How to make a sans serif ?

paulow's picture

Hi, wise people. Well, althought I has many fonts, in fact I never did a serious sans-serif in all my life. I desire to start now. So, I am here to hear your critiques and advices in this, yet unnamed, Intellecta First Sans. The sample here is the actual status of the font, still in work. To better resolution see the pdf.

AttachmentSize
IntellectaFirstSans_status_rev009.pdf21.23 KB
1996type's picture

Ok I'm gonna be honest with you. h, the first l m n u v w x y z have potential. The rest is simply ugly. Start all over again from the glyphs with potential and forget about the weird serifs you made in a, c, e, etc. Also forget about the weird curves you made in the bottom of the f, bottom of the second r, bottom of the first stem in the second m.

The best way is to learn sans-serif, is to study sans-serif. Read books (I never did this. Not a must.), but mostly, LOOK! Have a very close look at all the nice sans-serifs you can find and figure out why the designer made certain choices.

First real critique on the glyphs with potential: z looks top-heavy, w should be flipped, y maybe a tail curving to the left?

It's also important, to think of the purpose of your font. print, web, or both. text, display or both. Don't start with the other styles untill you've had some really positive comments from others.

Don't let this discourage you. My first designs were very ugly and this really does have potential. I found the most important thing in typedesign is to never be sattisfied. With the help of all typophilians and a lot of hard work you'll get there.

Jasper de Waard, amateur typedesigner.

paulow's picture

Thanks, Jasper. I will take in mind your vision. In fact, the "weird" serifs are not the final decision of this font. I add it to try different visions of the font, like you can see when I add four diferent versions of the "r". Believe me, is this the real first test of these lowercases. About your other question, I want create a text font, to be printed.

And I never lost the courage (with this attitude I did make over 400 ugly fonts), but I make fonts for my pleasure, only this. If they sells, it's a by product of my funny life.

In the last months I am trying read more about typography, but maybe I have a punk blood: "Make first, ask after". A last thing : is not easy to read techinical books in foreign languages, try yourself reading a portuguese book.

I like your point of view about the shapes. I will try some of them and back with more samples, soon, ok?

best regards

sim's picture

I don't want to discourage you, but I think there is too many ideas in it. You talk about sans serif and you place a “serif” in some letter (a, e, s, f). Concentrate your work on one goal, choose the h or the e for instance, which are different from each other in the drawing, to see which will give you satisfaction and make some draw from them. To help, you could take a closer look at some well known sans serif and observe how the letters are drawn. It takes consistency and similarity in the outline. Keep going.

paulow's picture

Hi, André. I agree, this sample have a lot of ideas, because I am not sure yet about the sahpes I want. Concerning the little serifs(a, e, s, f), many people don't like them, but I am searching for some kind of singularity to the font, understand? The "e" is totally wrong, I know, I will remake it. I agree wichh a look at good sans around the world can be help me to design this, but I am tryiing find the shapes to this font searching in the own font, if possible. I like to try first the things (self-taught) and before learn, if necessary. I appreciate (and respect) the knowledgement and the scholar methods and peoples, but I like, myself, try and learning using my way. This participation at Typophile is a attempt to learn with the other way. Thanks, soon I back with a revision of these lowercases

paulow's picture

Well, here is mt first revision. What you think?

PabloImpallari's picture

It's a step in the right direction.

The c still needs more work (Compare Helvetiva and Frituger for inspiration)
Once you have your c, made your e and s according.

The k triangles looks unbalanced.

The m can be just a bit more narrow

The r branch is too long.
Have a llok at this post: http://typophile.com/node/74824

The t tail is also to long.

paulow's picture

@Pablo, thanks for the advices. Follow here some modifications in the "k".

paulow's picture

for the "t"

paulow's picture

@pablo, I don't sure exactly what you try to say about to narrow the "m", but I show here two little variations in the "m"

paulow's picture

... a new "r"

paulow's picture

and, the last, the "c"

@Pablo, I don't agree to search solutions in Frutiger and Helvetica, in this case, because I think they are very clean, very regular, very classical, if you undertsand me. I am searching for a modern taste, something difficult, because the high number of fonts is the world, and plus difficult tome, because my low practices to make text fonts. here is, anyway, some new shapes to the "c"

PabloImpallari's picture

The new k looks nice!

