Duplo after having advanced to Lego. :)

For people who do (web)apps and want to target a very specific version of a specific device, sure it's nice to know there are so many pixels available. It's reliable in a way. But then, every increase of resolution will become an issue (admittedly, Apple "solved" it gracefully when introducing iPhone 4). And, as more and more such decives show up (leaving aside quality of hardware, quality of OS & its UI, and such) and all of them come up with different screen sizes and resolutions (pixel count information alone is useless), people better start thinking of an approach to express design in a way more flexible than just counting pixels. Now that Apple and Microsoft, after long period of reluctance, did this step with their "big" OS's UIs, it sounds like a bad joke that people should start getting back to do pixel-based design.

I thought that discussing resolution-independent design belongs to a past era. Hm.

Richard Fink's picture

@JH

Me:>Hinting is done "on the fly" in a sense.

You:>I think that is a misleading and potentially confusing statement. What Apple do is apply some vertical gridfitting to outlines as part of their rasterisation process, and then fuzz the whole lot with antialiasing. You don't need any hinting for this.

First of all, a corporation are not plural. But I give up. Uncle!
;)
Second, it all depends upon what your definition of hinting is. Probably yours is not mine and yours may be confusing to most people trying to understand it. "Hinting" is a terrible term. Non-descriptive, opaque inside lingo. What matters to authors is the result, not what the font creator needs to do to get that result.
Something somewhere has to map the outline to pixels. Based on the outline, a pixel outerform is created onscreen. In Windows GDI and DirectWrite, this outerform is influenced by instructions inside the font. On the Mac, it isn't. It creates the outerform on the fly. How on earth is this essentially different from autohinting on the fly?
If your point is that my statement might lead someone to think that the Mac does exactly the same thing as GDI or DirectWrite but does it automatically, point taken.
But other than that, are we unconfused?

But other than that:"Now publishers and designers can pretend there is only one device in the world and one fixed resolution for which one can design based on Actual Size. It's as if the past two decades and never happened, and the challenges of scalable dynamic display have been met by, um, not scaling or being dynamic. :)

Keep hanging with webbies, John! By George, you've got it!

@db

As is not unusual, I have no idea exactly what you're trying to say. And what is "Actual Size". (It sounds like something I'm not particularly interested in using, but what are you referring to?)

Si_Daniels's picture

>Do you have one by the way, John, Simon?

I used a first gen one when it came out. It really didn't do it for me, but 130PPI wasn't the deal breaker. I'll probably wait for V3 :-)

The only way CD's prediction will come true is if Apple ditches the larger form factor, lower DPI devices. eg. their 27 inch monitors are 108PPI - to bump that by 3x in 3 years seems far-fetched

Chris Dean's picture

I’m stickin’ to my guns.
(and that’s $1 plus compound interest @ prime)

Richard Fink's picture

"The hints can be generated either “on the fly” within certain rasterizers, or (in the vast majority of cases) they are developed as a part of the font production process."

Veteran font technologist Tom Rickner here.

Good article, too, BTW.

John Hudson's picture

Rich, I take Tom's point, but I'd prefer to reserve the term 'hinting' for the in-font mechanism, since this works equally well for both PS and TT fonts: it is the font giving hints or instructions to the rasteriser about how best to render it. Perhaps there is some other term we can use to describe in-rasteriser mechanisms for on-the-fly improvement of text rendering? Otherwise I'm going to have to spend the rest of my life typing ‘in-font’ or ‘in-rasteriser’ before every instance of the words ‘hint’ and ‘hinting’.

John Hudson's picture

David: The point was that key typographic ingredients are present, like browser irrelevance as far as displaying content and presentation, designer awareness of size and space, fonts that always match the device, and predictable scaleability.

Yes, I understand that in isolation these are very admirable qualities of the iPad device. My point, and Karsten's, is that the device doesn't exist in isolation. Yes, it is good for designers and publishers that when it comes to the iPad decisions can be made at the device level instead of at the multifacted browser/device intersection -- how long will that last, I wonder? --, but it still represents only one corner of a mobile Internet content display world.

Rob O. Font's picture

JH>...that in isolation these are very admirable qualities of the iPad device. My point, and Karsten's, is that the device doesn't exist in isolation.

