Measuring Font Legibility

But the abrupt jump to Windows vs Mac font rendering is very strange.

As a web programmer / designer I don’t think this strange at all. I am always balancing readability and aesthetics - on top of the DPI difference between Mac and Windows there’s the font rendering difference, RGB quirks, how many toolbars the user has on display etc etc…

Trying to convince a client that Century Gothic is virtually unreadable in certain colour combinations is a nightmare, let alone their penchant for non-standard fonts that will make pages look abysmal on systems without the necessary font installed.

I am reading this page in FF on a Mac - it seems perfectly fine to me, although it is displayed in Ariel.

On no! teh ev1l spammers have broken your highly cryptic code! :wink:

“Myriad Condensed Web” is the most aesthetically pleasing, and easy to read font bar none. http://www.linotype.com/10680/myriadcondensed-font.html

So if you choose a font that is more readable - people can read it easier - OK

If you use a more readable font of the Web people can read it faster - OK

So Windows font rendering is better - Maybe (a matter of opinion) But this does not follow, or is even relevant?

If you pick a hard to read font then no amount of tricksty rendering will make it readable, and an easy to read font probably won’t need it?

Looks rather like the Tiresias font
http://www.tiresias.org/fonts/design_report_sf.htm

I also prefer Gentium to Times Roman.
http://scripts.sil.org/cms/scripts/page.php?site_id=nrsiitem_id=Gentium

Interesting that here we have another example where designing for the needs of disabled people actually helps everyone.

“Linux and Windows are great.”

Linux…fonts…great? HAHAHAHA!!

Never fails, mention Apple and Microsoft in the same post and flaming starts.

BTW, Macs are for girls.

Great. Now all they need to do is make the lines on the road so they’re visible at night in the rain, and I’ll be a much happier driver. :slight_smile:

And that, my friends, is The Real WTF™. (Oh, sorry, wrong site – but I totally agree.)

Great. Now all they need to do is make the lines on the road so they’re visible at night in the rain, and I’ll be a much happier driver. :slight_smile:

And that, my friends, is The Real WTF™. (Oh, sorry, wrong site – but I totally agree.)

AMEN!!

The new Clearview highway font is specifically designed for a specific application, as a superior replacement to Highway Gothic. But would you recommend using the Clearview highway font for web browsing? Of course not.

ClearType is a very clever technology designed for a specific application, as a superior font hinting method on LCD displays. Very true. But “Microsoft’s font rendering strategy is ultimately smarter than Apple’s?” Oh, please. As long as you’re browsing the web on an LCD flat-panel, use ClearType all you want. But it is completely inappropriate for web-browsing on CRT or plasma displays, for text layout work intended ultimately for print, or for just about any other graphic application.

Fonts, displays, and graphics are deeply complex subjects. Artists and researchers can spend careers in the subtleties of color science, vision and perception. In all of these areas, it’s important to know when to apply a particular technique as well as when NOT to. Big sweeping generalizations are facile. Please stop insulting your audience.

“Linux('s fonts) are great”

As someone (quite happily, I might add) reading from a Debian box, Linux fonts look terrible. Not as bad as they used to mind, but still pretty bad.

Jeff, I’m afraid there’s been some good points made. Generalizations are like assumptions.

Interesting, but it’s going to take forever for the new font to become universal. Here in Upstate NY, there are plentiful signs with highly-reflective green backgrounds and faded white lettering. When headlights hit those at night, there’s barely any contrast on them. But as far as I can tell from traveling, it’s a problem that’s been solved for 25 years.

As long as you’re browsing the web on an LCD flat-panel, use ClearType all you want

Which is, what, 90% of the market these days? And climbing rapidly? I don’t know too many people using CRTs or Plasma displays on their computers these days.

[ClearType is completely inappropriate for] text layout work intended ultimately for print

I agree we should have a choice of font rendering strategy per-app. It’s mystifying to me why Apple forces this specialized mode on everyone; maybe they assume every Mac owner is going to do print work?

it’s important to know when to apply a particular technique as well as when NOT to. Big sweeping generalizations are facile.

I don’t think the big sweeping generalization of “try different designs and measure the difference”, which is my point here, is in any way facile. The title of the post is MEASURING Font Legibility.

Yes, the lesson is not to assume anything about “readability” unless you’ve seen the research.

Furthermore, readability is complex, and depends on typeface, setting, medium, and conditions. In certain applications, perhaps reproduction is more important than readability.

