Vainilla – Something Old-school

2»

Comments

  • OK, now that I'm on the right page...

    I agree about the arch of /n; the older, more symmetrical form is better. I might go for a more symmetrical shape yet. Either way, there's the blobbiness that you mention. It's hard to get a good transition from straight to curve. This may be why Slimbach uses points at... not sure how to phrase this... the corner positions of arcs? nw, ne, sw, se of a circle. But yes, a real blobitude. What's your solution?

    Quoth moi: "Hmm... I'm afraid you've lost me about /O /o diamonds and figs, but I know your insights far outstrip mine. I just see curves. "

    What should I be looking at?

    And agreed again on the /g (I'm insecure about all my /g shapes). I'll try reshaping the old tail as you suggest (and that of /y, since it's similar) -- I'll bet it'll work great. ty





  • OK, here's some shots of /g in progress. I don't think the last one (with the rightward curve) is quite right, but it's my favorite by far.

    Probably just need to start the curve higher. And make the bottom left half of the tail darker.
  • Christian ThalmannChristian Thalmann Posts: 1,938
    edited September 2015
    To avoid blobbitude, I would suggest using as few on-curve points as possible (the extrema alone will often do it), using the "Fit Curve" functionality of Glyphs as a good starting point for a smooth curve segment of the desired curvature, and give the end result a pass with the RMX Harmonizer.

    As for fig shape, I've recreated your /O in Glyphs and made a version with exaggerated "figginess" (left) to show what I mean. The glyph to the right is my quick-and-dirty fix with the abovementioned method.

    The two latter versions are quite similar in their outer boundary, but their counters are different. See how mine has a more continuous elliptical shape whereas yours is more diamond-shaped, with high- and low-curvature sections alternating?

    As for the /g, I agree that the last one is the most attractive. If you make the excursion to the right more subtle, it will probably fit into the typeface better. The other two versions still look like they want to fall over to the right, but the right version is stable.
  • Oh! I do see it! I think that moment of illumination is my favorite part of, well, everything. :)

    My method necessarily differs from yours -- no Mac, no Glyphs, no RMX Harmonizer, no money at all. Just TypeTool and FontForge, not even scripting capability. Circles with the minima of points and orthagonal handles I can do -- both of my apps let you draw one. What I do with /O and /o (and hence /C, /G and even /D) is shrink the circle until it's an ellipse of the right side, then rotate it counter-clockwise. There's always an extra point, so I delete that half of the part, then copy, rotate and join what's left. Lot of guesswork. "I am endeavoring, ma'am, to construct a mnemonic memory circuit using stone knives and bearskins." Still, now that I know what to look for, the results will be much better.

    As always, my thanks and gratitude.

  • Still learning, but big differences.

    vs.pdf 32.7K
  • Evie S.Evie S. Posts: 74
    edited September 2015
    Nothing
  • Hi, Evan. Could you elaborate, please? Nothing right? Nothing wrong?
  • Whoops. Had something that I wrongly wrote
  • Bold series completely redone. I had to learn the hard way that automagic emboldening blows, that it's; much better to do it yourself manually. The bold italic /w is definitely too dark, and the italic /w shapes are probably inappropriate.

    Some quirks:
    • All forms of /G are /C + /j (or /J) and /hyphen
    • /AE and /OE are transparently rammed together; if it's good enough for Zapf, it's good enough for me.
    • Similarly, ampersands are ligatures of /e + /t
    • Italic /ae is constructed from /e rotated and joined to itself
    • Tabular numbers are absolutely consistent with each other and math symbols across the entire family. I really do mean tabular.
    • Fractions are composed from smallified versions of the tabular numbers, and are also consistent across italic and bold variants. Totally tabular, dude.
    • Ligatures with /f /t /T are composed via glyph substitution. Sans designs make this trivial.
    • /S_T ligature exists in the roman forms. Maybe George Lucas will like it!
    At this point, it's practically a completely different font... so much has been redesigned.

  • Evie S.Evie S. Posts: 74
    edited September 2015
    Auto-bold blows, and is usually somewhat incompatible. Not to get too off-topic, but once someone did it and interpolating just was completely off.

    /E/F/H/I/J/K/L/N/P/R/T/U thick strokes too thin.
    /G is very quaint.
    A lot of stuff off in the lowercase with the thick strokes being too thin.
  • Hi, Evan. Agreed on the /G; probably best to make it an alternate. One day I'll have to make a font of all the quaint shapes I like.

    "thick strokes too thin" -- you're right, but wow, was it hard to let them get that dark. They look bloated already! :# Prolly just my obsession with weight loss. OK, thanks for the insights, and back to the specimen books!
  • Working on alternates. Still psyching myself up for the weight gain.

    I'll probably keep the second /G and neither of the other alternates. Hmm, that first /question has a weird terminal.
  • Still psyching myself up for the weight gain.

    Or declare that master a "Light" and make the heavy parts of round strokes (and some diagonals) ligher, rather than thickening the stems. The /M is the most extreme case.

    On /G: Definitely move the tailed version to a stylistic set... among the other two, I think I might slightly prefer the left one. In previous projects, I've found the /G often profits from being wider than /C; maybe that could also apply here? I would also suggest looking for other solutions for the top curve of /C/G; I'm not sold on the current one. Maybe raise the end a bit?

    The two /g designs could both work. Maybe try a version of the single-storey design with an almost vertical tail root? The two-storey version has a curiously shorter descender than the other one. The tail could use some refinement.

  • Or declare that master a "Light" and make the heavy parts of round strokes (and some diagonals) ligher, rather than thickening the stems.

    That seems sound. Checking out stem width to x-height ratios tells me that I have a light and a regular -- at least according to Bigelow & Holmes. Oddly enough, my "bold" has a 22.1% ratio, exactly that of Lucida Basic Normal (400Norm). Still looks bloated and obese to me, though!

  • Halfway is good.
  • Maybe ready for one point uh-oh?
  • Christian ThalmannChristian Thalmann Posts: 1,938
    edited September 2015
    Maybe ready for one point uh-oh?

    Most of the previously mentioned problems are still around. Weights are inconsistent (compare /f and /g, for instance, or the numbers to the letters), rounds are still clumpy (/O, for instance), and some shapes are just outrageously strange (/?, or the superheavy /M).

    Your serif project is in better shape, IMHO. I'd suggest focusing on them until you've trained your eyes to see these things. Sanses are much more sensitive to flaws. I also tried my hands at a sans when I was new on the forums ("Eau de Garamond", on Typophile), but failed due to lack of experience. I've only just started with one again now, some two years later.
  • Much better. Yes, I should do the serif project(s), but this one is eating at me.


    With a few alternates. Notes:
    • Methinketh the first (regular) /g should have its bowl reduced, perhaps by bringing its bottom up a bit, and that the second (alternate) needs a bit (ok, a lot) of work.
    • The first /question is made along a Minion model: basically, reversed and extended /s; the second was made by expanding a stroke.
    • I can’t decide which form of /ampersand I prefer.
    • I definitely prefer the alternates for /I, /G, and especially /Q. but normalcy quite literally rules.

  • Hmm, maybe /N should be narrower than /M. D’oh!
Sign In or Register to comment.