Classifying Cambria

Thomas PhinneyThomas Phinney Posts: 1,007
edited March 22 in Education
So, for those who are used to an Oldstyle - Transitional - Modern spectrum for serifed typefaces, with Slab on the side....

How do you classify Cambria?

The stress is vertical, or so close as to be imperceptibly different. But the contrast is quite mild, and the serifs are massively bracketed. I am inclined to say that it has elements of oldstyle and modern so it ends up being: (1) transitional...-ish; and (2) a fine exemplar of the problems of type classification.



  • Thomas PhinneyThomas Phinney Posts: 1,007
    I should add: very open apertures, for a serif face, unlike most modern serif faces.
  • James PuckettJames Puckett Posts: 1,536
    Medium contrast with vertical stress and bracketed slab serifs = clarendon.
  • Wes AdamsWes Adams Posts: 58
    Modern, owing to the letter widths and axis. Oldstyle may be an ambiguous term because it's been variously used to describe everything up to the 18th century. 

    So long as the details of one classification are grafted to the rhythm and proportion of another, the spacing still defines it function.

  • Neo-Utilitaranist.
  • A bit glyphic to I think. It's a lot controversal with several overlaping classifications.
  • Thomas PhinneyThomas Phinney Posts: 1,007
    edited March 22
    @James Puckett Agreed that it fits the “Clarendon” model pretty well. The schema I am trying to use (basically a descendant of the Vox-ATypI model) doesn't deal well with Clarendons... probably sticks some in slab and others in modern. Which probably means it's not a good schema for this purpose. Sigh.

    @Wes Adams Yes, if one relies on axis of stress alone (note: I am not saying that is unreasonable, just not my preference), “modern” for sure. The caps are a bit condensed, so they cease to be a reliable disambiguator between modern and old style proportions. The low contrast makes it seem not so much a modern to me—because I think of higher contrast being a feature of .

    After sleeping on it and thinking more, I am thinking the best thing to do for purposes of the job at hand is just to describe these variables, and downplay classification.
  • John HudsonJohn Hudson Posts: 1,377
    I've always classified it as Cambria.
  • Mike DugganMike Duggan Posts: 161
    this is the description from the Now Read This book, if its helpful 

