The bouma of space craft

raph's picture

We have a number of threads going about spacing. Please forgive me for starting a new thread attempting to tie them together.

In Creative Suite 2, we have a rousing debate about the merits of the "optical kerning" feature of InDesign (which is actually Kernus licensed from URW).

In the massive TYPO 13 thread, largely devoted to competing theories of letter recognition and the role of the bouma, there is a subthread on spacing, kicked off by five questions including "what letterspacing optimizes readability?" There is a new thread continuing the investigation into the human visual system, but I'd really like to see the spacing discussion continued.

There's a small thread entitled Theory of Spacing, worth referencing if for no other reason than Nick Shinn's beautiful graphic illustrating his own approach to spacing.

In context of Kernus, here are some observations and opinions I'd like to offer.

I think what Adobe is doing is slowly recovering from the loss of optical scaling in the transition from metal to digital. You see that with their release of optically scaled families (such as Slimbach's brand new Garamond Premier), but I think Kernus also plays a role, for the probable case where the font is not available in optically scaled variants.

To briefly sum up what I've learned about Benton's optical scaling technology implemented about 100 years ago, there are four major things that happen to a font as the size scales down:

1. The stroke weight increases, essentially implemented by stroking the outline with a fixed-width circular pen. Doing it this way also has the important effect of reducing stroke contrast. For example, let's say that the font at large sizes has a 2:1 stroke contrast, and at 6pt the thin and thick strokes are 5 and 10 mils, respectively. Let's also say that the radius of the circular pen is 2 mils. Then the resulting strokes are 9 and 14 mils, a contrast of approximately 1.5:1.

2. The width increases, based on linear nonuniform scaling (the same affine transform technology responsible for Squooshing, implemented using pantographs).

3. Extenders (particuarly descenders) shorten, or to put it another way, the x-height increases.

4 (and most important to this thread). The fit becomes looser.

I haven't collected serious data yet (I need to code up some tools to extract spacing from scans), but from eyeballing the ATF books, I'm sure that that the looser fit is not just a matter of adding the same hairline space between each letter pair (or, equivalently, adding half that space to the left and right side of each letter). Rather, I believe that both large and small sizes were fit individually.

At text sizes, I consider the ATF fonts a landmark example of good fit; in other words, I believe that the fit is better than the majority of digital faces available today. Most strikingly, the fit is good even for difficult words such as "power"; today you would expect even color in this word without additional kerning of the "ow" pair. See p. 536 of the ATF 1923 book (Century Catalogue) for example.

At larger sizes, the fit is not nearly so good (see, for example, the very uneven "power" in the 18-point sample on p. 537, but I believe it is still close to the best that can be done given the limitations of metal.

I believe that the perception of the space between letters is dependent on scale. At text sizes, I believe the eye sees mostly the space between the strokes (as is emphasized by Tracy's "Letters of Credit" approach and this tutorial). Serifs have a relatively minor contribution.

At larger sizes, I think the perception of spacing incorporates more delicate features of the lettershapes, particularly serifs. In other words, the perception of space between letters is more closely approximated by the actual length of the gap between the two letters. To set "noonn" tightly requires shaving a lot more space from the "no" and "on" pairs than from "oo". I'll illustrate:

example of tracking, 100 dpi


application/pdfexample of tracking, PDF
tracking.pdf (12.7 k)



To my eyes, at 12pt, Adobe Caslon Std looks just fine with the metrics, or with no kerning at all (the metrics call for -5 on the "on" pair). InDesign's optical kerning feature shaves 13 from "no", 14 from "on", and adds 6 to "nn". To my eyes, this is not an improvement; it makes the "on" and "no" appear crowded.

At 72pt, though, the story is altogether different. Not only is the fit too loose, but the "no" and "on" pairs appear gappy. Applying -30 tracking helps with the overall looseness, but makes the relative "no" and "on" gaps that much worse. The "optical kerning" feature here does almost exactly the right thing.

Combining tracking and optical kerning does not always yield good results. Adding +17 tracking (to keep the oo pair at the same spacing as the original) works well, but -20 tracking (for an overall tight effect) does not. The "oo" pair is too close for comfort relative to the rest of the word. I included a hand-spaced tight setting for comparison.

From all this, I draw these conclusions:

1. Optimal fit is highly scale dependent -- the smaller, the looser.

2. The perception of space is also scale dependent -- the larger, the more delicate features such as serifs affect the perception of space.

3. Adding constant tracking to a good fit at one scale yields a bad fit at a different scale. This fact strongly supports conclusion 2.

Unless one is designing a Multiple Master font or an optically scaled series, current digital font technology forces a "one size fits all" approach. Kernus is an improvement in that it will space a single font fairly well at different scales, but is still not as good as hand-fitting with the size taken into account.

I must say that this has been quite a learning experience for me. It was bad enough when I'd start tuning out the world around me because I was distracted by letterforms, but now I find myself constantly looking at the spaces between letters. A casual observer could easily conclude that I'm out of my mind. I look forward to continuing this discussion with the rest of you out there who are obsessed at looking at what isn't there.

enne_son's picture

You beat me to the punch Raph.
Besides pressing for a continuation of our investigation of the human visual system, I wanted to try and construct a bridge between the matters raised in my Typo_13 text and those under discussion in the http://www.typophile.com/cgibin/show.pl?30/68218 'visual communication' forum (as well as the 'creative suite' one you mention) and then relate that to your FontFocus whitepaper.

So I guess I'll put that here when it's ready.

The relevant statement in my Typo_13 text is: If spacing between letterforms is not in synchronicity with the space inside enclosed forms

billtroop's picture

Raph, although I agree with your approach to discerning correct optical values, that is, studying how ATF did it, I have several points of divergence.

