Type Design Project Opinion

Eryk Kosinski

I would appreciate if a professional Type Designer could give a very detailed opinion on my Sans Serif Typeface I am currently working on at stage of finishing 5 weights. All mistakes I am doing as well as advice on how to become more efficient. I use Fontlab software. Voice chat with screen sharing will be very helpful.

Christian Thalmann

Why don't you just post screenshots and/or a specimen here?

Eryk Kosinski

I started my project from light weight and sidebearings are only finished in this weight. I plan on using Variations panel to add more weights later when i properly fix these 5 weights and adding ink traps and smart corners later on.

Craig Eliason

In bolder weights, counterspaces of denser letters (particularly/a/e/s) get too compressed. 

Terminals of most letters are cut perpendicular but that’s not true of /t and /f. 

I’d spread the arm and leg away from the stem in /K and /k. 

In designs with long extenders like this (are you certain you want them?), I find old style figures that conform to x-height and extender length look a little goofy. Similarly, small caps can, and probably should, be taller than x-height. 

John Hudson

I started my project from light weight and sidebearings are only finished in this weight. I plan on using Variations panel to add more weights later when i properly fix these 5 weights and adding ink traps and smart corners later on.

That describes your working process, but what is your thinking process? 

I find it helpful to think of a typeface as a collection of answers to one or more questions, and being able to articulate the questions helps evaluate the answers. Some basic questions are

• What is this typeface for?
• How do you intend to use the typeface?
• What principles define the design?
• What makes this typeface distinctive?

Other questions emerge from within the design as it progresses, usually of the form ‘If this, then what?’, and articulating these questions helps the design to coalesce by suggesting answers that apply across multiple glyphs.

As a general comment, your lighter weights are doing better than the heavier ones. There are the makings of a strong skeleton in the lightest weight, but it doesn’t put on weight in a healthy way, and it all ends up a bit mushy in the heaviest.

There are proportional problems in the widths that will be easier to see in test words that put various letters in combination. The lowercase c looks too narrow, and the a and e a little too wide. The overhanging e is unstable and looks like it is falling forward.

The f and t stand out as an unusual shapes, but I think they can work and could be the key to the distinctiveness of the design. Possibly the crossbar should shift a little to the right, rather than being symmetrical.

Eryk Kosinski

John Hudson said:

I started my project from light weight and sidebearings are only finished in this weight. I plan on using Variations panel to add more weights later when i properly fix these 5 weights and adding ink traps and smart corners later on.

That describes your working process, but what is your thinking process? 

I find it helpful to think of a typeface as a collection of answers to one or more questions, and being able to articulate the questions helps evaluate the answers. Some basic questions are

• What is this typeface for?
• How do you intend to use the typeface?
• What principles define the design?
• What makes this typeface distinctive?

Other questions emerge from within the design as it progresses, usually of the form ‘If this, then what?’, and articulating these questions helps the design to coalesce by suggesting answers that apply across multiple glyphs.

As a general comment, your lighter weights are doing better than the heavier ones. There are the makings of a strong skeleton in the lightest weight, but it doesn’t put on weight in a healthy way, and it all ends up a bit mushy in the heaviest.

There are proportional problems in the widths that will be easier to see in test words that put various letters in combination. The lowercase c looks too narrow, and the a and e a little too wide. The overhanging e is unstable and looks like it is falling forward.

The f and t stand out as an unusual shapes, but I think they can work and could be the key to the distinctiveness of the design. Possibly the crossbar should shift a little to the right, rather than being symmetrical.

This typeface target is to be used in text, mainly in publications online as well as for print. Any advice on how to approach it if thats the target?

John Hudson

Test at the target sizes. Your x-height may well be a bit small for a text face.

My advice to new type designers tends to be that looking is more important than doing. It is easy to get wrapped up in moving beziér handles around, but time is much better spent contemplating the shapes, testing them in various combinations, and also looking at other typefaces with an eye to understanding what makes them work. You are working in a couple of well established genres, with elements of grotesque and geometric sans serifs. I would spend some time looking closely at the classics of those genres—say, Helvetica and Futura—, and in particular at how they handle modulation of stroke weight, how bowls and stems interact, and how they gain weight as they get heavier. Don’t work directly on your design while you are doing this looking, because the goal isn’t to import influences from those classic designs: take the time to really analyse the classics and learn from them, then go back to your design and figure out how to apply what you’ve learned.

