Display Serif w/ tall letters

Duarte Pinto

Hey, I've been interested in typefaces lately and decided to give it a shot.

I started using FontLab 7 and designed this typeface focused on beautiful letters with high contrast. I was struggling with the fact that it looked too "vanilla" so I decided to make the letters tall, as you can see on A, B, E, F, H, et cetera. I've just completed the letters A-Z but might develop some other glyphs in the future.

Anyhow, this is the result:
Doing these letters was a fun process, and I would like to get better at it, so any advice is welcome (not focused on letterspacing & kerning )

Craig Eliason

Your thick rounds need to get thicker to optically match the weight of the thick stems.
You're right to defer kerning but you should be setting up your default sidebearings as you design. That establishment of default interletter space will better allow you to judge if the intraletter spaces (a.k.a. counters) are working (is /D too narrow, is /W too crowded, is /C too light, etc.).
Is taper of /V and /W thicks intended?
In /G, the swooping interior contour up to the bar feels out of character with the rest.

Thomas Phinney

(Besides agreeing with Craig)
You need overshoot for your rounds. Look at how the bottoms of C, G, and U, and top and bottom of O, look like they are shorter than other letters.

Interestingly, not so much a problem for top of C and G as I would have expected. The particular serif treatment you have done creates a bit of a flat section, which greatly reduces, and possibly eliminates, the need for top overshoot with those letters. (Ditto bottom and perhaps top of S.)

Usual overshoot is 1–3% of the total height.

Duarte Pinto

Thanks for the feedback, Craig and Thomas. I actually did overshoot my rounds by 10 units, but I guess it wasn't enough. The tapering of V and W ain't intended, I thought it looked natural. Looking back now, I totally agree with everything you said. I'll make sure to make those adjustments.

James Puckett

Try making the upper bowl of S significantly smaller than the lower to echo the tall feel of other letters.

K Pease

The need for overshoot can be relative to size. I would argue your rounds don't need more overshoot because a face with such thin hairlines is meant to be used much larger than we see it here. The baseline points of /V/W will need more, though.

It seems /S/X/Z are exactly symmetrical. Because visual center is higher than true center, this makes them look bigger on top, especially in the context of the deliberately high-waisted letters. Typically you want them to look a little bigger on the bottom for a sense of stability. This is done by moving the stokes of X laterally so they cross higher, by shortening the left end of Z (also correcting an impression of overhanging left which unfortunately is always invisible in the pangram "LAZY" and will only appear in other contexts)… and for S, hours of nuance, though a more exaggerated approach as suggested above will do well here.

Alex Visi

Duarte Pinto said:

 (not focused on letterspacing & kerning )

This is a mistake. Letters are not only the black shapes but also the white shapes, both inside and between the letters. You can’t fix the current word “F OX” with spacing or kerning, you need to change the design of F to fill that gap. Or different width of D and O in DOG are also the result of ignoring the white.
So, do focus on both black and white shapes

Duarte Pinto said:
it looked too "vanilla"

I think this is because it has no original design decision, no matter what proportions it has. But it is actually a good thing for the first typeface, since your goal is to learn the basics and creativity almost always interferes that. And I think you’re on a right path here as it looks pretty good overall.

You can do some improvements if you step back a little and test your letters one by one, starting from simple to more complex. Start with testing and improving rhythm and proportions of HOIHOIHO, then add an E, then F, T, L, then perhaps D, P, B and so on.

It might be useful to compare your font to other high quality fonts of the same genre and design, as you work, it helps to see mistakes and how they can be improved. Something like Font Goggles is a good tool for that.

Duarte Pinto

Wow, I'm glad I posted here. Once again, thank you all for being so insightful. As a rookie, it's hard to pick up on what's off, and these answers gave me just that.

K Pease said:

The baseline points of /V/W will need more, though.

