Dosis variable font remastering: progress log

[Eli Heuer]

Progress report 2018-09-06:

I'm starting work on Dosis, the working Git repo is here: https://github.com/eliheuer/Dosis/tree/vf-mastering

With minimal effort, I was able to get FontMake to build Dosis as a variable font, but there are many errors I need to look into.

Here is a GIF that shows testing of the initial FontMake output:

[Eli Heuer]

Progress report 6-17-18:

I'm making progress on Dosis and think it's close to ready for review:

The working repo is here: https://github.com/eliheuer/Dosis/tree/vf-mastering

And here is the current specimen: 

For the most part, everything seems to be interpolating well. There are a few rendering issues with the outlines I need to look into and fix, for example "A" when used as a component has problems:

I've been using a shell script to help automate all the operations involved in making a variable font starting from a .glyphs source up to the font bakery test. The script is here: https://github.com/eliheuer/build-vf

Today I'm fixing check/65, equalizing kerning between ligated and non-ligated sequences. 

[Eli Heuer]

Progress Report 7-18-18, Part One: 

 

I'm working on fixing some visual issues, for example:

I also made some new issues on GitHub this morning: 

WARN: check/065 -- Is there kerning info for non-ligated sequences.

I was able to fix all non-ligated sequence kerning WARNs except one.

 

Running https://github.com/googlefonts/fontbakery/commit/a7335a9f3db284e5773e4a14b6479f4af99649c3 with python3 and the web UI I get the following error:

 

* WARN: GPOS table lacks kerning info for the following non-ligated sequences: ['N_o'] [code: lacks-kern-info]

 I made an issue about this here: https://github.com/googlefonts/fontbakery/issues/1977

 

WARN: check/064 -- Is there a caret position declared for every ligature?

I'm getting a warning about the font having 1 ligatures, but only 6 sets of caret positioning values.  Neither of these values is accurate and I have filed a bug here: https://github.com/googlefonts/fontbakery/issues/1976

[Eli Heuer]

Progress Report 7-18-18, Part Two:

Dosis originally used .vfb files as source, then the designer moved to project to Glyphs. When starting this variable font conversion I used the most recent Glyphs file.

While Using the GFRegression tool I noticed that some of the glyphs in the latest Glyphs file are significantly different from the version of Dosis currently hosted on GoogleFonts. 

I'm going to use the latest drawings because in most cases they seem like improvements, but if anyone thinks I need to stay closer to the original design, please let me know here. 

One example of this is the Latin uppercase "W":

[Eli Heuer]

Progress Report 7-19-18, Part One

I'm using GFregression today to test Dosis-VF against the version on GoogleFonts.

I needed to test some things with css font-variation-settings so I made a very simple web specimen for Dosis with CSS Grid.

A version is is online here: https://freefonts.neocities.org/dosis.html

And the source is here: https://github.com/eliheuer/freefonts.neocities.org 

[Dave Crossland]

I think having this as a ss01/salt would be nice

[Eli Heuer]

Progress Report 7-20-18, Part One: 

So, Dave noticed yesterday that the source I was working from(the latest glyphs source) was too different from the version on google fonts, so I made a new version with fontmake from an older source file. We discussed maybe making the latest glyphs source a new typeface called Dosis Display. 

Many of the drawings in the most recent glyphs files seem to be improvements to my aesthetic sensibilities, and I think if Pablo Impallari ever takes up working on this again he would want to start with the recent glyphs file. 

I updated the web specimen to show all three versions side by side: https://freefonts.neocities.org/dosis.html

Here are some of the glyphs with the biggest differences between versions:

And a basic view of all three fonts:

I'm moving ahead trying to get the version that is as close as possible to the version on font.google.com done, hopefully Mark will have time to do a review next week. 

Also, there is a slider half implemented on the specimen site, so look for that soon!

[Dave Crossland]

Nice! When I look at https://freefonts.neocities.org/dosis.html then I think the changes are maybe not worth having a Display family for... but I believe Marc's GFR layouts should be able to inform a better decision :) 

[Eli Heuer]

Ok, I'll just focus on testing and updating the new Dosis-VF.tff that fits the fonts.google design better.

I added a slider to the web specimen, so that should make testing much easier: https://freefonts.neocities.org/dosis.html 

[Eli Heuer]

There is an open PR for this here: https://github.com/google/fonts/pull/1733

Marc says it can't be merged until ttfautohint has better VF support, or it might be a candidate for VTT hinting.

[Eli Heuer]

Micah is working on VTT hinting for this now, so the PR should be ready to merge after that. Thanks Micah!

[Eli Heuer]

Micah is done with the VTT hinting and it has been merged: https://github.com/eliheuer/dosis-vf/pull/9

Felipe and Micah just need to look at com.google.fonts/check/072.

[WeiH]

Hey Eli, what are you using to make those animations?

[Dave Crossland]

Looks like axis-praxis recorded with a tool like https://getkap.co

--
You received this message because you are subscribed to the Google Groups "Google Fonts Discussions" group.
To unsubscribe from this group and stop receiving emails from it, send an email to googlefonts-dis...@googlegroups.com.
To post to this group, send email to googlefon...@googlegroups.com.
Visit this group at https://groups.google.com/group/googlefonts-discuss.
To view this discussion on the web visit https://groups.google.com/d/msgid/googlefonts-discuss/5e8e4c0e-8092-4c60-af20-a1210b3e4df9%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Cheers
Dave

[Eli Heuer]

On Monday, November 5, 2018 at 9:39:00 PM UTC-5, WeiH wrote:

Hey Eli, what are you using to make those animations?

DrawBot, here is the source code: https://github.com/eliheuer/dosis-vf/blob/master/docs/drawbot-sources/basic-specimen.py

I'm working on a repo structure that builds fonts and specimens at the same time, you can see a demo here: https://www.youtube.com/watch?v=Sy1nLdroOiM

And the repo source here: https://github.com/eliheuer/fully-automated-font-repository

PRs welcome!