TypeFacet Autokern

Ben Blom

But I cannot understand the opposite, that the spacing of a glyph informs the shape of that glyph.

The simplest example would be a monospaced family.
A less simple example would be styles that are required to match the metrics of each other, either for functional or aesthetic reasons.

These examples are interesting, although they can be considered as exceptions to the situation in most fonts. What characterizes these examples, is that the advance width of the glyphs are being determined before the actual design of the glyphs (shapes). So one could say that the advance width informs/influences the shapes which have to be designed, because each shape has to fit well within its given advance width.

However, the given advance width is no more than a constraint or boundary condition. Within the limits of this condition, one is free to design the shape in any way one likes. The fact that the advance width has a certain value, does not give any clue how to design the shape, except for its width. The values of the margins of the shape which is being designed within a given advance width, do not give any clue how to design the shape, except that the sum of the width of the shape and its two margins, should equal the advance width.

Ben Blom

but he makes every effort to stay close to the designer's original intention

Sure, but the original spacing is thrown out anyway. Staying close to the designer's original intention, can only be done by changing the parameters of the iKern program, so the spacing of the font is still put completely at the mercy of the iKern "auto" application. This throwing out of the spacing, obviously, will be hard to swallow for a typeface designer who put a lot of time and energy in the spacing of his fonts (and who might, perhaps, have a big ego — and who might, perhaps, consider the spacing as part of the design of his fonts — and who might, perhaps, have spaced his fonts for many many years himself in the same way, and cannot get used to the idea that an "auto" application will redo his spacing).

Separating the autospacing from the autokerning in iKern, would, I think, conflict with the original design philosophy of iKern. The fact that the spacing and kerning in iKern can only be done together, is, I think, not a "design error" of iKern, but is, I would say, the result of the insight that spacing and kerning are related, and because of this, are done by the same algorithm.

The phrase "as the program is currently written" suggests, that Igino is considering to rewrite iKern, so it can keep the existing spacing of a font, and only add kerning. That would really surprise me, because then he will abandon his original design philosophy of iKern. Did Igino say he is considering such a rewrite? (Perhaps he wants to extend his market by also offering his excellent service to those typeface designers who do not want their spacing to be redone, but who might consider to outsource the kerning.)

Ben Blom

And yes, there are universal principles of letterspacing.

Really? First I will quote William Berkson from http://typophile.com/node/31468, and then I will quote Chris Lozos from the same source. William Berkson:

You have already read Tracy's view of spacing, which seems to be the most generally accepted. According to him the side bearings of the control characters H and n should be slightly less than the width of the counters, and the other letters matched to these.

Frutiger says here that in a seriffed face, the width between the counters and the total of the side bearings should be equal in a seriffed face--wider than what Tracy recommends. I haven't read Frutiger's books, but he may state there his theory about sans faces--I'm sure he has such a theory, given how deliberate he is about his designs.

Other theories about spacing have been put forward by David Kindersley. His starting point was different, as his basis was the blacks of the letter, not the whites.

Hermann Zapf developed an automatic spacing system based roughly on the same approach of Tracy--looking at counters of control letters. He also took into account nearest proximity in letters like r, etc. The program he and others developed was bought by Adobe and became the basis of its 'optical spacing'.

There is also an Italian engineer who has developed an automatic system called IKern. I don't know how that works.

In spite of a lot being known about spacing, I think there is still a lot that is mysterious, and just has to be done by trial and error, with your eye and other faces you admire as guides.

Chris Lozos:

It is the eye that reads it, it is the eye that should judge the spacing. Don't go overboard with exacting schemes. You have to optically compensate for so much anyway, there is not much value in it.

From these quotes I conclude, that there are no universal principles of letterspacing — there are various approaches.

The "please-keep-your-hands-inside-the-moving-vehicle-at-all-times rule" (I love that name) is not a spacing rule, but a design rule. It is about how to design glyphs, not about how to "design" spacing. This rule is not universal, although I think it's a very wise practical rule. This rule is not universal, because not every typeface designer follows this rule: "see the leg of the R in Bembo and Trajan".

There are other practical design rules, like this one: Don't use serifs in a sans-serif font. Most designers follow this rule, but it is also not a universal rule. When Matthew Carter didn't follow this rule in the uppercase "I" of Verdana, many considered this as a brilliant idea. I guess that when a lesser known typeface designer did a thing like that before the publication of Verdana, the typeface design community might have considered that as a design error. Now it seems to be quite fashionable to design sans-serif fonts with a serifed uppercase "I".

Concerning letterspacing, I think that various approaches are possible, and many of them might lead to a good result. With Mao Zedong, I would say: "Let a hundred flowers bloom; let a hundred schools of thought contend".

