M75: going nowhere


Hello, all. I just can't make this (these) work... many of the individual bits are nice, but as a whole, they don't link up—much like the nearby eponymous highway, there are all sorts of things breaking up the flow of traffic. (Yes, there are issues with spacing and kerning, especially in the italic UC and all forms of the numeral 7, but these will evolve with the letters.) Can you give me some pointers? I'm past the point of objectivity, just from sheer exhaustion and overexposure!

Is roman /a too wide? Does roman /g seem too... sly? slinky? Is italic /S too wide? Is the roman /S too wide? The roman lc and UC don't entirely match; maybe it's Dutch vs. French influences? And the real issue: the italic lc... it seems too stiff, too static, too fugly, and the flags on the tall letters don't seem to match. Most of the individual letters are pretty, but they just don't want to work together.

«1

Comments

  • Dave CrosslandDave Crossland Posts: 1,040
    edited August 2015
    e, leans left slightly (less than s leans right.) Perhaps shift the lowest extrema of the tail to the left.

    Regarding the widths of s and a, you might also like to review Frank Blokland's Letter Model - http://lettermodel.org - which provides a sound theoretical model for determining such things :)
  • To my eye, some of the italic capitals look a little heavy against the lower case.
  • Hi Michael,

    Yes, there are issues with spacing and kerning, especially in the italic UC and all forms of the numeral 7, but these will evolve with the letters.

    There are a lot of details in your design that can be improved, but I think that your type will make better text if you improve the spacing first. Have a look at the a–e range and the gap after the f. The latter is somewhat too big, but compared with the cramped spacing of the preceding range it looks more worse than it is. Compare your spacing of the m and n with q and r, and you will realize that it will be difficult to improve the quality of your type by focusing on the letterforms first and next on the spacing.



    Based on my research I conclude that our illustrious precursors from the Renaissance defined the pattern first and adapted the details of the letters to this. Today most type designers do the opposite: they design letterforms and then they start to work on the fitting. Of course, they do not encounter the same technical restrictions as the early typefounders. Still, many of my Antwerp students start their type designs with defining the width/weight/contrast patterns first. And yes, in these examples from one of my former EcTd students the k is too wide too.



    Of course, it is possible adjust the widths of the letters during the spacing process. We use this tool, but I would not be surprised if there is an open-source equivalent somewhere.

    The short terminal of your f is interesting. Two decades ago Chris Brand told me that he never considered to sell Albertina to Lettergieterij Amsterdam because LA was affiliated with the Intertype Corporation. As you know, Intertype built linecasting machines and these did not support kerning. Chris did not want any limitation when drawing the f’s terminal. Eventually –with the help of Hermann Zapf– he sold Albertina to Monotype.

    F.
  • Have a look at the a–e range and the gap after the f. The latter is somewhat too big, but compared with the cramped spacing of the preceding range it looks more worse than it is.

    Indeed. (Heh, I thought it was just that I hadn't gotten around to kerning @f @g.)

    Compare your spacing of the m and n with q and r, and you will realize that it will be difficult to improve the quality of your type by focusing on the letterforms first and next on the spacing.

    Boy is that right. The rhythm thing of which you speak -- it's something I wanted to do, having always loved the 1:2:3 spacing of i:n:m in Trinité (and many Renaissance faces), but I was soundly talked out of it on Typophile. The letters haven't seemed as nice since then. It's not often one is happy to undo work!

    Still, won't this make /u too wide? It has the same width as /n, and people have said to narrow it. And don't differing widths help aid recognition? (But then again, I'm not trying to design a modern face; maybe it's just a matter of individual taste.)

    The short terminal of your f is interesting...Albertina...Zapf

    Yep, I do like a non-kerning /f. Maybe it's too much Zapf or Bringhurst, but I like to avoid such clashes altogether.

    Albertina is probably my favorite face, but as drooling over pdf spec sheets is the closest I've ever gotten to it, I'm basically just guessing that I like it. :)

    Anyhoo, back to the beat. ty
  • Craig EliasonCraig Eliason Posts: 926
    edited August 2015
    I'm not sure if I like the non-kerning /f/, but if you're committed to it, you might consider a thicker and wider crossbar to ameliorate the slightness of the glyph. (And crossbar of /t/ would follow suit.) Also not sure why it's so short compared to other ascenders.
    Lift of top serifs of /T/ and /Z/ looks really odd to me. In lining seven, looks like you've lowered the whole top part so that the serif doesn't stick up too far, but that looks odd to me too. 
    Many of the figures might be a little too narrow, but lining two is way wide. Its base is too thin too.
    Some diagonals such as those in /M/ and /N/ are too thick.
  • No, I have to be doing something wrong. Setting up just /n, with each leg equidistant (origin, ¼ the space of the character, center of first leg, ½ the space, center of second leg, ¼ space, right sidebearing) works ok, but seems wide. "jump" spaces ok, if wide. Spacing [b d p q thorn] to match /n gives me a huge amount of space after/before the bowls, but the stems are perfectly aligned with those of /n. It gets worse with the other letters. Kerning fixes everything, sure, but this is a lot of white space everywhere... the rhythms are technically perfect (same stems align exactly, same total character widths, bowls just a bit outside the appropriate stem) but they are lost.

