Kerning and extrapolation

How does the logic in G2 work when its comes to extrapolations and kerning ?

Example: f/a kerning. In a lighter weight you expect more negative kerning than in a heavier weight. Could it be possible that at one time, the kerning becomes positive when I extrapolate a heavier weight than the fattest master ?

Yes. If you have -100 in the Light and -50 in the Bold master and then extrapolate by 100% (which probably will give you pretty distorted outlines) you get a kerning value of zero. Beyond that you get positive kerning. If the difference of the kerning values is bigger, you will get it earlier. But you still need some big extrapolation that doesn’t give nice outlines.


I am well aware of the possible distortion :slight_smile: We wouldn’t want that !

If you imagine a case where you have -15 in the Light and -3 in the Bold, the point of flipping would be nearer.

It would be a 20% extrapolation. That is more than you can do without distortions in most cases. And the extrapolated glyphs would probably ‘need’ that kerning anyway :wink:

1 Like

And in case of three masters, it only extrapolates based on the two nearest masters ?

Example: Thin/Light/Bold and I want extrapolate a Heavy. G2 will use Light/Bold masters.