I am working on the editorial design of a book and Lucas Bihan of Bretagne Type Foundry was kind enough to give me access to a trial version of his serif typeface »Self Modern«, which I am now using for a draft.

A preview of Self Modern:

Self Modern Sample1 Self Modern Sample2

Unfortunately, the typeface does not include the Umlaute (ä Ä, ü Ü, ö Ö), which I need since parts of the book are in German. What I would like to do for the draft, is to use Glyphs and take the existing a A, u U and o O of the typeface as the basis for creating the Umlaute myself, of course in communication with Lucas.

Not being a fully professional type designer, is there someone who could provide me with some information regarding the design of the Umlaute, specifically in a serif typeface? I assume there are things to keep in mind and an Ä is not created by merely copying the A and putting a couple of dots taken from the i above it.

  • Does the typeface in question have an ß? Because if not, this will most likely be a much bigger problem (unless you are using Swiss German). Also, what is the point of having variants of single characters as discretionary ligatures?
    – Wrzlprmft
    Aug 21, 2017 at 12:51
  • @Wrzl I'd guess the single char ligatures are just labelled incorrectly and are actually just the mentioned stylistic set?
    – Cai
    Aug 21, 2017 at 13:04

2 Answers 2


You can find a lot of useful information and resources for designing diacritics here:

When designing a latin-based typeface, how are diacritics handled?

Specifically in this case The Diacritics Project. Which says:

The basic form of diaeresis and umlaut are two dots. The symbol thus may be created of two dots same as the one above i, placed in the same height. Sometimes, it is visually more pleasing to scale down the dot accents. However, even the scaled down dots should be of the same shape as the dot accent.

So in theory you can just take the existing characters and dots in the typeface.

Creating the glyphs...

In Glyphs you need to create new glyphs for the diacritics you're missing and set up anchors for those diacritics so Glyphs can auto-generate the combined glyphs for you.

Right clicking the relevant item in the Languages section of the left sidebar of the font view will show you the glyphs you're missing and let you generate them...

enter image description here

Note; the glyph you want is the "Dieresis", what Glyphs calls the umlaut is actually a double acute.

To make things easier, first generate only the diacritic glyph.

Then you need to create the paths in the glyph (so copy/paste some dots and scale/reposition as needed). Set the anchors while you're there too (cmd+U); the anchors are used to position the components in the generated combined characters so you may or may not need to adjust the anchors.

Then go back and generate the combined characters.

Creating the diacritics first allows Glyphs to auto-generate the combined characters as they are created. Otherwise it would have generated everything as an empty glyph and you would need to go through each and go to Glyph → Make Component Glyph to auto-generate them.

More info on setting up diacritics in Glyphs here:

  • Thanks for the input, really helpful. Generating the dieresis worked perfectly fine after defining the diacritic glyph, but is there a way for me to manipulate the dots individually for each dieresis afterwards? For example, I would like to move them in case of the »Ü« but I can only manipulate the diacritic glyph in general, thereby changing it above all the Umlaute.
    – JoSch
    Aug 21, 2017 at 16:07
  • 1
    If you set and change the anchors on the "U" you can control the placement for that character only. Alternatively you can right click the component in the "Ü" and select "Disable Automatic Alignment"; you can then move it as you wish.
    – Cai
    Aug 21, 2017 at 16:28
  • 1
    You can also right click the component and select "Decompose" to convert it to a regular path instead of using the component
    – Cai
    Aug 21, 2017 at 16:29

I assume there are things to keep in mind and an Ä is not created by merely copying the A and putting a couple of dots taken from the i above it.

While what you describe works surprisingly well, there are a few small pitfalls:

  • While lowercase umlauts mostly can inherit the horizontal metrics from their respective base letters, this usually does not work well when they are preceded by an f, some capital letters, or special characters¹. You can solve this by kerning or using a contextual variant for the umlaut (narrower dots) or the preceding character (e.g., an f with a less protruding ascender). Here are some examples and solutions from Linux Libertine:

    Examples from Linux Libertine avoiding umlaut collisions

  • Uppercase umlauts will often be taller than your otherwise tallest glyphs, in particular in a typeface such as yours where the capital are as tall as lowercase ascenders. With a small linespacing such as in your example, this may lead to awkward collisions with descenders from the line above. On the other hand, when lowering the dots, you do not want them to collide with the base letter. To avoid this, place the dots somewhat beside the base letter for the Ä and Ö or between the stems of the Ü. Below, you see Linux Libertine’s realisation of this. Note how for the Ä and Ö, the dots are further apart than for the lowercase umlauts, while for the Ü, they are closer together.

    Examples from Linux Libertine with low umlaut dots

Of course, depending on the length of your text, it may be worthwhile to check whether the above rare problems actually occur before you spend time on fixing them.

¹ Be aware that there are more potential problems in non-German text.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.