Typophile
Google getting into the Web Fonts game?
Oh great, now when I go to web sites I’ll have to wait for the slow Google Analytics servers and the slow Google fonts servers.
Rumors were true then.
Umm...I'm not positive by anymeans, but the domain listed appears to not be a google domain:
http://www.google.com/safebrowsing/diagnostic?site=googleusercontent.com/
If it is, it's odd that it'd be serving malware, but I suppose if it's unmonitored user content, that could be the case.
What is the "web fonts game"?
What is the "web fonts game"?
That’s the one where you design a font for web use and everybody places bets on what rendering method makes the font look best. You get extra points if designers don’t think you’re crazy when you try to explain that great web text fonts will probably require optical sizes customized for different rendering environments.
Yes, we're live now at code.google.com/webfonts. I'll post more in a bit, but I'm hugely excited about this. A huge shoutout to typophile as well, I reached out to a bunch of designers through contacts here.
@aluminum: It is a Google domain, but is split off into a separate domain because it does host less-trusted content, including uploads from users.
@James: Hey, I think Google servers are pretty fast :)
No doubt that leveraging Google's CDN capabilities can be a good thing speed-wise for lots of sites. Raph...what's your involvement with he project?
Paul Irish just posted some details too:
http://paulirish.com/2010/details-on-the-new-google-webfont-api/
@aluminum: It's basically my baby, with lots of help from the rest of the team :)
Raph, I eagerly await your more detailed post and hope you can offer insight into the project, it's scope, goals, benefits and drawbacks.
Also, would be interested to hear any opinions on the quality of these initial fonts posted here: http://code.google.com/webfonts
One thing that I've been thinking about/anticipating for a long time in regards to webfonts....
A whole lot of people are going to subject their readers to display fonts at body copy sizing. Which is even more troubling than in the print world because not only is it stylistically dubious... most display font on screen simply can't hold up to small sizes at screen resolutions.
Case in point... Josefin Sans Std Light. One of the inititial fonts on the google webfonts project. Looking at the specimen (http://code.google.com/webfonts/family?family=Josefin+Sans+Std+Light) it barely holds up below 30 px.
>...most display font on screen simply can't hold up to small sizes at screen resolutions.
It seems simple to recommend ranges of use per style. You can't stop them, but you can guide them.
I found this site very interesting. While bouncing simple samples off the server, it seems to 1.5X the size you ask for, points, pixels or ems? At least it measures that way on my Mac in Firefox and Safari. So maybe text is less of a problem that way... ;)
Cheers!
Also, would be interested to hear any opinions on the quality of these initial fonts posted here:
I wouldn’t say that any of these fonts is a bad screen type. But I really don’t see why Google offering free hosting of these fonts is necessary, or even especially useful. There are already other options out there for free font hosting and for converting such fonts for self hosting. And history suggests that the open-source crowd will eventually create many more offerings just for the sake of doing it.
More importantly, you also noted, most these fonts aren’t going to work at smaller sizes. If Google really wants to add something of value to the web fonts world it needs to go beyond offering the same free fonts that everyone else will. Commissioning type designers to build out open-source screen type families with a wide range of character sets would be much more valuable then hosting extant free fonts ;)
And just to be clear: that comment was entirely in reference to Google hosting those fonts. I do feel that Google putting another open-source web font API out there is probably a good, valuable thing.
Shout out to Ascender Corporation for the Droid fonts (well, a few of the weights). Google commissioned Droid for the Android OS, and should work very well in small sizes. (Saying this without testing it in a web context.)
Droid works great small. That's what it was optimized for.
Serif test page: Droid Serif Test.
Whenever there's a discussion on webfonts, I feel there's a missing point of view. And that is the point of view of Windows users who's only concern with screen typography is 'reading'. On Windows XP and earlier the default rendering of text is plain old Aliased Type (Unless of course you're running an application that renders type differently e.g. Acrobat Reader, Office 2007 etc...)
Aliased Type
Good old Aliased type works perfectly for reading on screen (Courier, Verdana, Georgia at sizes ranging from 8 to somewhere in the teens are pleasant to the eye and easy on the brain). Quality typefaces whether an FF Library font or a TDC Awarad Winner, whether FontShop dubs it 'Web' or it bears the name Droid, all of these and the thousands of print typefaces out there don't even come close to the quality of the "Core System Fonts" when it comes to reading aliased type in your web browser.