The new t looks nice also.
But still need to refine the bottom tail ending to make it like the rest of your fonts.
All other endings have one of the squares made soft. The t tail ending don't.

The new r looks nice!

Ok. For modern looking "c" have a look at Type-Togeher fonts:
http://new.myfonts.com/foundry/Type-Together/
They usually add weight to the top part, by adding a serif, or by making the top curve go deeper.

You also need tho make the "soft" ending to the c,e and s.
To be in line with the other letters.
Have a look at cronos from Adobe: http://new.myfonts.com/fonts/adobe/cronos/

Andreas Stötzner's picture

Forgive me if I aim to be honest too.

O.k., there is *some* idea in your draft, which does shine through in the n and m glyphs. But the whole set of glyphs you present reveals a total lack of basic typographic feeling and skills. I don’t know your educational background, I don’t want to step on your feet. Why do you think you must do letters?
If you mean it, put the whole thing aside, start studying type, and eventually return to this drafts, for heavens sake not earlier than after one year’s time.

paulow's picture

@Andreas, thanks for to be honest. One teacher, when I was a kid, teaches me which all questions deserves answers, so, here are: concerning my educational background, I am a self-taught. And, I make typography because I like to make typography, and, for me, nothing more is so important than like to do that we want. If we learn much, or a bit, at academic world or just trying, this is not really the most important thing, first comes our personal feelings about the whole thing, it's my position. I will not put the whole thing aside, I will still try, and try, and to do mistakes, and try again, ever. Anyway, I appreciate your point-of-view.

There are genius and happy people, craftsmen who take the hands in the mud. Maybe I Am the second kind.

kennyadam's picture

from what i could tell your "special" accent on the type faces the the fact on the tips of each letter you have on the corners, flat on one curved on the other. THEN on the opposing side you do the exact opposite. i see this as being most notable in the R. i like this distinct feature. Maybe taking and forming your letters based on this simple, geometric concept might be the key to making this font face?

Thomas Phinney's picture

When you have a curve that attaches on to a straight, as in the top left of b, h, m, n, you need to thin out the curve just a little before the join. Look at existing professional typefaces and see what I mean? Failing to do this will make that join look overly heavy.

Cheers,

T

PabloImpallari's picture

Another faces that have the same "one flat-one curved" feature are:
Overlock, by tipo.
http://new.myfonts.com/fonts/tipo/overlock/
Haptic:
http://new.myfonts.com/fonts/henning-skibbe/haptic/

paulow's picture

@Pablo, I make a new design to the "c" and a first try to fix the bottom tail of the "t".
Before I starts to fix the "soft" endings to the c, e and s I need to define the shape of the "c", and this sample here is the right direction, maybe. Thanks for the links (all them).

@kenny, thanks for the support. there are something more than critics in the world, indeed. I know which the idea of "flat on one curved on the other" is not really inovvative, but I really like this kind of shape since many years ago.

@Thomas, thanks so much by the teaching expertise. I will try to do like you says, but only after I confirm the overall shapes of the letters, you dont agree? I fell which the whole thing is not totally defined yet

well, so, here is the actual shapes from "c" and "t". the head of the "t" is not good yet, I will fix it soon.

kennyadam's picture

yeah to me it seems like thats your biggest accent for the font and you should show it through. how i imagine your C being the best is actually just to take the lowercase l and literally bend it into a circular shape. Then your tips would keep the same accent. then all you have to experiment with is how you want the actual tips to face.. (straight out towards each other at a 45 degree or even facing each other directly.) once you have this you should cup the bottom of your lowercase T in this same manner to give it uniformity.

paulow's picture

here is the actual version of the "t". I still working in the "c". It stills hard to define.

paulow's picture

@kenny, here is my actual "c". is not good yet. but I will test it compiling a font to test these lowercases all working. after, I will back with necessary corrections

kennyadam's picture

it looks like the curving of the bottom isn't as natural as the top.. maybe be more geometrically curved.

Thomas Phinney's picture

@paulow:

You wrote "I will try to do like you says, but only after I confirm the overall shapes of the letters, you dont agree? I fell which the whole thing is not totally defined yet"

I understand that you're still working out the general shapes, but to me this is really basic. They're just not properly drawn letters without these sorts of touches, IMO. I can't really get past it.

