Handles, Béziers & Points

Fabiouser's picture

Hello typophiles

Im learning the principles of designing a typeface via software assistant. So, when I need to control beziers to make a 'perfect curve' I fail again and again. I know that failing is good, but I think that the position of points and handles have some science on it.

Today I realize something that perhaps could help me, but I dont know if it make sense, have a look:

It seem to me that handles follow a certain logic that we can find in the structure of calligraphic variations…
This make sense to you?

Where do you [typographers] learn the 'art' of place perfect points and handles?

--------
Thanks
Fábio Santos

jonathanhughes's picture

I think the only way you'll learn the "art" of placing perfect points and handles is by doing it over and over.

In Fontlab, you ideally want your points to be at the extremes of curves.

Also, note that different software behaves slightly differently when it comes to things like deleting bezier points (Fontlab and Illustrator, for instance).

Ryan Maelhorn's picture

at last, a real question

LexLuengas's picture

The positioning of the anchor points is the most significant thing to achieve a desired curve. If your anchor points are in the wrong position, you may spend hours (yes) trying in vain to get the control points "right".

To understand how to place your anchor points you must understand at least some basic principles of chirography and stroke modulation. Almost all control points come in pairs: for every point there is an opposite point (cf. the brace in your image) at the parallel curve. If you draw a line across each pair of points, you will find out their angle follows a nearly ductal logic. With designs that go apart from chirography the game becomes harder, but to be good at that you first need to fail, again and again, and I can't help you with that.

The principle is the same for the control points: Always think in terms of two lines interacting with each other :-)

I have always seen the "put your points in extrema" convention as a restraint.

hrant's picture

This is exactly what you shouldn't be doing.

The "principles of designing a typeface" start and end with the mind's eye, not some arbitrary body part holding an arbitrary marking tool.

hhp

Fabiouser's picture

@LexLuengas really taught answer :) thanks;

@hrant my words about principles to design a typeface means in software aid creation system. I do search and drawi typeforms by hand, to train my mind and my knowledge, I'm addicted to it. I just want to know how I can do some experiences in that way. I do not ask this arbitrary :). I don't give up of drawing, and I do not like to sit every hour in front of the computer screen. But, I appreciate the concern :)

Thanks

hrant's picture

I'm not sure I understood everything you wrote (and I don't know what you keep smiling at) but:

Don't confuse drawing with -what can quite accurately be called- painting. I draw by hand all the time - it can be wonderfully useful. But we should be drawing the border between black and white, not pairs of lines enclosing the black (which makes no sense functionally, hence is an arbitrary design constraint, a lazy shortcut to nowhere).

Painting only "trains your mind and knowledge" about type's arbitrary past; when it comes to what function type needs to perform, it can only distract you. It doesn't matter if you're using a broad-nib pen, béziers or a ripe avocado tied to a stick.

Related:
http://typophile.com/node/31095

hhp

Fabiouser's picture

Conterpunch…

hrant's picture

Wait, I thought you said "The Stroke" was helping you achieve perfection... Please, no romantic blanket statements. {Now edited - thank you.} "Counterpunch" is a great book, but it's no type design bible. Luckily we don't have a bible, and hopefully we never will.

BTW it might surprise you to hear that the section about the "j" in "Counterpunch" was actually what started me on the journey that led to my rejection of calligraphic type.

hhp

Fabiouser's picture

Today I realize something that perhaps could help me, but I dont know if it make sense, have a look:

I refer calligraphic forms in this node because when I open a typeface on Fontlab, I realize that the handles have some kind of position that resembles the calligraphic experiences that Gerrit Noordzij mention on his book [like the image].

I do like some calligraphic experiences from the past.

Its nice to know how books can make us reject and accept things. Its a curious phenomenon.

Ryan Maelhorn's picture

Keep in mind when designing type, all curves are psychological.

LexLuengas's picture

But we should be drawing the border between black and white, not pairs of lines enclosing the black

Hrant, perhaps you misunderstood me in my last comment. What I intended to say was that, if your design is chirographical, then your anchor points should reflect that logic you have chosen. But there is not and should not be a fixation towards a chirographical logic, or any other logic at all. I completely agree with you in that aspect.