Anti-Aliasing
The Macintosh has had a great anti-aliasing engine for years now. The baseline and the x-height always align (Unlike PDFs) it looks phenomenal in small sizes even readable in size 6 for certain typefaces. It is part of the OS it is always turned on you don't have to worry about it. It works in all applications unless the application over-rides it. Conclusion: text on the Mac is beautiful to look at, the new @fontface websites look delicious.
On Windows it is optional to turn on Anti-Aliasing, but with all my respect to those who worked on the algorithm, it looks like a beta. The thing is the Microsoft guys were very resistant to the idea of imitating paper on screen.
So if you're publishing websites for Mac users to view then the future is now. Otherwise, wait for Microsoft to buy or steal the Apple Anti-Aliasing algorithm and for all PC users to switch to that latest technology, or figure out new fonts that look good when Aliased and yet are beautiful typefaces when it comes to their Contour and overall texture. "The" example to follow is Georgia. Oh wait Verdana Condensed sounds like a good start.
.
.
.
Google Chrome on Windows XP
.
.
.
IE8: The new Internet Explorer slaps Clear-Type on everything
.
.
.
Spiekermann.com // Google Chrome
.
.
.
IE8 fail, the only thing worse than Arial on Erik Spiekermann's website is Anti-Aliased Arial on Erik Spiekermann's website.
it is true... text rendering on windows is pretty terrible. "Clear-type" is a hack that is being marketed as a feature. I can consistently see the color fringes on clear type handled text... and it bugs me a lot.
The interesting thing—which gives weight to the argument that the type with which a reader is most familiar is the most legible—is that many a longtime XP user will complain that fonts rendered on Macs are too "fuzzy". It infuriates me.. but it's a fact.
@Johnny: Most of us (not counting those who are visiting Typophile only for ID requests) are not unaware of the challenge different rendering engines pose. The approaches to rendering differs, yes, but so does the opinions about which one’s the best. This terrain is hard to navigate anyways. There’s multiple rendering settings, multiple browsers, multiple OS’s and multiple font formats.
And a lot of drag too, with users still running old versions of OS’s and browsers.
Indeed; it's very hard to definitively say that any web font "works great small". Droid might work alright in one environment, but it also falls apart in other environments.
While a discussion over which rendering approach is best probably won't help much, I do agree with Johnny that many people – even those of us who should know better – often don't take the time to check how things look across the board. And understandably so; doing a proper test even in a minimal number of browsers can be a pain.
Unfortunately, it seems that a majority of the people who only check one environment seem to do so in one of the more forgiving environments (on a Mac).
@Nick, this is exactly what I'm talking about.
@Neumann, I've heard the "fuzzy" comment but I've also heard "Holy Shit, this screen has a really high resolution"
@Frode Support is eventually discontinued but XP is one of the most widely used. I agree there's way too many browsers with way too many surprising results. I don't know if this is the result of the free market or of the disregard to regulations and conventions.
Unfortunately, it seems that a majority of the people who only check one environment seem to do so in one of the more forgiving environments (on a Mac).
I don’t think that checking is the problem. It’s that not many people have the resources to develop well-hinted TrueType fonts, much less check to see how they look. Microsoft needs to get on the ball and make TrueType-hinting more designer friendly or the Windows rendering problems aren’t going away.
Cardo-nal Bembo.
Cute.
For those who haven't seen the new Google font directory site through the lens of Windows yet, these screenshots of some of the font samples with the default settings in IE8 on a Vista machine give Johnny's complaints a bit more context…
Regardless of your personal antialiasing preference, I think it's safe to say that the fonts above aren't very usable (of course I'm always open to other thoughts). I fully realize that text fonts which render well in Windows are hard to make (nevermind get as open source), but I'm surprised that most of the stuff in the current Google font directory made the cut.
@raph...thanks for all the info and participating in the discussion!
@james...the main benefit of using Google to host the web fonts (over any other option) is primarily ubiquity. Lots of folks are using Google as a CDN for things like jQuery where not only is it then being requested from a large (and localized?) CDN, but the odds of the file already being cached in the user's browser is increased.
@nicksherman
>I think it's safe to say that the fonts above aren't very usable
They'll get used anyway. Don't delude yourself.
It ain't what they is, it's what they're not. TNFG. (They're not f-cking Georgia.)
And they are free. Free counts for a lot.
Besides, betcha the look of these fonts can be improved greatly without much ado.
Gonna give it a try myself. Free, right?
Let you know.
