This copy was truncated by the archive that captured it, so the thread may be incomplete.
I echo what Christian said.
Additionally you might want to know that many things that are totally okay in illustrator in terms of point placement, Bezier angles, etc are not going to be so okay in a font. That is because the engine that interprets illustrator files is much more complex than the many & various engines that interpret font files. Disregarding this reality can cause weird bugs, printing errors or even no printing. Fontlab is more 'constrained' to help you keep your font from being a menace. ;-)
If you are going to just make one font you can look up Adam Twardoch's tutorial/explanation about how to make your letterforms go from Illustrator files into Fontlab and skip some of the learning curve. You will probably want to run some Fontlab scripts/routines on your outlines though. Like 'place points at extrema' & Maybe 'simplify' & so on. That will help.
However, if you want to make say... 5 fonts or more I suggest you embrace the learning curve and go 'all Fontlab'.
The other thing is yo might want to post some samples on Typophile for technical critique. Or other kinds if you want that too.
Eben,
I do want to embrace the learning curve in Fontlab but I'm finding it difficult to do what would be really simple in Illustrator.
In a post form early today I was asking about aligning two outlines, in that case it was the slash form the Oslash. In this situation is it a case of having to center the slash by eye or is there a method of centering the slash on top the O gylph.
The soultion I was using involved placing the slash as a seperate glyph and then adding that glyph as a compent which I could then center, then decompose and finally deleting the slash gylph. I'm sure there must be a better method than this.
Rr
Rachel,
It sounds as if you are new to type design. Once you are in the mindset to make a typeface (as opposed to simply drawing a bunch of outlines), more and more of FontLab will make sense. I repeat what I've said here many times before: You find FL difficult because you're used to Illustrator. But Illustrator can't make a font. FL can do everything. The simple solution is to commit; go through the learning curve with FontLab and discover the many many advantages to having all your type design tools in one place. Once you've got experience with both, you discover that niether is "better" or easier to use; it's just a dead end to keep drawing letters in Illustrator when you have to move them to FL to make a font anyway. Aligning, spacing, kerning, are all operations so basic to type design, and yet you can't do them in Illustrator. Don't believe that the same operation won't be as simple in FL; it's just that you're comfortable in Illustrator right now. In a week of intensive use you can get comfortable with FontLab.
Making an accented character is typically a more methodical process. Accents are useful for more than one accented character: make an accent and then use it in composites for a whole bunch of accented letters (agrave egrave igrave ograve ugrave etc.). You don't ever have to decompose them, and it's a time-saver to have the accent by itself, available as an ingredient to make more accented characters from. Composites allow you to make changes to the accent or to the base character that are propagated in the composite character. If you use your current method you would re-do a lot of work.
crossgrove,
I understand I need to leave Illustrator and start drawing in Fontlab, and I understand that fontlab containes everything I need, but I'm just struggling with some minor points.
The example I talked about using composites wasn't creating glyphs with accents as it would normally be used, but it was a work around I devised to try and solve the problem of centering the slash in the Oslash glyph.
Everyone ha sbeen really helpful here but I still carn't work out if Fontlab has a method of centering such things or if it's best doing by eye.
http://typophile.com/node/34540
Rr
Sometimes the reason things are different in FL won't be clear and then you will realize 'why' later & sometimes Fontlab may just suck in one way or another. The reality overall doesn't change though.
Moreover, while FL may have a tool for centering things, even if it does you may well be better off doing it by eye. Mathematically correct is often not right when it comes to the eye.
For more on this see this:
But especially this
With regard to the mathematical vs optical, the O and o slash actually are good examples. Note that they in effect have two of the 'x' style illusions in them of the kind that Briem mentions. As Mark Jamra notes here, to make it look 'right' you may want to rotate the middle line relative to the others. In this case what 'centers' can only be done by eye, particularly if your O and o are not vertically symmetrical.
I like the drawing tools in FontLab, but I'm much faster in Illustrator. Too bad there aren't tutorial videos for FontLab. It would be cool to see an expert work the tools in realtime. I may fully convert at some point, but I haven't really been frustrated with pasting the glyphs from Illustrator into FontLab. It's easy. It's fast. The only thing that frustrates me about FontLab is kerning. If I can figure that out, I'll be set.
