Improved Charis SIL with smallcaps

[DNSB]

Not sure where that 500/600 issue with the space and the non-breaking space comes from. I used fonttools to generate the xml ttx files from the 4 Charis files, the space and non-breaking space both show a width of 600.

    <mtx name="space" width="600" lsb="0"/>
    <mtx name="uni00A0" width="600" lsb="0"/>

[JSWolf]

Originally Posted by DNSB

Not sure where that 500/600 issue with the space and the non-breaking space comes from. I used fonttools to generate the xml ttx files from the 4 Charis files, the space and non-breaking space both show a width of 600.

    <mtx name="space" width="600" lsb="0"/>
    <mtx name="uni00A0" width="600" lsb="0"/>

No idea. I got the idea for adjusting the non-breaking space from a request earlier in the thread.

What do you think of CareInk7SP?

[DNSB]

Originally Posted by JSWolf

No idea. I got the idea for adjusting the non-breaking space from a request earlier in the thread.

What do you think of CareInk7SP?

Not bad. I switched to Charis 7 shortly after it's release. I prefer my variant where I used the Medium weight to replace the Book weight since it saves me having to adjust the weight when adding it to one of my Kobos after a reset.

Other items are that since my variants were generated using fonttools, the kerning tables are not affected where FontForge complains they are too large and the modified font file size is much closer to the original font file size though the zipped files are much closer in size.

I've attached a file with the space/NBSP width set to 500 to this message.

[JSWolf]

Originally Posted by DNSB

Not bad. I switched to Charis 7 shortly after it's release. I prefer my variant where I used the Medium weight to replace the Book weight since it saves me having to adjust the weight when adding it to one of my Kobos after a reset.

Other items are that since my variants were generated using fonttools, the kerning tables are not affected where FontForge complains they are too large and the modified font file size is much closer to the original font file size though the zipped files are much closer in size.

I've attached a file with the space/NBSP width set to 500 to this message.

I did try FontForge and get the complaint. So that was a no go. I also tried an older version of FontCreator that I had installed on my Surface. That did not like the font when saving. So that was out. I did install the free 30-day trial of Fontcreator 15. That worked no problem.

[DNSB]

Originally Posted by JSWolf

I did try FontForge and get the complaint. So that was a no go. I also tried an older version of FontCreator that I had installed on my Surface. That did not like the font when saving. So that was out. I did install the free 30-day trial of Fontcreator 15. That worked no problem.

When I started getting too many complaints from FontForge, I tried several alternatives and found that FontTools was the one I preferred. In some ways, a bit more pain since you convert the ttf font into a ttx XML file which can be edited (I use Notepad++) and then convert the ttx file back to ttf. OTOH, it does allow me to automate some of the processing between the font variants.

[JSWolf]

Originally Posted by DNSB

When I started getting too many complaints from FontForge, I tried several alternatives and found that FontTools was the one I preferred. In some ways, a bit more pain since you convert the ttf font into a ttx XML file which can be edited (I use Notepad++) and then convert the ttx file back to ttf. OTOH, it does allow me to automate some of the processing between the font variants.

Is FontTools easy to use under Windows? How easy would it be to use it to add weight to fonts?

[DNSB]

Originally Posted by JSWolf

Is FontTools easy to use under Windows? How easy would it be to use it to add weight to fonts?

It's easy to use for me since I use Python a fair amount. OTOH, you need to edit the XML file to make changes which is not likely to appeal to someone more used to using a GUI. Adding weight isn't that big a deal but testing added weight is the fun part (at least for me).

[Quoth]

Yes, well you need to test on different resolutions of display?

[DNSB]

Originally Posted by Quoth

Yes, well you need to test on different resolutions of display?

And at different font sizes on the same display. Not to mention that regular, italic and bold fonts tend to react differently to the same amount of added weight. Bad intersections don't help the situation.

[Quoth]

I was surprised to discover that it's relatively common for an italic font to use a different letter form for some letters, unlike an oblique font which really is the normal just with a slant.