Rich
@neumann:
A little OT, but I find interesting that at least two of the initial offerings are typefaces put up for critique on Typophile.
I feel that since the @fontface now google web fonts trains have left the station..
the best way to mitigate disaster might be as dberlow commented "It seems simple to recommend ranges of use per style. You can't stop them, but you can guide them."
Raph—if you're still reading this thread—any consideration of providing some recommended uses and sizes info in a prominent place for each font sample.
I stumbled across Typekit just now while looking at all this Google fonts business. Why would it be preferable to use Google's service for @font-face type hosting when they have a pretty crappy catalog of unknown faces when a designer could use Typekit for quite a nominal fee?
Forgive me if I sound uninformed, but I am.
Also, I use Mac OS and do not have the nasty rendering issues PC users experience as described above. I presume using a service like Typekit does not solve these issues of rendering disparities on different platforms, browsers, engines etc?
Awesome, hope they add most of the stuff Font-Squirrel has @font-face kits for
http://www.cnn.com/2010/TECH/05/20/cnet.google.fonts/index.html?eref=rss...
Now *that's* a url!
@jayyy there are technical/performance reasons to choose a hosted Google solution (CDN/caching, etc), but your criteria is broad commercial font selection at this very moment, then typekit is the better option for now.
Hey all, I took screenshots and compared of each of the typefaces in all three main Windows rendering styles: ClearType, Standard, and Anti-Aliasing Off. Unsurprisingly, the Droid Family is the only font in this set that I would use for body text. All others are only suitable for headers or display text. And I would avoid using Inconsolata, Tangerine, Josefin, and Cardo except at really large display sizes (36px or larger).
You can see my font comparison chart here. Each font listed links to it's corresponding screenshot to make it a handy reference for all you mac-only designers.
@James Puckett: Google is indeed commissioning type designers to do new work; I put Raph in touch with some MATD/T&M graduates, and Vernon Adams (MATD) produced http://code.google.com/webfonts/family?family=Nobile for the directory.
If anyone would like to contribute a font, complete this form:
http://googlecode.blogspot.com/2010/05/introducing-google-font-api-googl...
maybe this is off topic, but who else offers web font hosting now?
right now from a quick search its:
typekit
fontdeck
fontslive
typefront
fontspring
google fonts?
kernest
typotheque webfonts
any more?
[edit: added extras]
For the Typowiki:
List of web font services
Ascender Fonts
webfonts.fonts.com
Fontshop
* fontspring doesn't host, it's purely retail.
I'd bet there are a few more in the works, too... sigh
Nick,
I'm surprised that most of the stuff in the current Google font directory made the cut.
I guess we just have to accept that not all hits offered by Google are perfect. When you do a normal Google website search and 5 out of 10 hits are useless you don't complain, you just use the other ones. Google is pretty good at finding relevant results but you still have to filter out a lot of junk yourself. I don't have a problem with that, whether it is about websites of fonts. Maybe they will come up with a smart ranking system that helps you separate the wheat from the chaff.
@tim ahrens
When you do a normal Google website search and 5 out of 10 hits are useless you don't complain, you just use the other ones.
Maybe they will come up with a smart ranking system that helps you separate the wheat from the chaff.
Thank you for stating the obvious - it too often goes overlooked.
1) If a font looks really awful, don't waste time with it, pass it by. Plenty more further down the list, or on another site.
2) Ranking system - of course, there are many ways to handle this. User comments is one. A human editor whose job it is to prioritize is another one. (Wondering if this will be in Raph's remit.)
Typekit's gallery is terrific - because you can see their fonts as they are actually being used in the context of a complete design.
However, with typefaces, what looks good IS good. So, as long as the font is being shown using an in-browser specimen that you can examine, it's not like any secrets are being kept.
It will take awhile for all of this to sink in.
rich
@jayyy
Why would it be preferable to use Google's service for @font-face type hosting when they have a pretty crappy catalog of unknown faces when a designer could use Typekit for quite a nominal fee?
Two things:
1) I expect sharp improvements in the Google selections very quickly. When Typekit first announced, they had NO CATALOG. Nothing. They were in private Beta (if that) with nothing to see at all. Vaporware.
Google's style is different, they don't mind releasing with next to nothing and evolving in full public view.
And in that there is now precedent for font hosting and they don't have to deal with complex licensing issues from commercial font designers concerned with unlicensed copying, expanding the catalog is a lot easier and can be expected to move much more quickly than it did for Typekit.
