The rendering engine has the final say on how it interprets the hints. Since there are multiple rendering engines, interpreting both outlines and hints in different ways and for different devices, the 'rendering concoctions' are pretty diverse.

>the ’rendering concoctions’ are pretty diverse.

...and largely undocumented.

I often wonder how many designers just run the Adobe auto-hinter and forget it because trying produce and test well-hinted cross-platform fonts seems to involve a tremendous amount of trial-and-error. I expect to see users demand great hinting as web fonts get rolling, but how many type designers have the time to hint fonts and test them on IE, Safari, Firefox, and various cellular platforms?

If you're talking about PS fonts, then running the Adobe auto-hinter is going to produce pretty much the same results as manually putting in hints: there may be occasionally different decisions for certain features, but major stems will get the same hints. The ‘big’ differences in PS rendering will be affected by font-level hints: standard stem values, blue zones and blue scale.

However, despite significant improvements in PS rendering (and the smaller size of CFF fonts), screen typography remains dominated by TrueType fonts. And the diversity of TT rendering models is much greater than that of PS: everything from Apple Quartz rendering completely ignoring all hints to b/w bitmap displays that rely on deep hinting down to the individual pixel level.

I often wonder how many designers just run the Adobe auto-hinter and forget it because trying produce and test well-hinted cross-platform fonts seems to involve a tremendous amount of trial-and-error.

Actually, if you prepare your fonts well for hinting such as setting good values in the standard stems, the alignment zones and the blue scale, the Adobe hinter does a fantastic job! It is a professional tool. Afterwards you can test the fonts and if you see some strange things you can manually tweak the hints to make the rendering better.

And a prediction of the future: Hints will not be necessary at all :)

>And a prediction of the future: Hints will not be necessary at all :)

Agree, and I'd put that date at 1st July 2036.

Cheers, Si

Goran: And a prediction of the future: Hints will not be necessary at all.

According to Apple, the future is 2000. That's when we worked on a font for them and were told not to bother hinting it because OS X was going to ignore all hints henceforth.

Those Apple hinting patents expire this year, so we can expect a flurry of activity from Freetype implementers.

I know little about hinting, but from what I've heard TrueType hinting is such a complicated art that there are only a couple of people in the world who can hint hangul (Korean alphabet) TrueType fonts well. So if I decide to learn TrueType hinting and it takes me years to master TrueType hinting for hangul, what are the chances that it will all be for nought because those hints will have been rendered obsolete by then?

Where can I find a beginner's guide to hinting? I want to decide if it's worth the effort or to just stick to autohinting.

> a beginner’s guide to hinting

That might be an oxymoron ;-)

If the reason hinting might "become obsolete" is that device resolution will increase so much that the human eye won't notice a pixel here, a pixel there, I would argue that there will always be room for low-resolution technology. Will we allow this segment of typographic practice to suffer? Or will the selection of typefaces for this segment always be limited due to the overhead of hinting?

If the process of hinting is unnecessarily complex, we should criticize it. On the other hand, if it is what I think it is – many talented but autonomous professionals, working on their piece of the type rasterization assembly line – we should try to clarify it, or at least talk with these professionals (and listen when they talk with one another).

I'm not sure every type designer should need to learn hinting.

Having such tasks as hinting lifted from your shoulders is one of the many good reasons to sign with a foundry who will do most of the technical work on your fonts :)

@Sii
Really? The patents expire this year? You're meaning the patents within TrueType that effect hinting (the ones that forced Freetype to come up with their own method), not the anti-aliasing done by Quartz, correct?

If so, and Freetype gets to implement that, that's a boon to free OSes right? That would be one step in making them not-so-crappy looking.

Yep. Actually two of them must have expired last week. The last one on the 28th of this month... Party!

http://www.freetype.org/patents.html

If so, and Freetype gets to implement that, that’s a boon to free OSes right?

AFAIK the particular code is already implemented in FreeType, it's just disabled and you can enable it. E.g. if you are in some place where the patents don't apply or if you have a licence from Apple.

AFAIK the patented TT instructions are not useful. They have had no effect on the quality of anti-aliased type. OS X might ignore all hints, but in the end it is y hinting, like everything else. Learning to y hint is pretty much like learning to design type with strict alignments.

Cheers!

>They have had no effect on the quality of anti-aliased type.

Like I said, AFAIK the patented TT instructions are not useful. In the context of teaching someone what TT instructions can do for AA type, the patented TT instructions have no effect on the quality of anti-aliased type. FreeType is not waiting to 'flurry' AFAIK.

The patented diagonal instruction should not be interpreted by feeble rendering, which is what sii's displaying, undoubtably on a windows machine. My guess is that the graphics state is not being properly managed in relation to the use of reference point settings during the positioning of certain diagonal stroke boundaries.

Sii, inform us of the age of this illustration, the source of the font and hints, please. :)

>Basically: to what rendering concoctions are our web type specs subject?
Basically, unless you 'demand' the web user alight on a standard CSS size, like 'medium', (so you can take over the sizing from there, if you can), you have no control over size.

Cheers!

LOL, that picture is a persian extra's last look at life, where did you get it?

Cheers!

The (rendering) image is taken from the Freetype patents page referenced earlier ... http://www.freetype.org/patents.html

Cheers, Si

Well I wasn't going to post this version...

Source... http://www.allposters.com/-sp/300-Posters_i2114246_.htm

Miguel (Sousa) talks a little about hinting in his article about Times Reader 2.0.

Correction on the expiry dates. According to http://lists.nongnu.org/archive/html/freetype-devel/2006-11/msg00006.html - The expiration dates are:

5,155,805: October 13, 2009
5,159,668: October 27, 2009

>The (rendering) image is taken from the Freetype patents page referenced...
MS got screwed on Arial twice? ...the FreeTypers say these instructions are not an issue, and for this user's question, or any hinters issues, same goes.

Cheers!

"...It’s the portrayal of the historic Battle of Thermopylae in which King Leonidas and 300 Spartans battled to the death against Xerxes and his enormous Persian army."

Funny, it doesn't look like Gerry :-)

ChrisL

> Where can I find a beginner’s guide to hinting?

Bendy -- This time a little more helpful answer. I just ran across this article on the basic concepts of hinting, by the lead developer of Microsoft's Visual TrueType tool: http://www.microsoft.com/typography/tools/trtalr.aspx

You can be the judge of whether this qualifies as a "beginner's guide."

Thanks Kent!

Good stuff, Kent. Thanks!

I've always thought the Adobe Type 1 Font Format spec did a good job of explaining at least some of what's going on with Type 1 hinting.

I am biased, but I've always liked the Type 1 approach to hints. The key is in the word "hints"... they offer some advice about the dimensions of the font and its glyphs, which the renderer can use to make better decisions. TrueType instructions are cool (and there was a brief time many years ago when I did it "professionally"), but as many have observed, it's starting to look like a lot of hard labor for diminishing returns.

Even if one imagines displays becoming more common in higher resolutions, Type 1 (PostScript, CFF, whatever) hints still look useful. Remember, Adobe's initial concern was making type look good on early 300 dpi printers, not screens. Of course, those printers didn't have the benefit of grayscale antialiasing as screens usually do, but still...

http://www.microsoft.com/typography/hinting/tutorial.htm

and this one is a blast from the past not a leap to the Futura!