1. You are likely to see better examples of optical compensation in some Monotype and Linotype families because ATF fonts were often optically sized from a single master drawing using expert manipulations of the pantograph. Mac McGrew has been extremely informative on this point, though I don't think he has published everything he knows. I don't really think anything can beat the achievements of the Mono and Lino drawing offices _when they were at their best_. I would rashly and roughly place peak points as the late 20s for Mono and the mid 50s for Lino.

2. There has always been a distressing tendency amongst good designers and technicians to space o shapes too widely. Adobe Caslon is but one of a countless thousands of examples. By contrast, Benton, at the end and apotheosis of his career was spacing o's very tightly, as for example in his last completed typeface, Benton. This is the approach I prefer, but it is not a majority position.

3. Adobe Caslon is not a good example for setting at 72 points because it is explicitly designed not to be used about 14 points. (It would be interesting to see what Indy does to a variety of words with Big Caslon at 72 points.)

4. Confining the test only to no/on combinations does not begin to show the problems 'optical kerning' can get you into. Comparison should also be made at a variety of type sizes. It is not necessary to go to the trouble of illustrating every combination, since the curious can easily replicate the settings on their own computers.

5. Adobe has never in the past been able to make a proper optically scaled font because of the corporate mandate to ensure that 72 point type is usable at 6 point, and vice versa. This idiot proofing of the product is understandable but should be dropped. Has it been dropped in the new Garamond? I doubt it. I've had it sitting on the shelf for months and I suppose I will now actually have to look at it.

6. An ideal optical axis is not achievable with simple interpolation as we know it. Perhaps with EvB's superpolation, but that's another story. That's a technology I have been meaning to explore for some time. In any case, the new Garamond is of course a multiple master, even though just a few instances are being sold in the OT version. But notably, it was designed as two entirely separate MMs, one for text sizes, and one for display. That said, I do not see how even the most expert manipulation of MM technology can produce an ideal scale in the 6-12 point range. It is for this reason that Stone does not use it in his Cycles series, which is the closest we come in digital to a series of types drawn with the same attention to size optimization as was routine before 1960.

raph's picture

Thanks for your response, Bill.

1. More master drawings does not always equal better. In fact, one of the things I find appealing about the ATF Bodoni series is its butter-smooth consistency across the entire size range. Bodoni is well suited to this consistency, while of course for a Garamond you have to make the counters smaller at bigger sizes and so on.

What specific Mono and Lino fonts do you consider do be more technically skillful than the best of the ATF fonts? My Lino book is from the early '30s, and has some pretty good examples but nothing that really blows me away. I don't have enough Mono material to judge -- I've studied Centaur very carefully, but I certainly wouldn't consider it a better example of technical excellence than, say, Cloister. Among the refinements of the latter are variant short- and long-tail R's, fairly important in a Jensonian. Among the flaws is a very thin right side bearing for the 'e' in the 16pt size, and a fat one for the 'r'. (needless to say, the Centaur design itself is superior in many ways to Cloister, but that's as true of the Wiebking cuts as the Monotype)

2. I don't see the same trend in Benton's 'o' spacing that you do. The 'o' space in his 1900 Century Expanded is much tighter than Benton/Whitehall (I'm looking at 10pt samples in the 1934 ATF book). What I do see is consistently generous 'o' spacing at text sizes, and consistently tight 'o' spacing in display.

3/4. I chose AC solely because of the prominent serifs, and the noonn spacing solely because it draws attention to the philosophical underpinnings of spacing: are you spacing the stems and mostly ignoring the serifs (as recommended by Tracy), or are you mostly paying attention to the closest approach between the two letters (e.g. the serifs for the nn pair)? I am not at this point doing an exhaustive review of InDesign's kernus. Isn't that the job of, uh, writers in the trade press and industry consultants?

6. You're conflating a whole bunch of different things here: design for optical scaling, linear interpolation, Multiple Master technology itself, and Adobe's use of it. MM is actually capable of many tricks other than straight linear scaling, but suffers from the same problem as Metafont: while powerful, you have to be a computer scientist to effectively use all the features.

In any case, even though I am one, I don't plan on using MM as the basis for my optically scaled releases. Because I am stark raving mad, I am designing my own font technology from scratch. It bugs the hell out of me that I don't really grok spacing yet.

And now, a fun challenge:


application/pdfTake the challenge!
ccat.pdf (28.4 k)



Both columns are my Century Catalogue in 10 pts, and in one I've enabled InDesign's optical kerning (Kernus). Please print this out on a high quality printer, and post your answers to these questions: Which column is better? How much better? Why? Which one do you think is which is which? Why do you think so? I invite all typophiles to comment; among other things I want to find out whether most people see the same thing, or wether perception of spacing is hugely subjective. I'll post the answers and my thoughts in a bit.

Nick Shinn's picture

The one on the right is Kernus.
I prefer the other, because it is taut in the appropriate places, with the authentic spacing of a metal-era serifed type. The spacing is part of the type design.

(Nice face.)

hrant's picture

Note: Looking at that PDF is the only thing in this (entirely worthy) thread I've really paid attention to so far. The bills, you see.

--

The one on the right is clearly better, more even. If you actually do want fauve spacing for some reason, that's a different story. And I think Nick is hedging.

Before I can make a good guess which might be InDesign, I need to know how good a spacer you are! :-) But a "raw" guess would be the right one, since it's more regular, and that's what algorithms are good at.

hhp

Thomas Phinney's picture

InDesign's optical kerning is definitely the one on the right. No question. One can tell because of the cases where letter features from adjacent letters almost touch on the left sample but have been spaced out on the right, like "Wh" and "rn."