2) Because the fonts are freely licensed, there aren't any of the inefficiencies involved with obfuscation if you use Google. Typekit's appeal to font designers is that they have a strategy in place to discourage download-and-installation.
With Google (and Kernest, to name one other), you can use straightforward CSS. (Their javascript API is cool to have, but optional. Icing on the cake.)
Them's the differences that I can see right now.
I don't think the "catalog is crappy". The fonts offered are great in my opinion. I found Barry Schwartz's Goudy to be a closer match to my 24pt Goudy Old Style Italic metal font than Adobe's. Iconsolata is a personal favorite.
The Google team working on this is doing a great job, and the service they offer is great. The only problem is rendering and that's not Google's problem.
The @font-face model (delivering the fonts and not controlling the outcome) is one way to approach typography on the web. The other model is delivering rendered type (for example: GIF, Flash etc...) a great example of that is the Featured Face here on Typophile, another great example is the Preview on MyFonts and other font retailers; this reminds me of the 3rd party typesetter it's like working with Linotype lines and Linocuts instead of moveable type.
@font-face is like cast Glass Type. Unlike Lead Type(web safe fonts) it breaks it doesn't bend. But with appropriate ink(Browser) and an appropriate Press(OS) and above all with a skilled printer(Web Designer) it may very well produce great results.
does today's google doodle count as a 'web fonts game'? http://www.google.com/
So do any of these services allow you to upload your purchased type for online use that would comply with most EULA's and not allow end-user pirating of the type?
Cufon *seems* to fit the bill. TypeFront does but states you must use type with a "web-friendly license"...
>does today's google doodle count as a 'web fonts game'?
Maybe. Wonder where they got the idea... :-)
http://www.microsoft.com/typography/web/embedding/demos/7/about7.aspx
Cheers, Si
I fail to see why i would want load the @fontface from gooogle rather than my own server.
@jamespuckett
I don’t think that checking is the problem. It’s that not many people have the resources to develop well-hinted TrueType fonts, much less check to see how they look. Microsoft needs to get on the ball and make TrueType-hinting more designer friendly or the Windows rendering problems aren’t going away.
You've got it right. In a world where everything was funneled through the OS - mostly Windows, obviously - Microsoft's reliance on a small group of font technologists that already knew the ropes was tolerable. Now it isn't. But it's been neglected too long for it to be easily undone.
TrueType is extremely under-supported - especially hinting. There's been no investment in learning materials. This is obviously a frustration for professional type designers, so you can imagine how a novice feels.
How many expert TrueType hinters are there in the world? A dozen, maybe? It's kind of nuts.
Many of these concerns are so silly!
While I can understand the hubbub over which typefaces Google chose, cannot we all be happy that typography is once again in the limelight, with a massive player by its side?
I was just lamenting to a group of web developers that paltry choice of typefaces available and they all scoffed at me. So unimportant!, they said, get over it! they said. But web developers are getting more graphic tools in their arsenal and they have to pay next to nothing to do it.
Why is there no celebration over that?
The best way to kill something is to restrict it to the closed doors, in tall Ivy-draped towers. That's my opinion.
Why is there no celebration over that?
What do you want, an iChat circle-jerk followed with champagne toasts? Celebration isn’t going to accomplish much. Cycles of constructive criticism and hard work are far more necessary.
"I fail to see why i would want load the @fontface from gooogle rather than my own server."
Performance.
@James Puckett
Circle jerk—you mean that isn't happening right now? I see many accomplished typographers lamenting that an uninformed public will be utilizing fonts available to them incorrectly.
In other news, launching rocket ships propel forward.
And while you may think that there is constructive criticism to be had, it's more of a closed issue from how I see it.
Could better (better as in: superior design) fonts have been chosen? Purely the speculation of an elite few. Nobody cares. You could recommend typefaces until the sunsets; can we not all agree that there are 100s that were suitable for selection? A few typographers were lucky this week and received Google's watchful eye. Not much else to be had there.
Unless of course you believe that the typefaces chosen are poorly designed, period. In which case well... I won't disagree with your skepticism.
As the guy behind Crimson Text, I'd like to add that some of the typefaces just aren't perfect yet. Google has always been releasing early and often, and the font directory is clearly labelled "beta". Many of the faces don't have the most rudimentary hinting, some (like Crimson Text) aren't even really spaced yet. (And still, I saw a weblog using Crimson via the API yesterday. Wasn't too pretty.)
