On the logic of bearing for Italic

Let's start with the capital letters.
Bearing for Roman is relatively simple, or at least in theory it is based on clear principles.
The LSB and RSB in /O/ are the same; the LSB for /D/, /E/, /F/, /H/, /I/ etc are the same, and so on.
In Italic, on the other hand, I see completely different settings in various fonts. In GaramondPremierPro the measurements are completely different even for the same rods; in the Palatine the LSB is the same for /H/ and /I/, but completely different between /E/ and /F/, which nevertheless have the same design.
What then is the logic, the general conceptual approach of bearing, precisely for italic?
Thank you!

Comments

  • Correct me if I'm wrong, but those differences probably are an error. Italic spacing is a bit more tricky than Roman for two reasons:
    (a) the Italic needs to work alongside the roman (which can usually not be delivered without some kind of compromise – try typesetting “(f)” – you will need manual kerning),
    (b) font editors may not offer complete tooling to aid italic spacing (I saw there was some functionality for this in FontLab but I found it a headache, so I opted out). Because of the angle, if you don't use such features to calculate the sidebearings at an angle, some distances may seem non-obvious, and the whole process is more prone to errors.
  • mauro sacchettomauro sacchetto Posts: 353
    edited January 2021
    Given my inexperience, it is impossible for me to tell if Microsoft or Adobe fonts have similar errors.
    In fact, for example, the Minion Pro Italic has marked regularity.
    In GaramondPremierPro Italic instead, to give an example, the LSB of /I/ is -20pt, that of /E/ which has the same profile -42 and that of /F/ -28. I wonder the reason for these deviations.
    The LSB of Garamond Premiere himself is negative for almost all glyphs; in the Palatino it is positive for almost all glyphs. So I ask: what do we start from?

  • Ray LarabieRay Larabie Posts: 1,376
    Here's how I check my italics: copy to mask, slant upright, check sidebearings, swap with mask.
  • @Ray Larabie Even easier is deriving the Italic soberly from the Upright.  :-)
  • do you mean that you keep the Roman sidebearing?

  • Here's a thread from 15 years ago that mostly just demonstrates to me how glad I should be that modern font editors have slanted boxes and do the math for you, but also might explain why italics should have asymmetrical sidebearings: Typophile

  • John HudsonJohn Hudson Posts: 2,955
    In Italic, on the other hand, I see completely different settings in various fonts. In GaramondPremierPro the measurements are completely different even for the same rods; in the Palatine the LSB is the same for /H/ and /I/, but completely different between /E/ and /F/, which nevertheless have the same design.
    Are you looking at the sidebearing values, or at measurement from the stem at a standard height? Sidebearing values will reflect differences in serif length, which is why it is always better to make measurements from the stems.
  • edited January 2021
    Don't measure from the baseline, measure from the middle of the x-height (especially helpful in an italic).
  • thank you for your reporting, now I'll read the tread

    I considered the sidebearing. now I'll try another measurement, from a line placed (more or less) at half the total height of the glyphs.
    It seems to me that you consider it a more correct practice

  • Nick ShinnNick Shinn Posts: 2,131
    edited January 2021
    @mauro sacchetto
    Bearing for Roman is relatively simple, or at least in theory it is based on clear principles.
    However, several classic serif faces have notably wider sidebearings on /I. 

    Standardization may seem like a good idea, but I have to wonder whether this “cut and paste” ethos hasn’t been excessively promulgated by the hinting requirements established in the early days of digital, with low-res rendering.
  • Given that in the bearing, even for the Roman, there are different solutions, and despite the interesting observations I read in this thread, since I am at a very elementary level, many things continue to escape me.
    I understand that the practice is to measure the bearing more or less in the middle of the stem. It also seems to me that for some it is good to "center" /H/, while the centering of /O/, very common practice for Roman, is not followed by many.
    But, what then are the basic criteria for Italic, since in some fonts the location of the glyphs is very "shifted" to the left with a large amount of negative offset, while in others the location of the glyphs (obviously this will not be the case for the italic /f/) is much more centered in general?
  • Nick ShinnNick Shinn Posts: 2,131
    I think it’s best to work out basic criteria for yourself, in two ways.

    Firstly, by examining how others have done it, by opening up their fonts and seeing if there are any general principles that emerge.

    Secondly, by prototyping of your fonts in a variety of documents. A useful method here is to designate a benchmark font(s) that is well established, and see how your design performs in a rendering comparison of the same text and layout.

    For instance, when I was developing the Pratt fonts for a Globe and Mail newspaper redesign, they were using Utopia; therefore, my design had to have at least the same character count and apparent size as that, &c., &c.
  • John HudsonJohn Hudson Posts: 2,955
    I understand that the practice is to measure the bearing more or less in the middle of the stem.
    If you want spacing to be consistent across all the weights and styles of a family—such that e.g. you can set a word in italics between words in roman and it will be optically positioned evenly between them—then you need to establish a standard height at which bearing measurements are taken. In a bicameral script such as Latin, this height should be somewhere around half the x-height.

  • @mauro sacchetto
    Bearing for Roman is relatively simple, or at least in theory it is based on clear principles.
    However, several classic serif faces have notably wider sidebearings on /I. 

    Standardization may seem like a good idea, but I have to wonder whether this “cut and paste” ethos hasn’t been excessively promulgated by the hinting requirements established in the early days of digital, with low-res rendering.
    Not just classic proportioned ones: I always space in a specific way many glyphs, /I but also in some cases /N and /H, and /i, /n etc. Looking at well designed examples is precious.
  • Given that in the bearing, even for the Roman, there are different solutions, and despite the interesting observations I read in this thread, since I am at a very elementary level, many things continue to escape me.
    I understand that the practice is to measure the bearing more or less in the middle of the stem. It also seems to me that for some it is good to "center" /H/, while the centering of /O/, very common practice for Roman, is not followed by many.
    But, what then are the basic criteria for Italic, since in some fonts the location of the glyphs is very "shifted" to the left with a large amount of negative offset, while in others the location of the glyphs (obviously this will not be the case for the italic /f/) is much more centered in general?
    I second Nick's suggestions. Also: go by eye, train your eye. That is priceless, requires time but in the end helps you to judge in varied situations.
  • I understand that here the art overcomes the technique...
    The bearing difference then strongly depends on the font style. The MinionPro has consistent lines and graces, and consequently the bearing as well.
    The GaramondPremierPro is irregular, somewhat following the idea of ​​calligraphy and nib markings, and features entirely different designs and sizes.
    Even the width of the stems in some fonts is absolutely identical, in some others it has a certain variety.
    In any case, your advice is worth studying the best fonts ...

Sign In or Register to comment.