Re: Slanted Beams thickness

[Valentin Petzel]

Hello all,

@Abraham: In fact this is not done at scheme level here, but at C++ level. 
There reason is that we do not want to change the value of the thickness, but 
for Lilypond to think about this thickness in a more natural way. Also this 
example already handles the beam distance. For some reason the slanted beam 
are just slightly thicker than they should be, I’ll still need to look into 
why this is the case.

@Werner: In fact it is a bit too thick on slanted Beams. I just did not yet 
find out the exact reason for why this is, as mathematically they should be the 
same. I do not think it is a problem with calculation of the root, so I 
suppose is in some way related to the way lilypond draws these Beams.

@Jeam: I’ve stumpled upon a post in the Dorico dev blog which also mentions 
this Beam thing, which got me to look into this. One has to remember that in 
the old days Beams would have been cut into the pewter using special chisels. 
But these things cannot really cut diagonally, so you need to keep the chisel 
in line with the Beam, resulting in a constant Beam width regardless of slope. 
So this is something you will find in all old hand engraved scores. Furthermore 
if you look at the examples you must see that the way this is handles just 
look bad with high slopes. I don’t think you’ll find engraving literature 
stating that Beam thickness should apply to the actual thickness of the Beam 
and not the size of the vertical intersection as this does seem to be the more 
reasonable thing, but I think the other points make a strong case for why this 
is a legitimate claim.

Anyway, I do not think this is really intentional behaviour. The current 
approach as a big advantage for aligning Beams to Staff lines (which arguably 
is not that much of an issue for Beams with very high slope), but I guess the 
main reason for this is that it has little effect on the most common slopes 
you’d assume to find. Even with an 50% slope (which would be quite a high slope 
you’d find very seldomly)  we’d need only about 11% more thickness. With a 25% 
slope (which is much more realistic) we’d be at around 3%. So while you’d 
notice a slight difference between these it would be hardly noticeable.

Cheers,
Valentin

signature.asc
Description: This is a digitally signed message part.