I didn't even know the baseline points of /V/W needed overshoot, but it makes sense since they feel a bit taller. I had 10 units for the round overshoots, so maybe 15 will solve both overshoots. I was reading Practical Font Design With FontLab 5 by David Bergsland while doing this, and since he was doing a font with 10 units overshoot I just did the same without thinking. Maybe it depends on how big the grid is too.

K Pease said:

It seems /S/X/Z are exactly symmetrical. Because visual center is higher than true center, this makes them look bigger on top, especially in the context of the deliberately high-waisted letters. Typically you want them to look a little bigger on the bottom for a sense of stability. This is done by moving the stokes of X laterally so they cross higher, by shortening the left end of Z (also correcting an impression of overhanging left which unfortunately is always invisible in the pangram "LAZY" and will only appear in other contexts)… and for S, hours of nuance, though a more exaggerated approach as suggested above will do well here.

Good suggestions! And yes, I think going with the tall /S will work better on this typeface. 

Alex Visi said:

Duarte Pinto said:

 (not focused on letterspacing & kerning )

This is a mistake. Letters are not only the black shapes but also the white shapes, both inside and between the letters. You can’t fix the current word “F OX” with spacing or kerning, you need to change the design of F to fill that gap. Or different width of D and O in DOG are also the result of ignoring the white.
So, do focus on both black and white shapes

I was ignoring kerning because it's a process supposed to be done in the end, but I guess I should be considering it before getting there. As you said, some details can't be fixed with kerning. When it comes to letterspacing though, it was something critical that I ignored. As a result, the counters are very inconsistent, just as Craig mentioned.

Alex Visi said:

It might be useful to compare your font to other high quality fonts of the same genre and design, as you work, it helps to see mistakes and how they can be improved. Something like Font Goggles is a good tool for that.

This sounds interesting, but I can't figure out how to install it. Is it macOS exclusive?

Alex Visi

Duarte Pinto said:

I was ignoring kerning because it's a process supposed to be done in the end, but I guess I should be considering it before getting there. 

Yes, I meant spacing; But there are experienced people who also kern while drawing, but to me it seems rather tricky to keep track of it like that.

Duarte Pinto said:

This sounds interesting, but I can't figure out how to install it. Is it macOS exclusive?

Ohh, sorry, yes. But you can do the same in illustrator or anything else, if you just set the same word in each font manually.

Ori Ben-Dor

Very nice, good job!

– Diagonals look thinner than vertical strokes.
– U should probably be wider.
– I'm wondering about the low connection in the P.
– The extended bottom half of the C doesn't make much sense to me, it almost looks like a glitch...  

Duarte Pinto

Ori Ben-Dor said:

– Diagonals look thinner than vertical strokes.
– U should probably be wider.
– I'm wondering about the low connection in the P.
– The extended bottom half of the C doesn't make much sense to me, it almost looks like a glitch...  

Yes, I'll remake and adjust /U and /C based on previous feedback. The P is quite a different situation. It looks awkward if it's tall, but I will figure it out. Diagonals should be thinner than verticals, right? Because being the same width, they look bolder. I might have overcompensated them.

Craig Eliason

"High-waisted" and "low-waisted" might be clearer terms to describe your design. ("Tall" to me means condensed in overall width, which this isn't.)

Thomas Phinney

Duarte Pinto said:

Wow, I'm glad I posted here. Once again, thank you all for being so insightful. As a rookie, it's hard to pick up on what's off, and these answers gave me just that.

K Pease said:

The baseline points of /V/W will need more, though.

I didn't even know the baseline points of /V/W needed overshoot, but it makes sense since they feel a bit taller. I had 10 units for the round overshoots, so maybe 15 will solve both overshoots. I was reading Practical Font Design With FontLab 5 by David Bergsland while doing this, and since he was doing a font with 10 units overshoot I just did the same without thinking. Maybe it depends on how big the grid is too.

Yes. It depends on the em square—or more the cap height.
I have no idea how much “10 units” is without knowing the scale. (Although I would guess the scale is larger as the overshoot is not at all apparent. Perhaps a 2000 or 2048 units cap height.)

Duarte Pinto

I am using a cap height of 680 and the UPM is 1000. 