But there is a logic between the two lines that delimit the shape of a letter. If that interrelationship is meaningful in some way, then this shape will be of some value for "the mind's eye".

However, I don't understand why you would discourage a chirographical approach by someone who has presumably just started with typography. I assume you didn't wan't to discourage the chirographical logic itself (but I may be utterly wrong), but the fixation to that kind of logic. I think it is pretty useful as a starting point, and I also think one has to be careful by making such preclusive statements against chirography, in particular in front of people who do not have any judgement yet. After all, it's better to try to understand Vivaldi and Mozart before trying to understand Thelonius Monk and Keith Jarret.

hrant's picture

The "logic" you're talking about is specifically called the "moving front" (by Noordzij), and it does make sense, in its own world.

What bothered me about how Fábio expressed his intentions was the apparent view that the moving front is how typographic forms look "perfect". To me this is totally wrong.

As for whether a beginner should start with the obvious precedents or not, the problem is this precedent is a bad one (something Vivaldi and Mozart are not). I guess it also depends greatly on the individual beginner: some people are strong enough to turn a bad precedent to their advantage; but some people get hopelessly stuck for the rest of their lives.

hhp

Fabiouser's picture

What bothered me about how Fábio expressed his intentions was the apparent view that the moving front is how typographic forms look "perfect". To me this is totally wrong.

Now I understand your 'concern' @hrant. I think that the use of the word perfect by myself, was maybe a bad choice. But my intention with this node is just to share a point of view that pop up in my mind and, with that, know if my point of view make any sense to all of you.

rs_donsata's picture

Bezier curves inner logic is in no way related to natural geometric, calligraphic or chirographic constructions.
There is a book called "Type by design" which was an instruction book for fontographer. I contains a complete description of the nature of the Bezier curve and many guidelines to achieve the best results with them:

http://www.amazon.com/Fontographer-Type-Design-Stephen-Moye/dp/1558284478

I used to have a PDF of this book but I lost it with my laptop on a fire.

Fabiouser's picture

I already put my eyes on that book @rs_donata. A school library near my school have it on the shelf.
I start too look to the ways that characters were made, and I became fascinated, I think that book make me realize some hidden secrets of typeface design. I also brought a nice book: Learn FontLab Fast, by Leslei Cabarga.

Thanks

Ryan Maelhorn's picture

I prefer to build in TTF format, becuase it seems only TTF can have those kind of curves where the anchor point floats off the line, and is not connected via a control arm. I believe they are called splines, instead of bezier curves?

Also, when it does come to Bezier Curves, I seem to strangely be wanting a way to lock down the control arms while moving the node around, which is opposite of how one usually edits bezier, where the nodes stay in one place while you're moving around the control arms.

And then we can talk about spiro...

To say the least I still think we have quite a ways to go in "curve editing" as far as techniques are concerned.

sko's picture

Both use Bezier curves, but TTF use quadratic while PostScript uses cubic beziers.

Luma Vine's picture

Ryan, what you mentioned about moving the point while keeping the control wings is exactly what I have wanted to do in vector software for some time now. I end up using guides to locate the control points and then move the anchor and reposition the control points. It would be incredibly useful to do this for the type of vector work I do for logos and such.

5star's picture

Ya Luma I do the same in Illy too. To me it helps visualize the flow of the path whilst maintaining a comparable effect.

n.

Jens Kutilek's picture

I seem to strangely be wanting a way to lock down the control arms while moving the node around, which is opposite of how one usually edits bezier

Is it the opposite of usual? I edit curves like this all the time, and I think most outline editors can do that.

FontLab Studio: shift-drag
Robofont: cmd-drag

eliason's picture

Glyphs: option-drag

cerulean's picture

Always play with the modifier keys when you're learning a program, and you'll discover all sorts of things.

In Fontlab, holding down Alt/Opt while moving a cusp node will move the node while its handles stay in place. Holding down Shift before moving a smooth node will keep the handles in place and constrain the node to the line segment between them. (Shift after you've picked up the node is your usual rectilinear constraint.) Holding down Alt before moving a smooth node will scale and rotate the handles according to the node's relative position along the curve. Alt after you've picked up a smooth node will work like it does for a cusp node, creating a false cusp that will spring back to smooth the next time you mess with the handles.