Christian Thalmann

The Light looks reasonable overall in the basic alphabet, whereas the heavier cuts progressively look cruder, with bumpy curves and unbalanced stroke weights. It feels like you expanded the glyphs only to the outside, whereas the extra weight should mostly go on the inside. There's also way to little, if any, optical compensation.

As far as shapes go:

• D is very rectangular, M is saggy and narrow, S falls backwards.
• The cut on f doesn't work for me, but that might just be me. It's certainly out of character compared to the rest of the typeface.
• The figures strike me as very odd and out of character with the rest.
• The designs for /ß/ç/ð/ℓ/ are weird to the point of being unusable. Look at other typefaces to figure out how they're meant.
  
• Percent and fractions are too light.
• The accents in the light are microscopic.
  

Eryk Kosinski

John Hudson said:

Test at the target sizes. Your x-height may well be a bit small for a text face.

My advice to new type designers tends to be that looking is more important than doing. It is easy to get wrapped up in moving beziér handles around, but time is much better spent contemplating the shapes, testing them in various combinations, and also looking at other typefaces with an eye to understanding what makes them work. You are working in a couple of well established genres, with elements of grotesque and geometric sans serifs. I would spend some time looking closely at the classics of those genres—say, Helvetica and Futura—, and in particular at how they handle modulation of stroke weight, how bowls and stems interact, and how they gain weight as they get heavier. Don’t work directly on your design while you are doing this looking, because the goal isn’t to import influences from those classic designs: take the time to really analyse the classics and learn from them, then go back to your design and figure out how to apply what you’ve learned.

I took your advice and had a look at these styles of typefaces. Here are the images of changes to for now lowercase n.


As the typeface is for text and I have looked at text typefaces I have seen many of them have ink traps. Tell me what you think of the last n glyph with it added. 

John Hudson

These changes are definitely improvements.

Inktraps are complicated. I think they are best understood, and work best, as output-specific design features, i.e. whether a design should have inktraps and what kind and how large they should be really depends on how the typeface is going to be printed or displayed, and should properly also be size-specific. Scaleable digital type, by its nature, is largely output-agnostic: it is going to be printed or displayed in arbitrary sizes, resolutions, media, and technologies. So if you include inktraps in a design, you have to be aware that in some situations they are going to cease being a compensatory feature that helps avoid an area becoming too dark in some output and instead become a really obvious visible feature that may be distracting.

So I tend to use inktraps sparingly, and will not usually employ the kind you illustrate here. I am more likely to introduce a simple line segment in the crotch of a v-shape connection between two stems or between a stem and a bowl as in your n, just to open the space a little. I would certainly avoid the kind of impact of the inktrap that you have on the top right of the vertical stem of the n, where it becomes rounded, unless you intend to make that roundness a feature of the design.

Thomas Phinney

Yes, generally the changes are improvements! Yay!

Your curves are still a bit… lumpy. The outer curve of the n, the shoulder is weak at the top left, and the inner curve, the shoulder is strong. I don’t know which you intend, or if you want a normal/neutral shoulder, but it feels like the two sides are coming from two different places. For both of them, the apex of the curve is a bit to the left of where it should be—but especially on the outside.

Assuming you want a relatively normal/neutral shoulder:
- on the outside, move the apex to the right a little, and the handle (BCP) of the outside right stem up a bit while at the same time moving the on-curve point down on the stem.
- on the inside, again move the apex to the right a little bit. Move the on-curve point of the right stem down to keep it matching the outside. Move the handle (BCP) down as well, until you have a relatively smooth/symmetric look.

Alternatively, you could use FontLab’s “harmonize” functions to move nodes and handles. Even if you want to do it by hand, you could use “harmonize” functions on nodes/handles just to see what they would do. You could also turn on curvature visualization (View > Show > Curvature) to see where you have discontinuities.

Eryk Kosinski

Thomas Phinney said:

Yes, generally the changes are improvements! Yay!

Your curves are still a bit… lumpy. The outer curve of the n, the shoulder is weak at the top left, and the inner curve, the shoulder is strong. I don’t know which you intend, or if you want a normal/neutral shoulder, but it feels like the two sides are coming from two different places. For both of them, the apex of the curve is a bit to the left of where it should be—but especially on the outside.

Assuming you want a relatively normal/neutral shoulder:
- on the outside, move the apex to the right a little, and the handle (BCP) of the outside right stem up a bit while at the same time moving the on-curve point down on the stem.
- on the inside, again move the apex to the right a little bit. Move the on-curve point of the right stem down to keep it matching the outside. Move the handle (BCP) down as well, until you have a relatively smooth/symmetric look.

