Millennial Oldstyle: crit welcome

eliason's picture

Going to try my hand at a typeface suitable for extended text. Morris Fuller Benton is my mentor.


I'm looking to Century Oldstyle for general guidance on structure and proportion--it's a great model, in my opinion--but I'm drawing this from scratch and departing from the prototype freely in parts.
I envision eventually creating a broad palette of size-specific cuts and weights.

AttachmentSize
MillenialOldstyle01.pdf73.47 KB
MillenialOldstyle03.pdf87.64 KB
MillennialOldstyle04.pdf95.52 KB
MillennialOldstyle06.pdf112.12 KB
MillennialOldstyle12.pdf115.77 KB
leadingtest.pdf44.36 KB
MillennialOldstyle18.pdf122.18 KB
MillennialOldstyle20-1.pdf115.74 KB
MillennialOldstyle20-2.pdf120.47 KB
MillennialOldstyle20-3.pdf104.3 KB
MillennialOldstyle20-4.pdf106.31 KB
MillennialOldstyle20-5.pdf101.75 KB
MillennialOldstyle20-6.pdf118.41 KB
MillennialOldstyle29.pdf225.07 KB
hrant's picture

Second-longest Typophile thread ever. Really the longest, since the other one is non-typographic. And it's not just about quantity, since Craig is deriving real benefits. And he's not the only one, in this thread and others like it, as attested at conferences, "best-of" lists, etc. All the snide anti-Typophile naysayers can eat my finial.

As you were.

hhp

LexLuengas's picture

Just stopping by to say that /a/ is finally looking better than before. Also, I like the first version of that /g/. Looks natural. The other version doesn't seem to want to “embrace” the next letter.

Hope that I soon find the time to give a more thorough critique on this.

brianskywalker's picture

Hmm? Just sharpen the inner piece of the teardrop?

eliason's picture

"Just sharpen the inner piece of the teardrop?"
Yes, doing so can look pretty nice but I think it winds up looking too "swooshy" for this typeface. I also put together a barbed top that looks like the top terminal of /C/, but that too didn't feel fitting.

Catharsis's picture

The quotes look too small to me, at least at larger text sizes. And could there be more kerning between baseline punctuation and high quotes?

eliason's picture

Here's a closer look at the quotation marks (commas, apostrophes, etc.). I recently made them much curlier which improved them vastly, but I think you're right they could be a little bigger yet.


There's no kerning at all yet, which explains the punctuation spacing issues.

hrant's picture

BTW, I like making my single quotes bigger than each quote in the double quotes.

hhp

eliason's picture

Here are the quotation marks, commas, periods enlarged a bit. More space between the double quotes too. And I tightened the sidebearings, though obviously I'll still have to rely on kerning to perfect the spacing.

Catharsis's picture

Still on the small side for my taste... but maybe that's just me?

I don't think they need to be fatter, I'd just make the tails longer (by about 50%).

eliason's picture

Heh, I had a feeling you'd say that.
Okay, here's still bigger, and allowing more weight into the beginning of the tail for the regular and bold masters). Doubles have yet more space between, too.


It's funny that I'm somehow only able to enlarge these in small increments. With difficulty I'm pushing myself over the conceptual hump that a comma is kinda like a period in scale, when it's really almost more like a lowercase letter.

eliason's picture

Latest pdf:
http://f.cl.ly/items/1q3F192H1J3q0y3G3E0z/Millennial%20specimen%20146.pdf
Lots of tweaks since last time, only some of which I can remember
- loosened spacing of italics
- reworked figures quite a bit, including darkening osf's
- added alternate one-story italic g
- still more work on size and positioning of all quotation marks (and comma)
- developed unicase version further
- dropped ascending p alternate, and quaint ligatures
- made ellipsis much wider
- tweaked serifs of CEFGLSTZ

Sindre's picture

It is not easy giving critique at this stage, but I've had a look at the latest pdf, and here are some thoughts on the base design. I have not had time to look closely at all the weights, I'll get back to it.

I have not read all the posts here, so I will inevitably address things that are already well-addressed. Feel free to ignore any and all comments.

Roman

I'm wondering what the rationale for the sharp inside points of 'f', 'a', 'n' et al is? Is it too subtle to be a functional device, and shouldn't we also see them in the numerals and (more of) the upper case? This also applies to the very short straight line to the west of 'e'.

Is 'a' (very) slightly wide?

Is 'u' slightly tall?

Are 'w' and 'y' a little light and dark, respectively?

Why do 'i', 'l' 'h' et al have sanded-off right-tops?

Why are 's' serifs so dissimilar?

Is the right part of 'V's right serif too short?

Should '&'s arm be straighter?

'five.lining's serif is very heavy, and the top of the curve much lighter than its OSF sibling's. Also, the characters of the two fours are maybe too different for one typeface?

Italic

'a' a little tight at the top junction? Maybe very slightly too upright?

'z' lacks heft, I think.

Something a little stiff or unfluent about the 'u' shape I can't quite put my finger on. Italic 'u's are always difficult.

Unsure about the 'Z' crossbar, also because of uni01B5.

I really like the heavy weights, I especially dig the f/j treatment. But more on that later.

eliason's picture

Thanks a million for the feedback Sindre!

On the roman comments:

I'm wondering what the rationale for the sharp inside points of 'f', 'a', 'n' et al is? Is it too subtle to be a functional device, and shouldn't we also see them in the numerals and (more of) the upper case? This also applies to the very short straight line to the west of 'e'.

I can tell you the thinking, though I don't know that it serves as justification. At some point in the development I had the idea of cleaning up all the counters so that they had a simple geometry based on symmetry (so, for example, /b/'s counter is an oval rather that being straight on the stemside). I was particularly intrigued with what this meant for the /n/ (and other humped letters), which if there was to be a corner at the join, there would be a corner opposite, too. But the outer contour would simply arch around smoothly, and the mass as a whole would read as a pretty conventional arch--the corner in the arch would be "subliminal"


From there I decided to see how far the symmetrical-counter idea could be stretched. I wrestled it into the /g/, which I like. It was already there in letters like /e/. For the closed counter of /a/, the symmetry could divide the drop-shaped counter diagonally. The /K/'s counters could be up-down mirror images of each other, at least near the middle of the glyph.
From the /n/ I redrew the /r/ to have the same symmetrical arch underneath. I tried to do the same with /a/ and with /f/, but I probably hit the point of diminishing returns somewhere in there. The /f/ in particular feels like an outlier now. (The conceit is most obvious in "branching" structures which explains why it's less evident in figures and caps.)

Is 'a' (very) slightly wide?

I think it was, particularly in the display master. How's this?:

Is 'u' slightly tall?

It's the same as the similar stem tops in /n/m/r/etc., which raises the question whether they too are too tall. I have now lowered the top point of those stems a few points, and also taken out the bulge of that long segment at the the top. What do you think?

Are 'w' and 'y' a little light and dark, respectively? Is the right part of 'V's right serif too short?

Yes on all counts. How's this:

Why do 'i', 'l' 'h' et al have sanded-off right-tops?

I think I liked the idea of the top angle being a right angle, rather than acute. Almost like a mirror of baseline serifs, though bent over a bit. Does it bug you?

Why are 's' serifs so dissimilar?

I think that detail reflects my efforts to have the face be both classic and fresh. The full barb at the top feels more traditional, the cleaner solution at the bottom more modern, and I think maybe having them both be one or the other carries the letter too far over to that side.

Should '&'s arm be straighter?

What do you think of this?:

five.lining's serif is very heavy, and the top of the curve much lighter than its OSF sibling's.

Agreed on the curves; by the serif do you mean the teardrop at the bottom, or the one at upper right?

Also, the characters of the two fours are maybe too different for one typeface?

Hmm, maybe. If you were to make them more similar, which structure would you work with?

Responses to your italic comments later. Again, I really appreciate the careful look!

eliason's picture

Okay, this image shows some italic fixes based on Sindre's feedback:


- /a/ junctions are reworked a bit and counters are rotated very slightly
- /z/ is darker (and leans a little less than it did before)
- /u/ I didn't do anything with yet. I don't mind a little stiffness in this face, though I think it's a pretty homely letter on its own now. It's actually pretty directly a rotated /n/. And both are, again, based on a symmetrical counter (there's a canted, symmetrical, sort of parabolic arc in the counter). I'll probably try some more tweaking. You're absolutely right that italic /u/s are difficult!
- /Z/ in Zap shows the same glyph with the crossbar removed. I think at least in the higher contrast masters (#2 bold and #3 display) that long skinny diagonal gets awfully long without the relieving weight of the crossbar. So if just removing it doesn't work, one option would be giving up on the thin-diagonal structure (i.e. adopting the thick-diagonal structure of the roman). But I like the challenge of trying to get the thick horizontals to work :-)

Sindre's picture

The symmetrical counter idea is something I've wrestled with too, but I've so far always given it up. That doesn't mean it doesn't work here, but maybe it does the 'f' no favours.

