Does anybody really build faces like this anymore?

Ryan Maelhorn's picture

I've seen a lot of diagrams like this, 'explaining' how the glyph was drawn. I've never contructed a font in this manner, and it seems like it would be a pain really.

PublishingMojo's picture

No one ever built fonts this way (except gimmicky novelty fonts like this example). Even the most geometric-looking fonts, such as Avant Garde and Futura, deviate slightly from right angles and perfect circles.

Nick Shinn's picture

Here’s one I did recently, digitizing a 1970s design by Malcolm Waddell.
There are some very minor optical adjustments, but the circles are indeed circles.

Ryan Maelhorn's picture

**I misspelled its name in the original post. apparently those are uneditable now? It's true name is "Struktur"**

Novelty fonts are the future.

I have problems with this face {Struktur}, most notably the lowercase t, but I wouldn't call the face gimmicky. The ampersand is brilliant, if you ask me.

hrant's picture

That's one dorkmeister of a glyph.

hhp

Ryan Maelhorn's picture

be careful if you want to take a dip in the waters anyone, the sharks are out...

Maybe it needs to be seen with the rest of the face to see its beauty.

http://www.typographyserved.com/gallery/Struktur-Typeface-and-Posters/47...

damn, I didn't really want this thread to be about that face, that was simply the most recent diagram I've seen.

hrant's picture

That's one dorkmeister of a font.

hhp

quadibloc's picture

I felt you were really talking about things like Durer's geometric decomposition of the capitals, but in a failed search for that sort of thing, I found this:

http://opentype.info/blog/2008/07/08/traffic-sign-typefaces-poland/

LexLuengas's picture

Trade your freedom and creativity for circles and straight lines! Yep...

It's a task which seems fairly easy, because of it's obvious constraints: one ‘just has to be systematical’. Choose your radii and draw a shape over different circle contours that resembles the character you're trying to draw somehow. But I think the easier a task, the more careful and precise you must be with details.

Something I find very difficult about it is, for example, the transitions from curved to straight lines. The sample Nick shows does this pretty well, but Struktur shows a few flaws.

Nick Shinn's picture

… the transitions from curved to straight lines.

Yes, that was one place I cheated:

hrant's picture

But perfect circles look like ovals. Of uneven thickness.

hhp

Ryan Maelhorn's picture

everything depends on context

Ryan Maelhorn's picture

Hmmmm, I think I may have gone about asking this question wrong.

The point was not what that face looks like or if anybody uses perfect circles or straight lines, what I was trying to ask is if any of you start out with a bunch of circles, some of the circles divided into halves and quadrants; and then once they have drawn those how they like, then drawing the outlines on top of them.

So in the example above, the designer would start with all the shapes that are dotted lines, and then only after he had arranged those shapes how he wanted, would he draw the outlines by connecting the shapes.

I know I've seen a bunch of diagrams like this in books about type design, the idea is hundreds of years old.

hrant's picture

In the past theorists did stuff like that. These days only amateurs do.

hhp

quadibloc's picture

@LexLuengas:
Trade your freedom and creativity for circles and straight lines!

And that is different from trading your creativity for Bèzier curves how?

All right, so that isn't really fair; one needs, at least, ellipses, in addition to circles and straight lines, to have an adequately general set of primitives to easily construct a wide variety of typefaces.

Frode Bo Helland's picture

BP revided Romain de Roi, as BP Romain.
Here’s one of the original drawings.

Ryan Maelhorn's picture

Beautiful example, Frode.

I mean it makes sense that if you used a lot of elements (circles, it seems most often), a good majority of which were all the same size (such as circles A and C are in Frode's pic), the resulting type would hold pretty true to it's self. More homogeneous.

It makes sense technically. It just seems in practice it would be incredibly tedious.

hrant's picture

The telling thing with the RdR though is that when Grandjean was give the job of actually making the fonts, he deviated markedly from those drawings (with the diagrams for the Italic totally abandoned).

It just seems in practice it would be incredibly tedious.

