Feature suggestions

Post Reply
HugoJourdan
Posts: 11
Joined: 24 Jun 2021

Feature suggestions

Post by HugoJourdan »

List of suggestions :

KernOn specific master

Sometimes it could be useful to ask KernOn to launch on only one master.
For example, if I change the models on a specific master, I don't want KernOn to recalculate the masters that I did not modify models.

This would save a lot of time (especially when using KernOn on a font with a lot of masters)

______________________

Custom Kerning Interpolation

In the case of a font with 6 masters (Light Condensed, Light, Light Extended, Bold Condensed, Bold, Bold Extended), actually, if you want to interpolate Kerning from Width axe (that means for example to kern Light by interpolating kerning from Light Condensed to Light Extended) the kerning interpolation is not good.

This could be useful to add a feature to tell to KernOn which specific master you want to take as a reference to interpolate Kerning.
User avatar
Tim Ahrens
Site Admin
Posts: 404
Joined: 11 Jul 2019

Re: Feature suggestions

Post by Tim Ahrens »

HugoJourdan wrote: 26 Jul 2021 KernOn specific master
You are right, there is a lot of seemingly unnecessary processing taking place.

The problem is that the kerning groups are the same for all masters, they are generated after all masters have been autokerned. The generating of kerning classes is a kind of data compression, and it is not lossless.

If one master gets modified then the kerning classes might have to change as well, which means the kerning of these masters would get compressed again, with another small data loss. In other words, the result would not be identical to the current system, which always autokerns all masters, then you have the un-compressed, lossless kerning, which is then compressed.

In practice, however, this difference, i.e. the additional data loss, should be very small. I’ll think about how to implement that as a feature. The only thing that worries me is that we need additional user interaction, and I always hesitate to make the UI more complex.
User avatar
Tim Ahrens
Site Admin
Posts: 404
Joined: 11 Jul 2019

Re: Feature suggestions

Post by Tim Ahrens »

HugoJourdan wrote: 26 Jul 2021 Custom Kerning Interpolation

In the case of a font with 6 masters (Light Condensed, Light, Light Extended, Bold Condensed, Bold, Bold Extended), actually, if you want to interpolate Kerning from Width axe (that means for example to kern Light by interpolating kerning from Light Condensed to Light Extended) the kerning interpolation is not good.
Kern On’s “Interpolate master kerning” is exactly the same concept as Glyphs’ “re-interpolate”.

In your case, have you tried what happens if you re-interpolate the whole Light master, i.e. glyphs shapes and everything else? This will reveal the difference between your manual design and the interpolation, i.e. your manual corrections compared to the pure interpolation.
You could use Glyphs’ re-interpolate command to see how the glyph shapes change. Alternatively, you can set up an instance (or an export if you are using Glyphs 3) that has the same co-ordinates as the Light master but uses the “Disable Masters” custom parameter to disable the Light master. This gives you a re-interpolated Light instance/export.

You may find you that the width co-ordinate of your Light master is inappropriate. If the re-interpolated Light instance looks narrower than the real Light master then you need to adjust the width co-ordinate of the Light master. Otherwise you will not be able to use re-interpolation or “Interpolate master kerning” with the best results.

Maybe you can send em the file so I can have a look?
HugoJourdan wrote: 26 Jul 2021 This could be useful to add a feature to tell to KernOn which specific master you want to take as a reference to interpolate Kerning.
In your case, which masters would you use as a reference to interpolate the kerning for the Light master?
Post Reply