Where does Type 1 hinting make a difference?

butterick's picture

I understand what Type 1 hinting is and why it was important in 1992.

But none of the devices I use on a daily basis pay attention to Type 1 hinting.

So is there a class of modern devices that still care about Type 1 hinting?

Or do we pay attention to Type 1 hinting purely for backward optimization?

Thomas Phinney's picture

I assume that when you say "Type 1 hinting" you also include OpenType CFF hinting. I'll refer to them collectively as "PostScript hints" as they are associated exclusively with PS style outlines.

The devices that NEVER pay attention to PostScript hints are those using Apple's rasterizer, namely Macs and iOS devices. Depending on the settings used with FreeType, FreeType MAY ignore the PostScript hints.

That leaves all versions of Windows, most Adobe applications, and many things using FreeType as all being things that DO use the PostScript hints. Also many printing devices, but given today's print resolutions hinting isn't usually a very big deal.

The list is pretty much the same for using TrueType hints, as far as I know. Also, the PS hints may be used as the source for TT hints in conversion.



frankrolf's picture

In InDesign for instance, it makes a pretty big difference.
See below the same font (please no comments on the style :), first unhinted, then hinted (manually).

dberlow's picture

>Or do we pay attention to Type 1 hinting purely for backward optimization?

I believe, most Type1 hinting is performed in an attempt to get TT hints from them automatically.

Richard Fink's picture

Question regarding David Lemon's hinting tutorial:


It states:

“no zone should be more than 25 units deep, and any two zones should be separated by at least three units.”

Does this presume a UPM grid of 1000 or 2048?

frankrolf's picture

1000. I don’t think many fonts with a higher UPM were done when the document was composed.

Richard Fink's picture

I guess I"ll have to go digging into type 1 a little anyway but

Are we sure? Absolutely, positively?

(Type 1 was before my time with fonts. The few conversions I've got are at 2048 but I might have done that as an afterstep or by mistake.)

thx rich

Richard Fink's picture

Chris Slye at Adobe answered the question in an email.

At that time, 1000 units per em was the assumed standard.
thanks to all.


Thomas Phinney's picture

Yes, and to this day you'll find that >90% of Type 1 and OpenType CFF fonts have 1000 upm. Probably 99%. It's still the assumed standard, unless one has specific info to the contrary.

Richard Fink's picture

It makes the math easy, that's for sure.

John Hudson's picture

And there are still some PDF creation tools that presuppose a 1000 unit em for PS fonts, including CFF OpenType, and cause errors when embedding fonts with different UPMs.

dezcom's picture

limiting to 1000 does limit the drawing flexibility, though. I much prefer 2048 and the math is quite easy for it as well. Those of us from the now ancient 8 bit era are quite familiar with factors of 16 even if through digital images. Also, I like to assume even numbers for all dimensions unless I am really stuck. [Actually, I work hard to make everything a factor of eight because that gives more rotational possibilities to the handles that control curvature.] This makes it easy to spot errors because any odd numbered measure is a red flag. After the fun work is done and you are down to the nitty-gritty last tinkers, I am glad I have the option to do a tweener ;-) When I first started designing type, one unit of a 1000 upm seemed so tiny. Now it feels massive. Several years ago, Akira Kobayashi told a wonderful story about his visit with Zapf. While he was showing Herman some drawings, Herman took a look and said "This stem is a bit wider than that one." It turned out that even through advanced years and failing eyesight, Mr Zapf could still spot a one unit difference with assurance. It took me a few years at this kind of work but now I find that distinction quite normal even to me--but I am 20 years younger than Herman, too bad I have not yet achieved his skill ;-)

John Hudson's picture

limiting to 1000 does limit the drawing flexibility

Very much so, especially for designs with fine details and lighter weights in which the optical variations in stroke weight need to be precise. In the past few years, we've made two display types with 4096 UPM values.

Richard Fink's picture

I might be imagining it, but having worked with both 1000 and 2048, I find that for at least what I mostly work on, 2048 works out better.

Syndicate content Syndicate content