Re: font-size and accents, again

Tantek Celik wrote:

> Any other expert typographers or font authors on this list
want to speak up
> and settle this squabbling among amateurs? ;-)

I'm no expert, but I spec'd type for typesetting for over 17
years. A respected reference for production artists in 1973
was "Graphics Master" by Dean Lem. The book was not intended
to convey the historical source of printing and typesetting
terms, just the common usage of the day, as briefly as
possible. My copy is falling apart from years of heavy use,
but I keep it still. Here's what it says about em:

"An em is the term used to designate the square of any given
type size."

About the "Anatomy of Type":

"The three main parts of letters are ascenders, descenders,
and the 'x' height. The size of a type face is the measure of
the metal body, not the image of the printed letter. Thus a 48
point size letter may measure only 42 points from the top of
the ascender to the bottom of the descender.

"It is the 'x' height, rather than body size, or the length of
the ascenders and descenders, that conveys the true visual
size of a type face..."

The latter point is well-illustrated by a formal script face
such as 'Shelly'. To get the readability of 10-point Times, a
formal script may have to be set at 24 points or more because
the x-height is so tiny relative to the total height of the

"Set solid" certainly never meant that the top of ascenders
should touch the bottom of descenders on the previous line.
Type faces are typically designed with an ascender to
descender height less than the font size. A bold version of a
typeface might actually be slightly taller for its size to
accommodate the additional mass of the letterforms in the
vertical as well as horizontal direction. If the letterforms
did not exist at a well-defined position relative to a
baseline, the regular and bold versions might not align

It's a shame that 'The PostScript Font Handbook' misinforms
that "Point size is measured from the top of the ascender to
the bottom of the descender." That's contrary to Adobe's own
'Adobe Type 1 Font Format' manual and the data that can be
found in Adobe's AFM (font metrics) file for any type 1 font
(see AFM tidbit below).

And what of typefaces that exceed the boundaries of their set
size? 'Helvetica Inserat' is an example. Possible logic: If
you want to _really_ punch some word or phrase in a line of
Helvetica text, set it in Helvetica Inserat. The beauty is
that you don't need any special instructions to visually-align
these oversize, ultra-bold letters with the surrounding type.

Accents on capital letters might also be set above the height
of a font's ascenders, particularly on font faces with tall
x-heights. Should this change the line spacing of the line
where the accent occurs? I don't think so. IMO, that would
make for some very disconcerting line spacing. Better that an
author, knowing the nature of the typeface, allow some
additional line spacing to accommodate the accents.

Following is some selected data from the AFM file for Adobe's
Times Roman. The coordinate system for an Adobe font starts on
the baseline with up and right positive and is 1000 units per

  UnderlinePosition -98
  UnderlineThickness 54
  FontBBox -167 -252 1004 904
  CapHeight 673
  XHeight 445
  Descender -219
  Ascender 686

The FontBBox is just large enough to contain all of the glyphs
in the font if they were overlapped with coincident start
coordinates. Note that the vertical distance from top of the
highest glyph to bottom of lowest glyph is 1154 units, 15%
greater than the font size. Also note the vertical space
between ascender and descender is 905 units, 10% less than the
font size.

David Perrell

Digression: I experienced the change from hot metal to
computerized phototypesetting. The displacement of hot metal
was not welcomed by all, and I suspect that some type-savvy
nostalgics still scoff at "souless" digital phototype as vinyl
lovers decry the metallic shrillness of CD-ROMs. Each size of
a hot metal font was designed separately to preserve
legibility at small sizes and enhance elegance at large.
Letterform construction programs are just now getting a
semblance of that kind of sophistication. Of course, they will
never (hopefully) match the 'personality' of individual

Received on Tuesday, 23 November 1999 16:31:15 UTC