No magic number. Every font needs different leading depending on the line width, x-height, and other factors. Do what looks right in each situation.

Hi Billy

Stephen's right. There ain't no magic number.

The closest you might come to a pre-set formula is the (factory set) default value for auto leading in Adobe InDesign and Quark XPress – 120% of the type size. In other words if you set 10pt type the leading will measure 12pt from baseline to baseline. That's only two points extra to the body size – almost certainly too tight for sans serif faces, and not necessarily ideal for the majority of the serif faces either; I advise my students NOT to use auto leading.

120%. That's what I was looking for. Thanks.

That number won't work as well if you have a very wide or narrow column measure or your font's x-height is very high or low. Always proof and adjust.

I'd stay away from the 120% formula. There's no way that anyone can depend on a set formula, even if it "works" some of the times. X-heights, ascenders, descenders, style of the typeface, weight of the typeface, and the personality of the typeface are all factors. The secret is optical adjusting.

Dang! I shouldna told him...

Perhaps there's a reasonable formula we could generate through collaboration, however, that gives a fair approximation...
something that takes into account:
* font size in points (S)
* x-height (x)
* cap height (c)
* ascender length above cap height (a)
* descender length below baseline (d)
* and perhaps even line length (L)

Are there any other math fiends out there who also want to give this a whirl?

I'm imagining an equation that allows:
* leading is proportional to a, d, & L
(meaning fonts with longer ascenders & descenders and longer line lengths require more leading)
* leading is inversely proportional to S, x, & c
(meaning fonts with larger point sizes, x-heights, & cap heights require less leading)

To get started, let's establish some good target values. For example, what should the suggested leading be for:
Adobe Garamond
* at 8-point with 52 chars/line?
* at 10-point with 52 chars/line?
* at 12-point with 52 chars/line?
* at 8-point with 26 chars/line?
* at 10-point with 26 chars/line?
* at 12-point with 26 chars/line?
Gill Sans
* at 8-point with 52 chars/line?
* at 10-point with 52 chars/line?
* at 12-point with 52 chars/line?
* at 8-point with 26 chars/line?
* at 10-point with 26 chars/line?
* at 12-point with 26 chars/line?

Ouch, my nerdliness is REALLY showing.

--Michael.

------------------------------------------------------
// love what you do or do something else. //
Michael Ebert -- graphic designer, jazz saxophonist, horror movie devotee
------------------------------------------------------

1. Line length, or 'measure', is not a 'perhaps' but a critical variable.

2. Generally, some recommend more leading for sans serifs, and I personally think this is right.

3. How loosely or tightly the letter and word spacing is also has an effect.

Generally, there are too many variables and with computers and laser printers it is easy to test different samples. So I agree with Stephen. Just test it.

Oh, and leading is a critical factor in readablity, so all the more reason to test it in each case.

Damned network security, I had to rewrite this post.

I find it hard to get a working formula for leading because there are some factors that can't be accounted in it such as aesthetic preferences or design needs. Also we have to consider the specific typeface's rythm and weight (for example didones requiring greater leading than garaldes) and also letterspacing as William pointed.

I think that a list for a specific typeface like the one that Michael suggests or a loose formula for specific type styles (garalde, didone, modern, grotesque sans, etc) are the most feasible.

Héctor

Line Spacing is in the eye of the beholder--the better trained eye makes the better spacing to behold.

ChrisL

After a good hour of work on this, I've abandoned all hope of formulating a useful equation.

I felt like I was trying to impose math and logic on aesthetics: an exercise in futility.

Instead, I'd follow Hector's recommendation: create your own swatchbook for leading that contains various combinations of common fonts, sizes, line lengths, etc.

Now back to work. :-)

--Michael.

