Closed ischaap closed 1 year ago
The problem with the cross-bars in the italic is that some characters haven't yet been hinted. I'm taking care of that.
I think the fy and ty ligatures are impossible. I thought they were a good idea when I put them in, years ago, but there's no way to make the bar of f or t join up gracefully with the left-hand side of y. I think it's time to retire these.
There's something wrong with the OT features in italic, so that some ligatures aren't getting formed correctly. I'll fix them.
Junicode didn't actually have an f_f_r ligature, but after checking on the frequency of that sequence, I decided to add it. The _y ligatures have to stay, since they're in MUFI, but I doubt that I can make them look much better than they do now.
Thanks for checking on these. I think of the italic -y ligatures, the current fty looks quite good. If the others could match that I think it'd be a success.
I can adjust the italic, but the roman has the intractable problem of the bar of f or t being far thicker than the serifs of the y. One way or the other (thickening a y serif or thinning the f or t bar), it's going to look absurd. That may be why fonts in general don't ligature f or t with y.
Hmm, I see your point. But that said, I do think the current roman ty and tty look pretty good; if fty and the various f-y combos were like that I certainly wouldn't complain.
I'll think about it. But I've never been anywhere near satisfied with those roman _y ligatures.
I was crabby yesterday, desperately trying to hunt down a seemingly intractable bug. Still haven't located the bug, but these ligatures seem okay:
These look great in v. 2.000beta7. Thank you!
Whoops, premature closure, apologies. It looks like maybe the hinting isn't quite there yet in InDesign for certain italic f-f combinations (bottom row in screencap).
Also, I'm wondering whether the treatment of the f's crossbar is meant to be different in f-l vs f-b, f-h, etc.
People are used to seeing the bar of f meet the l in the fl ligature, but not elsewhere. So yes, that one is deliberate.
I'm seeing the hinting problem in LibreOffice on the Windows machine, but not in Word. It's hard to be sure what's going on without more info. What app am I looking at? What is the size of the type, and what is your screen resolution?
I take it you're using the static TrueType version of the font? That's auto-hinted. If the auto-hinter is messing up, I will override it for these glyphs.
The two screenshots in my last post showing the hinting problems in italic were taken in InDesign. The screenshots were taken this this morning in the static ttf 2.000beta7; but this evening I'm still seeing the same issue with static ttf 2.000beta8. Type size is 12 pt Medium; my screen resolution is 2560 × 1440.
It is the auto-hinter. It'll take a little while to fix: I'll close this issue when it's done.
This is done now (though the fonts are not in the repository yet).
I'm noticing a few issues with discretionary ligatures at all weights: roman ffy and fty don't look quite right; the horizontal stroke that should run through all three letters becomes distorted between the final two letters of each ligature. In italics, meanwhile, the final y of ffy and ftt is completely separated. Also in italics, the horizontal stroke of tty doesn't lead into the y as smoothly as it does in fty.
The above is from Word; in InDesign the problem roman ffy and fty is the same, but the problem with the italic ffy and fft manifests differently, with a misalignment of the fs:
Roman fy shows the same distored horizontal stroke as ffy; and italic ty shows the same issue as tty, but perhaps more pronounced:
Finally, in roman ffr, the r is detached; in italics (as seen in InDesign below), the fs are misaligned as well.
In word, fi, ffi, fj, and ffj look good in both roman and italic. But in InDesign the italics are misaligned:
Ultimately it seems that in InDesign (but not in Word), ff by itself and in all combinations seems to be misaligned. At extreme zoom levels, the misalignment is less perceptible, but I think it's still there. I haven't yet been able to test how these issues look in print, apologies.