[Deleted User]

The user and all related content has been deleted.

Max Phillips

Ben, neither I nor anyone else has suggested that the eye isn't the final judge of proper spacing, or that there aren't different valid ways to space a typeface. I did not say there were universal formulae for calculating proper letterspacing; I said that there were universal principles, the main one being that negative spaces within and to each side of a glyph must balance. There are many approaches to attaining balance, some of them automated, but you judge balance by looking at it. And—at the risk of another multi-paragraph refutation—if you do not take the requirements of good letterspacing into account while drawing glyphs, your typeface probably will not be well balanced or have a pleasing rhythm.

Nick Shinn

I recently designed a typeface for a major international branding company, and during the design process provided them with a completed (if not final) font to try out in the layouts they were doing for their client.

When they showed me the proofs, I was disappointed, they didn’t look good. On closer inspection, I realized that the spacing/kerning was different from that with which I had become intimately familiar. It transpired that the agency used InDesign’s “Optical” kerning as their studio default.

When a top professional outfit can so blithely believe in the merits of bog standard automation over and above careful “manual” crafting, this illustrates the dangers of an unquestioning acceptance of the benefits of new technology.

The graphics industry is in a constant state of technological makeover, as old processes and jobs are replaced. The old jobs vanished, and new ones emerge in which direction and management replace crafting, becoming a new way of crafting.

It’s likely that the inventors have only a superficial acquaintance with the culture and aesthetics of the task they’re automating.

Eventually, the aesthetics of the new technology catch up with and surpass the old way of doing things, which, not so long ago, itself utilised new technology to supplant old.

At the moment, I am extremely skeptical—in my own work—of the next generation of auto spacing and kerning, although I use classes, which is also a form of auto spacing—as is the simple use of a ruler!

I prefer to work in the manner to which I am accustomed. I spend too much time as it is involved in upgrades of one kind or another, which all take me back a step out of my comfort zone, before I can move ahead two.

Ray Larabie

Spacing is the reason letters look the way they do. Those minor adjustments I make after spacing has been finalized is part of the overall harmony of the font. As for kerning, how much I tighten the screws is a matter of taste. Sometimes, something you might think is under-kerned is like that for a reason you don't understand. If I'm going for more of a 1970's display font flavor, I might tend to over-tighten certain pairs. Imagine you're spacing and kerning Kabel Heavy. If you want to give it more of a metal type feeling, you might space it loose and under-kern it to make it look more believable as metal type and to create a harmonious rhythm. If your goal is more of a T Rex. Electric Warrior display font kind of feeling, you're going to get mathematically precise with the spacing so the gaps are consistent and then kern the hell out of it to bolster the gap consistency. It's not just a matter of turning a dial and kerning more or less. Those decisions are part of the flavor of a font.



However a better kerning diagnostic tool would be useful for spotting problems especially in fonts with unorthodox letterforms. Diacriticals colliding for parentheses . . . nobody likes that stuff.

Kent Lew

When they showed me the proofs, I was disappointed, they didn’t look good. On closer inspection, I realized that the spacing/kerning was different from that with which I had become intimately familiar. It transpired that the agency used InDesign’s “Optical” kerning as their studio default.

Nick — Christian Schwartz once told me a similar horror story regarding a magazine his studio was working with. And I’ve had the same sinking feeling when I’ve seen my fonts used in certain magazines, only to realize that they’re using Adobe’s Optical kerning. (I’m sure you’ve heard me rail against this before.)

Further, it seems there may have been (at one time, anyway) an Adobe consultant going around telling publication design teams that Optical was the recommended setting and yielded the better results. I heard this third-hand and have not confirmed it, but that would explain a lot. Grrr.

Jan Schmoeger

Before I became involved in type design I also believed that Adobe's Optical kerning was a good thing. I vaguely remember fiddling with the kerning tables of some Type 1 fonts in Pagemaker and Quark, namely Bembo (I think) because the built in metrics were (imo) atrocious. I recall publishers' editors complaining about the kerning of /fullstop/space/T, which practically obliterated the space (can anyone confirm this?). Then, the "Optical" setting seemed to help. Disclaimer: I was always in a hurry and paid very little—not much time for detail.

Robin Mientjes

I am heavily on the side of Ray, in that, if not used for outright creation, tools are most useful for inspection and diagnostics. Or that is how I see it, and it seems to align with what Ray is saying.

Tools like Just van Rossum’s ttx and Tal Leming’s Area51 and Prepolator do things that are tedious or simply impossible by hand. Much like fixing a motorcycle, sometimes you need the voodoo tools – like the current meter, which sees things I can’t see.
These tools do not impose any particular workflow, and do not care about it either. They just sit there, waiting for you to get stuck on a particular problem.