New 'a' width looks just right. But maybe the teardrop in the top example is slightly light?

About the 'u' height: I always end up with the 'u' top serifs slightly different from the 'inmh' top serifs, the left one a little shorter (and therefore a little steeper), the right one a little longer (to close the countershape a little). This approach might make it appear slightly less tall, although it obeys the x height. Just thinking out aloud here. Your new treatment looks better, though.

How's this:

Just right.

Does it bug you?

I'm not sure, but maybe it's a bit strange that the "sanded" stems are (optically) similar to the stems in 'm' and 'n', that are compensated for the tight junction.

Very good new ampersand! Much more in character, I think.

I'd strongly prefer the closed four for this typeface. I really like the open shape, but it is more Didot in character, making it a bit anachronistic in this environment, maybe.

eliason's picture

On the /f/, just to throw this out there, this


is what the regular and display masters of the bold /f/ look like. (They are used to generate bold and black interpolations, but they themselves don't appear in the lighter weights, currently.) There's something about them that I kind of like, and I wonder if I could get away with using them (or something like them) for those lighter weights after all. Or does the hook look too miserly? I rather like the idea (again, the challenge) of making a workable /f/ that doesn't make ligatures necessary or ensuing diacriticals problematic.

eliason's picture

On italic /Z/:


After all this I'm thinking for the capital, simple is better and the boring contrast pattern of #3 is the way to go. I guess I'm taking the long way to figure out why so many existing fonts follow that differing pattern for /Z/ and /z/.

I guess at least I can recycle that barred form for uni01B5! :-)

eliason's picture

Here's a question for you: must a small-cap Z match the contrast pattern of the full capital?

Catharsis's picture

They look a bit weird to me in juxtaposition — I'd expect small caps to follow capital logic. Also, I liked the crossbar in {Z}, but then again, I also like crossed sevens. ;o)

LexLuengas's picture

I second Christian. There is no reason not to transfer the capital treatment of /Z/ to the small-caps. I don't think the small-caps should relate to the lowercase.

Bold /G/ has a very upward-pointing upper serif which could be pushed a bit more to the right.

eliason's picture

Yes, you're right of course, about both Z and G.
Here's the latest pdf:
http://cl.ly/2p0V3W0R432a
In addition to fixing small-cap Z and G, changes include:
- spaced out double quotation marks
- tweaked italic /u/ and /n/ a little
- tweaked lining /3/
- trying out the single, narrow /f/ (redrawn a bit) for all weights instead of subbing it in for bold only (in both roman and italic)
- trying out closed /4/s across the board
- in the italic, I had a looping structure in the middle of B, R, and K. I've now made a simpler more conventional structure. The looping one will only appear in the display cuts (and not in any small caps).
- settled on italic /Z/ with a thick diagonal. A thin-diagonal one with a crossbar will appear with the SWSH feature.
- probably other things I can't remember

Birdseeding's picture

This is just out of curiosity: Why are the hairlines thicker in display light than in all the other display weights?

The only glyphs that at at a causal glance leap out as strange are the italic J and j and small cap J, which is a pity because I like the idea of the structure. They seem to balance very heavily towards the rather straight stem, which has almost no continuation into a tail that in turn is comparatively weedy... Might just a slightly more shallow curve there or a slightly bigger ball terminal make it more balanced?

eliason's picture

I thought the display light weight was more persuasive if it was almost monoline, which is why I thickened the hairlines a bit. I should probably generate a test instance with thin hairlines now to verify that early decision was a sound one.

Here's a bit of tweaking of italic /j/ (#0 is before, #1 is after).

Catharsis's picture

I think I prefer 0 for the regular/display, and 1 for the Black. They have more backbone.

eliason's picture

Here (#2) is a middle ground between 0 and 1 that I think works.

Syndicate content Syndicate content