Fonts like Vollkorn or Cardo have been around for a long time, and Google is trying to breathe new life into them (Vollkorn now has a Bold, for example). I'm sure you'll see huge improvements over the summer.
I see many accomplished typographers lamenting that an uninformed public will be utilizing fonts available to them incorrectly.
I see them too. But most of the fonts were freely available a long time before Google's font directory. Will the hype on the new API reach enough of those "uninformed" masses to allow them to ruin the web by setting text in 12pt-Tangerine? I doubt it.
How many expert TrueType hinters are there in the world? A dozen, maybe? It's kind of nuts.
Full ack. Microsoft should be paying damages for the pain TT hinting really is. Mac and FreeType seem to be doing fine without it anyway ...
>Google has always been releasing early and often, and the font directory is clearly labelled "beta". Many of the faces don't have the most rudimentary hinting, some (like Crimson Text) aren't even really spaced yet.
This approach doesn't work very well for fonts, where users expect the basics (eg. spacing) to be right, and more importantly not randomly change from version to version causing documents to reflow. That was one of the main reasons Adobe, when upgrading their library renamed fonts Std and Pro. It's great that Google is paying for freeware fonts to be upgraded but there's an argument for nailing the basics first before unleashing what amounts to core web font on the world.
> Microsoft should be paying damages for the pain TT hinting really is.
LOL! You do know who invented TrueType, don't you? :-)
>What do you want, an iChat circle-jerk followed with champagne toasts?
Uh, is there like a site or something where you can sign up for invitations to those? ;)
@all
To show you how quickly things change:
A lot of commenters here and elsewhere write like Typekit has been around forever and a part of the natural order of things.
Last year when they were a part of the Web Fonts panel at Typecon in mid-July, 2009 they had nothing but an alpha demo. Nearly vaporware. All they had was a proof of concept to show the type designers at the conference. After that, it took months for them to get out of Beta.
So they've been around well less than a year. (And now they're the gold standard - this shows you the power of being "first with the most".)
Google is doing with Google fonts what Google does: release at a very early stage and then evolve in full view. Plus, it's set up as an open source project from the get-go and so collaboration will be a big part of the effort. Plus, they've partnered with Typekit so the learning curve for them is greatly reduced.
Bitching about the "quality" of the few fonts they're offering beside Droid is beside the point. They're just getting started.
What's important is that they have started: @font-face is supported broadly enough to warrant Google's attention.
@sii
sii>LOL! You do know who invented TrueType, don't you? :-)
Yeah - I was thinking about this yesterday and couldn't help but smile and shake my head. Life and its ironies.
rich
LOL! You do know who invented TrueType, don't you? :-)
Haha ... I do. Ironic indeed :)
Sii >LOL! You do know who invented TrueType, don't you? :-)
Imagen?
Scosch> Mac and FreeType seem to be doing fine without it anyway ...
The Mac is a closed hardware environment where apple goes to great lengths to make sure fonts just work on the Mac, or are completely out of the user's hands as in safari and system fonts, iPhone as well.
And, I don't think freetype is doing fine without hints or hinting, unless you e.g. would be fine without you, e.g., or springtime would be fine without springtime. Like windows, Linux and other freetype-using systems, quality rests on a diverse technologically provocative hardware and software environment and freetype's usefulness there is varied.
Cheers!
Point taken.
I was oversimplifying. I find that non-hinted fonts usually look acceptable on Linux (FreeType's autohinter, that is) and Mac (which also ignores most TT hints nowadays), but are often completely illegible in Windows (with or without subpixel rendering). But all that has been discussed to death already.
I solemny swear to just shut up next time :)
>I was oversimplifying.
Me too.
The wikipedia needs editing re fonts, hinting, linking, em, embededding, and more.
The Mac either uses all the hints, as when the user sets smoothing to a size in the appearance menu, and then uses a size smaller than that setting. Or the user gets no hints, by choosing a point size larger than that same setting. Its simple, and the only bad match is a user with a setting in the appearance menu using an unhinted font below that setting.
On Windows, unfortunately, there are a lot more bad matches, and to bring it back to topic, the matches are beginning to swarm. A new and interesting question about font development, is whether corrections can be made to fonts fast enough in a live web environment, so bad matches are rematched to good type for users faster than new bad matches "arrive".
Cheers!
Coming back to the Google service:
Are there already any (good) real-world examples of Google webfonts in use?
.
Captured from web.archive.org on 17 Jul 2019. view this capture