First of all, Thank you very much for all the feedback! Sorry I didn't dig the archives myself, but initially I took the original Feta designs for granted and carved in stone.
Where did the original double-flat design come from? Music(X)TeX was a good guess, but the Music(X)TeX accidentals look completely different. Thinking of the introductory LilyPond Background Essay, there's a section about the flat symbol design <http://lilypond.org/doc/v2.18/Documentation/essay/engraving-details#music-fonts> and they mention Henle (computer typesetting) compared to Bärenreiter (hand engraving). When looking at the examples, the LilyPond flat seems to have been derived from the Bärenreiter stencil. *Conclusion:* Bärenreiter use standard flat widths for double-flat flats, so I don't see why we shouldn't adapt the existing Feta double-flat accidentals and create a matching (triple-flat, of course). Without Abraham's encouragement I'd never have dared touch the original Feta designs, though. Basically, nobody but LilyPond seems to use a compressed left flat. Double-flat alternatives in comparison In the old discussions mentioned in some of the answers, Abraham proposed a compromise that kept the original glyph width by applying an average compression to both flats. I've used MetaFont's proof sheets (with outlines so show how the flat symbols are put together on the left and the filled-in normal versions to check the visual appearance). There has been some manual cleansing of distracting outlines of superimposed parts and I've unified some parameters that made the counter (the small "hole" in the flat symbol) look slightly different in some accidentals containing flats. 1. Original Feta design The compressed left flat even "bites off" part of the lower stem and makes it look considerably thinner. And the compression of the left flat only reminds me of a rear-shunt car crash, sort of... ;) <http://lilypond.1069038.n5.nabble.com/file/t3887/double-flat-1-original.png> 2. Abraham's equally compressed flats Both flats are equally compressed as a compromise, thus keeping the original glyph width: <http://lilypond.1069038.n5.nabble.com/file/t3887/double-flat-2-abraham-equal.png> 3. Torsten's "real" flats with maximum overlap Both flats are "real" unaltered flats. Maximum overlap makes the double-flat glyph only marginally wider than the original design (cf. Dorico's Bravura font): <http://lilypond.1069038.n5.nabble.com/file/t3887/double-flat-3-torsten-real-overlap.png> 4. Abraham's "real" flats with minimum overlap Both flats are "real" unaltered flats, there is only a slight overlap. This is the widest of all the designs mentioned here (cf. Sibelius' Opus font): <http://lilypond.1069038.n5.nabble.com/file/t3887/double-flat-4-abraham-real.png> All the graphics have exactly the same height/width and glyph positioning, so you can download them and flip through them to directly see how the design slightly changes from image to image and gradually widens up. What do you think? Thanks for the support, Torsten -- Sent from: http://lilypond.1069038.n5.nabble.com/User-f3.html _______________________________________________ lilypond-user mailing list email@example.com https://lists.gnu.org/mailman/listinfo/lilypond-user