A tool that tries to force its hand on me is a tool I rather not have. The spacing and metrics of a font seem to be something that a tool can do, but that tool is blind in a way that is hard to put to words. It is blind for what we call taste. And I might fix my motorcycle pragmatically, but I insist that it looks good too. Something automated (or even somebody else’s hands, imagine that) could do a good job of fixing something, but I will verify and verify and verify until my eyes are satisfied.

Of course, eyes are only tools, a scientist would say, and then that scientist would do an immeasurable disgrace to the Buddha. But that requires a subforum.

Robin Mientjes

I would like to add that this discussion has never been about tools not doing a good job. It’s just that they never seem to do a great job. What makes something beautiful is, I think, a complex combination of simplistic principles, that form together to magically still be simplistic. This complexity is something that is usually best reserved for our hearts, not our minds, as it is whimsical and most certainly not scientific. Neither am I, or are these paragraphs of mine, so if you require the science, please just ignore me.

James Puckett

A tool that tries to force its hand on me is a tool I rather not have. The spacing and metrics of a font seem to be something that a tool can do, but that tool is blind in a way that is hard to put to words. It is blind for what we call taste. And I might fix my motorcycle pragmatically, but I insist that it looks good too. Something automated (or even somebody else’s hands, imagine that) could do a good job of fixing something, but I will verify and verify and verify until my eyes are satisfied.

I think that this is why iKern has been so successful. iKern isn’t an autonomous kerning program. It’s a program that derives spacing data and kerning pairs based on suppled parameters. iKern requires multiple rounds of designer interaction to get good results. I don’t know how that compares to DTL Kernmaster.

Ben Blom

A computer can easily do it if you feed it with the right algorithms and set the right parameters.

If someone can prove that, I’ll be delighted. But so far all auto-stuff I’ve seen including this one we talk about is not very impressive if you look at it at from a professional level.

The proof that it is possible to get a professional quality font when using an autokerning/autospacing program to do the spacing/kerning of that font, is like this. Let's consider Brandon Grotesque by Hannes von Döhren. Do we consider this as a professional and well-balanced typeface, with a good quality spacing and kerning? If we answer "Yes" to this question, we have the proof. If we have an example of one font where it can be done, we have the proof that it can be done.

This is, of course, not a proof that all "auto" programs are good; I guess many are not. Also, this is not a proof that all typefaces which have been spaced/kerned by an autospacing/autokerning program, are professional quality well-balanced typefaces; I guess many are not. Moreover, the fact that an "auto" program can be used, is not a reason to use it. It's just an option. I think that creating spacing/kerning without an autokerning/autospacing program, is a very commendable way of doing it.

The way is not in the sky. The way is in the heart. —Buddha

Jan Schmoeger

iKern isn’t an autonomous kerning program.

What's more, I believe Igino flashes the pairs on a screen and adjusts the iKern output by hand. His work always comes back as a suggestion, and the customer is free to adjust it or ask for changes, both global and specific. Igino's thoroughness and consistency were the main attraction to me.

Jan Schmoeger

Here is the Bembo "phenomenon":



I feel that this historical context is important for understanding the introduction of the "optical" setting in InDesign. It could have been there to globally address poorly kerned fonts in routine typesetting.

Ramiro Espinoza

.

Nick Shinn

Not all amazing new inventions catch on immediately, if at all. And often in ways quite different from intended.

The Newton was ahead of its time, more successful subsequently as the iPhone.

Faux styling of Bold, Italic and Small Caps in early DTP was a typographic train wreck.

Multiple Masters tanked as a consumer product, but is very useful in professional font production.

“Smart quotes” have been a disaster for the abbreviating apostrophe.

Alex Kaczun

Setting up the progression of spacing in any font family should always be done manually and visually. Using control characters in a string example like HHOHOOnnonoo00 is critical in making a good judgement. These square and round side bearing should be applied consistently throughout all remaining glyphs. If done properly, kerning is always the last step. And, always should be done manually. If you automate any of these steps you still have to review manually. So why not take the time to setup properly in the first place.
I feel that many type designers are spacing their fonts much too tightly. Especially in the last decade. Looks good at display sizes but suffers at small settings. Spacing should be effective for the average usage (12 point). You can always track minus or plus for extreme conditions.
When you find that you are kerning too many pairs you probably did not do a proper setup with the control glyphs throughout. I have, on occasion, experienced this problem, and it's a 'bitch' to go back and adjust all the spacing again.
Slow and steady wins the race.