FHWA was designed for reflective signs under various lighting and weather conditions, while moving at 100 kph. But Font Bureau spent big money to redesign it as Interstate, suitable for general design, and yet again as Whitney, suitable for body text in print.

http://en.wikipedia.org/wiki/FHWA_Series_fonts
http://en.wikipedia.org/wiki/Interstate_%28typeface%29

Is the type meant for editing code on the screen, or reading an article on a web page? Is it meant to be a GUI element? Is it supposed to be a typesetter’s preview of a print document, or a highway engineer’s design for signage? Is it meant to display fonts optimized for screen, or print, or both and then some. Or are we talking about a general-purpose rendering engine intended for all of these purposes and more?

Mr Atwood, it’s not clear whether by “Microsoft’s rendering strategy” you mean ClearType rendering, or fonts designed for RGB displays. I haven’t seen any empirical or theoretical research comparing Quartz to Cleartype rendering’s effectiveness. But a strategy that distorts the strokes of letterforms to fit the pixel grid too much is not the best for everything.

Woops, posting lag.

I agree we should have a choice of font rendering strategy per-app. It’s mystifying to me why Apple forces this specialized mode on everyone; maybe they assume every Mac owner is going to do print work?

Is it a specialized mode, or a generalized one? Quartz does do font hinting at smaller sizes, but never as hamfistedly as ClearType some would say. Monaco font in my text editor and terminal window renders with no antialiasing at all.

The title of the post is MEASURING Font Legibility.

Has anyone measured legibility of Quartz vs ClearType?

I still haven’t seen a scientific study which suggests which font rendering tech makes the screen more readable

The fact that Apple’s technique is explicitly optimized for print isn’t enough for you? Do you think it’s possible to render for print output and LCD output at the same time?

I’m not saying this is wrong, but I think it’s something of a zero-sum game… you can’t render for print accuracy without hurting LCD readability. Just like you can’t design a font for highway signs that’s also ideal for, say, newspaper print.

No, the “fact” is not enough.

“The title of the post is MEASURING Font Legibility.”

Oops, sorry to be glib, Mr Atwood.

But you are making a leap and taking it for granted. Your logical conclusion is based on false premises.

  • One characteristic of Apple’s technique is that it uses fractional character positioning on screen. One effect of this is that it lays out the same as print. Another is that it preserves the readability characteristics of good body-text fonts.

  • Apple’s technique does make use of font hinting, at least in some fonts at small sizes. You try to deny this by saying it “is explicitly optimized for print”, but that’s just incorrect. Again, it’s possible that more accurate rendering of well-designed fonts is beneficial for readability.

In my opinion, Apple’s font-rendering strategy is a generalized one, useful for content to be displayed in print, on LCDs, or both.

Windows’ font-rendering strategy is optimized not to piss off people used to seeing aliased text. ClearType is made for interoperability of Windows 98 and Vista. It’s meant for people who use Explorer 7 on Win 98.

That ClearType is more readable on an LCD than Quartz font rendering is supported by zero evidence.

The fact that neither MS nor Apple has put forward a comparison leads me to suspect that there is no significant measurable difference under general conditions. But if you say otherwise, you really should offer one shred of evidence.

@rupert: Jeff and I make our living developing software using the same convicted monopolists technologies. I don’t care that Jeff’s opinion is Windows focused, it’s his opinion and I read his blog for his opinion.

@Jon: I am not taking this personally at all, and by no means have I called Jeff a nut. See my previous reply, which I will expand on here. Maybe my comment was obtuse, but it’s pretty simple. In the paragraph where Jeff references his MS vs Apple font rendering post, he preceeds that closing sentence talking about and linking to scientific research on font and legibility experiments, and then links (literarially) to an opinion piece. It doesn’t matter if it’s his opinion or not, its the conflating of two unrelated sides of an idea, scientific research based and pure opinion.

The point of Clearview was to design a new font that was more readable. The font designers backed up their design with scientific measurement of how the font was going to be used. That has no relationship to either the MS way or the Apple way for font rendering because I still haven’t seen a scientific study which suggests which font rendering tech makes the screen more readable. I use both Windows and OS X all day long, and my opinion has been that the OS X font rendering is more readable, Windows makes fonts too skinny, but again just my opinion.

You might have seen the story in the new where a study was done that shows kids think food taste better with a McDonald’s logo, even though it was exactly the same.
http://www.forbes.com/forbeslife/health/feeds/hscout/2007/08/06/hscout607093.html

I would love to see an OS blind font “taste test” to eliminate OS bias, and then the same thing with OS identification thrown in to see which font rendering tech is truly more legible and readable.

They’re gonna update all of the highway signs in the states, isn’t this the perfect time to finally switch over to metric?

@Jeff: No, it’s not enough.

Michael Z said it well.