Alternatively, you could use FontLab’s “harmonize” functions to move nodes and handles. Even if you want to do it by hand, you could use “harmonize” functions on nodes/handles just to see what they would do. You could also turn on curvature visualization (View > Show > Curvature) to see where you have discontinuities.

I took your advice and made changes. I am still not sure if I should stay with the 1st one or 2nd one. As the typeface target is to be used in text and for printing i thought of adding Ink Traps and Round Corners for function as well as style of the font.

John Hudson

Test in print at the target sizes and as close as you can to target output resolutions and methods.

Craig Eliason

With your bolder weights, you'll want the whites to get smaller (smaller sidebearings and counter widths) in addition to the blacks getting bigger.

James Montalbano

I found this graphic by Adrian Fruitger helpful in my early days in type design. Perhaps you will find it helpful as well.

Eryk Kosinski

Christian Thalmann said:

The Light looks reasonable overall in the basic alphabet, whereas the heavier cuts progressively look cruder, with bumpy curves and unbalanced stroke weights. It feels like you expanded the glyphs only to the outside, whereas the extra weight should mostly go on the inside. There's also way to little, if any, optical compensation.

As far as shapes go:

• D is very rectangular, M is saggy and narrow, S falls backwards.
• The cut on f doesn't work for me, but that might just be me. It's certainly out of character compared to the rest of the typeface.
• The figures strike me as very odd and out of character with the rest.
• The designs for /ß/ç/ð/ℓ/ are weird to the point of being unusable. Look at other typefaces to figure out how they're meant.
  
• Percent and fractions are too light.
• The accents in the light are microscopic.
  

Thats my rough idea of Black Weight. I'm not sure how I should approach my "s" glyph, something is still not right. 

Thomas Phinney

The “conventional” way (though not the only one) to approach the very-heavy S is to put the full weight on the spine, and let the top and bottom be a bit thinner.

At the very least, you need to move _some_ weight to the spine. Try making it the dominant stroke—and if you don’t like how that turns out, experiment with something closer to what you have now, with just a little more weight on the spine.

Eryk Kosinski

Thomas Phinney said:

The “conventional” way (though not the only one) to approach the very-heavy S is to put the full weight on the spine, and let the top and bottom be a bit thinner.

At the very least, you need to move _some_ weight to the spine. Try making it the dominant stroke—and if you don’t like how that turns out, experiment with something closer to what you have now, with just a little more weight on the spine.

I have made top and bottom parts of s narrower and middle spine the same width. Does any other glyph need adjustments?

Christian Thalmann

The Black looks very different from the Light; it's hard to believe they belong to the same typeface. Unlike the Light, the Black looks very wide, blobby, and contrasted. Look at professional typefaces to see how they handle adding weight to a low-contrast sans.

In any case, I suggest you take the time to get your Regular and Bold right and worry about the other weights later.

John Savard

Christian Thalmann said:

• The designs for /ß/ç/ð/ℓ/ are weird to the point of being unusable. Look at other typefaces to figure out how they're meant.
  

That comment may be harsh, but, sadly, it is true.

John Hudson said:

The f and t stand out as an unusual shapes, but I think they can work and could be the key to the distinctiveness of the design. Possibly the crossbar should shift a little to the right, rather than being symmetrical.

To me, the f and the t did not stand out that much as unusual. Instead, the first thing I noted about the typeface was the y. Now, this is not a shape that is so unusual that it hasn't been used in other typefaces, but since this typeface looks a lot like Helvetica, News Gothic, or Akzidenz-Grotesk, it did not seem to be appropriate for that kind of typeface.

(Thus, John Hudson's point about the question of "What is this typeface for" and related questions needing to be answered did not occur to me, as I basically assumed that the answers to those questions were implicit in the appearance of the typeface: since it looks a lot like typeface X, obviously it's intended to be used where people use typeface X, and we can go on to the next question.)

Of course, though, the original poster should be warned that unlike the other commenters, I am not a professional typeface designer, so my opinions must be taken with a grain of salt.

Eryk Kosinski

Christian Thalmann said:

The Black looks very different from the Light; it's hard to believe they belong to the same typeface. Unlike the Light, the Black looks very wide, blobby, and contrasted. Look at professional typefaces to see how they handle adding weight to a low-contrast sans.

In any case, I suggest you take the time to get your Regular and Bold right and worry about the other weights later.

Here is my regular weight.

