Excellent topic, William. A typeface that dazzles is Vendome,
which is why I use it so successfully for display, but never text.
But I'm afraid I can't explain exactly why it dazzles, other
than the sharp serifs.

I think GN is totally right about the centrality of balancing the Black and White. I call it notan (a Japanese term). Ironically, notan is the reason his theories of type design don't make sense!

> "Real characters have built-in stabilizers that keep black shapes in the forground"

There is no foreground (at least not during immersive reading). We read notan, the balance, not the Black and White separately.

The relevance of notan to type design is complex, because we have to design "firm" atomic blacks, but half the whites (the sidebearings) exist on a more abstract level (since they only work in pairs). OpenType actually allows us to equalize the treatment of Black and White (by designing entire words which would get swapped in contextually), but I don't see many people doing that...

hhp

BTW, here's a nice little notan example from a book I have:

http://www.themicrofoundry.com/other/notanrocks.gif

hhp

Thanks, Hrant. Very interesting. I hadn't thought of symmetry as a factor.

I am hoping that this will lure some others as well into spilling secrets of 'notan'. As well as Kent Lew's accomplishment, mentioned above, James Montalbano's Clearview seems to be an example where this was thought out carefully.

He notes on his website that the typeface is designed to reduce 'halation' experienced by older drivers looking at lighted signs at night.

What is this about?

For myself, I think that fineness of lines and points can create dazzle, as well as evenly spaced parallel lines. I think there are psycho-optical effects going on that affect comfort in reading, but they do seem to be complex.

> Montalbano ... notes on his website that ....

He is in fact is a clear proponent of notan (although he probably calls it something else) based on something he's stated once or twice: that he modifies the black of a letter based on what might be called its "potential" spacing. Granted, many type designers do that, but not all, and not all who think they're good...

hhp

>reduce 'halation'

I believe the term halation refers to a sort of 'erosion' of a letter's outline due to the refelcted light from its counter. Or vice versa.

E.g., the letter b's counter looks like a circle - at night, on a reflective background - instead of whatever shape it actually is.

[edit: That doesn't seem very clear. What I'm trying to say is that the aura, or 'halo', reflected from the counterforms, eats away at the letterform. And reduces visibility/legibility.]

The term halation refers to this:

Thanks, James!

Do you think that this 'halo' effect with bright letters at night also applies to some extent with text on paper, and the problem of 'dazzle'in text?

If you are willing to give us some of your guidlines on avoiding halation and other kinds of 'dazzle', I would be interested to hear them.

pardon my ingnorance, but how does one go about
reducing the effects of halation? or is that a trade
secret ; )

Take a look at this article, "Univesal Solutions" (by Bill Dorsey, for How Magazine, Feb.97, page 80). There's a few words about James' Clearview, and a small ilustration (interestingly, James' name is not even mentioned here, but I guess it must be the same project, right?).

Its not a matter of reducing halation. Halation occurs, especially with the new high-brightness retro-reflective materials used in roadside signage. One tries to design glyph shapes so their inherent character allows them to be deciphered even with the halation.

In the example of Highway Gothic Emodified I previously posted, it is quite apparent that the counters on the lowercase e and uppercase B are too small. Also the bottom terminal of the e turns up and begins to close in on itself and form an oval. Also the descender of the g is too short and not clearly formed. The v-notch or crotch of the lowercase n is not destinct enough. All of these design characteristics become overwhelmed.

Emod was never designed for high-brightness, it was originally designed so the stroke width could support refective buttons that would reflect illumination at night.

The work on ClearviewHwy tried to minimize the distrotion of halation, by addressing these shortcomings in Emod.

ClearviewHwy's counters are gererous and the the round forms do not close in upon themselves (the worst example I can think of in regard to this is Helvetica).

All of ClearvieweHwy's "difficult" characters (a,e,g,s) have ample internal space. Again, you cannot minimize halation, you can only design the glyphs to survive halations effect. Letterspacing is a critical issue as well.

As to the notion of halation in text. I'm not sure. What I do know is when you are reading or deciphering highway signage you are doing it at close to or beyond visual threshold. Reading text on a page is done well within threshold. What I have found is that even in these different environments similar letterform constructions work.

Display designs fall apart at signage distances, they also fail at text sizes. The delicate finely wrought nature of a good display type appears too "dazzling as text. As signage it does not subtend a large enough arc across the retina. (translation: it doesn't appear large or strong enough).

Signage types (at least in my experience) work well as text. the same sturdy, clearly defined characteristics that allow them to succeed at the edge and beyond visual threshold, are the same qualities that allow them to succeed as text.

Blah, blah, blah.

James

May I add that the only way to design glyphs that survive haltion's effect is to continuously test the design, using the actual materials, in the actual environment. Drawing a glyph, putting it up on a computer screen, stepping back and squinting, is not recommended.

You have to make the sign at the proper size with the proper materials, put it up, get in a car, turn on the headlights and see for yourself. Perferably with an adequate test population along with you.