LexLuengas's picture

Has any application ever implemented rational bézier curves? I think it would be quite useful (not only for quadratic curves as shown in the article's image). Maybe it was in front of me all the time and I haven't noticed.

Ryan Maelhorn's picture

Thanks, that alt drag is handy

Fabiouser's picture

Its that think I like here: Put ignition on fire and have meal of it.

Cubic Béziers? Quadratic Béziers? Wikipedia have a really formal article on that theme; but mathematics aspects have to much emphasis… Like @Ryan said all curves are psychological.

Mathematics help us to forget psychological problems?

hrant's picture

Math? It's as real for us as the burin & file and workability of steel were for punchcutters.

modifier keys

One more - a personal fav: Alt-Shift-clicking on a curve (or even a line) creates "balanced" BCPs... although you really only want to do that to auto-add a BCP to a "half-dead" curve (which you might still have to tweak though).

hhp

Fabiouser's picture

Leslie Cabarga referes all that keys in the book I mention before. Great book

Ryan Maelhorn's picture

I doubt many of us are doing math. Anybody find themselves typing in the formula for a super ellipse lately?

I think what fabiouser is getting at is that there is a place where metrics can overpower expression, and vice versa.

Ryan Maelhorn's picture

Yep, Quadratic B-splines (2nd-order curves), is what I was talking about:

So much easier to work with. Wish all vector editors could use them.

The illustration here shows it with only one off curve control point between the 2 nodes, but you can put many more inbetween there.

Also, just checking, there is no way to work with quadratic curves in fontlab when editing open type fonts, correct?

Ryan Maelhorn's picture

Here's another editing idea: highlight the entire outline, showing all control points, and allow us to select all, or any lesser combination of the control points.

Usually you can only edit control points that are connected to a single node. even if you select all the nodes in an outline, once you click on any control point, your control gets limited to that point only, so no complex selection groups are capable.

hrant's picture

So much easier to work with.

Just FYI the great majority of people seem to prefer cubic béziers.

BTW could this be why your fonts are -what I see as- wobbly?
(Where's that thread where I was comparing your "o" to one that I like?)

hhp

cuttlefish's picture

I don't know about Fontlab, but FontForge allows you to choose between working with cubic or quadratic bezier splines, and also euler spiral segments (my preferred method, but that's a different animal that has to be converted to beziers to output a usable font). I do pity the fool who attempts to install FF. Give it a few weeks as it's undergoing a major rebuild.

Dammit there I go threadjacking again.

But anyway, my point is that there is a font editor that lets you draw with quadratic splines if you choose to do so. Seems like a pain to me, but that's what some others think of Spiro mode, so who am I to say…

Ryan Maelhorn's picture

why in the world does it still take at least a basic programming course to know how to install Font Forge? This is 2012. Ridiculous..

Ryan Maelhorn's picture

BTW could this be why your fonts are -what I see as- wobbly?
(Where's that thread where I was comparing your "o" to one that I like?)

Interesting, and a good catch. I would say that a lot of the wobbliness is due to the classic cubic spline setup, actually.

Though I prefer to build in Quadratic now, it wasn't always this way. I started off with the standard cubic spline and editing mode of the default adobe illustrator install, as I would guess most here have. Pretty much all my work that is out there was made in the cubic world.

I see quadratic as a way to remove the wobliness, actually. When I want to anyway. There is still something to be said for something that looks like it was done by hand and is not perfect, especially in todays digital world.

Fabiouser's picture

Anybody find themselves typing in the formula for a super ellipse lately?
I think what fabiouser is getting at is that there is a place where metrics can overpower expression, and vice versa.

@Ryan That's the point!

Ryan Maelhorn's picture

Well, Math, and what we are really talking about here, Counting, are ultimately abstractions. For that is their utility. That is what Counting is. A system of abstracting or reducing more complex information into a more easily manageable set. And data points entered into that system.

