Re: Chromatic clashes

[Jean Abou Samra]

Hi,

Le 16/04/2021 à 23:13, Lukas-Fabian Moser a écrit :
> Folks,
>
> while the following musical constellation is ugly in the first place, 
> I think we can agree that LilyPond's rendering is less than ideal:
>
> \version "2.22.0"
>
> { <c' cis'> 4 }
>
> yielding
>
> [image]
>
> If we force an accidental to c, we get something that might have a 
> slightly higher probability of being interpreted as intended:
>
> \version "2.22.0"
>
> { <c'! cis'> 4 }
>
> [image]
>
> My questions are:
>
> (1) Do you folks agree that the default behaviour qualifies as a bug?
Indeed. There is also an ugly-looking dependence on the input order:

{ <cis' ces'> cis' }
{ <ces' cis'> cis' }

We should probably have a special case when there are several 
accidentals in the same staff position. I suggest to open an issue on 
GitLab.

> (2) Is there an easy way to force LilyPond to add the additional 
> natural sign automatically?
>
> (Of course I know that in manually entered scores, one would simply 
> add one ! and be done with it; but that is more difficult in my 
> context, where I'm working with algorithmically generated chords.)
>
> For what it's worth, I think I found a not-quite-easy way for (2), 
> which I'm happy to share of course (see below), but I am doubtful that 
> this is the most elegant solution...
At any rate, that's pretty much how I'd do it personally.

Best,
Jean