Tesserae — an experiment in modular variable fonts



  • I think the original /el-cy looks fine. It's the tucked-in tails of /de-cy, /che-cy and /shcha-cy that irk me. Try giving them their own column.

    Oh, and the huge breve on /iishort-cy looks like a cap on an African military dictator.
  • Christian ThalmannChristian Thalmann Posts: 1,772
    edited August 2017
    Hi Samuil,

    I've tried out a few implementations of your suggestion, but I'm not too happy with them. In particular, the typeface was able to do well without any kerning at all, and now I'll have to start kerning /De-cy etc with other things...

    Is the previous /El-cy merely usable or actually preferable over the more sloped newer versions?

    The problem with shortening the breve on /iishort-cy is that it can then no longer be centered on the base glyph. Would you still prefer this...?

  • I'm trying to eliminate that almost-collision you can see in a few /e in the last line of the previous image (e.g. in «alphabet»). The following solution is cleaner, albeit a bit less e-like... probably preferable, all in all.

  • David SudweeksDavid Sudweeks Posts: 29
    edited August 2017
  • The rightmost /De-cy, but with shorter teeth (move the topmost quarter-circles one step down and get rid of squares) looks to me like the real /De-cy the most.

    However, I've tried the same treatment with /che-cy and /tse-cy, and their teeth (tails?) look limp. They look much better when row 0 is square and row -1 is flipped so that the sharp angle points inwards.

    But this approach applied to/De-cy makes it look absolutely bloodthirsty. I guess squares in row 0 and sectors turned outwards in row 1 are the only compromise that can work for both.

    The misaligned breve looks fine, to my surprise. I think even a single square over the middle column will look fine.

    Could you try setting some running Cyrillic text in the latest edition of Tesserae? This would help a lot to catch any shapes that don't work that well.
  • Christian ThalmannChristian Thalmann Posts: 1,772
    edited August 2017
    Like this? I'm not too fond of the huge amount of whitespace the low sides leave. I also don't like the square top edge to the sides much, do the rounded edges work?

    Would the taller versions (2nd in row) still work with some good will from the reader? I've also made quite a few concessions in the Latin alphabet to ensure they all look as blocky as possible.

    I actually don't mind the bloodthirsty teeth. :grimace: But I do prefer the outward-curved ones.

    As for LC, wouldn't it be viable to point the descenders outward for /de-cy but inward for the one-sided ones?

    Running text:

  • @Christian Thalmann That's surprisingly legible! Even /de-cy is legible, to my great surprise. I guess the surrounding letters support it well enough for the brain to make the correct guess. I guess this answers your question whether taller teeth will work or not.

    However, I've stumbled over /pe-cy and /tse-cy. I have to idea how to make the former more recognisable (beyond flipping it and /en-cy around).

    The latter will probably need that aggressive tail to be legible. Or maybe it's my own idiosyncrasy and everyone else gets /tse-cy without any problems. I would wait for another opinion or two before roughing that tail up.

    Could you please set this text in all caps as well?
  • Flavor of the day: A tip of the hat to Albers. :grimace:

  • Horizontal is perhaps even a better. There's a lot of optical illusions in the vertical approach that make the stroke width seem inconsistent.

    Samuil: What do you think of these two options for /en-cy? I've chosen a very radical approach for /pe-cy, but I suspect it might work out...

    I don't expect all-caps to work well; it doesn't do so in Latin either. Much too similar shapes.
  • Actually, /pe-cy is fine now. I don't like either /en-cy, especially the second one. It interferes with by brain's language detection. The first one is too skeletal and jumps out of the line because of that. I think it will be fine if /pe-cy and /en-cy have dissimilar widths and the original /en-cy is used.
  • Christian ThalmannChristian Thalmann Posts: 1,772
    edited February 2018
    OK, I've finally wrapped my head around how Python scripts work for Glyphs and made my first variable font out of the Tesserae genome (thanks for the help, Rainer! :grimace: ). It's based on the original block aesthetics, and it allows you to vary the size of the blocks, the rotation of each block, and the slant of the grid on which the blocks lie:

  • [like]
  • Christian ThalmannChristian Thalmann Posts: 1,772
    edited February 2018
    Some more non-variable variants. I guess they could be animated via variable font technology; e.g., the reflexes on the shiny bubbles could be made to change the angle of incidence, etc...

    EDIT: Actually, is this a better solution for the bubbles?

    EDIT: Last one for today...

  • EDIT: Last one for today...
    I guess it's a new day now! :grimace:

  • This whole exploration has been fascinating to watch. Cheers 
  • Craig EliasonCraig Eliason Posts: 1,250
    @Christian Thalmann stumbled across this thread and am wondering if these designs ever got released?
  • Christian ThalmannChristian Thalmann Posts: 1,772
    edited July 2021
    No. I think it would make sense to release them under the OFL so other people can play with them (I'm way too busy with other things right now), but I haven't gotten round to it. Maybe something for the vacation to-do list.
  • Nick CurtisNick Curtis Posts: 117
    If your aim was a typeface with all the simple molecularity of pixel fonts but little if any of the legibility, I would have to say that you have succeeded admirably.
  • My aim was to create a simple modular testbed for special effects made with Variable Fonts on the pixel level. Legibility was not an issue. (And yet its legibility does surprise me every now and then, especially when seen from afar.)
  • Nice experiment. A grid with height=5 and width=2-4 is limited. But nice to see what's possible.

    In my collection of pixel fonts the smallest one is 3x3. 8x6 was most used in the old days. 8 allows vertically 2+4+2, and 6 allows horizontally 3 black and 3 white spaces.

    With the slanted version I would try to adapt the 90 degree angle of the black boxes to the angle of the slant.
Sign In or Register to comment.