Closed kateliev closed 4 years ago
So far we know for sure that /K and /k related shapes suffer from kinks.
/K This is in both directions: CndCtr -> :Ctr
and BlkExpCtr - > BlkCtr
...but not beyond. Also interesting enough the effect is not "that" visible in /k
/k
this whole problem will affect more or less the following glyphs: /K /k /uni0416 /uni0436 /uni043A /uni0496 /uni049A /uni04A0 /uni04C1 /uni0497 /uni049B /uni04A1 /uni04DD /uni0436.smcp /uni043A.smcp /uni0497.smcp /uni049B.smcp /uni04A1.smcp /uni045C.smcp /uni04DD.smcp /uni04C2.smcp (... and more)
Submitting three solutions for /K to explore /K.fix_01, /K.fix_02, /K.fix_03.
For me neither works, so I am now more inclined to leave the /K unfixed - simple and more natural than having to outsmart ourselves....
Every suggestion is welcome!
I spent a while trying another fix, editing the form to reduce the difference in deltas between the offending points... and as expected it's horrific. Sacrifices any semblance of decent from for a smooth transition.
So yeah, I'm with @kateliev. We leave it and chock it up to the limitations of the technology. Can't have everything.
I think I am in favor of an alternative: leave overlap in the /K I think the bad interpolations are bad enough that they are worth leaving overlap in. The question is, how much overlap is required? I think we could merge either the arm or the leg with the vertical and not have any problems. We just can't merge the arm and the leg with each other.
I ended up using internal loops to fix /K Interestingly, they were not needed as far as we have noticed in /k Anyway, this is better now. If we spot more issues, can re-activate.
This is to serve as a placeholder and collection bin for problems that we find with the interpolation of the "finished" outlines!