Open twhb opened 2 years ago
Mimicking humanist sans-serif is considered out of scope.
I’m not 100% clear on what it means for a font to be humanist, but here are a few that seem to me not to be, and also aren’t on Wikipedia’s list of humanist fonts.
This issue is stale because it has been open 60 days with no activity. Remove stale label or comment or this will be closed in 15 days.
Built Iosevka with [.widths.normal] "shape = 610", and [.metric-override] "cap", "ascender", "xHeight" and "parenSize" scaled down by 0.90 (10% shorter glyph size) to make glyph size close to Source Code Pro. Font size 12 for example:
And, yeah, in this case I would agree with @twhb - more open "c" would be good, but not as open, as @twhb pictured. Just a little bit. MsPaint "simulation":
This issue is stale because it has been open 60 days with no activity. Remove stale label or comment or this will be closed in 15 days.
I'm still interested in a more open c
.
I support more open 5
and c
👍
Some as yet unmentioned monospaced fonts with way more open c
: Ubuntu, Victor, Overpass
I don't think Ubuntu Mono's c
looks very good though.
My concern: Having a more open aperture (or, "separation between strokes") is a feature of humanist sans-serif. In general, Iosevka's geometric style is still Neo-Grotesque (and somehow having DIN features, like vertical lines at the sides of O), therefore introducing such a variant may look inconsistent with other letters & variants.
But if it worsens the readability of glyphs... Perhaps there is a compromise solution? What criteria are used to define the size of the separation? Just wonder.
I don't know if this carries any weight, but some fonts considered neo-grotesque but with especially open c
s nonetheless:
OCR-B is also considered neo-grotesque and it has an aggressively open 5: The tail doesn't even curl up at all, but almost the opposite -- it is almost pointing down, like it wants to avoid looking like a 6 at all costs. I like it a lot.
I think the problem with Iosevka here is that it attempts to make the curved corners in the loop of the 5 almost symmetric (left side being a mirror of the right side), which results in an excessively closed angle where the curve touches the vertical stroke on the left, and a tail that curls up too much. On the other hand, the lowercase "r" does a great job with its hook, which is asymmetric (right side has a sharper curvature than the left side) but doesn't clash with the overall font aesthetics. (I think the hook of the "r" dips down just the right amount, while the "c" dips down / curls up too much.)
Did a little experiment on using completely flat hooks on either end, kinda copying FairfaxHD and a few other pixel-based fonts
I suppose it would be a compromise between staying geometric and having an open aperture, though the results looks kinda meh:
Having a flat top but a curled bottom makes the shape look more like a G, while the others look sorta fine in Heavy but gradually looking out-of-place in the thinner weights (especially Classical + Flat).
The Flat+Flat combination looks like it would go along with variants like #1269, but it also feels like a completely different style at that point.
I'll just leave this here for visualization/reference; I don't strongly recommend these takes.
Flat+flat might look better wit a larger curvature radius on the other side, but that might make the c not fit in the style of other characters. Any chance of making the c opening slightly more open, not even changing the shape, just "shaving" a few pixels off each side?
Like this:
I just copy-pasted a bunch of Iosevka c
s on an image editor and erased one pixel at a time.
0
is the original; too closed in my opinion.
-1
removes one pixel from each end (keeping the ends horizontal).
-2
removes a second pixel, and so on.
(1 px = 1/30 of the character width, at the font size I used.)
I think something between -1
and -2
(or maybe even up to -3
) provides more readability than the original (it "looks more like a c", or that was my impression), without clashing too much with the overall font style.
afaik (and to answer @susinodan's question) c
hooks are controlled by the parameter Hook
(y-depth of normal-sized hook shapes), if you know how to build with metric overrides, you can try shrinking that to see if it works. No guarantee that other things won't break though
Here's what happens if I replace it with SHook
(depth of hooks in s
, etc):
Which does look more open, but not without its own issue: This might be the same problem shared by some Epsilon-like glyphs though. idk.
Issue aside, I think it looks much better that way! What do others think? Also, does this apply to 5 as well? (I'm particularly interested in that one and not as much in the c, because I think 5 kinda looks like a 6 if you're not paying attention)
Hook
affects the bottom hook of 5. The upper one is controlled by AHook
instead.
Here's how it looks like if I only change Hook
:
The same as c
probably applies to the letter e
, which is even more closed (especially bold)
By the way, if mirror the current e
, there will be a +- more open version of c
.
Just to add my 2 cents.
As mentioned by @vladkryv, the legibility of e
(and also of s
IMHO) is much more problematic than c
or 5
. Especially on bold, smaller font sizes, lower resolution screens, due to the shape itself. e
almost looks like a short θ
(theta) on a low-res screen.
A more open e
and s
would be excellent, for those of us with less than perfect screens / eyesight.
P.S. Thank you very much @be5invis (and contributors) for this amazing font!
There might be two sub-variants:
May need to be careful about the GID impact -- we have over 40K glyphs now again.
@be5invis thank you for your work. I'd like to ask a couple of clarifying questions about the proposed sub-variants:
Would these metric overrides affect only the characters e
, c
, and 5
, or would they be applied to all glyphs in the font?
Regarding completely removing the hooks:
e
appears without a hook in many typefaces, so that seems normal5
would look, and especially the c
, without hooks...
- Would these metric overrides affect only the characters
e
,c
, and5
, or would they be applied to all glyphs in the font?
It will affect other glyphs (unless something is changed), like 2369
for example. Like I mentioned in my previous comment, no guarantee that other things won't break, but maybe you can raise an issue if something did. (We can't really make sure that all param changes work properly either because they are not really constrained)
Regarding completely removing the hooks:
- The
e
appears without a hook in many typefaces, so that seems normal- However, I'm not sure how the
5
would look, and especially thec
, without hooks...
See this for a very rough demonstration for c
(probably needs refinement)
But otherwise, here are some other fonts without hooks as reference: FairfaxHD (the one i tried to replicate):
(this one removes ALL hooks)
Having the bottom left of the “5” curl less upward makes it more distinct from a “6”, especially at small sizes, while still being recognizable. Similarly, having “c” a bit less closed makes it more distinct from “o” while still being recognizable. A more open variant is already available for “g” (and appreciated), and a somewhat similarly edited variant for “t”.