I see that your letter shapes are such that the "o" is slightly egg-shaped, and bigger at the top than the bottom. I assume this is on purpose. It doesn't bother me in the "o" so much, but the attempt to use the same shapes as the basis for, say the "d" is not working so well. The "d" just ends up looking lumpy. (I'm using the same word Robert Slimbach used a lot in critiquing some of my early work, btw, not meaning to sound too harsh.)

Cheers,

T

paulow's picture

@Thomas, many thanks until here for your high valuable advices, ok? I back here with a revision at the "a", "o" (less egg-shapped now), and the "t" (and soon the "f") with your horizontal stem in the level of the head of the "a".

Concerning the case of the "curve that attaches on to a straight, as in the top left of b, h, m, n" I will post a new sample in few minutes, showing my new shape to this curve

The question of the same basis for letters like "a", "b", "d" I will study in the next step. You dont seems too harsh, au contraire, you are doing like a gentlemen, and this is very important between colleagues. I think some personalities in this community are, in the last weeks, so much exalted.

ps. : what means "IMO"?

paulow's picture

here is my solution (if I understand correctly your advice, Thomas), about the curves on the top of the vertical stems.

@kenny, I still working in the "c" - a great problem until here

Thomas Phinney's picture

Paulow,

That is a major improvement, yes, for the joins. You could probably do the same thing just a *little* more, but that's pretty close.

When you show the letters large like that the "lumpiness" issues become more apparent.

For example, in the "b" the top right and bottom left parts of the bowl seem thinner than anywhere else. This feels particularly "wrong" because it is the exact opposite of the effect of chirography, the idea that if a calligraphic implement was involved in drawing that bowl, those would be the thickest parts and the top left and bottom right would be the thinnest.

Those same parts of the bowl are also flattened for some reason.

Another issue I'm seeing in letters like "t" and "o" is that the horizontals look thicker than the verticals. The verticals need to be slightly thicker than the horizontals to "look" the same, to achieve the feeling of being monoline (having even line thickness). I'm assuming that's what you're going for here.

Cheers,

T

P.S. "IMO" = "In my opinion"

paulow's picture

(...) in the "b" the top right and bottom left parts of the bowl seem thinner than anywhere else. This feels particularly "wrong" because it is the exact opposite of the effect of chirography, the idea that if a calligraphic implement was involved in drawing that bowl, those would be the thickest parts and the top left and bottom right would be the thinnest. Those same parts of the bowl are also flattened for some reason. (...)

Well, here is the actual status for "b" "d" "a" - to better resolution see the pdf attachment in the top of this topic

paulow's picture

"another issue I'm seeing in letters like "t" and "o" is that the horizontals look thicker than the verticals. The verticals need to be slightly thicker than the horizontals to "look" the same, to achieve the feeling of being monoline (having even line thickness). I'm assuming that's what you're going for here."

a new revision here (to better resolution, see the pdf in the top of this topic)

Thomas Phinney's picture

Getting better all the time (as Lennon & McCartney said)!

Still some definite "lumpiness" to the curves in b & d. Also g has the same stroke merging problem b and d used to have.

Cheers,

T

Igor Freiberger's picture

Paulo,

you have very interesting elements here. They point to a nice sans. The challenge is to articulate all them to create a coherent design. I see you already made huge improvements since the first post.

I think you must define a global criteria for each glyph element and, after you have the whole A-z built based on this criteria, maybe introduce some particularities to extend font character. Otherwise, you will be builting a group of glyphs which together does not work harmoniously.

For example: the curve on top of stems have two opposite directions. Some goes at left (h, t), some at right (n, m, b). Define a direction, build all glyphs based on this and them you have a good base for improve. You can even substitute a whole piece in FontLab (so it becomes very easy to change ALL the tops with curves to left with tops with curves to right, if you decide to try this).

I like your |f| very much. But |t| is strange due to very large bar at left. Its bottom terminal seems also undefined. Bottom of |g| seems thicker than remaining stems.

A word about curve quality. This is your b (or something similar to it, as I traced it in Illustrator and grought to FontLab):


.

The problem here is that you don't have a standard. In each quarter you have one kind of curve. Together, it seems you |b| has some irregularity. Of course, you don't want an exact circle but actually an unequal ellipse to compose the b. Anyway, this ellipse needs to be regularly build to get fluidness.

This is how I'd do this:

Firstly, I make a copy of original glyph and put it at mask layer (all this in FontLab). With Ctrl+Shift+O, I make the mask visible. So I did edit control points in one quarter to get a more circle-like curve there:
.