Tedious to the point of not being worth it, especially for tricky glyphs. The lc "g" is missing from the RdR diagrams for a reason. :-)

hhp

rs_donsata's picture

Romain Du Roi was a rigid rationalization true to the spirit of it's era. It was a time when people expected children to sit stiff for hours.

hrant's picture

Don't get me wrong though - I feel of itself it was a valid piece of cultural progress. And we might have never had Baskerville (and then Bodoni, who was an admirer of Baskerville) without it.

hhp

LexLuengas's picture

And that is different from trading your creativity for Bèzier curves how?
(...) one needs, at least, ellipses, in addition to circles and straight lines, to have an adequately general set of primitives to easily construct a wide variety of typefaces.

It's true you could theoretically get along without Bèzier curves, using only arcs to approximate your curve (like one esentially does with Bézier curves). Not even ellipses are necessary.

My point: from the practical point of view it makes no sense. Everything that restricts a designer technically in his creativity, everything that makes a class of shapes more likely than other, is mistaken as soon as it is presented as an ‘aesthetic method’. Just saying, more than often this idea is interpreted as a suggestion to a specific way of drawing. Like someone who looks at Romain de Roi's drawing and says: “So this is it! It was circles and ovals all this time! And I've been struggling with this weird Bèzier thingies for so long...”.

oldnick's picture

everything depends on context

Ryan, you are such a freaking genius…

quadibloc's picture

@LexLuengas:
Everything that restricts a designer technically in his creativity, everything that makes a class of shapes more likely than other, is mistaken as soon as it is presented as an ‘aesthetic method’.

This is absolutely true.

My unhappiness with the Bèzier curve is that some designers might want to specify traditional designs directly in terms of conic sections, at least in part.

One very early electronic typesetting system actually provided only one curve, an Archimedian spiral, in addition to the straight line. Ikarus, I believe.

Nick Shinn's picture

The question was once famously asked, “How long is the coastline of Britain?”—the answer being that it depends on the amount of detail in the measurement method. (This was the primary consideration that led to the invention of fractals.)

A similar principle applies to type outlines; any glyph shape can be rendered with straight lines and look sufficiently smooth if the curves are divided into enough straight segments, and rendered at an appropriate resolution (Licko’s Journal springs to mind). The idea is that a complex line may be broken down into an appoximation by using shorter components of a less complex type. There are four or more levels of line complexity (according to the sophistication of one’s maths): straight, circular, oval, quadratic…

Certainly, if one restricts a type design to the simpler lines, there is the danger of banality.
It helps a designer working in a simple mode to be familiar with complex curves, because this facilitates a subtlety of expression (by hand and eye) which mitigates the potential for crudeness and banality; two examples—for instance, firstly, how much should be “shaved off” the transition point between straight and curved? (See my post of Oct. 16) And secondly, what is the relationship between the radius size of the curves used in /m, /n and /o?


Thirdly, how big should the overshoot alignment zone be?
And so on.
There is arguably less leeway in determining these subtle parameters of a type design in types based on simple modular forms, than there is in more complex types.

hrant's picture

This was the primary consideration that led to the invention of fractals.

Are you sure? I thought Benoit Mandelbrot invented it for kicks.

any glyph shape can be rendered with straight lines and look sufficiently smooth if the curves are divided into enough straight segments

Yes, enough to fool the eye and/or overshoot the output resolution. Some early font formats did that, and in fact in the end RIPs* still do this AFAIK.

* http://en.wikipedia.org/wiki/Raster_image_processor

hhp

quadibloc's picture

I know that a book about fractals used the question "How long is the coastline of Britain" to introduce the idea to the reader.

Some fractals existed before Mandelbrot, such as the "dragon curve"; but Mandelbrot can be credited with the investigations that led to fractals being recognized under a unifying concept. This led to the conceptual framework that let mathematicians come up with such things as Hausdorff-Bescovitch dimension, for example.

Fractals are relevant to such things as painting trees and leaves; fonts tend not to need to be natural and wild. But mathematics did touch on some fonts - such as Piet Hein's superellipses giving rise to Melior.

John Hudson's picture

Hrant: And we might have never had Baskerville (and then Bodoni, who was an admirer of Baskerville) without [the Romain du Roi].