Overall, I think I like your spacing better. But there are a few cases where I prefer InD's work. I'm curious: which treatment do people prefer for the e-apostophe-s combo in "nature's" on line 5? I think I like the way InD has tucked the "s" in nice and tight, but I can imagine this won't be to everybody's taste.

Just a couple of comments on InDesign's optical kerning:

1) Personally, I tend to use well-spaced fonts with decent kerning, and I only turn on optical kerning on the fly for specific problem combinations that were not kerned at all in the font. For example, the "nT" combo in OpenType frequently gives me trouble. Dang interCap words!

2) Remember, it's just another feature. It's not on by default. You don't like it, you don't have to use it! Personally, I find there are some fonts whose overall spacing seems improved by turning it on, but they're generally not fonts I would use.

Finally, I would just note that Adobe bought, rather than licensed, the relevant patent and source code (when the original URW went bust, I think). I know that some further modifications were made, so calling it "Kernus" is not really quite accurate - though it's certainly descended from the Kernus technology.

Cheers,

T

Nick Shinn's picture

>If you actually do want fauve spacing for some reason, that's a different story.

"Wild beast spacing"? I would say that prefering the original spacing paradigm for metal era fonts is more like a preference for "original instruments" in listening to Baroque music. In terms of glyph finish, Storm's work is along the same lines with his funky versions of the classics.

"hedging"

No way, dude.

>e-apostophe-s combo

IMO, the tighter fit of Optical looks better here, but the reason is that Raph's space character is quite narrow. In general, the appearance of "apostrophization" should be related to word spacing. When I first started kerning fonts, I tended to overdo it, and the apostrophe is one instance where that's easy to do. However, more recently, I've come to appreciate that there are certain glyphs where ample sidebearings are part of the character -- the apostrophe and the figure 1 being the most obvious. In fact, getting really precise, it may be a good idea to have separate glyphs (although the BCPs could be identical), differently fitted, for the apostrophe and single quote. And of course, with OpenType, that's possible! Christopher Slye or Adam Twardoch (Zapfino Extra) have probably already done it.

Are the apostrophe and the quotemark separate Unicode characters?

>Just a couple of comments on InDesign's optical kerning

I agree with you completely Thomas. But I woud add that I am critical of "Optical" spacing for two reasons:

(1) The myth of poorly spaced indie fonts. It's tempting for many people to think that there is a technological fix for the amateurishness of fonts from "boutique" foundries. That's the myth. The reality is that the virtues of "Optical" technology are quite modest, and that spacing is an aesthetic and cultural quality of types which shouldn't be given the "one size fits all" treatment. Many Indie foundries do brilliant spacing. Some may do "dodgy" spacing, but it is nonetheless a quality of their work, which should not be quantified and eradicated by a megacorp bot.

(2) By engaging in critical discussion of Optical, we gain a better understanding of the threshold between what is appropriate for automation, and what still requires manual work. While I am aware of the benefits (and inevitability) of progress, I favor a balanced approach somewhere between "the old way was best" and "let's automate everything".

dan_reynolds's picture

Some may do "dodgy" spacing, but it is nonetheless a quality of their work, which should not be quantified and eradicated by a megacorp bot.

Wait a minute

Nick Shinn's picture

>why can't you call a spade a spade

Dan, you're a Philistine ;-)

Which particular standards body would you refer to in determining whether or not a font is bad? You have seen from this thread that there is disagreement between typohiles as to what constitutes good spacing. Why do you propose that one person's, or one corporate committee's idea of what good spacing algorithms are, should be implemented as widely available software? That kind of de facto standardization would be pretty dire, like a cheesy Photoshop/Fireworks filter pre-set.

Ultimately, there is no such thing as a bad typeface, only inappropriate usage.

Besides, I've yet to see conclusive proof of "Optical" fixing up the spacing of a "bad" font. Hrant thought it would be a piece of cake with Mrs Eaves, but it didn't turn out like that.

dan_reynolds's picture

Dan, you're a Philistine ;-)

That's probably the nicest thing I have been called all year!


Ultimately, there is no such thing as a bad typeface, only inappropriate usage.

This is just a platitude. While there *may* not be such a thing as a bad typeface, there are certainly bad fonts. If I make a font and sell it to you, and it doesn't even function properly, then it is a bad font. No brilliant use on your part will be able to save it,* and chances are that brilliant use with such a poor quality tool would be highly unlikely.

So, I'm saying that there IS such a thing as a bad font: this is not a design issue, but rather a production one.


Besides, I've yet to see conclusive proof of "Optical" fixing up the spacing of a "bad" font. Hrant thought it would be a piece of cake with Mrs Eaves, but it didn't turn out like that.

I don't have the depth to objectively evaluate InDesign's optical spacing feature, although I do use it occasionally, and have been pleased with its results. However, I didn't say that the InDesign feature was good, I said that if a manufacturer were to pull off something which really would improve poorly spaced fonts, that would be praiseworthy. I'm sure that many graphic designers and software users would agree with that.


*Even if you, with your excellent capabilities, could make good design with a bad font, that doesn't mean that the average user would be able to discover enough work-arounds.

hrant's picture

If Raph wants to produce anachronistic work, then of course you wouldn't consider Optical spacing. But I give him more credit than that. Many people who do decide to revive adopt the position of trying to take the intent of the original (not its every flawed detail) and make something usable today. Nick, you of all people should appreciate that.

Bad fonts? We're surrounded by them!

hhp

Nick Shinn's picture

>So, I'm saying that there IS such a thing as a bad font: this is not a design issue, but rather a production one.