James

> Reading text on a page is done well within threshold.

I would instead say that [immersive] reading pushes thresholds just like highway signage, just one of lateral resolution loss (the parafovea) as opposed to things like halation, atmospheric interference and distance. Maybe that's why the two conditions have similarities (unlike display type).

BTW, there's been some interesting discussion of Clearview on Typo-L in the past.

hhp

Thanks so much, James; that is really interesting.

Hrant, your hypothesis on why the same considerations apply to text sizes seems plausible to me.

When you look at a page of small text in Helvetica or Bodoni, the problem is not that you can't read it. It is that it is noticeably more effort to do it, and it probably slows you down also. If we do, as you (and others?) hypothesize gather significant information outside our center of focus, then that would explain it.

The sharp points and hairlines - and more closed counters etc - would fuzz out before more beefy points and lines, and open counters, so you would lose that information.

It may even be that this is the cause of the 'dazzle' experience. When you look away slightly from part of the text with sharp or thin lines, they fuzz into vanshing, and move your eyes they 'pop' back into appearance. This would explain the 'twinkling' experience.

I do have the feeling that there are also some 'optical illusion' effects as Noordzij hypothesizes. -and the two types may interact.

With Whitman, the 'resting' effect is noticable to me at least, but it is a very different face from Clearview - and less beefy than James Montalbano's Rawlings.

I hope we can rope Kent Lew into commenting on this also.

Very informative!

>test the design ...in the actual environment.

This is very important for newspaper fonts. One's typefaces (both headline and text) may look perfect on a laser proof on white paper, but they will be dull printed with dark grey ink (aka black, but nowhere near the density of toner) on light grey newsprint paper. So a little pre-emptive dazzle is required.

So to answer the original post, if I'm designing a typeface for a specific use, I try and get what I consider to be just enough sparkle, but not too much dazzle, for that context, and, as James says, do in situ tests.

However, with more "experimental" typefaces (and perhaps Vendome may be so described), I may decide to crank up the dazzle to 11 just for the hell of it, and let typographers figure out what to do with the font. Hence Eunoia.



With a dazzling font, the typographer has the option of playing off the dazzle by printing it in subtle colors, or on rough stock, or very small, or filtered, or moving fast, or low res, in which circumstances the dazzle can be considered a pre-emptive sharpening. Or the typographer may use the font precisely for its eye-popping effect.

Here too, you're talking about display, not text.

hhp

Interesting, Nick. Here close parallel lines create dazzle - a different effect from fine or sharp serifs, though you've got fine horizontals here too.

You seem to indicate that you have a box of tricks to turn up or down the dazzle. What are your guidelines for going one way or the other?

I notice that Nick introduces another variable here, the contrast in brightness of the paper and ink. - A design might work well with low contrast, but dazzle with high, or be too low contrast for easy reading, but fine with high.

Are different factors involved here in design issues?

This seems to get more and more complicated.

Incidentally, can anyone report on what Gerard Unger said about readibility at ATypI? I saw a picture with him standing in front of a slide about brightness, so he may have touched on these issues also.

I am amazed at how readable are the PDFs (at his site) of his sans serif Vesta. He sure knows something about text faces and readibilty.

>Here too, you're talking about display, not text.

No, both.

>a box of tricks to turn up or down the dazzle

Generally, I don't think about it too much. There are so many design axes (or elements in the graphic language of a typeface), and once you get the knack of type design, they all get mashed together, into a global skill.

I would say that dazzle is a question of contrast and precision, so it can manifest in any parameter -- which is why Bodoni is so dazzling across the board, as it has strong thick/thin stem-width contrast, and also sharp serifs -- sharpness being a strong figure/ground contrast at the level of detail. Bodoni also has a strong contrast between straight and round forms.

The dazzle that is caused by an equivalence of positive/negative forms, where there's ambivalence about which is figure and which ground, is apparent in Vendome, where the sharpness of the serif shape is "mirrored" in the sharpness of the acute joints.

Another kind of dazzle caused by equivalence and precision is the "picket fence" effect of evenly spaced vertical stems.

Again, on the subject of equivalence, when a typeface has a very precise evenness of spacing between characters -- I mean the almost imperceptible negative "hairline" space between adjacent serifs -- this creates sparkle, if not dazzle. You can also see that in Extra Bold House Gothic.

>The dazzle ... where there's ambivalence about which is figure and which ground, is apparent in Vendome, where the sharpness of the serif shape is "mirrored" in the sharpness of the acute joints.

Hmm. This is what Noordzij was writing about. But you are saying the figure/ground dazzle is not just a matter of proportion, but also shape?

Yes. If you "turn off the preview mode" when examining a font character, you are left with just a thin perimeter. In Vendome, the line is extremely angular, whether it goes round the outside or inside of a corner -- creating positive and negative shapes that are similar. You don't get this in say, Times, for although it has sharp serifs and high stroke contrast, its serifs are bracketed.