    Or am I just taking things too literally? Here's the rest of the (roman) lc:


    in which the rounds are the culprit. Still, in LeMo's examples, the n-stems and overall character widths all line up. Or seem to... is it just a matter of adjusting sidebearings for optical correctness?

    Wow, my /a looks thin and weak, and a claw terminal might not be the best fit.

  • You should get a closer look at the red rhythm overlapping image (uh… the first one Frank posted). The rounds are compensated in that design – you seem to have taken mathematically identical space between the glyphs. See how the red n shape overlaps with the initial b–d–g sequence for what I mean.
  • I concur with Rob. You have managed to mess up the spacing completely now. IMHO, the problems you encounter are due to the short route you have taken by looking (too) closely at typefaces such as especially Bembo. You added (unconsciously?) some French-Renaissance and Baroque details and as topping some Zapf-inspired stuff. I would like to discourage that a bit, because the highest achievable will be a pastiche.

    Of course, you could argue that this is in the best tradition of the métier, because Renaissance punchcutters often used the type of their precursors as basis. If one looks at the lowercase of Van den Keere’s Parangon Romain, then he surely must have transfered the image of Garamont’s Parangon Romain directly to his punches before adapting details. I know of a later description of such a technique that could well have been used during the Renaissance.



    However, we live in a time in which one constantly stumbles over type designers. Yesterday I met someone who isn’t a type designer and I was actually relieved. Unless one makes a revival, the only thing that distinguishes one as type designer more than ever before in the history of type, is a strong idiom. To develop this one has to study all aspects of type design from scratch. But you can call me highly old-fashioned. Some related aspects were discussed in this topic BTW.

    Your flirt with the Renaissance makes your typeface very suitable for artificial spacing. If you send me <PatternMan[at]dtl.nl> your font, I will cadence it in a split second. I will use for this the LS Cadencer tool –built by Lukas Schneider on my theories and algorithms– under RoboFont then.
  • > I know of a later description of such a technique that could well have been used during the Renaissance. 

    Off topic, but I’d like to read that! Can you tell me where it was written?
  • Hi Dan,

    All you need is a little patience; you can read it in my dissertation later this year.

    F.
  • You should get a closer look at the red rhythm overlapping image (uh… the first one Frank posted). The rounds are compensated in that design – you seem to have taken mathematically identical space between the glyphs. See how the red n shape overlaps with the initial b–d–g sequence for what I mean.

    Yes, that's exactly what I've done. So it is ok not to be mathematically exact, to adjust things so they look right. That's... actually something I needed to be explicitly told. (Too much autism, too little self-confidence.) Cool. This means a solution is possible.

  • IMHO, the problems you encounter are due to the short route you have taken by looking (too) closely at typefaces such as especially Bembo. You added (unconsciously?) some French-Renaissance and Baroque details and as topping some Zapf-inspired stuff. I would like to discourage that a bit, because the highest achievable will be a pastiche.

    Indeed :( And consciously if unwisely added; I tried to simply include a bunch of features that I liked in various faces.

    Unless one makes a revival, the only thing that distinguishes one as type designer more than ever before in the history of type, is a strong idiom. To develop this one has to study all aspects of type design from scratch. But you can call me highly old-fashioned. Some related aspects were discussed in this topic BTW.

    Intriguing!

    OK, I'm off to do some serious rethinking and redesigning. Many thanks!
  • However, we live in a time in which one constantly stumbles over type designers. Yesterday I met someone who isn’t a type designer and I was actually relieved.

    Heh, where I live, you can't even meet artists, much less type designers.


  • Much better already.
  • Michael, if you tell me the advance width of the /n, I can recommend you the advance width of all the other lowercase letters. Then you can play and see if you like the results (or not).
  • Michael, if you tell me the advance width of the /n, I can recommend you the advance width of all the other lowercase letters. Then you can play and see if you like the results (or not).

    Sure, sounds fun -- I've been wanting to use your spacing macro for a long time now. Umm... 32 on the left, and 22 on the right, stems centered on 140 and 420; the serifs are symmetrical, but equal space on each side looked wrong -- probably because the flags are shorter than the feet.


  • Michel, the spacing macro is old & abandoned, it was an experiment, but I've never get to finish it completely.
    Now I have a new system, based in part in Frank's theory, and the other part in data processing.
    You gave me the side-bearing and stems values, but I don't need those. What I need is the advance-width (the full width of the glyph).
  • PabloImpallariPabloImpallari Posts: 527
    edited August 2015
    Took the liberty to grab your font from the pdf to perform some test. Hope you don't mind.
    Find the results in the attached pdfs.

    In the edited version, I've only expanded (or contracted) the the advance-width, and nothing else.
    I think you can get even better results if you keep the width of the glyphs as they are in the edited version, and modify the contours to better fit into those widths.
    Some glyph can also be better visually-centred, like the /e.

  • Hi Michael,



    I took the liberty to ask Pablo for the grabbed font and to cadence it in a jiffy using the LS Cadencer extension in RoboFont. You will find the outcome in the PDF below.