I agree completely. As you noted, I said there is no such thing as a bad TYPEFACE, only inappropriate usage. That's not a platitude, it's a philosophy.
"There are no wrong notes" -- Monk

>the intent of the original (not its every flawed detail)

It's usually the "flawed" details that produce the most useful developments.

enne_son's picture

About Raph's challenge:

Mostly what I see, I see at a glance. There is a perceptible distance on what I called in Typo 13 the 'gestural-atmospheric' axis. The text on the left is more even and regular and quietier, the one on the right more assertive and active and generally a tad darker, but having said that, I have to say the distances between them are quite slight. I cannot say the one is right, the other wrong, or the one is good the other bad.
Both are readable on my screen. Readability and legibility are, I am coming to believe, gross measures. Legibility of type is a matter of thresholds (perceptual discrimination affordances at a distance, or small size or in inferior lighting); Readability of a type-in-use a matter of degrees (visual wordform resolutional affordances, which seem to be quite generous). Spacing and rendering are matters of tolerences. Beyond a certain point (and the allowances might be quite sensitive) space manipulation alters the gestural-atmospheric consistency or character of the typeface.
Does optical spacing cause that kind of shift? and if so, do we prefer the gestural-atmospheric 'attack' of the altered protocol?
For the purpose of continuous reading both spacing implementations seem tolerable. On closer inspection the column on the left shows a level of refinement the one on the right doesn't have. I see this in the 'Wh' combination of the opening word, and the decision about where to break the line in line 9&10 of the second paragraph. However, the spacing of powers' at the end of line 3 seems worse. So how to fix (with optical kerning) the "Wh" of the second without touching 'power' if we want it's gestural-atmospheric mood?
Which 'attack' would I prefer? Depends on the character of the text I am setting, and the medium (newspaper / book / magazine / etc).

hrant's picture

> That's not a platitude, it's a philosophy.

Yeah, a bad one.

> It's usually the "flawed" details that
> produce the most useful developments.

No, it's "fringe" concepts.

When a metal font has an overly-loose "v" because its longish serifs might break in usage because the alloy is cheap, reproducing that in digital is not of this age. It's cloying.

--

Peter's observation that the left one (presumably the original spacing) is more "spiky" in its quality than the right one (the presumed Optical), meaning that it has some especially bright and especially dark spots, is in line with what one would expect from heuristics versus algorithmics.

hhp

Nick Shinn's picture

>Yeah, a bad one.

No. just different.

pstanley's picture

I don't know enough to say which is which. My own subjective view (taking cover):

The left hand column is better.

Only slightly better.

This is based almost entirely on one letter combination "ff" which seems to me to be too widely spaced in the right hand column. An argument, I suppose, for ligatures. "ff" is a common combination and it bothers me perceptibly that it looks wrong.

I think the right hand column spaces "it" too close (compare "it" and "in" on line 1); also perhaps "ie" and "le". To my eyes at least the left is more even on these particular combinations.

On the other hand, the right hand column seems to me to get the "he" space in "the" better: in this case the left hand column seems to put excessive space between those two letters, so that "the" tends (fractionally) towards "th e". (Not helped, to my my eyes, by the "t" falling forwards.)

But (except for the ff, IMHO), neither could be said to be bad.

hrant's picture

Yeah, and Bush is "just different" than Mother Theresa.

hhp

raph's picture

First, a couple of errata. In the "noonn" graphic, the values for the "optical +17" line should be (-14, 0, -16, 6). This actually changes my evaluation of InDesigns optical kerning somewhat; it is sensitive to scale, but in this example only microscopically. These values vary by only 2 units (out of 1000) from the 12pt optical setting.

Second, in comparing ATF Cloister with Monotype metal Centaur, I want to make clear that the fit flaws I described are in the latter; in my post 236, that was very unclear.

You guys are good! The column on the right is indeed InDesign's optical kerning (and thanks, Thomas, for pointing out that it isn't quite correct to simply call it "Kernus"). On the left is my own attempt to space the font. In fact, it's my third from-scratch stab. This time, I tried really hard to get the side bearings right, then kern with a very light touch. In particular, I made it a point not to kern round-round or serif-round combinations.

I find it fascinating that there isn't unity on which is better (or maybe Hrant is just the outlier whose judgement should not be trusted). Even Paul Stanley's preference for my spacing is suspect, because the lack of the 'ff' ligature is a bug in the font that will get corrected before release.

My spacing could no doubt be improved. In fact, I now consider a comparison against InDesign's optical kerning an important tool in discovering places where spacing can be improved.

The word "powers" is definitely spaced more like a metal face (optical kerning hairline spaces po and kerns owe quite a bit). Because this is a historical revival, I'm not sure that's a bad thing, but InD's optical is also absolutely a valid stylistic choice. I cannot agree, though, for cases such as "liberty", which to my eyes has "ib" too tight and "rty" far too loose.

It's interesting that Peter perceives InDesign's optical kerning as "darker". I fiddled with the tracking values (+3 for my sample, -2 for InDesign) so that they'd take exactly the same amount of total width. To my eyes, there isn't a significant difference in color, so I wonder if this is something that other people see as well.

If you look at just the stems, my sample is far more even (look at "institute", line 15, for a striking example). Here is a fundamental question: is InD's variation in stem spacing signal, or noise? In the frequency analogy, my stems form a stable carrier frequency, with features such as arches added on, while the more varied stem spacing of InD is like frequency modulation of that carrier wave; it will occupy a broader peak of spatial frequencies. (hey, I should run a Fourier analysis!)