So in art, unlike as in science, measurement will always take a subservient role to what i conveniently call the "psychological intent" of the curve.

hrant's picture

Wash your mouth out - I do not make art.

hhp

Ryan Maelhorn's picture

I'll quote you on that. :D

cuttlefish's picture

re: why in the world does it still take at least a basic programming course to know how to install Font Forge?

That is a damn good question.

Fabiouser's picture

I think I find something that have a close relation with my original question.


It seems like the handles have some logic in that perspective…just like I point in the original question I made.
Thanks @rs_donsata to remind me that book!

Fabiouser's picture

@hrant I think too that typography isn't art. I think that sometimes we need a little of subjectiveness.

THE SHAPES OF LETTERS
The shapes of letters do not derive their beauty from any sensual or sentimental reminiscences. No one can say that the O's roundness appeals to us only because it is like that of an apple or of a girl's breast or of the full moon. Letters are things, not pictures of things.
—Eric Gill, Autobiography, Jonathan Cape, London, 1940.

Even knowing that letters are not pictures of things.

Luma Vine's picture

Is anyone else thinking that those guides all over the /S picture above are typical post rationalization? I am not really seeing the relationships they are trying to highlight, rather it looks like they found some points that happen to be almost in line and connected the dots. This is actually a much more useful version of such an idea: http://dribbble.com/shots/446385--s-construction

Luma Vine's picture

Looking again, I see they included such a disclaimer. However isn't the intersection in the middle of the /O just because it has strong symmetry? A perfect circle would exhibit this characteristic, but I doubt anyone here would say that a perfect circle is a good /O.

ahyangyi's picture

@Ryan:

I don't think you need any programming knowledge to install FontForge.
FontForge is shitty anyway. But I have seen very active development in recent days. Maybe it could be a bit stabler in the future...

Fabiouser's picture

@Luma Vine; no one is talking about if you use that way, you get the perfect character. Is just a way of seeing things that could make some sense (or not)

dberlow's picture

"I am not really seeing the relationships they are trying to highlight"

And yet, they so often exist.

"However isn't the intersection in the middle of the /O just because it has strong symmetry? "

Yes.

"A perfect circle would exhibit this characteristic, but I doubt anyone here would say that a perfect circle is a good /O."

If you look at the difference between this and the intersecting post-rationizations of a perfectly circular O, you'll see the differences are the difference between this O and a perfectly circular O.

"Just FYI the great majority of people seem to prefer cubic béziers."

Just FYI, the great majority of people have only ever worked with cubic béziers. So a "preference" for any-such only-thing they've ever worked with, is little more that a Hrantaggeration in the greater scheme of better things.

hrant's picture

It's certainly true that deep cubic habits aren't happy being moved to quadratic (human nature) and cubic had a big head-start, so maybe it's not possible to do a fair comparison. But that doesn't mean I'm wrong. :-)

hhp

dberlow's picture

"...aren't happy being moved..."

And to whom has this happened? do you ever read what you write, think about it, and then edit it?

hrant's picture

You really haven't heard anybody complain that working with quadratics is more of a pain? I admit I haven't heard it as much lately* but it used to be a common refrain. Now, are there occasional technical advantages? Of course, and nobody wants to rely [too much] on an automatic conversion from cubics.

* Which I have to wonder why. Is it because people got used to it? Maybe.

Re-read what I write? Every time. Think about it? People accuse me of thinking too much. Edit my mistakes? Only if I had a time machine - otherwise it's cheating. :-)

hhp

dberlow's picture

I asked who has "been moved" to Quadratics. No one apparently.

"You really haven't heard anybody complain that working with quadratics is more of a pain?"

About four years ago, back when TT still had secret patents Apple was using to "troll" for violators (not), and TT instructions could "Bluescreen" a PC (not), and hints were an unfathomable fur ball (sure, why not), I heard that working with FL Quads, and FL Quad conversions was lees than optimal.

But that's a FontLab thing piled on an Adobe thing. Good quads are cool and give you options you never knew you needed. Besides which, getting used to Adobe technology is just not as good an idea as it used to be.

Syndicate content Syndicate content