  • You gave me the side-bearing and stems values, but I don't need those. What I need is the advance-width (the full width of the glyph).

    D'oh -- sorry, thinkographic error. I still had it in my head that the way to space a letter formed on /n is that the stems within the letter and with the following letter should be identical.

    Anyway, it's 550 (560 for full symmetry of sidebearings).

  • I took the liberty to ask Pablo for the grabbed font and to cadence it in a jiffy using the LS Cadencer extension in RoboFont. You will find the outcome in the PDF below.


    Nifty -- looks great! I was surprised at some of the values, seeing /i and /l spaced the same, for example, or how /v has zero sidebearings; no wonder my efforts were so lame.

    How does it do on italics?

    I've done some glyph redesigning, hopefully to the point where further design won't affect spacing (ha!); here are the current files.

  • Evie S.Evie S. Posts: 56
    Pdf?
  • attarattar Posts: 209
    > how /v has zero sidebearings

    That is because v has a lot of whitespace on its sides, so it needs far less spacing than a plain vertical n.

    Anyway, spacing like anything in type design is not so much a matter of numbers but of how things look like. nnvnonovoo is the spacing template (going from plain to round, the most primitive shapes). Make uniform whitespace between the letters then change v for the other letters and do the same, that is how I do it.
  • Pdf?

    The italic is still a mess, but here it is without kerning. Very much wip!

  • Anyway, spacing like anything in type design is not so much a matter of numbers but of how things look like.
    Learning that the hard way... amazing how often I see optical correction everywhere now.
  • > So it is ok not to be mathematically exact, to adjust things so they look right.
    It’s not just ok, it’s one of the bases of type design.

    But also math is at the basis of type design. From my research I conclude that Renaissance punchcutters standardized and systematized type because this was a prerequisite for the production (see also my article in The Eternal Letter). Of course, if the standardization would not have resulted in optically appealing structures, the punchcutters would have adapted the system.

    The standardization of roman type structured the handwritten pattern and directly influenced details and proportions of letters. All of you have been conditioned by these letters and are capable of optically reproducing and judging the patterns. Hence, you could conclude that the origin of the patterns is purely an optical one. I think it is slightly more complex than that. And the fact that the cadence-units system, which has been distilled from Renaissance patterns, artificially generates an optically appealing fitting of type in a jiffy underlines this.

    >I was surprised at some of the values, seeing /i and /l spaced the same, for example […]

    The reason for this is that for the Renaissance punchcutters an equal stem interval was important, if only because it made production matters simpler. They adapted details, such as the lengths of serifs, to this pattern. Today type designers only look at the equilibrium of white space. Consequently they adapt the space to the details of the letters. In your case you could make the left side bearing of the l slightly more narrow. But you will interrupt the stem interval then.

    Please find a PDF with the cadenced italic below. TBH, I did not have a Cadence Units Spacing Table (CUST) file for italics yet and consequently I adapted a roman-type one (which probably is historically correct anyway). The outcome is not perfect but it gives you a good basis, I reckon.

    F.
  • Please find a PDF with the cadenced italic below. TBH, I did not have a Cadence Units Spacing Table (CUST) file for italics yet and consequently I adapted a roman-type one (which probably is historically correct anyway). The outcome is not perfect but it gives you a good basis, I reckon.

    Looks great! Amazing, the difference that spacing makes. I forget who it was, but somebody here (Nick Shinn?) once said that it was the spacings and proportions that matter, rather than smaller details like serifs. I'm convinced of this more and more as this thread lengthens (not that my design is particularly good except as an example of n00bish pastiche  :) ).

    Anyway, I have a few questions. First, the file cuts off after /p. Second, what do the suffices _SS _SX _XX (Latin words should have Latin plurals, dammit) mean? Most importantly, what do the numbers mean? If I treat them as absolute sidebearings, I get different results than your graphics show -- much tighter spacing overall, and positive sidebearings to the right of tall letters (especially UC). Then again, the individual glyph cells you show are easy enough to reproduce manually.

    As for adapting roman-type spacing, I'd bet that's how they did it in the really old days (long before Linotype) -- it's pretty clear that this was how they adapted blackletter to roman (long before fussy French design books), so there was already a precedent. Man, your dissertation is going to be badass!
  • Hi Michael,

    Second, what do the suffices _SS _SX _XX […] mean?

    ‘S’ stands for ‘stem’ and ‘X’ for extreme. Either the measurement is made from the stem or from the extreme on the x-axis. Therefore the capital I has ’SS’, the D has ‘SX’ and the O has ‘XX’.

    Most importantly, what do the numbers mean?

    These indicate the number of cadence units. You can find more info about this system I have developed (actually distilled from archetypal models) in paragraph 3, The relation between the harmonic system and the rhythmic system on this page.

    Man, your dissertation is going to be badass!

    Yes! :smile:

    F.
  • Latest version; italic spaced (probably too tightly, especially after /G). Also, added some niceties -- can't help it, 'cause ligatures and alternates are fun!
Sign In or Register to comment.