You can also look at the spacing, in terms of the "features" concept of Peter's Typo 13 essay. In particular, I space the rvwy class tightly, much more so than InDesign. So, in my spacing of "ty", these two letters visually connect (they're not quite touching, but presumably the gap is so narrow that it wouldn't be seen in the parafovea), while in InDesign the gap between them would count as a visual feature. From a readability standpoint, which feature is better, connection or gap? Are they both functionally useful for recognizing the "ty" pair, perhaps, so the choice is merely aesthetic?

For pairs such as "vi", I strongly prefer my spacing, because it avoids creating a large triangular gap. This is especially so for a word such as "evinces" (4th line from bottom), where the large gap contrasts with the relatively tight ev and nc spaces. (this effect is even worse for the yp pair, as can be seen in one of the nonsense words)

Even so, I'd have to agree with Paul that InDesign's optical kerning isn't bad. In many cases, such as a poorly spaced text font, or if you have Hrant's eyes, it would definitely be an improvement. So I think we can put to rest any assertion that optical kerning should not be used for body text.

It amazes me that spacing is so hard. After all, it's just a handful of numbers. Yet, we often find that people who draw beautiful letter shapes stumble over the problems of spacing (Mrs. Eaves is a commonly cited, and dramatic, example). I wonder to what extent it's possible to understand spacing systematically, or whether it will always remain a black art, with all but the most talented and experienced masters doomed to forever crank out poorly fitted fonts.

billtroop's picture

We're not going to get anywhere with Indy/Kernus until we stop treating as a mysterious black box which somehow may, or may not, do the right thing to spacing in some particular case, but instead, approach it as a bad program which generally does bad things in an entirely predictable way, no matter what its input.

This is pretty easy to illustrate. Take the behaviour I illustrated in Adobe Garamond Pro. My objection is that it destroys the font by inserting unwanted space between u and m in the word jumps at 12 points. This is wrong because in any normal font, the distance between u and m has been perfectly spaced. This is not conceivably a letter pair that under any normal circumstances needs spacing repair or additional kerning. To make matters worse, Indy/Kernus also subtracts space between j and u (again, a letter combination invariably perfectly spaced under all normal circumstances) thus entirely distorting the spacing of this most common word.

OK. So what do we do? Stop, for Heaven's sake, acting as if this were a one-off event. Instead, test this same letter combination with a wide variety of fonts.

What I found at once is that for every serif font I tested at 12 points, Kernus/InDesign invariably subtracted space between j and u and added space between u and m. I tested Times, Georgia, Mrs Eaves, Mrs Eaves small and petite caps (slightly different), Palatino, Minion, Courier, Didot, and most illuminatingly, ITC Bodoni Ornaments.

Now, instead of wasting our time looking at nono combinations in a particular font, let's look at a simple row of mmmmm and nnnn

These are again rows that could not conceivably need alteration of any kind.

Yet InDesign/Kernus invariably tampers with them. It generally adds positive space. For example, for both nn and mm, it adds 14 units in Minion at 12 points. (Is there anyone here who can convert this into PS font units? In any case, by any measure, this is a shocking alteration to the most fundamental spacing property of a font.)

For Adobe Caslon by contrast it adds 4 to mm and 6 to nn. For Copperplate Gothic it subtracts 32 units.

I'm sorry folks, but this is a Frankenstein come to life.

This is not behaviour that can possibly be justified by any rational theory of typography or graphic design.

>Finally, I would just note that Adobe bought, rather than licensed, the relevant patent and source code (when the original URW went bust, I think). I know that some further modifications were made, so calling it "Kernus" is not really quite accurate - though it's certainly descended from the Kernus technology.

This gives the erroneous impression, Thomas, that Adobe initiated the relationship with URW. It did not. It inherited the relationship from Aldus, which licensed jack-in-the-box from URW for PageMaker 5. Not for the first time, I express the wish that you would get your Adobe corporate history right! Why should it be up to me to remember these ridiculous arcana?

Is anyone seriously going to tell me that a program which invariably tampers with the nnnn combination of ANY font can be defended on any grounds whatsoever?

raph's picture

I don't think it makes much sense just to look at the numbers. You run the risk of making the same kind of mistakes that newbies often make by paying too much attention to the "side bearing" numbers in their font editor; you see way too little space to the right of the j all the time.

Instead, I prefer to follow Peter's approach and concentrate on the visible features, in this case the gaps where the serifs visually connect letters together, and the blocks of space separating the vertical stems.

visual features of "jump"

I've spaced this font according to Tracy's philosophy of making the space between stems even. For this word, the blocks of space are all rectangular, so are easy to quantify. With vee-shapes and rounds, you'd also get triangles and concave lens shapes. With perfectly even stem spacing, the blue blocks would all be square.

Given the pattern of serifs at the top and bottom of the lowercase line, even stem spacing leads to very uneven gap sizes (visualized here as tan circles). In this example, medium, tiny, huge. Indeed, at display sizes the mp pair looks gappy.

From what I can see, InDesign's optical kerning simply applies a different balance to the two visual features, making size of the gaps more consistent, at the expense of making the stem spacing more uneven.

I propose that, at text sizes, stem spacing is more important, and that at display sizes, the gaps at closest approach are also important. Using existing tools conservatively, this suggests designing the metrics of fonts to optimize text sizes, while turning on InDesign's optical kerning for display. Of course, expert spacing by hand is always best, but in the cases where that's not practical, I think what InDesign does is useful.

Of course, what I really want to do here is space the different optical sizes differently, according to the principle of varying relative importance of blocks of space and gaps at closest approach. But that's also clearly a lot more work; otherwise I think most font releases today would be optically scaled.

hrant's picture

The bottom "jump" is spaced much better than the top one. Since Proof can only be positive*, that's your proof right there that Optical spacing can be useful. Not that I think this kiddie-benchmark of proof that some people espouse deserves any attention.

* You can never proove something doesn't exist.

