Wednesday, August 13, 2008


Even more unsolicited advice to young composers: if you use a computer engraving program, don't use the default text fonts.

Use of the out-of-the-box settings may be taken as a signal the composer is unengaged with the appearance of the score and those New Roman Times are very 1995, if you get my drift. Why not try, instead, to use fonts that visually complement the sensibilities of your music? If you have the skills, make — or have made for you — your own fonts, ideally for both text and notation. Composer-novelist-all-round talent Carter Scholz made a suite of text fonts based on calligraphic designs of Lou Harrison (they're available from Frog Peak Music). A font based on John Cage's best text calligraphy has been available for some time, as well (it's here; I believe, however, that some of the glyphs in the "extras" set are actually from drawings from Thoreau's Journal that were used by Cage in works). The scores of Karlheinz Stockhausen, from his own home-based publishing house, use both music and text fonts designed, to Stockhausen's specifications with the text based upon his own handwriting, by James Ingram. I believe that the Stockhausen Verlag uses Optima as its house titling font, another part of the distinctive look.

In my own scores, I'm slowly getting closer to a satisfactory default engraving style. For some time, I've had a font for titles which I like a lot (Cochin Archaic) and I've come round so far to the principle of distinguishing markings within the stave systems (lyrics, tempi, expressions, various devices; all items that require player interpretation) with a serif font from text external to the stave systems (information and instructions that are given rather than interpretable: names of instruments, measure numbers, clocktimings, the word "score", the copyright line) which is done sans serif. I will admit that, up to now, I've put these in out-of-the-box New Times Roman and Arial fonts, as placeholders for my dream fonts, but am finally settling on a pair of alternatives with which I will retrofit all of my scores. One of my next big projects will be tweaking my music notation fonts to get a set of symbols which are somewhat closer to my own manuscript style, but fonts are tricky to make and there's a lot of learning to be done before I get there.

Websites remain, however, the one ubiquitous media in which a choice of fonts is seriously limited. You might really want to have say, Palatino, but not everyone has a Palatino on their machine, so the CSS or HTML code will include a list of acceptable fonts in order of preference (for example: Palatino, Garamond, Georgia, "Times New Roman", Times, serif, to steal from the style sheet of a prominent young composer-blogger). If you really want a particular font, you have to place the text in question in a graphic of some sort, and that can mean some bandwidth costs.

On the other hand, it can well be argued that fonts are unimportant, and as long as they don't get in the way of the information they are supposed to help communicate, they're just gilding the lilies and time spent on such details is time not used otherwise more wisely. But even with that argument, there are many big unknowns concerning the readability, clarity and efficiency of typefaces, a complex topic about which I have too little expertise, so any definitive statements about whether a particular type face helps or hinders score reading (or blog reading, for that matter) are probably premature. In the meantime, I enjoy the few gilded lilies that come my way as little tokens that someone cares...

Here are some related earlier posts: Tools, Handmade, Sins & Fibs, Drafting

Also this: anyone else seen Gary Hustwit's documentary film Helvetica? My least favorite typeface — that tail on the "R" really does it in for me — but still a film to watch. The money quote comes from designer David Carson: "Don't confuse legibility with communication".


Anonymous said...

I always used to prefer Book Antiqua and have sedway'd into Georgia (the numbers are more easily distinguishable to my wizened eyes). But I hand-letter
my scores, partly because I worked my way through art school as a signpainter.

Civic Center said...

Just make it clearly readable first, and exquisitely pleasing second, and this is coming from an old typesetter.

Elaine Fine said...

Helvetica is a great documentary. We own it, and I have watched it many times. Helvetica, the typeface, doesn't seem to work for music though. Perhaps it is too neutral. I just go for "out of the box" Times New Roman for words that I use in music, though I would love to get my hands on the fonts that Alphonse Leduc used for both notes and text in the early part of the 20th century.

Daniel Wolf said...

rbonotto --

I think that Georgia is a very good web font. Maybe I'll change this blog to Georgia.

sfmike --

I agree entirely.

Elaine --

I'm fond of Leduc editions as well (together with the more Art Nouveau Viennese editions, they represent a high point for me in engraving). The titling font I use, Cochin Archaic, came from a hunt for something along Leduc lines, eventually settling for a cheerily archaic look. I don't think that Helvetica works for music, either, but I do want a sans serif font for score instructions that are non-interpretable. For this, Arial, being even more neutral and consistant in appearance works better, but I'd like a bit more style. Stockhausen's Optima was a good choice, but I'm now tending toward Candara (I love the "x"). For a lyric font, I'm now looking for something like either Constantia or Georgia, as a serif font that is less harsh than a Times.

Perhaps we can make a joint project of hunting down early-20th century French sheet music text fonts?

Unknown said...

I've been in love with fonts since I got my first USGPO font handbook more than 40 years ago. But I leave my scores at the defaults for familiarity (people 'learn' to read these fonts), since they're not designy (the font becomes an invisible quantity), and because most of my clients want the Finale files and won't accept fonts that aren't the defaults ... right down to having to create some symbols as graphics to avoid using them.

It's also worth noting that the source files are much more usable if common fonts, which have a greater likelihood of being up-converted at least for a few years into the future, are used.

I do have some special items I've designed (Roman numerals that are conjoined, quarter-tones that I like better than those I've seen, etc.) but they're all based on the default fonts. Unfortunately, I can't talk my clients into using them.

In truth, Times New Roman is not a bad font, but it's unattractive in Finale because it has no kerning. It's amazing how nice this font can look with kerning. (And Unicode would make some titles much easier.)

Robert Bringhurst's The Elements of Typographic Style (2004 edition) is excellent. He finds Times New Roman bland but useful (and gives an example of well-kerned usage on pp.204-5) but notes that it and other commonly used typefaces do not pair well with others. Some faces mix well, but not Times New Roman -- which may account for its lack of appeal in scores, where mixed fonts will turn up.

(Helvetica is in my Netflix queue.)


Anonymous said...

My daughter once did a study on serif vs non-serif fonts as her undergraduate senior project in psychology. She tested for comprehension using articles written in Times New Roman and Ariel. It was all very scientific - she had a test group of about 40 people and did all the proper statistics.

What she found was that a serif font like New Times Roman increased comprehension measurably. Why? The thinking is that all the extra hooks and lines in the font grabs the eye and holds it longer than the more slippery non-serif arial.

So I have avoided New Times Roman partly because it is a default font cliche and partly because I don't want the performers looking at the dynamic markings, tempo changes etc, for so long that they lose the notes in the process.

Seems like a simple precaution...