Wow. I was thinking about asking a very similar question. Yeah, there are some very respected type designers on this board and I, too, would love to know what your thought process and workflow looks like.

For example, do you start with a sketch or go straight to the vectors? Do you create individual glyphs in a program external to fontlab et al. (Illustrator, Freehand) then import the eps into fontlab? Do you start with any specific glyph? Do you cut and paste components of one glyph into another or draw each separately (serifs, stems, ascenders, things that in some faces, anyway, are modular)? Come on, spill it!

Aaron, mechanically you seem to be on the right track (although honestly that's not very hard), but ideologically, or perhaps even spiritually in a way, to me the key to "real" type design is grasping the part-versus-whole nature of the way in which individual signs come together to create "visible language", the way in which the balancing of positive and negative (notan) leverages human cognition to full effect. Don't ask me to elaborate any more, because I'm still in the forest myself. But at least I know I'm lost. :-)

Without this sort of meta-understanding you can still make type that many people will love and even purchase, but it will be a shadow, or better a silhouette, of what type really is.

--

BTW, there have been some Typophile threads in the past with great mechanical insight - hopefully one of our archive hounds will bring those up. I tend to have lousy luck with the archives here.

hhp

I'm not sure anybody wants to hear about my process, because I haven't actually released anything yet, but here goes.

Almost all of my fonts in the pipeline are digitizations of existing faces, primarily from the ATF catalog.

The first step in the process is to make extremely high quality scans. I'm fortunate to have an Epson 4870 for other work. I usually scan at 2400 dpi and save as JPEG with quality 90. The resulting files are around 40M per page, and are thus a bit cumbersome to work with, but I figure I'm not losing much.

Next I decide which font size is going to be considered the "master". All these fonts are optically scaled, so there is a very significant difference between a 12pt text showing and the same font in display at 48pt. Often, I'm able to make an educated guess as to which size was the original master; for example, for Century Catalogue and Bodoni Book, I'm pretty sure it's 18pt.

I then take a careful series of measurements, largely to understand the inner logic of the optical scaling. If I do this right, then I get a set of scaling parameters I can use to make the images of any size line up with the master. Of course, enlarging small sizes won't have the same quality as working from larger masters, and the corners in particular can get pretty beaten up, but I think it works. The Century Catalog is drawn from originals ranging all the way from 12pt to 36pt, and I feel like I've mostly evened out the inconsistencies.

To start work on an individual glyph, I start by choosing one or more exemplars. This is actually one of the toughest parts, as the printing quality can vary widely between the different samples of the same font, especially if there are differences in paper.

There's a new trick I usually use these days, which is to carefully align multiple scans of the same letter, then average them together. Currently, I do this by hand in the Gimp, but I expect I will write a tool to automate this for me. It's labor intensive, but the results can be striking.

When I finally get an image, I start tracing. Most of the fonts are done in FontForge, but I've just now started using my prototype Cornu spline based curve editor. However, that tool is lacking much functionality, so I'll probably be stuck moving the outlines into FontForge for a while to make them into a real font.

I spend a lot of time nudging the outlines until I get the curvature to look just right. I usually make test prints, both at the size of the original so I can compare them, and at around 200pt so I can audit for problems. I have become somewhat fanatical about matching the original. For example, I invite anyone with a 1929 Centaur specimen booklet to compare the 72pt printed metal with my draft. I'm inclined to soften the points on the 'a', and make the stroke on the 'e' heavier, but both of these quirks can be found in the original.

I do not tend to cut and paste a lot. Most of the fonts I'm tracing have enough individual variation that copying doesn't really save time.

The last step is the spacing. Most of the fonts in my pipeline need a lot more work on this. I have been doing it pretty much by seat-of-the-pants, but I expect to be following the procedure outlined in Chapter 10 of Tracy's "Letters of Credit".