Thomas Phinney

Hmmm! I would have expected your overshoot of 1.5% to be enough or very nearly so. But then again, I haven’t done such a high-contrast face (at least, not with normal rounds). You're at 10 units, perhaps 14 (which would be 2%) will do the trick.

Duarte Pinto

It makes sense that high contrast fonts need a little more overshoot on the thin strokes.

I decided to go with 15 units of overshoot. Lately, I have been changing many details thanks to your feedback, and this is the result:

I think you guided me in the right direction. It feels a lot more like a typeface, in my opinion. As you can probably guess, I haven't started kerning yet but is the spacing good enough? I'm sending another pdf, in case you wish to judge it closely

Jasper de Waard

Very nice! Great improvements overall. I'm not sure about the new P, I think I liked the old one better. The key in type design is to know when to stick to the rule (in this case: a high waist) and when to break free from it. The more pointy M and N are better, but NVW could still go pointier. Take special care of the serifs at the end of thin strokes: top-right of K feels light, Z feels light compared to E, T feels light next to S, and come to think of it, the vertical ones such as on M feel light next to horizontal ones such as on L (look at LM). 

I think the proportions still need some work. If tall is what you're going for, I would make some of the letters (CEFGOQUV, to name a few) a little more narrow so that the proportions are more natural. Alternatively, you could get adventurous and make some letters narrow and others wide, like TypeTogether's Fino for example. The latter option is (I think) more interesting but also more difficult to pull off. Either way, you have to do it with confidence, which will come with experience

Perhaps you first need to learn 'normal' proportions before you are able to design adventurous ones...

Focus on proportions and spacing for now, and forget about kerning. In my view, you need good proportions for good spacing (although it's more of a two-way street), and you need good spacing for good kerning.

Anyway, best of luck! I like where this is going

Alex Visi

The problem with small-headed P is that it leaves a huge gap underneath, which is what you want to avoid for good spacing. Was going to write about the thin strokes serifs the same as Jasped wrote. A’s crossline can go lower to open that counter, raising it that high feels too naive

Also, pangrams are not very useful for testing fonts, this could be a better starting point:
https://www.typography.com/blog/text-for-proofing-fonts

Duarte Pinto

Thanks! I'm not sure about the new /P either, probably changing it back for the same reasons you typed. The thin strokes serifs are indeed inconsistent. I will give them a few more hours.

Jasper de Waard said:

I think the proportions still need some work. If tall is what you're going for, I would make some of the letters (CEFGOQUV, to name a few) a little more narrow so that the proportions are more natural. Alternatively, you could get adventurous and make some letters narrow and others wide, like TypeTogether's Fino for example. The latter option is (I think) more interesting but also more difficult to pull off. Either way, you have to do it with confidence, which will come with experience

I apologize for portraying my concept poorly since you took the time to give me all these suggestions; The font I'm going for isn't tall. As Craig better described than me, I was trying to pull off a high-waisted font, not tall.

Supposing I was going for a tall font, something around the TypeTogether's Fino would be very interesting. I like how its rounds stand out from the other letters.

On the other hand, from the high-waisted standpoint, I feel like the proportions are natural already. I agree that the /A's crossline should be lower tho.

Alex Visi said:

Also, pangrams are not very useful for testing fonts, this could be a better starting point:
https://www.typography.com/blog/text-for-proofing-fonts

Very interesting, funny subtitle. Ty for linking this

Ori Ben-Dor

Good progress!

I'd try a high-waisted-but-not-that-much P before settling on the original low-waisted one. Yes, the white space is indeed a problem, but it's not a huge one, given that it's a high-contrast display type.

Also, as the design is overall pointy and sharp, so I'd make the connection in the K thinner. That's going to be tricky, since the combination of high and thin connection would mean a very long leg, but still.

Drawcard

Like others have said, the first thing that jumped out at me was the /P design, I would probably interpolate it with your original design and go for a 50% mix

Otherwise it's looking good and coming along nicely.