So certainly what works for Normal might not be so good for italic. Perhaps the bold version of normal and italic helps when designing heavier weight? Also some font look terrible in bold or on most displays (even as normal) as they need more than 300 dpi.

https://en.wikipedia.org/wiki/Garamond

This shows that the same revived typeface can be in quite different weights and have variations. The "open source" Garamond I have works on paper, but is poor on even 300 dpi eink. OTOH the Georgia I have works on most displays, as there was a version tweaked for maybe 72 dpi (certainly for lower resolution CRTs). MS has different versions of Georgia.

[nabsltd]

Originally Posted by DNSB

Adding weight isn't that big a deal but testing added weight is the fun part (at least for me).

In most cases, adding weight creates errors in the curves, like overlaps, handles that aren't in spec, etc. If the font is designed well and uses composition properly, then fixing time isn't usually too bad (you fix the base "a" once and it is fixed in all glyphs that use "a" as part of the composite), but without that, a font with just 300 glyphs can take nearly an hour to fix.

[JSWolf]

Originally Posted by nabsltd

In most cases, adding weight creates errors in the curves, like overlaps, handles that aren't in spec, etc. If the font is designed well and uses composition properly, then fixing time isn't usually too bad (you fix the base "a" once and it is fixed in all glyphs that use "a" as part of the composite), but without that, a font with just 300 glyphs can take nearly an hour to fix.

I'm not about the fix the problems adding weight can bring. I found ChareInk6SP with added weight to be OK at the sizes I used it. I don't use large enough sizes for it to be an issue. Besides, all 4 variants would have to be fixed. I'm not spending all that time fixing them.

What I will do is start with the medium versions so I don't have to add as much weight.

[DNSB]

Originally Posted by nabsltd

In most cases, adding weight creates errors in the curves, like overlaps, handles that aren't in spec, etc. If the font is designed well and uses composition properly, then fixing time isn't usually too bad (you fix the base "a" once and it is fixed in all glyphs that use "a" as part of the composite), but without that, a font with just 300 glyphs can take nearly an hour to fix.

That was why I used the Charis 7 Medium fonts to add weight to regular and italic. The last time I added weight to a regular font, it took way too much time I could have spent reading to fix the damaged glyphs. The italic variant was no better. Admittedly when I opened the original font files in FontForge, it whinged about issues before I even modified anything. Then there the sheer joy of displaying all the glyphs at various font sizes to check for issues.

[JSWolf]

Originally Posted by DNSB

That was why I used the Charis 7 Medium fonts to add weight to regular and italic. The last time I added weight to a regular font, it took way too much time I could have spent reading to fix the damaged glyphs. The italic variant was no better. Admittedly when I opened the original font files in FontForge, it whinged about issues before I even modified anything. Then there the sheer joy of displaying all the glyphs at various font sizes to check for issues.

I did try FontForge with Charis Sil 7. But when I went to save, it complained about something it should not have.

[nabsltd]

Originally Posted by DNSB

Then there the sheer joy of displaying all the glyphs at various font sizes to check for issues.

High Logic Font Creator will find all the issues and fix many of them automatically. After that, it's a manual process, but at least it shows you exactly where the problem lies.

Originally Posted by JSWolf

I'm not about the fix the problems adding weight can bring.

One of the issues with not fixing the problems is that font renderers will use more RAM and spend more time drawing. This isn't really an issue on modern desktop computers, but on a device with limited RAM and CPU, it's more likely to cause problems. Also, some of the problems like contours with the wrong direction will cause pedantic renderers to do exactly what the font says to do, which results in filled-in openings in letters. Crossing contours (which happen a lot when adding weight) can often cause a contour to become the wrong direction.

But, if you don't care if the font displays correctly on all renderers, that's fine. For me, though, not fixing the errors in a font is like not using CSS that renders properly on all readers. It's lazy and sloppy.