Multiscript critique

guifa's picture

I've been working to add a number of new characters to my font. In the file I'm attaching, I've got one page each from Homeric Greek (polytonic), Golden Age Spanish, Old Church Slavonic, Anglo-Saxon, Vietnamese, and Armenian. How well does the font hold up in each one? Is there anything that strikes anyone as odd in any of them? The last page includes some random quotes from Wikipedia to show off how each script "feels" in the presence of the others. Any suggestions?

AttachmentSize
TypeSample.pdf76.09 KB
TypeSample2.pdf81.13 KB
glyphList.pdf620.58 KB
guifa's picture

And hrm, it appears font forge messed up a few Cyrillic letters again. So a few of them like л or и have some odd distortions. Also hinting (at least on the Mac) seems to be a bit off, but zoomed in looks fine.

«El futuro es una línea tan fina que apenas nos damos cuenta de pintarla nosotros mismos.» (La Luz Oscura, por Javier Guerrero)

Number3Pencils's picture

So these are meant as complements to Berling? Okay. Let's see what I can say about them.

I'll start with the Greek. The contrast is one thing that's getting you. You have thins that are way too thin (iota, mu, delta, psi, tau, epsilon, eta, finial sigma) and thicks that are too thin (phi). A lot of your letters look warped, like the mu, upsilon, delta, alpha. The kappa has overhang. The nu has a terminal with no relatives in any of your other letters; the gamma and the terminal of the bottom beta bowl are similarly alien. Also, I think you should have just one stress scheme in the font--the lambda and chi have old-style stress, but the rest of the alphabet is modern. One thing that's bugging me throughout the alphabet is the outstroke/instroke, the one with a little bit of thickness on the end (iota, tau, pi, eta, upsilon, and others).

In all the scripts, I'm seeing more thickness on the right side of the o than on the left. (omicron, o, Russian o).

I can't really offer any critique on the Russian until I see a sample that isn't garbled.

The ð should be designed of its own accord, not put together from an inverted 9.

I can't really say anything about Vietnamese or Armenian, though I imagine Hrant might come by and say something about the latter.

guifa's picture

Nice catch on it being based on Berling. I've fallen in love with the font and been working on it since.

What is a better source to use for the tails in Greek? I was working with a slightly smaller O and chopping at the extrema on the right. It's my first attempt at anything non-Roman so I figured I'd still screw up a lot, although most of the native Greeks (granted, not type face designers, but do programme localisation) that saw it didn't notice anything. How would the chi and lambda look in a "modern stress"? A much thinner top left to bottom right stroke?

Incidentally, the os are unchanged from the original Berling (they are just set as references to the Latin o).

«El futuro es una línea tan fina que apenas nos damos cuenta de pintarla nosotros mismos.» (La Luz Oscura, por Javier Guerrero)

Number3Pencils's picture

1. Garamond Premier has classical stress. 2. Minion has modern stress. (Note which stroke is the thick one in each letter.) I don't know if these are the broadly accepted terms for the two kinds of stress, but they work okay for me.

As for a "source" for your tails, I can't recommend copy-pasting it from anything in particular; much better is designing it freehand. The "a" might be a good place to get inspiration, but definitely don't copy it. Design it to look good; that's all I can really say.

guifa's picture

Okay I've uploaded two new files. One of them is the test file without cyrillic distortion as well as another sample text in modern-day Bulgarian since the old Cyrillic doesn't quite look the same.

As should be obvious in the glyphList file there are a number of characters still in the process of being made, I just took the output directly from FontForge's print entire font function.

I think the tails look a lot nicer now, I realised I had them only about half as thick as they should be (basing the thinnest part of the tail on the thinnest part of the lowercase O). I also gave it an ever so slight rounding off on the edge and that's made a big difference from my POV. The eth has also been updated.

«El futuro es una línea tan fina que apenas nos damos cuenta de pintarla nosotros mismos.» (La Luz Oscura, por Javier Guerrero)

hrant's picture

Berling is actually a pretty "progressive" design... for Latin. Which is why Bill Hill chose it for the original MS Reader. But its extreme vertical proportions are pretty much guaranteed to cause problems in other scripts, including Armenian.

BTW, I know I still owe you a proper review of your Armenian.
Especially since I admire the fact that you're working on it at all!

hhp

hrant's picture

What the hell, things are quiet on the home front - I'll just do it now.

BTW, I'm going just by your TypeSample2; your glyphList isn't coming up
for me, which means I'm almost certainly not seeing all your characters.

The main structural problem is where you're missing parts, like the rightward bars at the bottoms of za and liwn. A much less obvious but almost as important structural issue is the severe shortness of the descenders (coming from Berling); besides making the result less readable and less culturally authentic, this also results in some letters looking pretty horrible, like the sha, cha and jheh.

Also, the following letters are malformed:
- yiwn is way too wide.
- ca is just mangled.
- eh needs to be wider, with a stronger descending curl.
- ja is a Latin pastiche.
- What's that flick in the AYB?

There's issues, like the ho being too Latinized,
but those are on another level of refinement.

Sorry this took so long!

hhp

Syndicate content Syndicate content