Tweak handling of ligatures with tracking

[garrettm30]

Put briefly, I suggest tweaking the threshold at which increased tracking splits ligatures into their corresponding characters.

 

 

Here's the longer version, which was a learning experience to me and which some might find interesting. While I was doing some proofreading of a text I had prepared in Publisher, I noticed some odd spacing around ligatures (this is a screenshot, but I spotted it while I was looking at the print at 11.1pt):

 

As I thought about it, the problem was understandable: since the ligature is a single character, there could not be any spacing added between the letters in "ffi" to match the spacing of the others. But then that made me wonder why I had never noticed this before though I had set similar text for several years in InDesign. So I decided to see if InDesign is doing something special. I discovered that as tracking is increased, InDesign switches from a ligature to regular characters.

A fancy trick, to be sure. "Maybe I should suggest it for Publisher," I thought. But with some further experiment, I discovered Publisher is doing the same thing. Kudos for the attention to detail in this first release, Serif!

So instead, I only suggest tweaking that threshold a little bit. Here is the same line in Publisher (top) and InDesign (Publisher) with settings as near as I can make them:

In this case, I think the spacing is better without the ligature.

 

Addendum:

I did a little more testing, and it seems that paragraph justification may be complicating the issue. There may be a calculation error somewhere. In the below shot, the black text is from the same document, with justification word spacing set to 80% 100% 120% and justification letter spacing set to -2% 0% -4%. The blue text above and below is the same word with a left align and 17‰ on top and 18‰ on bottom, which is the smallest smallest distinction I can make in Publisher to show where the threshold is. To my eye, both of them look good, but the version that is in text with the justification applied does not.

 

 

[Pšenda]

Just a question - did you try another font? This seems to me some optically unbalanced (see the big gaps between "on", while "or, ou" is fine).

[garrettm30]

I had not tried another font, so I tried again at your suggestion. The font in my original post was Georgia, but I have tried again with Times New Roman, just to pick an example of a very common font (though Georgia is rather common).

However, I now believe this is a bug rather than a matter of tweaking. I previously discovered (to my delight) that Affinity intelligently splits ligatures when tracking is increased beyond a certain threshold, but it appears that this does not happen when extra tracking is automatically applied as a result of paragraph justification, which would have been the case where I first noticed it.

Here is an exaggerated example where I put my theory to the test. I have a narrow text frame (45.8mm wide) with Times New Roman 18pt. The paragraph is full justified, with justification letter spacing set at -2% 0% 40%:

 

Moderators, I think this thread could be moved over to Mac bug reports.

 

 

[Alfred]

16 minutes ago, garrettm30 said:

Here is an exaggerated example where I put my theory to the test. I have a narrow text frame (45.8mm wide) with Times New Roman 18pt. The paragraph is full justified, with justification letter spacing set at -2% 0% 40%:



 

It’s rather curious that the ffi ligature in « affirment » is only partially decomposed, with the fi surviving as a ligature.

[garrettm30]

8 minutes ago, Alfred said:

It’s rather curious that the ffi ligature in « affirment » is only partially decomposed, with the fi surviving as a ligature.

That's a good point, and it does seem to contradict my conclusion that the feature to split ligatures (I don't know what to call it) is not working at all in the context added letter space because of paragraph justification.

[Dave Harris]

Thanks for reporting this. It's supposed to split ligatures with justification the same way it does if you apply tracking manually, but currently it's broken.

I don't think Times New Roman has the ffi ligature.

[garrettm30]

2 hours ago, Dave Harris said:

I don't think Times New Roman has the ffi ligature.

It appears you are correct. That explains what we thought was a "partially decomposed" ligature.

Thanks for commenting and taking note of the problem.

[garrettm30]

I have tested the document that appeared in the screenshot in the first post, and the issue now appears to be resolved as of beta 1.7.2.458. Thank you, Serif.

This particular issue has helped me to understand that even if Publisher can't be everything from the start (naturally), Serif really does care about even the little details. This is a good sign for the quality of Publisher as it matures.