white space compensation or other reasons in N, M, W etc.

Lukas HornLukas Horn Posts: 14
edited September 2019 in Technique and Theory

when I was drawing Ns, Ms, Ws etc. symmetrically (in the first row of the first picture) I've been told to put for example the right Stem of the N nearer to the diagonal stroke (second row). So the white space between the left stem and the diagonal is bigger than the white space between the right stem and the diagonal. And voila: The N looks better now! 

Then I realised that the same thing occurs in other letters. So it's about balancing the white space, right? But I don't get how it really works or if there is some kind of theory behind it I didn't heard of. And maybe I misunderstand something and now it's a good time to get it right.

And is it the same 'problem' here? Or is this a type history topic?

I'm happy for every piece of information, big thanks!


  • In the bottom image, you just reinvented Cormorant. :grimace:
  • If I well understand your request, I think you will find something interesting in this book. To me it has been really helpfull when I drew my first typeface. https://www.amazon.com/Designing-Type-Karen-Cheng/dp/0300111509
  • The M and W are all about balance. You (usually) want them to look like a structurally sound building. Imagine them carrying some weight and not breaking.
    That's a good visualisation for me I think! Thanks Jasper :)
  • Depending both on style and on conditions of reading, a and e can properly have quite even or quite uneven counter ratios. But within a font, whichever ratio one has, the other should relate. 
    This is interesting, because some fonts got an a with a middle bar visually positioned in the vertical center of the letter but also got an e with a middle bar positioned above the center. I thought it was wrong but I didn't had a consistent answer against it. Thanks Craig!
  • Rafael CasesRafael Cases Posts: 31
    edited April 2020
    It has to do with notan and with letting the counters activate the negative spaces.  At least this is what I learned from Adrian Frutiger's mentor, Alfred Willimann.

    With your lowercase e, try experimenting with a diagonal crossbar or with a curved crossbar.  Find which one activates the sheet of the white paper better.  Or play with Photoshop's Gaussian Blur and set it to a low value.  In my experience, Stone Sans and Myriad Pro are negative examples, whilst Quadraat by Fred Smeijers is a good example.  Frutiger typeface is also a good example.
  • K PeaseK Pease Posts: 100
    I think I have some insight about your experiments with M and W here. With N, you balance two spaces, which is as simple as making them seem equal. But balancing three spaces bilaterally does not require making them all equal. Making the middle space significantly different helps the whole thing cohere as a unit. It keeps the reading mind from breaking it into smaller pieces. The left half doesn't balance, the right half doesn't balance, but all of it together does, which lets you recognize quickly that this is one letter with one center.
  • Craig EliasonCraig Eliason Posts: 1,126
    That’s a clever theory but I don’t believe that “making the middle space significantly different helps the whole thing cohere as a unit.”  To the contrary, overly large middle counters in M and W and w will more often break apart those letters than unify them. 
  • Nick ShinnNick Shinn Posts: 1,706
    I’ve always admired Zapf‘s Ws (Aldus, Palatino, &c.) which apart from looking magnificently calligraphic, have certain spacing benefits:

    (kerning off).
  • Igor PetrovicIgor Petrovic Posts: 105
    edited May 2020
    In general, this is the intersection area between typography and cognitive psychology. We subconsciously apply "expectations" from the real world to what we see on paper and screen. 

    One of the most obvious here would be the "gravity expectation". In other words, we expect that there is a force from the top to the bottom, where the baseline acts as a ground. @Jasper de Waard mentioned stability in the sense of carrying the weight. The letters with a rounded base like S are in danger to roll over left/right, for the same reason.
    Indirectly, we expect that the "pyramidal" distribution of the weight is most stable. That's why most of the letters tend to slightly conform to this: Bases of Z and X are slightly wider than the top, the same goes with terminals/counters of S, counters of N, B, terminals of C, E etc. 

    The other expectation is that horizontal movement is more important than the vertical. For the evolutionary reasons (like hunting i.e.) we better control what's going on in "our layer". That's why we have a "cognitive bias" and make horizontals actually thinner in order to look the same as verticals.

    Expectations like these, shape our aesthetics in general. I remember from my architecture studies that sometimes columns are made thicker than calculated by the mathematical analysis because people feel uncomfortable if they look too thin. 

    Gestalt psychology pays attention to visual cognitive effects and biases, illusions, etc. This area should be more explored for the purpose of typography and graphic design in general. If I had a chance to be in Hague or Reading I would probably pick this as a thesis. If someone has articles/books to recommend on this topic I would appreciate it!
  • Craig EliasonCraig Eliason Posts: 1,126
    @Igor Petrovic did you catch this thread from last year?
  • Igor PetrovicIgor Petrovic Posts: 105
    @Craig Eliason No I didn't, many thanks for the link! I'll check the article and thread.
  • Chris LozosChris Lozos Posts: 1,342
    I’ve always admired Zapf‘s Ws (Aldus, Palatino, &c.) which apart from looking magnificently calligraphic, have certain spacing benefits:

    That  W looks awkward to me, as if someone had pulled its legs too far apart.  Also, it would be a problem to kern with a cap "A"

  • Nick ShinnNick Shinn Posts: 1,706
    Also, it would be a problem to kern with a cap "A"

    You don’t kern Aldus and Palatino caps, you letter space them!

  • Chris LozosChris Lozos Posts: 1,342
    great, that helps
Sign In or Register to comment.