Now maybe one of the great designers can weigh in. Heck, I'll settle for a good one :)

How is the typeface on your tea box ?
Aaron, you have posted here, so this means design
is not dead. Also it is not corrupted by death because
wit is the sudden marriage of ideas, which, before
their union, were not perceived to have any relation.
Think positive. How would you design the weather today?
Here is a document on Vox's classification:
http://ellie.rit.edu:1213/pdf/appendixc.pdf
Check Kris Sowersby's site : http://www.klim.co.nz

Alessandro


PS/

Designing a typeface
http://www.typophile.com/cgibin/show.pl?4100/47342

History

abc Typography

Aaron and Noah,

Like Ralph, I have not released anything yet. I am just a beginner, but from what I

Question: Fontographer is old and hasn't been updated in ages. For this reason it is probably easier to learn than Fontlab, which I've had a chance to mess around with and found very confusing. But by the same token, is FontLab the de facto standard with capabilities not present in FOG, like Open Type support? Which program, once learned, makes for the best tool in terms of an ideal end result with a minimum of fuss?

Noah, I strongly recommend that you forget about Fontographer. It simply is not worth learning this I suggest that you either start with TypeTool (the cheaper, stripped down version of FontLab) or go direct to FontLab. (You could also download the demo of DTL FontMaster, in case this appeals to you more.) I recommend getting a copy of Leslie Cabarga's Learn FontLab Fast, which will make learning to use FontLab much easier.

You'll find that different type designers have different approaches to the design process, and also that processes might vary depending on the typeface itself. For instance, I do almost all my work directly in FontLab, with very little drawing on paper, while some other designers always do quite detailed drawings. If I do drawings on paper, they tend to be quite rough, and are just to help me get possible shapes clear in my head.*

If you are going to start your design on paper, which is a good idea when you are beginning, do not feel that you need to produce high quality, 'finished' artwork. It is nice to look at lovely outline or inked drawings by great draughtsmen like Hermann Zapf, but these need to be understood as artifacts of industrial manufacturing processes. My recommendation for digital font development is to follow the older practice from the days of the manual punchcutter: make only rough drawings, and allow the refinement to come out in the outlines you make on the computer.


* If I'm designing for a script I have not worked with before (I do a lot of multilingual type design), I tend to spend time writing it with appropriate instruments, in order to understand the development of the script relative to the tools with which it has been written.

Fontographer is not capable of producing what is today an "ideal end result." FontLab on the other hand probably doesn't produce results with "a minimum of fuss," either. DTL FontMaster is rather expensive, placing it beyond the reach of most hobbiests and folks just getting started.

Overall, I'd suggest getting FontLab, or getting TypeTool if you can't afford FontLab.

Beyond that, what John said. Particularly about the design process.

For myself, I fundamentally design in my head. I may make some sketches and such to clarify my ideas, but I don't actually bother scanning these in. Like John I tend to then sit down and work directly on the computer.

I have noticed that *on average* the question of how one approaches design is a function of age and background. Type designers who were born before 1960 and/or have significant calligraphy or hand-lettering background almost all tend to draw things out pretty carefully first and digitize later (perhaps much later). Of type designers born in 1970 or later, without any hand-lettering background, many work directly on the computer without starting a design in analog form.

Cheers,

T

>*on average* . . . how one approaches design is a function of age and background. Type designers who were born before 1960 and/or have significant calligraphy or hand-lettering background almost all tend to draw things out pretty carefully first and digitize later<
Just to show you that the average Thomas speaks of isn't an absolute, I was born in early 1944 and have a caligraphy/handlettering background. I rarely draw letter-forms on paper at all anymore. I either start straight in FontLab or Illustrator with a blank screen and just draw. Some of us geezers just mess up the rules :-)
The point is, it does not matter a hoot how anyone else works, do what ever works for you. Try several ways. However you start, you will learn from it. You can change your work process later. I did, I used to work with just 2 brushes and black and white paint. The important thing is to learn to "see". Work, then look, then fix, then look, then tinker. You will learn the infinitesimal amount of possibilities that exist in a few points of space.

ChrisL

I reckon the first thing you should do is figure out what it is exactly that you want to do. Jumping in and fiddling is good for a while, but pretty counter-productive in the long run. Set yourself a brief, give yourself some restrictions. Find a problem; treat it like any other design problem. Look very closely at what others have done. The list goes on!

Best of luck Sir.

kris.