Duarte Pinto

Hello, once again. I haven't posted in a while, but I'm still working on this typeface. I just wanted to thank you all for the feedback. Looking back now, I can easily spot so many mistakes that I wouldn't at the time.

I spent some time working on the proportions and other refinements. I also decided to do the kerning since I couldn't find anything else to change on the typeface. I'm showcasing it with the pangram as the other posts to ease comparisons.

I'm attaching a pdf to the post once more, but this time with a long string of text, so you can see how the font behaves. I probably ended up missing something, so any feedback is always welcome.

Jasper de Waard

Not bad, but if you want this to really be a good and useful typeface, you should probably slow down a little.

Proportions still need some work. A K M N W X and Y are a bit narrow.
S might be a bit wide (and top-heavy). L is usually a little narrower than E.

The J looks like it descends a bit too much.
The middle joint on B and R could maybe be made a little lighter
The leg of R looks a bit feeble. I would get rid of the upward curl and end it horizontally. Also the curves on the leg could be a bit stronger.
Same for the tail of Q.
V and W could use more overshoot
The bowl on P is falling downward
The S looks out of balance. Compare its terminals to C for example. I think it would benefit from more horizontal terminals (so, don't make them curve in so much, but keep closer to the baseline/capheight)
Many of the curve-to-straight connections look a bit off, not very smooth. For example the top of D and P.

Some characters look okay, but would space better with some changes. For example, it would help if the top-left of N alights with the bottom-left. Once you got this, spend some time on spacing, and spacing alone (yes, that means removing all of the kern pairs you already made). Try different things, make printouts, see what works. Only once you've absolutely nailed everything else should you do a final kerning round.

Also, maybe time for some numbers, punctuation, language support?

Craig Eliason

Spacing isn’t there yet (BROW N, JU MPS). Your W may be a good example of a letter that looks terrific on its own but proves very difficult to space well without some alteration. 

I don’t agree with everything @Jasper de Waard said (in particular, in a face with “quirky” vertical proportions (high waists), one needn’t “normalize” the letter widths too much), but listen to him about the S (and the kerning). 

@Ori Ben-Dor’s suggestion of “high-waisted-but-not-that-much P” is worth trying. This halfway version is a missed opportunity to fit the particular style of the face. 

Alex Visi

I agree, too early for kerning.
For letters with diagonal strokes, such as A V W etc, there’s a spacing benefit in making outer serifs shorter to allow the letters getting closer. 
I think the S can get stronger if you make the spine more curved, which would make the negative spaces more rounded and stable.
Try heavier bottom serif on N?
Agree about the R, try making that leg without the points in the middle.
Feels like the descenders can be a little shorter to avoid line spacing problems.

Also, highly subjective, but I think the high waist here is one of those cases when you keep peeing against the wind just because “that’s the idea” which “makes it different”. Let it go and it’ll go smoother. Or make both versions and see which you and people like more.

Duarte Pinto

Well, there goes my kerning.
I started this typeface to learn solid basics that I can use in the future, so training my eyes to good proportions is a must.
I understand that while I was learning this, my creativity got in the way. I did not want to do something that already exists. I decided to do a high-waisted typeface for that reason, but I am willing to give it up if that helps me improve.

Craig Eliason said:

in a face with “quirky” vertical proportions (high waists), one needn’t “normalize” the letter widths too much

Maybe that's a case where the high waist idea is giving me more issues than benefits. For now, I will create another version of what would be considered a normal-waisted typeface. This version might help me to focus more on the understanding of type.
As Jasper de Waard said:

Perhaps you first need to learn 'normal' proportions before you are able to design adventurous ones...

Perhaps I do. That's also another reason I didn't want a tall/compressed typeface before doing a normal one. I am thankful for the extensive feedback and will keep updating you with the changes made. At least, I plan to make the rest of basic Latin, but I want to have a solid A-Z before proceeding. 

Ori Ben-Dor

I liked the second version (July 26) better than the third one (August 2). It had more character, and the S was way superior, both stylistically and technically. New S looks skewed.