.

With a quarter closed, I simply copy-n-mirrow it to get the ellipse. Add the stem and you have a good start point:
.


.

You can see I like to work with points (BCPs) and handles at a 4x4 grid. It's just a personal preference.

As I said, this is just a startpoint. Actually, you don't want a so regular b. The advantage with this structure is you can make experiments with one piece without interference in another. Fot example: handles {436,294} and {498,258} would be slowed to make upper conection more soft and less geometric. You can experiment with these handles while the other quarters remain ok.

At the end, you will probably modify severely the curves of quarters sw and nw, while ne and se will receive just small changes regarding height and width.

Well, this is how I'm working. Since I begun to produce with pieces and use FontLab mask for my own designs (and not to show another font as an example, as I thought it was designed for), the workflow improved a lot.

Following this method, I did some adjusts (for my taste, of course) and the glyph resulted this:
.


.

Even if you search for a very different result, you can see there is a regularity between the curves. And if you need to make then more oval its easy to move just some points or some handles.

The final result still needs work, but shows my idea about regularity:
.

Igor Freiberger's picture

Paulo,

other things you can consider in your development:

.


.

Overshoots usually are different for uppercase and lowercase. I use 12pt to lower and 16 to upper. I do position BCPs in overshoot just for elliptical shapes. As you see in |a|, belly goes up to –12 (full overshoot). But for serif I put the bottom BCP at –8 because the serif shape is not straight nor elliptical.

A global ruler at the vertical middle of the glyph kernel (I mean, without asc/descenders) is useful to control mirrored shapes. In this case, I have rulers to mark middle of lowercase, small caps and uppercase.

Marks named top, bottom and ogo are anchors to automatically position diacritics. Most times to position these anchors exactly at the horizontal meddle is not the best. For |a|, I put top anchor at geometric meddle of this glyph without the serif (because serif is at the bottom and has small effect on optical middle at the top). In other hand, bottom anchor consider the whole glyph width. To position anchors is somewhat related to adjust sidebarings, IMO.

Finally, you can see I used as few BCPs as possible. Every time it's possible, I use tangents and smooth connections.

This draft goes at the mask layer. I copy-n-paste it to main layer, merge contours and verify if further adjustments are need. Sometimes, FL changes slight some BCPs after merging, especially because it needs to round all positions to integers. After this, I manually add links (I prefer to use links instead of hints because them move with your BCPs in case you edit something later).

Please forgive me if you already know all this. I know you are a master in script fonts but, as this is your first sans, I believe some of these techniques may be not completely known – not to mention the possibility some generous fellow come here to make improvements on what I'm saying.

paulow's picture

I was not satisfied with the "c" development. So, I do a overall remake in it. Thanks, Freiberger, for the high valuable (and detailed) informations. What I am doing now? I am compiling the informations learned until here in a new method process (based in your advices), to resume the development of the font direct in the FontLab. The problems related by Phinney and Freiberger will now to that software, to work considering the balance and the exact rules to the standard shapes recquired to the font. I am sure, at this point, about the overall feeling which I want to the font, need, so, work to get what I want within the new problems who appear, with the help of this forum, off course, Here the new "c"

To better resolution, see the IntellectaFirstSans_mocau_18oct.pdf at the top of this topic.

Igor Freiberger's picture

Two more ideas to your |c|:

sim's picture

I'm off thread, but I'd like to say many thanks to Mr. Freiberger for is so valuable lessons. :^)

Igor Freiberger's picture

André, sorry if I annoyed you. I'll avoid to post informations which must be obvious for most Typophilers.

paulow's picture

The overall status of the font now is it. I am confortably now to go to FontLab to fix the accurate definition of the unbalanced curves, the top heads (like Phiney says), the proportion of the horizontal stems in comparison with the verticals, and other things. The last problem is the "s", which I still working, anyway. To better resolutionsee the pdf in the top of this topic.

cdavidson's picture

There's something about the d, p and q that makes them look a little odd... they don't seem to 'flow' around the corners and some of the key points on the curves look a bit square.

Also with the f and the j, I'm not sure I like how they've been 'pinched' at the 90˚ corner.

Finally with the a, try to fix the bowl, it looks like the top of it has been stretched out.

I think you're on the right track.

Syndicate content Syndicate content