Bill, stop worrying about archaic, non-linguistic control strings,
and optimizing fonts for one point size - that's so 20th century.

hhp

John Hudson's picture

Hrant, I disagree that the lower 'jump' is better spaced. Go and stand on the far side of the room and look at them. In the second, the u is drifting to the left in the space between j and m.

hrant's picture

Near or far, I'm seeing the bottom one quite close to what I would do by hand, and the top one deviating from that more.

hhp

William Berkson's picture

I agree with Raph and John on this one. I guess the question is whether you think Tracy had the right approach.

John Hudson's picture

Remind me not to ask you to space any fonts for me, Hrant :-)

hrant's picture

> I agree with Raph and John

?
You mean about the "jump"?
Where are Raph and John agreeing, for you to agree with both?

--

> Remind me not to ask you to space any fonts for me, Hrant

You certainly wouldn't want my White to upstage your Black. :-)

hhp

Nick Shinn's picture

>It amazes me that spacing is so hard.

It's as hard as you want to make it.

The subject is rhythm, not space, and it can't be reduced to words, numbers, or algorithms, it has to be felt.

To train yourself to feel the rhythm as completely as possible, do a revival (ie facsimile) from scratch with a magnifying glass, rather than tracing scans.

William Berkson's picture

>Where are Raph and John agreeing, for you to agree with both?

They agree that the top example is better spaced for text. I concur.

enne_son's picture

John, Raph, William: in the best of all possible worlds, shouldn't the top 'jump' have even a slight bit more space between the 'j' and the 'u'? (Taking a screen shot of the samples and zooming out in photoshop to text size suggests it might.) Hrant: on what do you base your preference for the o.k. version?

hrant's picture

There is no rhythm in text type. Because there is no flow.
Rhythm is a romantic product of our consciousness, inventing things about our subconscious.

William, Raph wrote "at text sizes, stem spacing is more important", but if you read the previous stuff (not to mention look at the top "jump", which is not spaced purely by stems) you might realize that he [probably] meant relatively more - and I might agree with that. John on the other hand made an outright endorsement for the top one.

> on what do you base your preference for the o.k. version?

"I put my middle finger on the monitor port and feeeel the viiibes, maaan."

No. I use my gray matter to analyze the information from my sense of vision.
Think With Your Brain.

Nick Shinn's picture

Type design is a craft.
All fonts are hand-made.
One learns by doing, not by theorizing.
Designers think with their hands, drawing on the left side of the brain.
If you think only with your right brain, your designs will lack grace.

Isn't this why you are studying calligraphy, Hrant?

hankzane's picture

Spacing is hard

William Berkson's picture

>shouldn't the top 'jump' have even a slight bit more space between the 'j' and the 'u'?

Yes, I think you're probably right, but I'm not experienced enough to predict how it would look to me without actually trying and looking.

hrant's picture

> One learns by doing, not by theorizing.

One learns to refine existing norms by doing.
One uncovers future potential (ie marks progress) by theorizing.

I study calligraphy to understand. Somewhat like insurance investigators study car crashes...
And one big reason I want to understand is for people to stop using it as an excuse against me, as a diversion for their own failings in the realm of progress/theory. By assimilating the practices that other people (in particular people who resist good theory) know, I can exert more effective pressure.

--

Speaking of theory:
Raph, your theory about the relative balance of stems versus serifs in spacing (which is quite good - I'll be using it) fits very well with "known" theories about optical scale in general: that smaller features have more weight in proportion to size. Just like how serifs need to be stronger for smaller sizes. It makes sense that this applies to the White as well.

Furthermore: The validity of this theory does not preclude InDesign from doing an "adequate" job for both text and display, but -depending on the algorithm- it could indeed mean it's better at one more than the other. If InDesign does not use "feature weight" in its calculations (and I would actually be surprised if it did - it seems quite difficult*), it makes great sense that it would indeed be better at display.

* Too expensive to solve for this industry.

Holding More Water All The Time.

Nick Shinn's picture

Pardon me Hrant, if I implied that your designs lack grace. Like everyone, you use the left side of the brain when you draw, write and design, thereby thinking with the hand. I studied calligraphy as a practical discipline because it was interesting, enjoyable, and because I thought it might give me a better "feel" for letter forms -- as an adjunct to more academic study. I'm assuming you have some of the same motivation, despite the premium you place on intellectual theory

William Berkson's picture

>One learns by doing, not by theorizing.

Making theory and practice a dichotomy is a fundamental mistake. Theory and practice should interact and correct each other. If theory were useless, science would be useless instead of conquering the world, as it has.

"There is nothing so practical as a good theory." - Kant

That being said, in my view the state of theory in reading and typography is not very strong, so the main guide currently for practitioners has to be the eye and experience.

Nick Shinn's picture

Um, make that drawing on the Right side of the brain.

enne_son's picture

Hrant, I asked: on what do you base your preference for the o.k. version?
You replied: I put my middle finger on the monitor port and feeeel the viiibes, maaan.
I ask myself: am I to take this as a hostile comment or a trivializing move? Should I see it as directed at myself, or Nick?

Then you say: I use my gray matter to analyze the information from my sense of vision.
Should I read you as being sarcastic here? Or bitter?
I do not see how this helps me understand your preference for the second. Do you mean you think about what visual inspection yields? And is the thrust of this that others don't? Or am I to understand that in theory you prefer the o.k. concept because it address more adequately what the visual cortex's needs? And don't you need to show then that there is a mismatch between what the visual cortex needs and what our obsessions, if that is what they are, with colour or rhythm on a craft level lead us to strive for?
You have not shown there is a mismatch. If I am to trust your perceptions about a mismatch, I will need you to work things out at ths level. You say immersive reading is counterintuitive, but yet you use your intuition to elaborate your understanding of it. And if your intuition is superior, or better grounded, you have an ethical imperative that governs how you behave on line.
I was asking for you to pinpoint specific things about the spacing of the o.k. version that could help me see what you see in the second 'jump'.