    Cambria has been designed for on-screen reading and to look good when printed at small sizes.
    It has very even spacing and proportions. Diagonal and vertical
    hairlines and serifs are relatively strong, while horizontal serifs
    are small and intended t0 emphasize stroke endings rather than
    stand out themselves. This principle is most noticeable in the
    italics, where the lowercase characters are subdued in style, to
    be at their best as elements of word-images. This font is suitable
    for business documents, email, web design.
    A sturdy typeface for business
    When the call for proposals came from Microsoft, type designer
    Jelle Bosma says, he had been doing heavy-duty programming
    for four or five months straight, and he was eager to get back to
    design work. Some of his colleagues at Agfa Monotype had just
    been in a meeting with Microsoft about the need for a set of
    fonts for a variety of purposes, all of which would be designed to
    take advantage of ClearType rendering to look good on screen,
    while still being useful in print. “So my wish was granted,” says
    Bosma, “and I was put to work.” He found, however, that he had
    come into the process a bit late: Microsoft wanted proposals by
    the end of the week.
    “I tried my hand at a proposal for a monospaced and a
    serif font in two variants: one variant as an e-book font, and
    a version adapted as a business document font: what could
    be a Times New Roman replacement.” It was the latter that
    turned into Cambria. “The other bit of information was a list
    of fonts, mostly traditional old-style fonts, which Mike Duggan
    recommended looking at, because they worked well. I must
    confess that once I got started, things developed in another
    He studied how different geometric shapes were rendered
    by ClearType on the screen, and what the effects on them were
    of hinting. He also looked at the fonts shipping with Microsoft
    Reader, since he knew they had been hinted for ClearType.
    “The result was that at the smallest sizes, the effect of hinting
    did make a difference, but for most types of shapes, except
    long diagonal lines near horizontal, shape detailing mattered
    little. However, starting at slightly above 20 pixels per em, some shapes worked better than others; it seemed best to use
    curves that turn away from the horizontal extreme as quickly as
    “The next step was to draw a lowercase n in all sorts of
    variants, hint them, and select those properties with the ‘best
    “At the smallest sizes, ClearType gives the effect of ‘dressed
    up’ bitmaps. I do have some experience in that field, including
    bitmap design proposals that were the subject of a legibility
    study where various design parameters were tried out. So
    once the n had been established, I had some idea about the
    proportions and spacing which would create a suitable pattern
    of black and white shapes that scores well in a legibility test.
    So starting from an n, I gradually built the rest, trying to keep
    things in harmony.”
    Designing typefaces for the screen, Bosma points out, is very
    different from designing for the printed page – or from designing
    for the screen just a few years ago. “Within the TrueType format,
    hinting is part of the design. The outline drawings you make
    in Fontographer or FontLab are the outlines used when you
    have no pixel limitations. With hints, you draw the outlines
    that express the design at lower resolutions. In the ClearType
    rasterizer with subpixel positioning, you have fewer options to
    influence the appearance with hints, because in the x-direction,
    hints are mapped to a make-believe higher resolution. But it is
    possible to force some things, if need be. For example, when the
    top horizontal of the 5 and 7 has to be one pixel, because two
    pixels is much too heavy, I compensated to make the horizontal
    hairlines a bit heavier, to prevent the top half of the characters
    from becoming too light.
    “In black and white bitmaps, you are by nature forced into
    certain proportions. But if your font doesn’t have these, you can
    hint them to be that way. The price for that is that the screen
    appearance may give a false impression of what it will look like
    in print. In ClearType you still have the limitation that you need
    to separate vertical features clearly. Therefore proportions are
    even and spacing open, even between round stems (oo). So
    in that sense, you are a bit more limited than when you draw
    outlines for black and white hinting, because there the hints
    take over completely.”

    The features that make Cambria uniquely suited to its role as a
    robust, all-purpose workhorse text face have been carefully and
    precisely thought through.
    “For ClearType,” explains Bosma, “one designs monoline fonts
    best. At small screen sizes, hints make the horizontal hairlines
    heavy, and if your horizontal hairlines are thin, they remain
    thin and look silly. You also want to make narrow characters
    relatively wide. In metal type terms: your serifed ClearType font
    should be made from the drawings for the 7-point. But that
    looks rather dull at large sizes, and not much like a Times New
    Roman alternative.
    “So in order to have a bit of both worlds, the design has a
    relatively low contrast in the x direction, and a high contrast
    in y. With the given proportions, this makes the image heavy
    enough for it to survive being used at 8 pt printer sizes, while
    not looking dull and heavy at 16 or 18 pt. At these relatively
    small sizes, the perception of equally thick horizontal and
    vertical hairlines is making the verticals look lighter. A bit of
    exaggerated compensation for this goes unnoticed. Of course
    when used at really big sizes, it will start to look strange – but
    with such wide spacing, I don’t see a big future for Cambria as a
    headline font anyway.”
    Bosma has done something unusual in the f-ligatures, by
    using the possibilities for contextual glyph substitution in
    OpenType. “The implementation of the OpenType ligature
    doesn’t use the drawn ligatures, but an alternate drawing of
    the f only. This has a narrower and lighter top and is used as a
    substitute not only in front of the i, b, h, k and l, but also when
    it is followed by any other lowercase glyph that has something
    sticking out in ‘f space’: narrow characters with top accents,
    anything with a left-side ascender.”
    Cambria comes with a large extended set of mathematical
    glyphs, to support math setting in Microsoft Word. This
    supports an additional 2,000 math, scientific, and technical
    characters from Unicode 4.0. Bosma describes some of these: “a
    full set of combining marks, additional punctuation marks, the
    letterlike symbols, arrows, math operators. There is also a set of
    mathematical alphanumerical symbols which contains a variant of the italics. The lowercase has diagonal stroke endings rather
    than horizontal serifs. These characters are drawn to stand on
    their own, rather than as part of a word image.”
    The design of Cambria’s Greek and Cyrillic complements was
    the most difficult of all the new ClearType typefaces. Although
    the x-height and the cap-height are consistent across all three
    scripts, conflicting ideas about consistency dogged the project.
    (Which is more important: keeping details the same across
    scripts, or making each conform to the standards of that script?
    There is no one answer to this question, nor to the questions of
    style in designing a successful text face.) In the end, Bosma was
    responsible for the Latin design of Cambria, and the Greek and
    Cyrillic were designed by Robin Nicholas and Steve Matteson
    at Agfa Monotype with contributions from, respectively, Gerry
    Leonidas and Maxim Zhukov. The result of this joint process
    should be a family of fonts for business, math, and technical
    documents that will function seamlessly for users in many
    languages and all three scripts
  • I think this type has a strong relationship to the original Cheltenham, which has similar stroke contrast and serif structure. Cheltenham is classified as an oldstyle (its original name was Boston Oldstyle), and I don’t see any reason why Cambria should not be classified in that category, which is pretty broad. 

