Re: shortened flags affair, part 3 - 32nds stem length

[Trevor Daniels]

Janek Warchoł wrote Friday, March 04, 2011 11:42 PM

> this is the next step of making stems and flags more beautiful:
> i suggest making unbeamed 32nd stems a bit shorter than they are 
> now.
> The main reason for doing so is to better match the stem length of 
> the
> beamed notes.
> As we know, the optimal situation is when unbeamed stems are just 
> a little
> bit longer than beamed stems, as in attached optimal.png.

Agreed

> Unfortunately, due to beam quanting it's not always possible to 
> achieve this
> optimum.
> However, i think we agree that unbeamed stems which are between 0 
> and 0.5
> staffspace longer than beamed stems are acceptable. Unbeamed stems 
> which are
> more than 0.5 staffspace longer than beamed ones are not good, and 
> when the
> difference reaches 1 staffspace it becomes quite ugly, see 
> ugly.png.

Agreed

> On the other hand, unbeamed stems shorter than beamed ones are not 
> desired
> too.
> I have evaluated current output and the results are here:
> http://www.sendspace.com/file/alsfo0

I was surprised to see the quite wide variation in the lengths
of the beamed 32nd notes.  Some seem too long and some
too short.  By that I mean moving them to the next quan position
to make them shorter or longer respectively would seem to be
an improvement.  I wonder if the default quanting parameters
are optimally tuned.  Perhaps this should be investigated first?

> (red - unbeamed stem is 1 staffspace longer than beamed stem, 
> orange - 0.75
> staffspace longer)
> As you can see, there is quite a lot of red and orange there.
> Now what would it look like if we changed the length of the 
> unbeamed 32nd
> notes to 4.25 ss (instead of 4.5)? Look here:
> http://www.sendspace.com/file/2mzt4a
> Looks much better to me - no red, only orange. Unfortunately it 
> introduces
> some yellow (unbeamed stem shorter than beamed one), but it's just 
> a little.

Agreed, irrespective of my comments on quanting above.  Increasing
the length of the very short beamed 32nds would remove some red,
but reducing the length of the very long ones would introduce more,
or at least more orange.

> A lot more yellow exists in case of 16th notes, see here:
> http://www.sendspace.com/file/g96nuo
>
> In addtion this will make the flagged 32nds look more balanced - 
> the center
> of mass would be a little lower, and the (upstem) notes wouldn't 
> look like
> tipping over (16ths and 64ths give very stable impression now).
> Also, the sequence of lengths of 16th, 32nd and 64th notes would 
> be more
> even, see "stems new" vs "stems old".
> Therefore i call for shortening 32nd unbeamed notes by 0.25 ss. Do 
> you
> agree?

Yes.

> cheers,
> Janek

Trevor



--------------------------------------------------------------------------------


> _______________________________________________
> lilypond-devel mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/lilypond-devel