You say there is no rhythm, and I say this is like saying there is no God. I am more interested in understanding belief, or what 'range of convenience' a word like rhythm can have 'vis a vis' type than I am in outlawing it. Perhaps you do not find it useful. Others do. And it helps them do good work. So I want to see how the concept of rhythm or colour works itself out in the details of construction, contrast manipulation and space craft. And not just for writing with prefabricated letters, but all writing.

hrant's picture

Nick, I actually do think my forms are not very graceful, but they do have "feeling" - I'm human too after all. But type design can't be all feeling, and some aspects of it -in particular spacing- need to be quite analytical.

Peter, that thing was directed at Nick's "it has to be felt". And not really hostile, more like snide.

To [try to] elaborate more (on the phrase right after that), I would say that I look for good notan (balance/tension among the White/Black), which however depends on intended point size (and I've been assuming that's what the "18" is in Raph's illustration). I squint, move back & forth (like those charmed snakes), and try to see holes & clots. BTW, are you guys squinting (adequately) when you gauge spacing?

> And is the thrust of this that others don't?

I don't know what others are doing. We're each born with slightly different eyes, and of course quite different objectives. The trick is to guess what users want/need.

> You say immersive reading is counterintuitive

To the layman. By looking and thinking, a person can still unlearn one's intuition (because it's not infallible). For example how point sizes that seem most easy to read are in fact a bit too large. But I certaily can't "proove" the validity of my opinions - which however shouldn't stop me from thinking and acting. And each of us has to constantly and dynamically draw a line concerning who to believe about what; based on the aggregate of what somebody has learned about us, a person could choose to believe John, or me; it's the decisions we each [have to] take.

Specfics: for 18 point, I think the top one needs adjustment for each of the three pairs("ju" a hair tighter, "um" a hair looser, and "mp" a bit tighter), while the bottom one just needs the "um" a hair looser. (BTW, I can't quantify "hair" and "a bit", but I can tell you that the latter is more.)

> You say there is no rhythm, and I say
> this is like saying there is no God.

I don't think so. I think it's more of a terminological confusion caused by the romance of the consciousness (which forgets we're animals). The reason there's no rhythm is quite simple, mechanical even: to me (and I think dictionaries agree) rhythm is a pattern in a flow. But there is no flow in reading (except at the level of linebreaks, but that's far removed from letterform notan), so there is no rhythm. There is Pattern - let's please use that instead, no matter how unromantic it is.

> prefabricated letters

I think this is also a misleading (or at least archaic) term.
Please read my "reply" in Typography Papers 4.

--

BTW, I'm supposed to finish getting ready for my Beirut talk, so if I become silent it's not a slight or a mark of detachment.

hhp

raph's picture

On brain hemispheres: here's an accessible yet scientifically grounded article. I'm personally a very unbalanced person; my ability to improvise is extremely weak compared with my logical and analytical reasoning. As such, I have a strong tendency to seek a deep theoretical understanding in areas where many would insist that "if you gotta ask, you'll never know" (Loius Armstrong). Will this imbalance prevent me from making beautiful fonts? Either way, it's my path to follow.

In any case, I stand by my assertion that (good) spacing is hard. If it were easy, then everybody on this thread would be agreeing, and even amateur shareware fonts would be well spaced. The empirical evidence suggests otherwise.

Hrant: no, the "metrics -18" notation refers to the spacing settings; my metrics minus 18 units (of 1000) between each pair. The intended size for that example is "large", but my metrics were very much tuned for 10/12 point setting. I am not claiming that the top line is better for display sizes; the "mp" plainly looks gappy.

Hrant is of course correct in assuming that my metrics aren't based on mechanically making the spaces between the stems equal. Maybe the "j" could use a little more space on the right (if you're Peter or William), maybe a bit less (if you're Hrant). Tracy's guidelines side with the former camp; his 'j' right stroke space is the same as that of 'i'.

Good luck with your Beirut talk! Your snideness is forgiven, just this once anyway.

enne_son's picture

"can't be all feeling,... "
"in particular spacing need[s] to be quite analytical."

"good notan ( [=] balance/tension among the White/Black)
"I squint, move back & forth (like those charmed snakes),
and try to see holes & clots"

(I'd bet we all do the squinting / seeing holes and clots thing.)

So is gauging notan analytical or by feeling?
I think the squinting and moving / seeing holes and clots is the 'analytical' end of your practice, but in doing this you 'feel' tensivities / equivalencies or balance /[and yes] rhythms.

Random house: "a patterned repetition of a motif, formal element, etc., at regular or irregular intervals [pe: can be of space] in the same or a modified form."

I don't see how putting space in the 'um' of 'jump' brings the o.k. version into notan; it just destroys the visual integrity of the wordform by making the 'ju' bond and the 'mp' bond strong and weakening the 'um' bond disproportionately. Your proposals for the metric version have the same effect. Pushing this any farther leads to visual_wordfrom_resolutional response_bias collapse.

William: "the state of theory in reading and typography is not very strong."
Exactly, so let's do what's necessary to strengthen it.

peter_bain's picture

For myself:
top "jump" is better, for text sixes anyway
bottom is useless, either text or display

I'm firmly convinced that at display sizes, difference in personal taste for spacing exists and is unavoidable. This variety is consistent across type composition and hand-made letters, and comes forward in tighter spacing at larger sizes.

peter_bain's picture

For myself:
top "jump" is better, for text sixes anyway
bottom is useless, either text or display

I'm firmly convinced that at display sizes, difference in personal taste for spacing exists and is unavoidable. This variety is consistent across type composition and hand-made letters, and comes forward in tighter spacing at larger sizes.

enne_son's picture

"can't be all feeling,... "
"in particular spacing need[s] to be quite analytical."