The Romain du Roi was commissioned in 1692, and not actually in use until 1702. By that date, John Ayres and George Shelley had already perfected the style of the English Roman on which Baskerville's types are based. The earliest expression of that style that I have so far found is John Seddon's of 1690, i.e. two years before the commissioning of the Romain du Roi. The line of Baskerville's development is so obviously in this English tradition, that I don't think any connection can be made between his types and the RdR.
______

And then there's this, which I came across in a recent Oak Knoll catalogue. This is the title page of a prayer book written by Louis Senault and dedicated to Marie Anne Christine Victoire of Bavaria, who in 1680 married the French Dauphin.

The book is undated, but indications are that it was produced in the 1680s. That lowercase in 'Sainte Ecriture' shows clear indications of being written with a split nib, resulting in the pattern of stroke modulation and vertical stress that characterises the letters of the Romain du Roi (and which are not determined by the construction method of the RdR; as Durer, Tory and numerous others had long since demonstrated, geometric approaches can just as easily produce diagonal stress).

James Mosley, in his essay on 'Les caractères de l'imprimerie royale', has already drawn his attention to the influence of the written romans of Nicolas Jarry and Louis Senault on the design of the RdR, some of the former's dating from as early as 1668. I'd go so far as to say that Grandjean's actual RdR types owe more to Jarry and Senault than they do to the committee's constructivist experiments.

So in terms of the development of neo-classical typography of the mid-18th Century, I posit two lines of development:
French : Jarry/Senault -> Grandjean -> Fournier
England : Seddon -> Shelley -> everyone else* -> Baskerville

* By the 1730s, the 'English roman' shows up in the standard repertoire of every writing master whose work I have examined. No exceptions.

The interesting question, then, is whether these French and English developments were independent pursuit of the same goal -- development of a style of type-like lettering better suited to neo-classical layout and ornament than the actual types of the 17th and early 18th Centuries --, or if Seddon was aware of what was happening in France. I'm inclined to think that the developments were independent, at least in terms of the evidence of the actual writing that resulted. There are distinctive differences in the solutions for some letters that persist in the writing and types of the 18th Century that are identifiably French and English.

hrant's picture

Thanks for the correction/elaboration.

hhp

Nick Shinn's picture

… such as Piet Hein's superellipses giving rise to Melior.

More likely the other way around.
The squircle of Microgramma and Melior (both 1952) preceded Hein’s superellipse (1959?)
IMO the vogue for this shape derived from the CRT screen, where it was incidental.
Commercial TV programming started in 1948 in the USA.

oldnick's picture

Nick—

Your divining the CRT Effect is unexpectedly brilliant…

Ryan Maelhorn's picture

Seems this RdR face is pretty important, does it exist as a digital font today?

Ryan Maelhorn's picture


This is the most interesting point to me. I myself (who does not have any of Nick's skill), probably would have placed that point further north along the curve, or possibly not have had a point there at all. you could stretch out the control arm of the point at the top of the curve, stretch it out to the right and probably get a smoother curve there. Usually less points = smoother curve. Also, It's hard to tell from the shot, but it looks like Nick may have wanted to create a slight bulge between the point I'm talking about and the one just south of it. Is that the case? I'd love to hear Nick talk about that point a little.

HVB's picture

@Nick Shin - "Commercial TV programming started in 1948 in the USA."

Not sure what that had to do with the topic, other than the CRT, but ...

We got our first TV set, a 7 inch RCA, in 1947. By then there were three broadcasting networks, NBC, (WNBT), CBS, and ABC (WJZ-TV), each carrying one of the New York City baseball teams. Licensing for commercial TV started on July 1, 1941; the first US advertising was on WNBT as part of its test pattern (Bulova Watch Time) that same day. I first saw TV in 1940 at the 1939 NY Worlds fair, but not again until early 1947.

- Herb

Ryan Maelhorn's picture

Actually I heard it was the windows in the then new passenger jets..

$0.02

HVB's picture

Neither particularly funny nor correct. The first passenger jet was put into service in 1952.

oldnick's picture

Herb—

http://en.wikipedia.org/wiki/Television#History

The first set which my birth family had—circa 1948, just like me—was a console model in which the CRT faced up, and was viewed in a mirror attached to the lid of the console. My mother’s father—a professional scenic artist—managed to equip the box with a Fresnel lens, which projected the image onto a wall. The world’s first projection TV? Who knows?

