Problem with > in Autometrics

[Mike Thompson]

I have used the Optical Auometrics for all my fonts with apparently good results.
The exception is the greaterthan sign. In all the fonts the bearing on the left is
far too wide and the bearing on the right is maybe a little too small.
The lessthan sign is ok.

Mike

[Bhikkhu Pesala]

IMO Maths Symbols should be omitted from Optical Metrics. (>_<)
(not Autometrics, note, which is applied to selected glyphs).

Maths symbols should be a fixed width — the same as the digits.

[Mike Thompson]

Bhikkhu,
I agree and furtheremore I would like a way of choosing which characters to omit.
Maybe the Optical Metrics could be a Wizard with the first tab asking if we want to
omit any characters, and if the answer is YES, Font Creator would make select
all the characters it proposes to process.
We would then at least see what CF is about to do!
The user could then deselect some of the characters.

I may want to deselect some of the characters used commonly in both Hebrew and Latin
so that the sidebearings can be set by me to work (as a compromise) for both
left to right and right to left typing.

Mike

[Erwin Denissen]

We will further improve optical metrics in the near future, but it would help if we receive more feedback.

We would like to know what parts you feel are already working very good, and where you see room for improvements.

[Mike Thompson]

I'm using Autometrics again and have searched for the list of characters that are processed
by default. Pretty sure I saw it somewhere...
Could someone please post the list again.

Autometrics is a bit overzealous and I'd like to write a transform script to undo the (few) results
I disagree with.

Mike

[Erwin Denissen]

It is not one list, but several sets of glyphs that are processed.

[Mike Thompson]

Erwin,
Sorry I did not understand you. What I meant to ask was
what glyphs does optical autometrics process? True I can see them in
the window as they briefly pass by, but I might miss seeing some of them.
Mike

[Erwin Denissen]

I meant to say numerous glyphs are analyzed in order to calculate the best possible left and right side bearing for each individual glyph.

The list of characters that will be processed by the optical metrics feature is:

abc..xyz
ABC..XYZ
0123456789
!"#$%&'()*+,-./:;<=>?@©€

[Mike Thompson]

Thank you

[mkv]

So the Optical Metrics only handles characters used by the English language, right? Accented and other funny characters are omitted?

Let's assume I have a ready-made font and add characters â,š,đ and ŋ - I will have deal with the metrics manually, anyhow?

I just want to make sure which version of FontCreator to purchase in case I end up in such a dreadful situation

Markku

[Mike Thompson]

Mkv,
I'll just add this to your query so somebody more knowledgeable than I can answer a broader question.

If â is a composite of a and has â has 'Use this glyphs metrics' checked in its composite window, then when a is adjusted by optical metrics, â will also be adjusted in the same way. Can someone confirm that?

(My Hebrew font doesn't have any characters like â, so I can't be sure of thsi answer)
Mike

[Bhikkhu Pesala]

If your font includes composite glyphs with accents like àèìòù, then after running autometrics, select all composites, make them simple, and complete composites.

If they are not composites already, you will need to select them manually, and use complete composites.

Note that kerning will also have to be changed after running optical auto-metrics (or regular auto-metrics).

So, run optical metrics a few times until you're satisfied with the results, before completing composites, or doing any kerning (auto or manual).

See this feature request thread to see how different values for optical metrics are needed for different typefaces. Don't assume that the default value is right for your font, or that the right value for the regular type style is right for italics or other type styles.

[Bhikkhu Pesala]

Complete composites is designed to create dcroat from d + combining short stroke overlay. If short stroke overlay doesn't exist in the font, then underscore will be used instead. The metrics of dcroat will be taken from lowercase d, so after running optical autometrics, it will be the correct width. Some editing may be needed.

dcroat.png (4.21 KiB) Viewed 10907 times

Complete Composites composes lowercase eng from lowercase n + comma. This is a design aid to use the same metrics and basic glyph shape as lowercase n.

lowercase eng composed.png (4.44 KiB) Viewed 10907 times

However, the designer may decide that the metrics need to be different for lowercase eng. Either way, if you run the autometrics first, then design the glyph, it will match the lowercase n. If the glyph is already designed, then don't run complete composites, and manually adjust the metrics to match lowercase n or to suit the design of lowercase eng.

lowercase eng.png (3.6 KiB) Viewed 10907 times

[Erwin Denissen]

So the Optical Metrics only handles characters used by the English language, right? Accented and other funny characters are omitted?

Let's assume I have a ready-made font and add characters â,š,đ and ŋ - I will have deal with the metrics manually, anyhow?

I just want to make sure which version of FontCreator to purchase in case I end up in such a dreadful situation

Markku

Right now there is no easy way to set the metrics of those at once. We will most likely improve the auto metrics feature. We are thinking of making it group based, e.g. the a and â can both be in the same group, which will get the same left and right side bearings. However before we actually decide to implement this, we first need to have more feedback from current users of FontCreator. Maybe people have other needs, or a better approach, so do let us know your opinion!

If â is a composite of a and has â has 'Use this glyphs metrics' checked in its composite window, then when a is adjusted by optical metrics, â will also be adjusted in the same way. Can someone confirm that?
Mike

No, the â will keep its current values, and the 'Use this glyphs metrics' will become unchecked.

[mkv]

Thanks to all of you for the answers so far.

The "grouping" sounds like a pretty attractive alternative. If it can be implemented as an optional way to carry out this task (and maybe others) it should not break anything, either.

I have not done any font design for two decades but what from what I have seen I cannot say very much has changed since Windows 3.11, really

Markku