search

googlefonts / roboto-flex

SIL Open Font License 1.1
466 stars 32 forks source link

Cyrillic Spacing and Letter suggestions wght: 100-900 | wdth: 100 | slnt: 0 | grad: 0 | opsz: 14 #184

Open vikagrabowska opened 2 years ago

vikagrabowska commented 2 years ago

Below are my observations for wght: 100-900 | wdth: 100 | slnt: 0 | grad: 0 | opsz: 14

The paragraph set in Cyrillic LC looks slightly more dense than Latin LC. Combinations like "нн" or "пик" where verticals meet create a feeling of density, this is specifically Cyrillic issue. Adding +1 or +2 units to the sidebearings of the verticals might help to achieve more even color between Latin and Cyrillic.

Meeting descenders in цд or дд ДД also create dark spots. The space is also very close in pairs like гъ and ГЪ. Since there is no kerning in Cyrillic these and many other issues will be kerned away.

Here are more detailed comments:

wght: 100 | wdth: 100 | slnt: 0 | grad: 0 | opsz: 14 1) I'm wondering if п, ц, н could have just a touch wider proportions? This would help achieving more even color between Latin and Cyrillic. And also with some groups of Cyrillic letters, such as б,р,о,ю and л,д (they are on the wider side, but closer in their width relation to Latin) and в,а,е,я. 2) б feels slightly too large in bowl comparing to -o. 3) Looking at the relation with H I wonder if П could be just a little narrower? 4) I think ЛДШЩ also look a little wide in comparison to the rest of UC and comparing to the darker weights. This is especially noticeable for ШЩ. 5) the spacing of Latin C (tight sidebearing on the left) determines spacing for Э. However the pair "СЭ" becomes pretty tight. Same happens with "сэ" pair. 6) Ы - right stem is too thin

wght: 400 | wdth: 100 | slnt: 0 | grad: 0 | opsz: 14

1) similar observation and advice as in point 1 above. Here -н actually feels slightly narrower than п, ц. 2) и and especially ж could be tiny bit lighter on strokes 3) Ж as well 4) є,э could be just a touch lighter on the curve which is crossed by a bar 5) the the rest of busy skeletons like в,я,з,к is awesome 6) Cyrillic breve looks a little too dark

wght: 900 | wdth: 100 | slnt: 0 | grad: 0 | opsz: 14

1) combinations like нн or пик where verticals meet create a feeling of density. 2) в - has pretty thin horizontals. 3) и - could be slightly thinner on stems. It looks a little dark in text 4) м - I would look for a way to make it touch lighter by widening or/and letting the stem on the right be thinner. However I realise that this one comes from the same features of UC M, so idk if that would be possible o change.
5) у - is very narrow, it comes as component from Latin where it had related in width -v. IDK if this is possible to touch, but feels a little alien in Cyrillic. 6) Cyrillic breve looks significantly lighter when everything else 7) ж - bottom legs look darker than bottom leg of к 8) Д - left descender could move slightly to the right. 9) К - bottom leg looks a little dark, on its own and also in comparison to Latin K. 10) Ж - same comment as for К 11) У - the top of it looks pretty light (see the word ЖУЖЖА or УХА)

dberlow commented 2 years ago

Thanks for these. I made a little sample to look at. Cy and La ppgs.pdf I'm waiting on some other changes to the Cyrillic, but then the changes (2, 3, 4 and 5), to 400wght will be made.

I am not sure about the width changes being requested.

dberlow commented 2 years ago

wght: 100 | wdth: 100 | slnt: 0 | grad: 0 | opsz: 14

  1. I'm wondering if п, ц, н could have just a touch wider proportions?

    Screen Shot 2022-02-18 at 2 16 40 PM

    This would help achieving more even color between Latin and Cyrillic. And also with some groups of Cyrillic letters, such as б,р,о,ю and л,д (they are on the wider side, but closer in their width relation to Latin) and в,а,е,я.

  2. б feels slightly too large in bowl comparing to -o.

    Pasted Graphic 5

  3. Looking at the relation with H I wonder if П could be just a little narrower?Pasted Graphic 6

  5. I think ЛДШЩ also look a little wide in comparison to the rest of UC and comparing to the darker weights. This is especially noticeable for ШЩ.

    Screen Shot 2022-02-18 at 2 23 45 PM

  6. the spacing of Latin C (tight sidebearing on the left) determines spacing for Э. However the pair "СЭ" becomes pretty tight. Same happens with "сэ" pair

    1. Ы - right stem is too thin repaired direction I feel that most of this 1/3rd of an issue is wishing the metrics of the family were different, either the Cyrillic itself, or the Cyrillic alone. It’s very easy to imagine other proportions in a sans, (any sans actually). But, the reduction of weight on this style, was done very systematically to maintain the proportions of Roboto regular’s widths.

wght: 400 | wdth: 100 | slnt: 0 | grad: 0 | opsz: 14

  1. similar observation and advice as in point 1 above. Here -н actually feels slightly narrower than п, ц.
  2. и and especially ж could be tiny bit lighter on strokes
  3. Ж as well
  4. є,э could be just a touch lighter on the curve which is crossed by a bar
  5. the the rest of busy skeletons like в,я,з,к is awesome
  6. Cyrillic breve looks a little too dark Aside from #1, I made all very small changes.

wght: 900 | wdth: 100 | slnt: 0 | grad: 0 | opsz: 14

  1. combinations like нн or пик where verticals meet create a feeling of density.
  2. в - has pretty thin horizontals.
  3. и - could be slightly thinner on stems. It looks a little dark in text
  4. м - I would look for a way to make it touch lighter by widening or/and letting the stem on the right be thinner. However I realize that this one comes from the same features of UC M, so idk if that would be possible o change.
  5. у - is very narrow, it comes as component from Latin where it had related in width -v. IDK if this is possible to touch, but feels a little alien in Cyrillic.
  6. Cyrillic breve looks significantly lighter when everything else
  7. ж - bottom legs look darker than bottom leg of к
  8. Д - left descender could move slightly to the right.
  9. К - bottom leg looks a little dark, on its own and also in comparison to Latin K.
  10. Ж - same comment as for К
  11. У - the top of it looks pretty light (see the word ЖУЖЖА or УХА)

Thanks. There were quite a few incorrect angles in diagonals this style that were not compatible with 100 and 400. Those were corrected, helping solve some of your comments and then I checked and changed the rest, except for в

vikagrabowska commented 2 years ago

Hi David, many thanks for all your comments about updates. I will be glad to have a look at the changes in text. Are they already included in a current font file available in GitHub?