    To my eye, Cambria has a kind of studied anonymity. That’s not necessarily a bad characteristic, though in this case it feels extreme to the point of unpleasantness. The terminals of the a, f, and especially r are clumsy and look as if they weren’t fully worked out. The lowercase is somewhat narrow, but the caps are even more so. The lowercase s looks expansive compared to the capital, which leaves me thinking that the type was edited by a committee.

    No love here.

  • Thomas PhinneyThomas Phinney Posts: 1,007
    edited March 23
    Mike: Thanks! I do have Now Read This, but I am sure it is of interest to many who have not seen it.

    Scott: I agree that it is rather like Cheltenham in most of the serifs, although Cambria's chiseled terminals (on a, r, f and s) contrast rather strongly with Chelt's rounded terminals. But although Cheltenham may have once had “oldstyle” in its name, that does not make it an old style typeface from a classification POV. The vertical stress pretty much rules that out, in my opinion. Cambria’s cap proportions are not particularly old style either, being condensed more than classical (and Cheltenham itself having pretty even, modern cap proportions). That said, I suspect our disagreement on this may be emblematic of why it is troublesome to classify.  ;)
  • Nick ShinnNick Shinn Posts: 1,273
    edited March 23
    To build on what Christian and Thomas Höggren have said, to me it looks like an agate version of a glyphic style. As Jelle notes, it’s not a display face.

    It’s important to consider how a style would scale optically.
  • John HudsonJohn Hudson Posts: 1,377
    It’s important to consider how a style would scale optically.

    A subset of characters in Cambria Math do scale optically, but only in a downward direction.

  • Belleve InvisBelleve Invis Posts: 230
    I would more like:

    "Serif, low contrast, vertical stress, open aperture, high superness"

    (I'd like to add a "superness" parameter to describe the blockiness of the O shape. The idea comes from Lp space.)
  • Thomas PhinneyThomas Phinney Posts: 1,007
    “Superness” is not a broadly used term in type, but “super-elliptical” is fairly well-known, I think.
  • Ray LarabieRay Larabie Posts: 737
    Super-ellipticity or super-ellipticalness?
  • Thomas PhinneyThomas Phinney Posts: 1,007
    edited March 25
    Heck if I know. I didn't try to go there in the report I am working on. I'm just using adverbs like “slightly,” “moderately” and “highly” to modify “super-elliptical.”
  • Simon CozensSimon Cozens Posts: 292
    edited March 26
    Knuth called the parameter "superness" in his Computer Modern sources:

    superness:=1/sqrt2;    % parameter for superellipses
    more_super:=max(superness,sqrt .77superness);
    hein_super:=max(superness,sqrt .81225258superness); % that's $2^{-.3}$

  • Thomas PhinneyThomas Phinney Posts: 1,007
    edited March 28
    Thanks, Jelle! I am very happy to get your input. 

    And I can’t object that you essentially agree that it is transitional...-ish.  :)
Sign In or Register to comment.