I enjoy doing detailed sketches, I find it

I would like to recommend a good book:
"Creative lettering today" by Michel Harvey.
It's good. It also have a chapter on type design and even stone carving.

As pointed out above, we all tend to go about this in different ways.

I draw exclusively in FontLab with a Wacom tablet. I say exclusively
because it's important to avoid the Illustrator/Freehand route. I know,
I know. Some of you will have a beef with this, but it's a terrible habit
that only leads to sloppy outlines and even worse interpolations.

When I'm stuck, I use any number of broad nibbed pens that I keep
around to work out the logic of forms. Most often the italics. Any
sketching I do is usually very rough and I can't remember ever using
a scanner to start an outline.

As for the age connection, I was born in 1974 and had my first
Mac around '84. I think my elementary school had computers by
around 1980 or so?

Mine is a somewhat confused process...

I was born in '75 but studied architectural drafting fairly extensively; There's no way I could ever draw an entire typeface on paper (I wish I could though), instead I've found myself using a technique where I'll draw (on paper) two or three or four "key" glyphs, scan them, put them as a background in FontLab, and simply manual trace from there. At that point (and after some polishing of the shape in FontLab) I can interpolate the shapes I want in the other characters from the shapes I drew and scanned.

This works for me, and I find it very relaxing and meditative (nothing beats seeing those smooth-as-silk vectors take their place over that rough bitmap ghost).

I've been doing something kinda wacky lately, too. I'll photocopy-magnify my handwritten (ballpoint and broadnib pens) glyphs, polish them up with a black marker and a whiteout pen, scan them, and do the routine that I laid out above.

Again, it works for me, but we're all different.

here's a great thread:
http://www.typophile.com/forums/messages/4101/36168.html?1087537507

>> Again, it works for me, but we're all different.

Absolutely. It's really interesting to compare the approaches above.
I mean, Raph compared to Thomas is pretty massive.

One more thing. It's really important to draw all the time - with
a pen or a computer. It sounds simple, but by constantly drawing
you're always re-thinking your decisions and hopefully refining them.
Type design is after all a series of decisions.

Unless I'm digitizing an existing face, I start with lots of sketches and doodles. Nothing very finished-looking usually, and I often use lined paper (legal pads) or graph paper. At some point, I have a solid idea in my head what the font looks like (or ideally looks like) from doing the doodles and sketches.

Then I start working directly in FontLab, sometimes putting scans of the sketches in the background as a reference. I generally don't literally trace the scans, but they are a big help in getting proportions correct. Plus it helps psychologically not to have to start from a blank screen and to have some concrete idea what I'm shooting for. I am capable of doing very tight drawings--my early lettering work was all ink on board--but I find it much more efficient and flexible to do the tight stuff on the computer screen. The sketches are really more like visual notes than artwork.

I worked in Fontographer for years using Illustrator for drawing, and then copying and pasting between them. This was mainly because I disliked Fontographer's drawing environment. I'll second Eric's advice to draw directly in the font program. Since I started doing it, I can't believe how much more efficient it is. (And thanks to Eric and some others for encouraging me to give it a try. I used to be so hung up on the difficulties of moving art from Illustrator to FontLab, it kept me from dumping Fontographer for years.)

Oh, and for the record, I was born in 1955 and started using a computer to design type in 1985 (Fontographer 1.0!) after trying to do it on paper since the mid-seventies. (Although I did dabble a bit in screen font design before 1985.)

BTW, Raph: I like your taste in the fonts you chose to revive.

This is becoming quite the keeper thread!
Very nice of people to share all this detail.

hhp

When working on the computer do you prefer a tablet or a mouse? I've found that I am most comfortable doing graphic work with my wacom, but most of my experience is with bits not vectors (photoshop rather than illustrator).

I usually feel very dissapointed when I can

thank you very much for all the comments and guidence in some cases, that you have given me...

I have really enjoyed reading all the posts - I find it so amazing that there is a wealth of experience brought together by this website...

Oh yeah - I ended up doing the course - under articles and courses on this site - GREAT stuff -

__________________

thanks again to everyone that has posted - hopefully when I get some type designed I can show you all...

Until then take care