Miscellaneous experiments, opensource fonts and tools.
Machine Learning Font
Could a machine learning model be trained to generate realistic (or just interesting) letterforms? As it turns out, yes, albeit with mixed results. A styleGAN model was trained in RunwayML on a dataset of 2674 Google fonts organised as individual image-per-glyph in Drawbot. Runway-generated images were then piped via Python in to GlyphsApp to process the final font.
Full project and downloadable fonts released under the SIL Open Font License:
<Confusify>, Python and GlyphsApp Scripts
A confusable is a glyph that is so similar to another that it may be confused for it. Confusify.py generates fonts that when typed automatically swap confusables for the original. All confusables generated are existing characters in the source font. In that sense it’s a remix.
Select pre-processed fonts released under the SIL Open Font License: