Feldspar (Nee Mixteca)

Bendy's picture

Ok, so I thought it time to come back to the crit forum after a bit of a break from here working on other designs.

Here is my new design, Mixteca Serif. You'll see in the pdf a bit of background about the design concept. Any crit would be very welcome; I'm hoping this can turn out to be really useable so tell me if there's anything you don't like or looks wonky. This has been six evenings' work, so I'm sure there's lots more to do. It reminds me a bit of Charter and a bit of FF Olsen, though I was thinking more along other lines: it actually started from playing with my first font, Eternal.

For some reason kerning didn't export with the otf file so the spacing looks a little gappy.

Look forward to hearing your opinions! :) Thanks!

Feldspar Specimen.pdf490.07 KB
Feldspar with Italics.pdf949.05 KB
riccard0's picture

I like it, but that K, with its open joining, stand out a little too much.

Bendy's picture

You know, I was attached to the idea too, but agree it may need to come out. The overall glpyh looks too wide.

riccard0's picture

It's not a bad idea per se, but I think the poor K feel a little lonely, being the only one... ;-)

eliason's picture

Even before I read your description, I was already marvelling at how well this design works at both display and text sizes. It looks great. Your proportions are rock solid.

The spacing seems tight to me but I'm growing to think that's your style, and I don't see any major misfittings to trouble it.

The large serifs on the descenders work in concept, but they emphasize the stumpiness of the descenders on /p/ and /q/ and maybe get a little tight between serif and bottom of bowl on those letters. Would you consider lengthening your descenders?

/S/ kind of reminds me of Palatino.

Cap /J/ is an unhandsome glyph to my eyes.

Is the arm of /r/ too light and/or short?

/y/'s curving stroke looks too soft I think. Perhaps (but less certainly) Q's tail join too.

Some combination of shaving weight off or slightly widening /w/ might be good.

/x/ is awfully funky - it looks somewhat swastika-ish - but may be rare enough that you could get away with it.

I love the bottom of /t/ - cool stuff!

Glad to see Eternal actually isn't over!

hrant's picture

The "K": one very smart thing Zapf has said is that you must throw away your most favorite glyph.

The vertical proportions are indeed very good.

> The spacing seems tight to me but I’m growing to think that’s your style

I would worry about style intruding into something so technically delicate, especially in a text face. I would say it's indeed too tight, considering the small target point size.

"g": Very nice. I'd make the right side of the bottom counter segmented.

"w": Why so narrow?

"y": Descender way too heavy.

"G": Left curve too dark.

"Q": Too dark.

"S": Slightly too dark.


nina's picture

Like I said on Flickr, I really really like this one. I think you've got a very promising foundation here; I think it does still need a bit of time, and a bit of agonizing over detail decisions, and the diamond file; but if you allow it to have that, I think this can become seriously great.

Salt ready? :)

First, what are "conventional proportions"? You mean horizontally or vertically?

I'm not sure about the notch action in things like the "c". Especially in the "a" that device seems to be quite the hype recently, and I wonder if it really helps or if it's just stylistically pretty.* Especially in your "c", it seems to add a bit of broad-nib pen flavor that I'm not sure matches the angularity of the glyph, or the overall tone of the font. Dunno. In any case, you might need a concept for where to put these (or I'm just not seeing it). Why doesn't the cap "C" have the notch? Or the "o"?

I'm also still not sure of the slanting half-serifs on diagonals. For one thing, the left counter of the "w" looks larger than the right.

"f": Very cool!
"g" seems a little dark.
"r" needs a more determined terminal. Look at "kerning": This "r" is prime "combine with 'n' and track in to make an instant 'm'" material. :-)
"s" looks too dark and/or needs more contrast.
"w" also looks too dark.
"x": Not sure, it is very funky. Looks a bit like an aleph. Hm. Craig, of course it's a rare letter – but OTOH, it looks like's in the font name!
"y": Counter might be too wide. Also not a fan of the soft tail.

"A": I guess that taper in the crossbar has carried over from Eternal, but why is it there? It's barely visible now.
"E": Does the middle crossbar need more weight? The one in the "A" is much heavier. Also, why not put the flares the other way round, away from the stem – kinda like in the "Z"? (Also in "F" and "L".)
"F": Slightly narrow?
"G": Seems dark.
"J": That terminal is maybe a bit overkill, no? :)
"M": Yum! :)
"R": I think this could use a less soft / more determined leg.
"T": Why no flare?
"W" and "Z" look dark.

And I think the question mark is too oldstyle.

* I think trying to balance functionality and style in a text font that also works for display is very ambitious; it might be fruitful to set clear priorities, like if you want to make this work as a text font, do that first, and then see what you'd need to change to make it also work for display? Maybe.

eliason's picture

(Nina) “E”: Does the middle crossbar need more weight? The one in the “A” is much heavier. Also, why not put the flares the other way round, away from the stem – kinda like in the “Z”?

Oh yeah, I meant to say that too: the /E/ and /F/ look too heavy on the left/light on the right.

nina's picture

What is the target point size?

hrant's picture

I'm seeing 7-8.


nina's picture

But the PDF says it's for books.
Maybe that x-height is a bit on the large side, no?

hrant's picture

Oh. Wow, you actually read the text... Hey lady,
spend enough time designing fonts, and you won't. :-)

Well, at 9 point maybe - but that's pushing it, both in terms of the font's proportions and the reading experience. To me this is much closer to being a news face.


1985's picture

What is the exact quote from Zapf, Hrant? It's profound!
What does Mixteca mean, Bendy?

Good effort, too good for me to crit.

1985's picture

Apart from the K!

Jos Buivenga's picture

Nine should be fine :) I'm still a bit puzzled by the sharp node in the counter of c.

hrant's picture

Andrew, I don't remember the source* so can't look it up.
I'm afraid you'll have to wait for me to run into it again some day...

* Hmmm, where's that online interview
from a few years back done by Linotype?


hrant's picture

It might be somewhere in this exhaustive (but pretty fun and interesting)
10-part interview: http://www.linotype.com/1494/theschoolyears.html


Bendy's picture

Oh gosh! I was out playing in the snow and missed all this! Thanks guys, I'm delighted to have such positive feedback from people with such talent :)
Right, a few comments for now.

Target size: I was thinking 8-10pt. I realise that might be quite a large range to cover and am learning more about these considerations only by listening to what you guys have to say on the matter...so if there are things I can do to make that work better then great, and if I need to be more specific then I will say let's make it the upper end of that range since it seems to take more technical experience to design really tiny faces? Although Hrant might have a point that this is already better suited for news. I'm really no expert.

Hrant, that is indeed a great quote. I was already thinking of throwing out the K and k but so far didn't due to other people suggesting it was a strong glyph. I think it was the first glyph I drew. Maybe I can keep it for something else.

The lowercase c: Right. That notch and the one in a were the start of an experiment that looks like it didn't work. Following Jos's advice on Flickr I drew the rest of the glyphs with smooth curves and hadn't come back to correct a and c; also wanted to see what feedback they would elicit. Still, I like the textured feel they bring in so need to think about how that could be maintained without the curve notches.

Spacing: wow, this is where my noneducation is seriously getting in the way! I thought I'd made it pretty loose already! You know, I was hesitating about making a separate display cut since I really like chunky rather than overcontrasty and light faces. But I've rethought this now and there's no reason a display face shouldn't keep the characteristics I like so perhaps I will after all. I like Nina's footnote and agree that approach sounds right for this face.

Nina, conventional proportions: I meant that it wasn't as condensed as many of the trendy text faces (like the ones that come out of Reading's MATD); however, on reflection it seems to be following that model quite closely.

J: Aargh, I knew this was the weakest link and it'd be obvious I hadn't decided quite what was best for it. I tried various different treatments of the terminal and couldn't settle on any particular decision. Has anyone seen any fonts with a really good J? I was looking for something with unconventional treatment, again chunky and perhaps reversed stress?

E,F: Yes, I see what you mean. Those serifs on the right need to be much bigger. May try reversing the flares too — thanks for the suggestion Nina. I did try flaring the T and it looked rubbish. Let's see.

hrant's picture

> I think it was the first glyph I drew.

Ah. That reinforces the dumping suggestion. :-(
IIRC, according to Carter a text face starts
with an "n" (meaning: something bland).

> I thought I’d made it pretty loose already!