Nick Shinn's picture

Sorry for the inaccuracies Herb, I got my TV dates from Wiki, not personal experience!

Nick Shinn's picture

That’s post facto BS.
A veneer of promotional tech-porn.
Microgramma is squircular in form, i.e. (super-) elliptical, not circle-based.

Ryan Maelhorn's picture

A great book by a great type designer.

HVB's picture

@Nick and Nick - The 1948 date in that Wiki article refers specifically to network broadcasting. The same article gives the 1941 date for the first US commercials, albeit without the WNBT reference.

- Herb

quadibloc's picture

Mind you, some of the earliest TV sets had round picture tubes; others, because the picture tube was quite long, had rectangular faces with only small round corners.

The shape we associate with a TV set came about a bit later on - although it was around from the early days of black and white, before deflection angles got even wider and tubes were aluminized.

oldnick's picture

The 1948 date in that Wiki article refers specifically to network broadcasting. The same article gives the 1941 date for the first US commercials, albeit without the WNBT reference.

True enough, but a trivial distinction: a mass medium does not become a mass medium until it is accessible to the masses. As opposed to the ubiquity of print, television required the purchase of a receiver, in order to mesmerize with its Dancing Scan Lines—a different kind of cool, if you dig McLuhan. Today, it is pixels which do the dancing, constantly seducing us to part with our money to assert our precious individuality: Lord, what fools these mortals be.

From 1981 to 1984, I worked as an Art Director for KXAS-TV, the NBC affiliate in Fort Worth-Dallas. The station came on the air as WBAP (We Broadcast a Program) in 1948 as the first TV station west of the Mississippi, the love-child of Amon Carter, publisher of The Fort Worth Star-Telegram. A few years before I arrived, the station was purchased by Lin Broadcasting as a cash cow: the station routinely generated a 55% profit margin. BTW, Lin was one of the pioneers in developing cellular phone technology.

At the station, I had the privilege of working with professionals who had been there since the beginning—people like Dick Smith, who was Chief Engineer. I became a Grasshopper, and learned well at his feet. State-of-the-art technology at the time also occasioned my first foray into font-making—creating bitmap fonts for the station’ Chyron IV character generator.

Nick Shinn's picture

Chyron IV — sounds like a planet in a distant Star Trek nebula.

hrant's picture

Ah, the Chyron... Sorry this story is a few decades less old :-) but when I made the first ever Armenian fonts for video titling in the late 80s (deployed via Broadcast Titler on the Amiga) the Horizon TV editor guy was very skeptical at first, saying we should instead send in my bitmaps to the Chyron company for burning on a special chip or something at a cost of a few thousand dollars... Yeah, right.

hhp

HVB's picture

A bit later (c.1960), I was at an aerospace company doing Fortran programming, creating graphs for aperture cards (IBM cards with space for inserting a 35mm film frame), using an IBM 740 CRT recorder to make the film inserts. The 740's built-in alphabet existed in only one size and had no lower case. I programmed a set of machine language subroutines, including very simple point-to-point vector-defined characters that could be called from a Fortran program. The user could call the text subroutine, passing it information defining a text string, its location, direction, and size (in mm, if I remember correctly, but it might have been in hundredths of inches).

The IBM 740 CRT Recorder

- Herb

Té Rowan's picture

Heh. A neat hack, I gotta say.

Ryan Maelhorn's picture

"That’s post facto BS. A veneer of promotional tech-porn. Microgramma is squircular in form, i.e. (super-) elliptical, not circle-based."
- Nick Shinn

The thing is though, that book, Alfa Beta, was actually written by the man who created Microgramma. Could he be lying about it's construction, throwing out a red herring if you will, in an attempt to confuse and slow down his competition? Sure, but it seems a bit much, considering no one was forcing him to write that book, and that writing that book probably took a good deal of his time, and that when it was published, Novarese might only have had a handful of real competitors. I believe he really was explaining how he drew and constructed faces.

Frode Bo Helland's picture

Replica from Lineto is grid based, fyi.

hrant's picture

Good point Ryan. Although I still think Nick is right. :-)

hhp

Syndicate content Syndicate content