Asymmetric spacing of italic sidebearings

MatteoBologna

I'm working on an old style italic and I was checking how Adobe does the spacing.
I noticed that there is a difference of around 50-60 units more on the left sidebearing compared to the right sidebearing of each glyph on most of their fonts.
Can somebody tell me what is the rationale behind that.

Craig Eliason

There's some good discussion at this old Typophile thread that might be helpful:
http://www.typophile.com/node/14966

Scott-Martin Kosofsky

This is how it was done traditionally, in metal, and how it has to be done still, especially with oldstyle italics. Here's why: In these designs, the right-most elements of the capitals, the arms of V, W, Y, and T were made to overhang the body of the type, fitting over the left sidebearing of the following letters. (The overhanging elements were called "kerns.") This allowed for a reasonable fit to x-height lowercase letters, such as "a," "c," "o," etc. While those caps would have fit perfectly well to those lowercase letters without extra left sidebearings, you certainly needed more left sidebearing to fit those caps to "b,""h," "l," etc. The purpose, then, of the extra left sidebearing on the round-sided x-height letters was not to fit to the caps, but to balance with the other lowercase letters. In metal, one never fit the caps first and then the lowercase, but rather regarded the cap-to-cap fit as a matter to be adjusted visually, with thin copper or brass spaces. They could have adjusted the left sidebearings of the caps to achieve better cap-to-cap fit, but that would have resulted in too many mould adjustments. In systems such as metal Monotype, one could sometimes play with the row arrangements to create better cap-to-cap fit.

This is not unique to italics. In roman type, the lowercase stem letters—"b," "h," "n," etc.--are not visually equal, so the left side requires a bit more sidebearing than the right. This has to be balanced through the font. There are, however, no kerns on the caps in roman letters.

The old guys were extraordinarily disciplined and didn't have kerning tables to fix their sloppiness. Yet, if you look at the fit of Jenson's Eusebius type or Bauer's metal Futura (or any number of others), you'll see how perfect spacing could be achieved by clever handling of sidebearings. There's no reason to do anything differently in digital type--it's the same set of issues, even though we now have the habit of kerning the "problem" combinations, such as "Wo" and To." You can't achieve good fit with kerning alone.

If you're looking for more information, try Walter Tracy's book "Letters of Credit."

MatteoBologna

Thanks guys for your great help. Wish me luck.

Max Phillips

Good luck, Matteo. At least you're using an editor that does cells w/ slanted sides.

MatteoBologna

Max, is there any other application to design fonts besides Glyphs?

Max Phillips

Max, is there any other application to design fonts besides Glyphs?

Nah.