Have you printed it?
Because a text face always seems too loose when viewed large on-screen.

BTW, I personally like the "J".


Bendy's picture

Right, that sample is 10/13pt and to me the spacing when printed looks generous. Need to develop my eye a bit it seems :)

I'll set a sample in smaller sizes and take it from there.

nina's picture

"I meant that it wasn’t as condensed as many of the trendy text faces"
Ahh. Yes. I like the boxy/not-so-narrow proportions.
Maybe they're part, though, of what makes this seem to target smaller point sizes than you wanted? Not sure.

"I did try flaring the T and it looked rubbish."
It's probably good. It was more of a gut reaction on my part upon comparing with "L" and those others. FWIW I just looked at Chaparral, which has similar flares (if gentler), and that pulls it back quite a bit in the "T" too.

Good "J"s: If I was you I'd probably start a new thread to collect those. :)
For the record, I've found those very hard to make good, too.

BTW, I still think the "k" is a great glyph, taken by itself… but yes maybe it's too lonely in there. On the other hand though, I think it's the structure that makes it so nice, not the gap alone. Meaning it'll still rock without the gap.

dtapia's picture

love it

quadibloc's picture

> The lowercase c: Right. That notch and the one in a were the start of an experiment that looks like it didn’t work.

Doesn't the lowercase g also have one of those notches?

Gary Lonergan's picture

This looks a very assured design and I would love to see a hairline version.

Bendy's picture

Right ho.

I've followed all the advice you've given me — thanks for some very good ideas. I think this is getting better. In addition to the tweaks suggested for g, k, r, s, v, w, y, A, D, E, F, G, K, Q, R, S, V, W, and Y, I've changed the descender of y and moved the serifs on Z.
The new pdf shows type at 10, 9, 8 and 7 points then various display sizes. I've put in lining numerals as well as oldstyle and I'm actually preferring the cap height ones and find the OS 5 isn't quite fitting with the rest yet.
There's also basic punctuation: I tried various styles of comma but this one looks most elegant IMO. Not sure about the questionmark.
As for other family members: yes, a hairline will look great. I'm really excited about making a lyrical but not soft italic.

hrant's picture

Good stuff.

First thing: the spacing is too tight for the target size range (although the spacing is still to uneven to be sure by how much). BTW, if your target is around 9 (instead of the 8 I'm seeing now) I might raise the ascenders slightly (especially since the descenders are still a bit long in comparison).

g: heavy at the bottom.

y: superb!

Q: much more tail please! Maybe a serif at the end.

U: buttom-right needs liposuction!

lining nums: 3 and 5 too narrow.

OS nums: 6 too wide, 8 too narrow.

asterisk: very nice.

parentheses: too dark.


nina's picture

! A sample without gradients!
SCNR. ;-)

This is looking super Ben!
Just a couple of impressions/questions:

Sorry to bring this up once more, but I'm curious – on horizontals, have you tried putting the flares the other way round (so that they get thicker on the non-stem end, like on the "Z")? Not sure it'd work, but I keep wondering.

Agree with Hrant that the "g" is dark in the tail, and the "y" is very cool. About the "g" I now wonder about the notch in the tail – it looks funky, but now that the other notches are gone, I wonder how much it still belongs in the font.

I still wonder if the "r" could be made stronger – its foot serif is almost as strong as its, uh arm?. Maybe a terminal of some sort? But maybe it's OK.

"k": I'd raise the crossbar. Maybe in the cap too.

The slanted half-serif on the top-left of the "x" feels alien to my eye.

Not sure what to think of the OSF "5". It's a bit unexpectedly old-style with that upward top serif. But maybe that can actually work quite nicely.

I like the big serifs on your nums (although the one on the lining "5" might be a bit much?), just wondering, have you tried a foot serif on the "7"? (You might have to "Romanize" its contrast for that to work, but then it might be Very Cool.)

> making a lyrical but not soft italic

That sounds… interesting :)

Bendy's picture

Glad this is meeting your approval, chaps!

Nina, the horizontals, do you mean specifically EF and L to take the flare pattern from Z and 2? Sounds like a good idea to try.

The g: I agree. I put the notch back in but not sure it worked. May remove both notches but am a bit worried it'll make the counters too small without some other solution.

