Support Web Typography, Support ttfautohint

abattis's picture

ttfautohint is the new autohinter in town, which I hope will make hinting fonts faster and easier, so all type designers can make web fonts that look good on Windows :-)

I made a fun little promotional video, http://youtu.be/81ioae5XNew

There's more details on the ttfautohint homepage, http://www.freetype.org/ttfautohint

And Werner is accepting donations to help fund the development at http://pledgie.com/campaigns/15816

Richard Fink's picture

@db
Not. But not worth pursuing the line of thought. I'm sorry I wrote it already. More weather I'm complaining about.

Spent a lot of time reading off my iPhone yesterday.
It's an existence proof:
TT hinting will eventually go the way of the dodo - once you have the pixel density it's obviously not worth the trouble. Text clear as a bell, at any size. Print-like in a lot of ways.

If you'd like to take on THAT thought, I'd love to hear what you have to say.

Forever Dumbly,

rich

abattis's picture

@TP: If the made it public free software, they would absolve themselves of the costs :)

@RF: Sure, everyone acknowledges that this is all a temporary measure until those screens are on tablet, netbooks, laptops and desktops, and this will take a few years.

John Hudson's picture

Dave: Sure, everyone acknowledges that this is all a temporary measure until those screens are on tablet, netbooks, laptops and desktops, and this will take a few years.

The manufacturing costs of high resolution screens increase exponentially as screens get larger, due to the high attrition rate of dead pixels and other problems. Screen resolution will likely improve for smaller tablets and netbooks but not for laptops and desktop displays. Manufacturing an iPhone 4 display with 614,400 pixels for a smart phone is a different proposition from making displays of 8,516,520 pixels for a 13" laptop.

The "Devices sorted by size" chart here is instructive. Note that with the exception of the IBM T220/T221 'Big Bertha' displays (discontinued) all displays of 200ppi or higher are no larger than 8", with most being below 5". There is a fairly steady decrease in resolution as screen size increases.

I remember back when ClearType was first announced in the late 1990s, the indication from several major laptop manufacturers was that they would be moving to a baseline 120ppi resolution within a couple of years. It didn't happen. It seems to me that 'those screens' are always 'a few years' away.

It may happen, but I stopped holding my breath around 2002.

Frode Bo Helland's picture

John, isn't the typical Macbook somewhere around 130 ppi?

John Hudson's picture

Frode, the Macbooks range from 100-130 ppi. My comment re. 'it didn't happen' was that 120ppi+ didn't happen in the time frame of a couple of years that was expected in the late 1990s, and it didn't happen across the board.

[Note also that 120ppi was only the baseline necessary to hit the ClearType 'sweet spot'. There's a long way to go from there to a place where one doesn't need hinting.]

Jack B. Nimblest Jr.'s picture

Not enough DPI soon enough... I told you so in 2005!? ;)

John Hudson's picture

You did indeed, David. And you were right.

I was lulled into false confidence by the first signs of 120+ ppi laptops appearing in the early 2000s, which I took as a sign that the manufacturers really were going to follow through. Instead, such screens became speciality items for high-end machines -- such as my 133 ppi Toshiba and 145 ppi Dell -- and there wasn't a general increase in average display resolution for either laptops or desktop displays. In trying to understand why it didn't happen, I learned more about screen manufacturing and the problems of attrition as you increase resolution. So now I'm sceptical about the likelihood of significant increases in resolution for larger devices.

Richard Fink's picture

Re: Hi DPI screens.
Gonna be a long time comin', that's for sure. abattis might be as gray as me by then.

In the meantime, all Werner's got is 377 bucks according to Pledgie.
And a not insignificant percentage of that is from me.

All talk no action so far, ladies and gentlemen.
I mean, I don't want to call anybody cheap but, uh, I heard that one of the guys on this thread is so cheap he turns off the gas while he flips the burgers on the barbecue.
But then again, being cheap can get results. Copper wire was invented when two of the guys on this thread found the same penny at the same time in the hotel lobby at TypeCon.

Once again the link is:
http://pledgie.com/campaigns/15816

Jack B. Nimblest Jr.'s picture

> "All talk no action so far, ladies and gentlemen."

Lol, I wish. Unfortunately, the unenvisioners had diagnosed the problem as needing a wrapper, and 'ave put it in place. On the heels of this suckcess, they now endeavor to stuff more Barf into it so that WOFF is going to become a tourniquet around the neck of web publishing ;(

lemzwerg's picture

David,

your last comment is completely cryptic for me. Could you elaborate, please?