"good notan ( [=] balance/tension among the White/Black)
"I squint, move back & forth (like those charmed snakes),
and try to see holes & clots"

(I'd bet we all do the squinting / seeing holes and clots thing.)

So is gauging notan analytical or by feeling?
I think the squinting and moving / seeing holes and clots is the 'analytical' end of your practice, but in doing this you 'feel' tensivities / equivalencies or balance /[and yes] rhythms.

Random house: "a patterned repetition of a motif, formal element, etc., at regular or irregular intervals [pe: can be of space] in the same or a modified form."

I don't see how putting space in the 'um' of 'jump' brings the o.k. version into notan; it just destroys the visual integrity of the wordform by making the 'ju' bond and the 'mp' bond strong and weakening the 'um' bond disproportionately. Your proposals for the metric version have the same effect. Pushing this any farther leads to visual_wordfrom_resolutional response_bias collapse.

William: "the state of theory in reading and typography is not very strong."
Exactly, so let's do what's necessary to strengthen it.

hrant's picture

> the "metrics -18" notation refers to the
> spacing settings; my metrics minus 18 units

Oh. I figured that was an en-dash. :-) Especially since you didn't state the intended usage size (while I know you know that's important). Tsk tsk tsk... ;-)

> The intended size for that example is "large"

Well, 18 is close to what I would call "large" myself. To me large is anything above 40 (this is based partly on the work of the Bentons, as well as the span of the fovea wrt [typical] viewing distance).

The question (wrt my previous statements) becomes: is a global tracking adjustment adequate for bringing your 10/12 spacing to 18? I think considering the particular forms in "jump", it's probably fine. Which means my statements stand.

> Tracy's guidelines

BTW, I think you're ascribing a degree of accuracy to his guidelines that they can't enjoy - and I don't think Tracy intended them to - he was a very "eye is the final arbiter" guy. Which I agree with to some extent, although I don't see it precluding a higher degree of "method" than what some people [like to] stop at. It's always easier to just eyeball - you just can't reach the pinnacle that way.

--

> So is gauging notan analytical or by feeling?

I'm not sure.
But it's clear we're seeing (in the useful sense of the term) different things... The question reverts again to what [we think] users need/want; and how much we trust the senses/analysis of others versus our own.

> Random house: "a patterned repetition of a
> motif, formal element, etc., at regular or
> irregular intervals [pe: can be of space]
> in the same or a modified form."

I think this is a non-productive definition (assuming your "can be of space" interjection), because it reverts to Pattern again; pattern and rhythm become interchangable. I think it's more productive to restrict rhythm to something that exists in a directed continuity (like time). That way you have two distinct, focused concepts/terms. The problem is our consciousness needs to take a cold shower.

hhp

raph's picture

(hey, I should run a Fourier analysis!)

I see I'm not the first one with this idea. Hrant just now posted a link to someone who's done just that, starting from a page of Aldine italics.

Anyway, here are Fourier transforms of the two 10-point settings above. First, mine, then Adobe optical kerning.

FFT of Century Catalogue, my spacing

FFT of Century Catalogue, Adobe optical kerning

The ladderlike forms sorrounding the vertical line at the center represent the line spacing. The dark splotches to the left and right are the main stroke rhythm. I speculated that the peak would be harper in my version, more spread out in optical kerning. I think that is true, but the effect is very subtle:

histograms of FFT

Interestingly, the peak of the main stem frequency is at a very slightly higher frequency in the optical kerning sample; perhaps that corresponds to Peter's perception of more darkness.

This was interesting, but the differences between the FFT plots are subtle, and not easily accessible to interpretation, so I have yet to be sold on the general usefulness of the technique for understanding spacing.

John Hudson's picture

There is no rhythm in text type. Because there is no flow.

How many times do we have to have this conversation? How many times does it need to be explained to you that visual rhythm does not imply flow? How many times does this have to be demonstrated with diagrams? How many times does it have to be suggested that you are confused by notions of musical rhythm, which is temporal and hence sequential (flow)? How many times does it have to be explained that visual rhythm is not temporal, but immediate? How many times does it have to be explained that rhythm in type design is about the relationship of shapes to each other, and not descriptive of the reading process which we know is saccadic? How much clearer can we make it, than to repeat that the relationship of any shapes in a group can be more or less rhythmic, and that this doesn't imply anything about flow or, indeed, anything about the reading process at all? How often do we need to go over the same ground before it sinks in?

enne_son's picture

Raph, what happens with perceptably terrible spacing? I mean, very irregular. Also, what happens with very very tight spacing, and very very loose. That might tell us something about tolerances, wouldn't it?

John Hudson's picture

When other strategies fail, deploy concrete poetry. I call this number 'Got Rhythm', hepcat. Dig it.
Got Rhythm

John Hudson's picture

Raph, what happens with perceptably terrible spacing? I mean, very irregular.

There ain't no such thing as irregular spacing, Peter, 'cause irregular ipso facto and all that implies regular, and regular is rhythmic. And you heard the man, there ain't no rhythm in text type: it's all a 'romantic product of our consciousness' and that ain't cool. Unless, of course, the man is wrong, and all the hepcats know what rhythm just means regular in its arrangements like. And sweet.

[The device employed in this message is known in classical rhetoric as berlowia]

Syndicate content Syndicate content