Hrant: I'm wondering about the relative proportions of extenders. I thought the extenders should get longer at text sizes and shorter at display sizes, but it seems you're saying that a 9pt-optimal cut should have longer ascenders than an 8pt-optimal cut. Can you explain? And to me (and Craig I think), the descenders look quite short already, yet I think you'd like them to be somewhat shorter? Is that because of the large x-height and/or low cap height?

Here's a quick sketch of a hairline cut. Hairlines are DIFFICULT! The a, b and u look a bit wide I think, but it's fixable!

nina's picture

Hey Ben, yes I meant the E/F/L. Guess that was a bit unclear, sorry.

Bendy's picture

Actually, I've just seen you did explicitly specify the EF&L in your first post above, sorry! Will try tomorrow ;)

Bendy's picture

Here we go:

Nina, I think your idea works better than the original. Thank you ;)

nina's picture

Interesting. I wonder. I now see that the other way round was probably more intuitive vis-à-vis the serifs, but this seems to better «define» the non-stem ends of the horizontals… Could you maybe show it in text?

hrant's picture

First, something I missed before: The bottom of the "t" is out of character. Maybe try a terminal like in the "f" and "j".

> I thought the extenders should get longer
> at text sizes and shorter at display sizes

In my view extenders do something mostly functional at text sizes, but mostly aesthetic at display sizes. So for text sizes they need to vary depending on the target point size, while for display they need to vary based almost entirely on your taste!

Especially in a text face, vertical proportions, color and spacing are very strongly inter-related. If you have a certain color, that implies a certain narrow range of proportions and spacing that can result in good readability. Usually it's best to decide the target point size (since that's what users mainly choose); the other two mostly follow from that.

The smaller the target text size, the larger the x-height needs to be (which means shorter extenders)*. What I'm seeing here is around 8-9 point. If you want to lean smaller I'd make the descenders shorter; if you want to lean larger I'd make the ascenders longer. So I'm not actually recommending one over the other; the main thing is for ascenders to be sufficiently longer than descenders. But of course this is not an exact science.

* Don't worry about cap height too much here.

BTW, I'm not good enough to be able to recommend small tweaks based on a PDF - I might be able to do it if I saw some hi-quality printed output. But really, such small degrees of change start wading into aesthetics (something no font is entirely devoid of) so you might easily disagree with somebody. My own taste BTW leans towards darkish fonts.


eliason's picture

I'm not sold on the serifs on lining 5 and oldstyle 2, 3, and 5. Outside of the serifs, I think the drawing of all the figures (with exception of the lining 5) is very strong. (Lining 1 looks too tall but that may be rasterization.)

Question mark should be rethought.

I agree that /x/ is even stranger now that v and w terminals have been normalized (which needed to be done).

/g/ gets too pinched in three places: (1) where the top bowl comes up on the right near the ear, (2) where the notch remains at the lower right, and (3) at the attaching spot in the middle. (I think (1) and (2) are the "notches" you just said you're thinking of removing.)

Is /u/ a touch too wide?

The new /y/ treatment is surprising, and I think I agree with the others, ingenious! You may have to rename the font to something with a /y/ in it! :-)

Optically, to my eye neither /X/ nor /x/ looks like a true intersection but I think they work well nevertheless.

Yes, I'll be interested to see (at text sizes) if those EFL changes will require an adjustment of the angle of the serifs on the other side of the stems.

Bendy's picture

Hrant, I think I disagree about the t. Whilst it is an unrepeated feature, to my eye the square cut harmonises really well with the other baseline serifs, especially with the lc r and s. I think it also balances nicely with the angularity on the head of t. I'd say perhaps the foot could be slightly wider and perhaps darker too though. I also think I don't want to add too much interest to the baseline, and the trumpet-snout terminal (as Nina christened it) would bee too much.

One of my original premises was to add interest and emphasis only to the areas of glyphs away from the baseline. Since there's always a lot of activity at the baseline I figured the eye might find it easier to read if the less frequent parts were emphasised. Hence the flared ascenders and the weird g, j and y descenders.

Am fine-tuning the numerals and will look at that g: is it the left-pointing curve at the base that's too dark?

hrant's picture

Good point about the "t".

The "g": looking at your second PDF up top, I'd say that the contrast is good, but the bottom bowl is too dark all the way from the end of the bar to the tip.