Jack B. Nimblest Jr.'s picture

lemzwerg, ttf auto hint outputs ttfs, I assume. Since ttf is the actual default font format of the web accepted I think by all the major browsers, including the latest visions of IE, you have no need to concern yourself with decrypting my comment.

lemzwerg's picture

Yes, ttfautohint emits TTFs, but this could be changed to any outline format which uses TrueType bytecode for hinting.

Si_Daniels's picture

Perhaps more money would be raised if this were publicized on the web design blogs and forums, or perhaps associated with fonts in the Google font directory itself - "don't like the quality of this font's hinting? Pledge here!"

Jack B. Nimblest Jr.'s picture

I think Sii is on the right track, but... I think the idea in a lot of cases if for the message to be, "don't like the quality of this font's hinting?", click here, and the ttf will be auto-hinted, re-manufactured, re-located and re-served.

abattis's picture

dberlow: Don't put the cart before the horse :-)

Sii: Yes, I'm doing what I can to publicize it, and I'll speak to the team about that idea too :-)

Richard Fink's picture

@dberlow

"... and the ttf will be auto-hinted, re-manufactured, re-located and re-served."

Well, that's exactly what's beginning to happen at Kernest Konstellations with the RasterBRIDGE™ line of fonts. I agree with abattis in that, without good tools, what's the point of reprocessing? (I think that's what he meant.) It's the first requirement.

Have you donated yet?

Jack B. Nimblest Jr.'s picture

>Have you donated yet?

Lol, that information is only available in the smoke-filled back-rooms of the type mafia cabal. You want to cut off a pinky, you're welcome to apply.

Jack B. Nimblest Jr.'s picture

abattis quoting me> Don't put the cart before the horse :-)

I have a pretty clear notion that the horse and cart arrangement is seriously in need of rethinking, from one end of web typography, (what the font?) to the other, (what the output device?). At http://t.co/EzTowt1, one can read the early public emergence of the notion that it's more than fonts that need hinting. To be as clear as possible on this, people outside of my circle of fiends are starting to realize that a three column grid with two gutters and two margins has a lot in common with a lowercase "m". The sooner the developers of auto-hinting catch on to this, the sooner the idea I just spotlit will spread from fonts to where hinting needs to begin. Have at it ;)

John Hudson's picture

David: ...ttf is the actual default font format of the web accepted I think by all the major browsers, including the latest visions of IE...

There isn't any single default font format of the web today, because compatibility with the large number of older versions of browsers still in use requires serving at least two different formats.

IE9's raw TTF webfont support requires installable document embedding permission and is explicitly targeted at libre fonts. Personally, I wouldn't want to set the embedding bits on a font in that way just to be able to serve it to IE9 in that format, since that embedding bit can be legitimately interpreted by any piece of software as allowing the font to be installed on the recipients system, independent of the action or intention of the webfont licensee. IE9's raw TTF support really only makes sense in the context of libre font licensing.

With the release of Safari 5.1 in July, current versions of all the major desktop browsers now support WOFF, as do a growing number of mobile browsers. If we eventually reach a point at which there is a single served format equally supported everywhere, that format is going to be WOFF, because that's the only format on which browser makers are all concentrating, the only format that is on the road to being a W3C web standard, and the only format for which there is a web-specific implementation spec and test suites.

The font data content of WOFF files is likely to remain predominantly TTF for text for a long time to come, I reckon; but we've already seen an increase in the use of CFF for display web type.

Jack B. Nimblest Jr.'s picture

>I wouldn't want to set the embedding bits on a font in that way just to be able to serve it to IE9 in that format,

How you say in Canadian? The horses are out of the barn on the embedding bits.

>WOFF, because that's the only format on which browser makers are all concentrating

Proof is in the pudding, and the pudding says otherwise.

And on the side, like a slab of quivering bacon, if that's not true, then what's all the nonsense about absorbing SVG into OT?

dezcom's picture

..with a hint of lime, it saves time ;-)

John Hudson's picture

David, we're clearly eating different puddings. As I pointed out, as of July all the major desktop browsers support WOFF, all the major browser makers have been actively involved in the specification of the WOFF format. I don't see effort being expended on other formats, and IE9's raw TTF support is so heavily laden with caveats about the installability of the fonts on remote systems that I doubt if many of your colleagues will want to license for it.

And on the side, like a slab of quivering bacon, if that's not true, then what's all the nonsense about absorbing SVG into OT?

In general terms, it is about finding a way to add colour and animation to scalable fonts. In specific terms it is about supporting emoji characters.

Syndicate content Syndicate content