------------------------------------------------------
// love what you do or do something else. //
Michael Ebert -- graphic designer, jazz saxophonist, horror movie devotee
------------------------------------------------------

Yeah, plus it depends on the layout - line width, page margins, etc. Just make sure if it's a multi-column layout that the leading doesn't overwhelm the gutter btw the columns. You also have to consider the weight of the type, serif or sans, and how dense you want it to feel.

While a formula (a mathematical one) isn't a solution, a method should exist. I'm sure that most of you have a method that you use in order to find the best leading. „It has to look right.” Of course, but what does right look like. What should you look for? I would really appreciate if someone took the time to describe a method like this and really exemplify how the leading should be modified on different criteria. How to achieve that balance between the lines of letters and the lines of space.

There are guidelines that tell you to have the linespace grater than the wordspace, but not having it too big or the lines will fall apart. But how does this translate in relation to the inner spacing and such? When is the leading too small? When is it to big? I would like to be able to answer these questions just by looking at the way that the different spacings relate to one another, but I don't know how to check this.

Better than “what looks right” is to do “what reads right”. Typesetting is really all about producing text you can read and comprehend without distraction. If you don’t read and care about the content you’re setting you’re much less likely to set it well. So, experiment: try various settings using various texts (that you aren’t already familiar with) and see what you read most comfortably.

I wasn't using the correct words, but that's what I meant.

So you're saying that the solution is to print the text(s) in different increments of leading and just see which one reads better? :)

In other words, the solution would be to try all the increments of leading between just larger than the word space and too much that the lines falls apart and find the sweet spot based on which one reads better. Right?

Language should affect leading.

In German all nouns are capitalised, and I think “k” is more common than “c”. This means there are more tall letters, so the leading should be greater.

In Latin literature as printed nowadays, capitals are usually only used for proper nouns, not for new sentences (which may mean extra space between sentences); “k” is very rare. There are fewer tall letters, so less leading is needed.

tomita: Sure, this is one approach that might work well.

Bernard: Good point. This is why proofing the actual content is more affective than relying on some formula.

Something else affects leading. Suppose you have a paragraph of block capitals.

If it is a copy of a Latin inscription you want to make it as easy to read as possible. The language is not the first language of the reader. This may mean generous leading.

If it is an end user licence agreement, the object is to make it unpleasant to read without being illegible, so that the end user clicks the “I have read the EULA and agree to it” button without having read it.

Hi,

That might not be a good idea to discuss this old thread, but I feel something needs to be added after this.
If:
—there's no magic formula;
—there's nothing like a trained eye;
—the one who's asking the question must not have such an eye, per hypothesis.

then I think what lacks here and must be of great use would be some various examples of what is OK what is not, in order to get, little by little, the “ruling eye”. I have found very few examples, and I feel some variety would be useful.

Thanks for attention (and for linguistic indulgence).

It's certainly true that 120% is not going to be always right. However, maybe the poster was never going to use any other typeface but Times Roman, in which case a simple number, whether it was 120% or not, would work.

Could one generate a magic formula if one knew the x-height of the typeface in question, in addition to its point size?

I don't see why not. It still wouldn't yield perfect results in all circumstances, but as a rule-of-thumb for a first cut at typical body copy, with all the usual qualifications, it might work.

I suppose the idea would be that Caslon and Jenson don't need leading, and other typefaces need as much leading as makes up for the missing height in their ascenders and descenders.

However, if the x-height is given as a ratio between x-height and cap height, you still need the ratio between cap height and point size.

Another consideration, in the case of book text, is how many pages the publisher is willing to extend to. If there are limits there would have to be trade offs concerning font choice, point size, leading, margins. I do find that generally speaking InDesign's default leadings or Centaur work very well for book text and poetry. Historically there have been famous type and book designers who have not liked leading at all. But I think that works only in quite specific cases.

Could one generate a magic formula if one knew the x-height of the typeface in question, in addition to its point size?

I don't see why not.

If it wouldn’t also use the line-length as required parameter, it would be as good or bad as just starting with 120% as default.