I'm so very used to (guess what?) CorelDraw I always draw vectors in it. Its quite easy to import into FontLab if you follow these steps:
1. Measure your type in Corel appropriately (to FL's 1000 units)
2. Draw all your glyphs
3. Export each glyph into EPS
4. Open FL
5. Edit font properties - all guides should be easy if you measured glyphs in Corel appropriately.
6. start importing glyphs. Vectors should be perfectly preserved if your scale was 1:1
7. Do the font stuff like kerning etc.
It's the fastest way for me because I've been using CorelDraw extensively since version 3.
Rachel:
I was like you the first time I launched FL to draw letters. I was used to work with Illustratror since is first version (88). I did some switch in the beginning (draw in Illustrator, import in FL, go back and forth between those two of them...). However, I learn FL and his tools, with the Cabarga manual and now I find FL is very in first place and works better to draw letters.
Rachel,
Your recent message is the best post in the thread! Congratulations. You work quickly!
Not-so-comforting thought: FL has some severe bugs and non-intuitive functions, but overall I marvel at the power it has, for things I always wanted, like putting an entire font in the background, and building all my composites in a single operation. It's deep like Photoshop is. And I do expect FL to improve; the programmers are very nice and very talented and have been absorbing a lot of feedback since Studio 5 came out.
Russ,
The time saved is not drawing in Illustartor and then having to copy into FL, it dosen't sound like much but when I'm copying 200+ glyphs, the time adds up. The points in fonlab will only snap to the grid (I hope so anyway), but in Illustrator even with snap to grid on and the grid set up correctly I found that some points will move when copied into FL. I found it really frustrating that I spent the time drawing in Illustrator to then have something slightly differnt in FL.
I've only just started drawing in FL, so I'm no expert but I found it really simple to pick up. I think I've finally realized that drawing in FL cuts out the middle man. Crossgroves first post on this thread sums it up for me.
Rr
Thanks for the response Rachel. I've noticed only very slight differences in the shapes of my glyphs when I paste them back to Illustrator from FontLab - nothing drastic enough to worry me. Are you seeing big differences? Pasting the letters in does take some time, but it actually goes pretty fast for me. It's a drop in the bucket compared to the time it takes to kern. I also like the wide open space that Illustrator gives to work in, especially when I'm experimenting with alternate characters.
Leslie Cabarga explains that the basic difference between Font drawing programs and Illustrator is that when you delete a point, the line doesn't collapse in, but springs out. This is useful for shaping curves, as the 'sprung' curve will have a simpler geometry, with no lumps. You can then decide whether you want the extra node or not.
I really like the editing tools in FontLab, like removing a redundant point and still keeping the integrity of an arc. I also like how you can nudge bezier handles with arrow keys. On the other hand, I'm pretty good at drawing curves accurately the first time, without extra points. Cleaning up the inevitable mess of an autotrace is not my idea of a good time.
> But there's no way you're pixel-accurate.
Pixel-accurate? Compared to what? My scanned pencil sketch? You're absolutely right. I draw the glyphs to the point where they look good to my eye. There is a very slight shift of some points when I paste the glyphs into FontLab. The amount of shift is generally acceptable to me, and if it isn't, I can tweak it.
> Using Illustrator to draw glyphs is virtually never a good idea.
This is a general statement I might agree with too, if I could get someone to elaborate beyond general statements.
Thanks for the clarification Hrant. When possible, I tend to recycle parts in order to maintain consistancy. In your example I would probably clone the bottoms of the stems on the "m" to keep the widths consistant. I'm guessing other people do this too.
I mentioned the curves to show that I am not autotracing. I draw the glyphs with as few points as possible, so I don't end up with a lot of editing and cleanup.
One more thing. FL is probably much better when you do the outlines point by point. But I normally draw glyphs using lots of welding/trimming/intersecting. I can reposition each part separately and then do the weld on all of them and do the final tricks on the outline itself.
So when it comes to some geometric sans serif it's rather easy to take one part of a letter and use it with another. I don't really know if this would be the case with FL. But if you draw your letters on paper so good you scan them and do the outlines in FL that's as you all agree much better and easier with FL.
" I still carn’t work out if Fontlab has a method of centering such things or if it’s best doing by eye."
" the problem of centering the slash in the Oslash glyph."
I was talking to Petr about this recently and realized Fog had this right. Fog allowed one to select any number of points, type cmmd-= and the "reference point" would jump to the x y center of the selected point(s). With this simple tool, centering one shape on another was near instantaneous with either a visual or mathematical precision in mind.
FL does not have anything like this, does it?