I'd definitely make S & Z narrow like they used to be in the 2nd version. I'd make E, F & L similarly narrow as well. You want width variability, it compliments your design. Sure, it shouldn't be random, but you weren't very far from getting it right. You can always look at Roman capitals (e.g. Trajan) for reference.

There are still inconsistencies in the weight department. For instance, the vertical leg of the Y looks thicker than the thick stroke of the X. And the K looks too dark (it has two thick strokes close to each other, so you need to compensate by making them a bit thinner, especially the vertical one). The thick stroke of the Z looks thinner than any other thick stroke. The left stroke of the V looks thinner than that of the W (where if anything it should look thicker, as V doesn't have a second thick stroke you should compensate for).

Also, you've lost some of the dynamic nature of the 2nd version by adding so much mass to the serifs of the thin strokes. Yes, the more pronounced serifs bring balance, but if it's supposed to be a display type—which I think is the case—then you need to be careful not to cross the border between balanced and boring.

Why do E & F have a rounded inner corner on top while D, P, etc. don't? That's not necessarily a mistake, but make sure you're aware of this inconsistency and happy with it.

Drawcard

My personal opinion @Duarte Pinto but don't overwork this alphabet too much. It's really 95% good right now, and too much fiddling might reverse that (I have been there too many times ). Try having a go at numerals, punctuation, diacritics etc. and then come back to these letterforms with fresh eyes.

One feature I love about your design, is how a lot of the serifs touch each other, creating these nice little arches in between letters. And coupled with the high-waisted look, it has this lovely 70's dry-transfer vibe that's honestly hard to faithfully replicate. So well done on that!

Maybe as I get older, I'm getting a bit looser about conforming to 'the grids' and 'the spirals' - but your proportions and stem / stroke thicknesses don't look that bad to me. Sometimes those imperfections add a personality to the design that makes it special.

If you do have to make adjustments, ask yourself why - what problem is this solving? If there isn't a good answer, let it go. 

Just one technical thing I want to raise though - Perhaps the top and bottom of the bowl in /D is a touch too flat? If that's part of the identity, no worries. But my 'why?' rationale for this change would be that the /G and /O have smoother stroke transitions, and the /D could benefit from that too.

Duarte Pinto

I finally did the numbers. Before getting into that, I kept working on the capitals for a little longer. I also changed my mind about doing "normal" waists (at least, for now). These are the changes I made: 

Proportion Changes

• /A /K /M /N /V /W /X /Y - Made them wider

• /L - Made it narrower

• /S - Changed it back to the previous version (narrower)

Weight Changes

• Took some mass off the light serifs because they were a bit exaggerated

• Made the longer serifs (like the ones on /Z and /S) lighter to fit this new change

• /K - Made it lighter

Technical Changes

• /J /Q - Descended too much, which could cause line spacing issues. Since its a display face I wasn't too worried about those but still decided to make them a little shorter

• /V /W - Added a little more overshoot (5 units)

• /N - Aligned top (left) with the bottom (left) to improve spacing

• /P /B /R /D - Improved stroke transitions

• /Y /P - Changed their waist to fit the typeface

• /R - Changed its leg

• Took off all-rounded inner corners as I think they were unnecessary, and this makes for better contrast too

Currently, I am not working on these letters since I moved on to numbers and punctuation, but any constructive opinions on the capitals are always welcome and will get noted down for another day.

Here are the numerals:

I noticed that in some typefaces, the /4 has a similar top to the /A. After all, they share that triangle look. I thought that detail would be interesting in this typeface and decided to do the same. Any tips?

Christian Thalmann

The bulging top and bottom overshoots in /B/D/P/R/ are a step back IMHO. Was that in response to Drawcard's comment? If so, I think it went way past the original intention.

Yes, the top of /4/ feels quite out of character.

Adam Jagosz

Yeah, the new /B/D/P/R are certainly not a technical change — it’s a design change. What I think was meant to change was just the easing of the transition: