« Box o' Crack | Main | Archives | Passport Control »

Settling on Proportions

We had an informal group crit a couple of weeks ago, and it was good to get feedback from more people after so much time alone with my typeface. It was no surprise that there are still a bunch of glyphs I need to finesse, but overall the comments made it clear that I needed to take a good, hard look at the overall proportions before I got locked in any further. I had played with the ratio of the x-height to the ascenders and descenders a bit in the earlier stages of development, but it was hard to evaluate the differences when I was just making manual changes to a limited character set.

Since I’m mostly happy with the overall design of the alphabet — and since other people seem to agree that the proportions aren't quite right yet — I decided it was time to conquer my fear of multiple masters and run some tests. Luckily, setting up the masters was as easy as I’d hoped (although there's probably a more efficient way to do it than I did). I took the most recent version of the font for Master 1, and then made a copy with a few adjustments for Master 2:

  1. I increased the overall size by 110%

  2. I dropped the ascenders, descenders, and caps back down to their original heights.

  3. I decreased the overall size by 90.91% to restore the original x-height and horizontal metrics, but with short ascenders, descenders, and caps.

I combined the two versions into a multiple master with an optical scaling axis and started generating some instances for comparison, extrapolating a bit past the original masters just for good measure. Here are a few sample glyphs to show the range (highlighting the ones I’m leaning toward at the moment):

Overall comparison

Text was still easy enough to read at the far end of the scale, with the very short descenders and caps, but I think it would be too challenging for lengthy book text. Also, assuming that mathematical and technical texts will mix a lot of material in with typical text, a little bit more detail above the x-height ought to make it easier to pick out the symbols and formulas mixed in with the words. Also, I’d still like to leave a little room for the ascenders and descenders to contract for small optical sizes.

I noticed looking at the tests there seems to be a point at which individual letters start to become tricky to recognize on their own. They might still combine into recognizable word shapes, but as stand-alone symbols, they could be a little tricky to identify readily. At a certain point the extenders become so out of scale with the x-height counters that they look more like awkward stems than anything else. I think the sweet spot falls between the original and my test:


Around this point the extenders hold their own and the caps stop looking quite so condensed at they did in the original. I think an instance with slightly taller ascenders will be the best point from which to move forward. Once I pick the right one, I'll need to fix the curves and details and scale the whole set up to fit the em square better. I want to get these vertical proportions settled before I move onto the figures and italics, though, since those will be harder to scale.

« Box o' Crack | Main | Archives | Passport Control »
Powered by Movable Type 5.2.13Creative Commons License