Cheers
> with [...] mathematical precision
My background is in mathematics, so I will jump in. If the counter is perfectly circular then there is a center. If not, the counter has a "center of gravity" i.e. a point such that if you cut the counter in straight line through it, both parts have the same area, the same "amount of white". If you select a few points on the boundary of the counter and get their center of gravity, it won't be that of the counter. Moreover, even if you have a program that gives you with mathematical precision the center of gravity of the counter, as soon as you draw a thick line through it, the two white pieces in general stop having the same area, i.e. the same amount of white and I understand (please correct me if I am wrong) that what is important is that the amount of white under and below the slash be the same. Moreover, other issues due to the Poggendorff optical illusion need to be taken into account. So the question is not a simple mathematical problem. If we forget about the necessity to correct the optical illusion, then we would need a program able to calculate the area above and under the (thick) slash; is there such a thing on the market?
"(please correct me if I am wrong) that what is important is that the amount of white under and below the slash be the same"
Okay.
The amount of white below the slash is greater, slightly.
This is what made the fog method perfect.
The slash was centered on the inner counter,
and then nudged to visual certainty.
Cheers!
I guess it would have been better to say that the amount of white above and under should be perceived to be the same, but for me "visual certainty" is problematic: in Briems' site cited above it is said that for most people the 8 that is checked has a lower half that looks smaller than the upper. For me they look the same and in the fourth, the lower part looks much larger.
I have lived all my life with eights having an upper part that looks smaller than the lower part...
>I have lived all my life with eights having an upper part that looks smaller than the lower part...
In general whenever a character is divide in two, it looks more 'right' when the top is smaller than the bottom. Partly, there is the visual illusion Briem notes where eg the bar of the H (or the O slash) has to be higher than the center to make the top and bottom look equal. But more than this I think we react as if the letter is an object with weight, and things are more physically stable when the top looks smaller, not equal. Whatever the reason this holds with many horizontally divided characters, such as the B, S etc. Visual equality is what you may want in an H or O slash, but often not in an E and 8, etc. In almost every character, the mathematically simplest shape and proportion looks awkward or ungainly.
>But critically, not during immersion.
You have a good point, but I think 'aesthetically' rather than 'critically' is more accurate.
Readers don't generally set out to criticize a type face. They normally take it totally for granted. But they are strongly affected by aesthetics during the first glance at a page, and particularly the titles, headings, and layout. Then when they read extended text the face should 'disappear' as they focus on meaning. During orienting glances is when aesthetically pleasing shapes and proportions have their impact.
Yes, there is no pure, timeless immersion, and aesthetics always matters. But we should not ignore the "dark half": that during immersion "the aesthetics of gravity" is moot (which is something besides "readers don’t generally set out to criticize a type face"). This distinction is critical because it helps us strike better balances depending on what kind of thing we're making, like a text face vs a display face.
hhp
>kind of thing we’re making, like a text face vs a display face.
I guess here is where we have some difference of opinion.
Unless a face is intended for tiny sizes only, like an agate face, text faces also have an element of 'display'. Even if they are at text sizes, the aesthetic aspect, including proportions, has an impact each time we have a first look at a page, or perhaps even a paragraph. I agree that the beauty of type at heading and titling sizes is more important. If the face is done in optical sizes, then the heading and titling sizes can successfully put a bit more priority on looks and a bit less on readability.
Yes. ScanFont 4 (and upcoming 5 which will be also available for Windows) can import many glyph images from single AI file. You can also customize scaling which will round to UPM grid only on very final stage without losing any precision.
Both BitFonter and ScanFont 4+ have outline layer - it is not editable, but can be sent to FontLab (or TypeTool).
I just came across this link in another thread.
http://groups.msn.com/fontlab/tipsandtricks.msnw?action=get_message&mvie...
anyone have anything to add? or does this pretty much cover what needs to be done?
Drawing letters in Illustrator is very simple if you have lots of practice. I've been doing it since 1989 on a Mac of little substance. If you have the "eye" for it, you need not use many beziers, only the minimum necessary.
I have worked with artists who use many to quite a few bezier points, so I know what many of you might be suffering through. I would love to give lessons on the minimalization of the bezier-drawn letter. It is so easy once learned.
Having said that, how well would a letter with the minimum number of beziers transfer to Fontlabs? I am interested in creating some of my own ideas.
Captured from commoncrawl.org on 18 Apr 2015. original URL