Cory Maylett

Regarding your latest regular weight, the relationship between the thins and thicks, to me, seems a bit haphazard. For example, the middle horizontal strokes of the a and e are different thicknesses. In addition, the bowl of the a is squished in a way that seems at odds with the roundness of the eye in the e and other letters.

The thin strokes where the round sections of several glyphs merge with the stems look too thin. (In heavier weights, reducing the bulk in those areas is necessary, but in lighter weights, that isn't as much of a consideration unless it's a consistently applied design feature throughout the typeface.)

The arm of the r seems to stretch out too far and will create spacing issues with whatever glyph follows it. The x strikes me as a bit too light. The curves of the s seem a little unnatural. The point at which the arm and leg of the k touch the stem creates a tension point that doesn't exist elsewhere in the font. The z seems just a bit too narrow. The apex of some of the bowls seems slightly off, preventing the roundness of the strokes on either side from flowing naturally. I'm unsure about the large-radius curves in the f and t. I'm also uncertain about how some terminals end at unusual or inconsistent angles.

Many of these quirks I've mentioned could be design features if they were applied more uniformly throughout all the typeface weights.

Eryk Kosinski

Cory Maylett said:

Regarding your latest regular weight, the relationship between the thins and thicks, to me, seems a bit haphazard. For example, the middle horizontal strokes of the a and e are different thicknesses. In addition, the bowl of the a is squished in a way that seems at odds with the roundness of the eye in the e and other letters.

The thin strokes where the round sections of several glyphs merge with the stems look too thin. (In heavier weights, reducing the bulk in those areas is necessary, but in lighter weights, that isn't as much of a consideration unless it's a consistently applied design feature throughout the typeface.)

The arm of the r seems to stretch out too far and will create spacing issues with whatever glyph follows it. The x strikes me as a bit too light. The curves of the s seem a little unnatural. The point at which the arm and leg of the k touch the stem creates a tension point that doesn't exist elsewhere in the font. The z seems just a bit too narrow. The apex of some of the bowls seems slightly off, preventing the roundness of the strokes on either side from flowing naturally. I'm unsure about the large-radius curves in the f and t. I'm also uncertain about how some terminals end at unusual or inconsistent angles.

Many of these quirks I've mentioned could be design features if they were applied more uniformly throughout all the typeface weights.

Middle strokes of e and a same now. Bowl of the a adjusted. Thin strokes are a design feature which will be applied in other parts of the font. Arm of r made shorter. x adjusted to be darker. Lower s terminal lifted for more balance. z made wider. f and t curves adjusted to be thinner and more in line with n glyph. In regards to terminals all of them end in the same angle.

Thomas Phinney

“Thin strokes are a design feature which will be applied in other parts of the font.”

It will be interesting to see how that evolves as you propagate it elsewhere. I would expect the thinness of those bits to be reflected in (at least!) the thinnest parts of stroke joins such as those in hmnu, and the thinner/sharper joins (of the two in each) of bdpq.

John Hudson

Possibly also in the relative thickness of diagonal arms, and crossbar of f and t. It is a tricky task, though.

Eryk Kosinski

Thomas Phinney said:

“Thin strokes are a design feature which will be applied in other parts of the font.”

It will be interesting to see how that evolves as you propagate it elsewhere. I would expect the thinness of those bits to be reflected in (at least!) the thinnest parts of stroke joins such as those in hmnu, and the thinner/sharper joins (of the two in each) of bdpq.

I have finished designing uppercase and lowercase glyphs with comma and a period. I have added files with pages for each weight in different sizes with 300dpi for a4 size.  

Eryk Kosinski

Thomas Phinney

Well, this certainly gets dramatic and strange in the heavier weights. Having the mid-height bar or portion of ABEFPR and aeft as a hairline, but other horizontals are not, and then diagonals are mildly stressed (MNVWvwx and maybe Zz) … pretty unusual!

Definitely the H is inconsistent with other treatments. But even if you fix that, I am not convinced this is “working” for me, as yet. Is “being weird” part of the plan?

Eryk Kosinski

Thomas Phinney said:

Well, this certainly gets dramatic and strange in the heavier weights. Having the mid-height bar or portion of ABEFPR and aeft as a hairline, but other horizontals are not, and then diagonals are mildly stressed (MNVWvwx and maybe Zz) … pretty unusual!

Definitely the H is inconsistent with other treatments. But even if you fix that, I am not convinced this is “working” for me, as yet. Is “being weird” part of the plan?

I have made adjustment to the black weight for now reducing contrast in lowercase and uppercase. Your thoughts?