Hardly are the few qualities I mentioned in isolation.

>Yes, it is good for designers and publishers that when it comes to the iPad decisions can be made at the device level

Good? Lol. That's kind of like ignoring the last two decades. And "device level" is a bad term, I think, for the user experience.

>...instead of at the multifacted browser/device intersection -- how long will that last, I wonder? --,

Yes, well how long can the multifacted browser/device intersection last more costly than print? I don't think you even have enough experience in that ugly intersection to answer, do you?

As for hints, when your terminology evolves beyond the imprecise term rasterizer, to scaler and renderer, it can be explained simply, as has been pointed out previously.

miha's picture

After having seen the iPhone 4, every other display makes me feel like going back to Duplo after having advanced to Lego. :)

Reminds me of this picture.
Subpixel Typography
__

I am certain that when iPad gets a higher resolution screen it is going to be exactly 2× higher than original, having 264 ppi. Which is less than iPhone 4, but a viewing distance is probably a little bigger.

Si_Daniels's picture

The sub-pixels in that picture are messed up... GB/RGB/GBR/B !?

josefrichter's picture

Thanks everyone for all the info. So no point in buying hinted fonts just for iPad, right?

@Tom Gewecke – to be more specific, it seems roman cut of Neue Helvetica on iPad does include Latin Extended A, while the bold one does not, for some reason. And certainly there's no light nor any other cut.

josefrichter's picture

fyi response directly from Linotype:

IPad and IPhone do not need any XSF-Hinting. Hinting is only necessary for Windows-Displays.
For IPad and IPhone-Apps we offer a special SVG-Fontformat. You will need a standard license and a license extension for apps.

Richard Fink's picture

db>
Yes, well how long can the multifacted browser/device intersection last more costly than print?

More costly? Stop trying to emulate print onscreen and take the savings.
As for how long can it last - probably forever. Because it's a good thing, not a bad thing.

Jens Kutilek's picture

IPad and IPhone do not need any XSF-Hinting. Hinting is only necessary for Windows-Displays.
For IPad and IPhone-Apps we offer a special SVG-Fontformat. You will need a standard license and a license extension for apps.

How do you embed an SVG font into an iPad app? Is it even possible? Or do they rather mean web apps ...

Rob O. Font's picture

> Stop trying to emulate print onscreen...

You mean like sites that use ugly typewriter faces for heads, and just go on ppg after ppg as a spew of text, code and links, I assume. ;)

mike_duggan's picture

I think this thread could have been condensed to

Q:Font hinting on iPad?
A:No

Nick Shinn's picture

Well, all threads could be condensed:

WTF?!
Actually, No.
IMO, Yes
Whatever.

Jens Kutilek's picture

Q:Font hinting on iPad?
A:No

If someone would write an iPad app for hinting fonts, then perhaps I’d buy one ;)

Christopher Slye's picture

"Hinting" is a terrible term. Non-descriptive, opaque inside lingo.

Actually, I think it's an excellent term. I don't know if it was created or merely popularized for the Type 1 font format, but it's quite appropriate for that. T1 hints do not dictate how device pixels behave, but simply provide more information ("These stems should be the same", "These features should all align") about font and glyph features that a PostScript rasterizer can use to create better output.

Rob O. Font's picture

MD>I think this thread could have been condensed to...

But then we wouldn't have found out a lot, which makes the forum a good thing, No?

Q:Is font hinting executed on an iPad by the OS or apps?
A:Not in most cases, but I think it's up to the app.

Q:Could hints in fonts be used to improve the appearance of type on an iPad?
A:Not in some cases, but at smaller sizes, yes.

Q:Are people confused about the trending resolution/device scale that's formed around user distance, i.e. 80, 100, 200, 400 dpi, for TV, laptop/desktop, tablet and phone respectively?
A: Seems so, with folks pining for 400 dpi tablets and drooling for 300 dpi desktop screens. These are as unlikely to acquire a mass of users as 88 dpi tablets or 200 dpi TVs.

CS>...that a PostScript rasterizer can use to create better output.

Yeah. Until there are not quite enough pixels to go around (25 ppm and below), and then, as the last 20 years prove ...hints need instructions:) In fact, you and most others want to leave the instructing, composite positioning and kerning entirely up to the rasterizer... and just thinking about all that random rounding together in text at low res gives me the qualitative heebeegeebees.

mike_duggan's picture

that was just bait for you Dave, got ya hook line and sinker! :-:)

John Hudson's picture

Not so very long ago, there was an industry that made the great bulk of its money by selling machines for setting type, and a smaller portion of its money by selling fonts of type to be set on those machines. Given the amount of time and labour necessary to manufacture the fonts, I suspect quite a lot of them served as loss leaders: the availability of an extensive library of typefaces was a selling point for the machines, even if some of those types were seldom actually purchased. Competition in the machine market took place in the areas of real or perceived quality of output, price, and typeface availability (the latter open to various forms of circumvention, from similarly styled types to outright copying of designs, and inviting the eventual development of the machine-independent typeface business model of ITC).

How does this relate to the iPad discussion? It seems to me that what is missing from some of the analysis of rendering variables in the webfont market is the recognition that it is still with the machines where money is to be made, and where competition takes place in the area of real or perceived quality, and hence in the area of rendering. In the period of greatest predictability in text rendering, i.e. the initial years of TrueType, there was essentially only one machine, manufactured by various companies from largely interchangeable components; machine differences in device resolution and gamma had minimal impact on what one could expect from the behaviour of a given TTF. But that is very obviously no longer the case. Now there are lots of machines, and there is once again value to companies in trying to compete in terms of real or perceived quality of output, or 'user experience' as it is now reckoned, including in the area of text rendering.

This competition comes with a massive support cost, of course, but it is one that the machine makers have successfully shifted onto the publishers, i.e. the cost of trying to target multiple devices with different rendering environments. So, David asks: ‘...how long can the multifacted browser/device intersection last more costly than print?’ And I suspect the answer is that it will last for as long as there is more money to be made from making and selling a machine than there is to be made from either publishing content to be read on it or making fonts to display that content.

Rob O. Font's picture

MD>that was just bait for you Dave, got ya hook line and sinker! :-:)

Surprise! ;) and now "You're going to need a bigger boat."

JH>...it is still with the machines where money is to be made

Really!? I don't think so.
1. Most hardware prices, incl. Apple's are pretty close to the bone.
2. Apple's strategy has long been to make money on software.
3. and that's shifted to software and content.
4. companies like Google, MS and others don't even make any machines

And the question I asked, ‘...how long can the multifacted browser/device intersection last more costly than print?'? was asked in relation to the astonishing statistics showing that north of 95% of the money made on the web is made by south of 5% of web developers.

I.E. the cost of trying to target multiple devices with different rendering environments is way too high for most all benefits.

blank's picture

companies like Google, MS and others don't even make any machines

In this case the machine doesn't need to be hardware—it could be software or a service. For example, Google does not sell much hardware, but subsidizes creation of fonts so that web developers will link to fonts on Google servers and supply Google with tracking data for sites that aren't using Google Analytics. Of course, even Google isn't willing to pay for creating great screen fonts, so they're taking advantage of designers with a penchant for open-source software.

I.E. the cost of trying to target multiple devices with different rendering environments is way too high for most all benefits.

Which makes me wonder if web fonts will really take off for sites not targeted at hi-res phones. Because right now, even among the handful of well-designed screen fonts, few web fonts come close to Georgia and Verdana.

John Hudson's picture

James: In this case the machine doesn't need to be hardware...

Right. I had originally referred to physical machines and virtual machines in my post, but removed this sentence as I thought the analysis was already getting too complicated. The point is that the bulk of money to be made is neither in the typefaces nor in the content set with the typefaces. Which puts the makers of type and the makers of content largely at the mercy of the competitive strategies of the hardware/software makers, who are playing for the really big stakes.

David: ...the astonishing statistics showing that north of 95% of the money made on the web is made by south of 5% of web developers.

Well, that's better than the overall distribution of wealth in the capitalist system.

The interesting analysis would look how closely that 5% of web developers is tied to hardware/software developers. I don't think the disparity can be explained in terms of ‘the cost of trying to target multiple devices with different rendering environments is way too high for most all benefits’. This disparity has developed in a market in which everyone was using Verdana regardless of rendering environment. Web fonts are too new a phenomenon for their cost factor to in any way account for the failure of most web developers to make more than a pittance.

Richard Fink's picture

Dunwich>"Which makes me wonder if web fonts will really take off for sites not targeted at hi-res phones. Because right now, even among the handful of well-designed screen fonts, few web fonts come close to Georgia and Verdana."

Take off?
Not quite sure what hi-res phones have to do with it, but the Google Font API now serves roughly 50 million daily requests, across roughly 800,000 unique websites.

I started out assuming that visual quality - at least what that term means to me - would play a big role. Seems I'm wrong. And that's been obvious for awhile.
Whatever "value" is to these users, it doesn't have anything to do with comparisons to "well-designed" screen fonts like Georgia and Verdana.

The web has it's own decorum. More casual and experimental than print, for sure.

O would some power the gift to give us, to see a font as others see it

Rob O. Font's picture

For anyone reading the above post now, of in the future, I was not trying to "explain" the disparity Mr. Hudson quotes above as a result of either web fonts or development cost/benefits. I was simple pointing out a general fact of the web's gross revenue profile.

>This disparity has developed in a market in which everyone was using Verdana regardless of rendering environment.

John, everyone was not using Verdana regardless of rendering environment, that is why @fontface was needed.

>Which puts the makers of type and the makers of content largely at the mercy of the competitive strategies of the hardware/software makers, who are playing for the really big stakes.

I love the "competitive strategy" drama dairy — endless milk for discussion while we go over the hump.

John Hudson's picture

Sorry if I misinterpreted your point, David. This is what you had written (my emphasis):

And the question I asked, ‘...how long can the multifacted browser/device intersection last more costly than print?'? was asked in relation to the astonishing statistics showing that north of 95% of the money made on the web is made by south of 5% of web developers.

What is the relationship between the question and the statistics?

Tom Gewecke's picture

@deberlow

I am puzzled by your statements regarding Apple's hardware prices and strategy. I have always thought it was exactly the opposite, e.g.

http://www.computerworld.com/s/article/9150045/Apple_makes_208_on_each_4...

Nick Shinn's picture

@Richard: I started out assuming that visual quality - at least what that term means to me - would play a big role.

It will, it just never does at the first wave, when mere novelty and the cheapness thereof rules. This goes back to digital day one, when typesetter organizations decried the quality of Apple's 300 dpi typesetting, fat lot of good it did them.

Then once Linotronic output resolution was high enough, and we had 600 dpi laser printers, we didn't worry about bad hinting on monitor screens for 10 years, until it became an issue with the WWW.

I don't have a problem with the fuzziness of Apple's rendering—after all, as a connoisseur of letterpress printing, it reminds me of good old ink spread on a lovely high-fibre rag stock. And haven't we been decrying the emaciated quality of so many digital body types, and craving heftier fonts, for most of the digital era?

John Hudson's picture

Nick: I don't have a problem with the fuzziness of Apple's rendering—after all, as a connoisseur of letterpress printing, it reminds me of good old ink spread on a lovely high-fibre rag stock.

It reminds me more of nth generation reprints of mass-market paperbacks in the check-out aisle of the supermarket.

Rob O. Font's picture

>I am puzzled by your statements regarding Apple's hardware prices and strategy.

That link is a list of parts and $10 for manufacturing that add up to 3/5ths the cost of the device. In most retail electronics, the retailer needs 2/5ths the cost of the product.

So on one end of the Apple chain, (your link) you have empty iPad hardware, and on the other end, ownership of the retail outlets with total responsibility for complete costs of integration in between.

Just the parts and normal retail demands add up to 5/5ths of the sale price.

This, perhaps completely, explains the second half topic of your link?

Rob O. Font's picture

JH>What is the relationship between the question and the statistics?

If few websites make money, and web typography becomes more expensive than type in print, (along with many other things about the web), then the iPad is a good model for publishing, and better looking than anything you can point to.

JH>It reminds me more of nth generation reprints of mass-market paperbacks in the check-out aisle of the supermarket.

It's either that or what reminds me of nth generation hot metal printed by an ink miser after the round tops and bottoms of each such letter were first filed flat.

That's competition;)

Si_Daniels's picture

"ink miser"

Isn't that one of Roger Black's new start-ups? ;-)

Syndicate content Syndicate content