Bendy's picture

Well well. I think the numerals are just about there. It's been pretty intense work getting them to sing together. I think they took longer than the 26 uppercase letters! Somehow the lining numerals wanted more serifs, to look more 'Roman' somehow, but I took some of them off for the oldstyle ones.

Please see the new pdf up top and let me know your comments.

eliason's picture

Drawing of five still not there IMO. It looks to me like the bottom half has been rotated ~5 degrees CCW, and the top half skewed about the same amount in the other direction.

The eights look a little doughy compared to the solidity of the other figures. Maybe the bottom 1/4 of it could somehow be firmed up?

Is is possible that lining zero is a touch too heavy in the upper right and lower left quadrants?

I really do love that lining seven with the serif!

Something I really admire about what you've done here is capture the style I associate with oldstyle figs in those figures, and the style of lining figs in those (after all, there's more of a difference than simply presence or absence of extenders), while keeping the two sets stylistically uniform as a font. Really nice work.

Bendy's picture

Craig, thanks. That's really nice to hear. When I was working on these I realised that the two sets could be 'emancipated' (as Nina would say I think), so the lining ones have more in common with the uppercase and the oldstyle ones share characteristics with the lowercase. Hence the lining ones have slightly more vertical stress as well as more serifs. I'm really pleased you think this has been successful, except for a minor tweak to a couple of glyphs. You're right about the 5. (The lining one, yes?) That's going to be hard to get right, but coming back to look at it now, looks like it needs widening and the stress rotated like you say.

eliason's picture

Yes, the lining 5.

Thanks for that LF:caps::OSF:lowercase explanation - total light bulb over my head!

Bendy's picture

Yeah, it was like that for me too. The lining ones are also slightly darker.

I wonder, would a case feature switch oldstyle to lining?

Tomi from Suomi's picture

That uppercase 5: traditionally the serif on top right is facing up.

In fact I would just lose it, and copy the top of the lowercase.

Bendy's picture

I tried flipping the serif but that didn't suit the font at all.

I might try removing it but I'd like to get it to work since the uppercase alphabet is pretty serify. I'll twiddle a few dials and see what happens.

Bendy's picture

Is a tilde allowed to do this?

I haven't found examples where the terminals are heavy.

Tomi from Suomi's picture

That would be a guestion for the Spanish-speaking. For me that is just fine, if you just get it balanced; right now it seems to try to take flight. Lower the right side a little.

hrant's picture

Allowed, shmallowed.
I like it (although the thin parts should be thinner).
You could also try a gradual thinning from the tips to
the center instead of a solid-medium-thickness middle.


Bendy's picture

New pdf up top.

/g/ has the notches taken out. I may need to make the top of the loop squarer/sharper/heavier?
E and F have crossbars with heavy side on the right.
Q has a longer tail. Too silly?
/x/ probably needs changing. X needs more weight?
Lining 4 may need a higher crossbar?
Lining 5 tweaked: are the proportions better?
I can't figure out how to make the bottom of 8 more defined.
Eszett looks rubbish.
How are questionmark and capital J?
There's some tapering on the octothorp, slashes, percent and phantom dollar sign. I like it a lot but is it too bizarre?
Are the brackets still too dark?

Bendy's picture

I've had interest in this design from a new magazine that's starting up, so getting back to speed with it. :D Feel free to critique harshly! (I'm thinking spacing and kerning might go to IKern to do.)

eliason's picture

Exciting to hear, Ben!
The number sign looks big and splayed out to my eye.
I think the hood of /f/ gets spotty - I'd shave of some weight.
I think question mark looks great. /J/ I still don't love, but I'm not sure what to advise on it.
Yes, /ß/ isn't there yet.
Did you try an upright implied stroke for the /$/?
I think the upper bowl of /g/ may need just the tiniest bit more contrast.
I think /Q/'s tail is fine as is.

Bendy's picture

Well, the magazine went all quiet after I told them there were no italics or bold weights! Still, been working on spacing and refined some of the shapes. Ascenders are a few units taller, crossbars are now at x-height, e is not leaning backwards, x is revised, head serifs on VWXY are now sloped like the H.

Let me know any thoughts. :)

Syndicate content Syndicate content