On 23.05.2015, at 22:06, Carl Sorensen <[email protected]> wrote:
>
>
> On 5/23/15 6:34 AM, "pls" <[email protected]> wrote:
>
>>
>> Here is another test case for your patch:
>>
>> \version "2.19.20"
>>
>> <<
>> \new FretBoards {
>> <d-1 a c'-1 fis' a'-1>1
>> <g,-1 d f-1 b d'-1 g'-1>1
>> \transpose g a { <g,-1 d f-1 b d'-1 g'-1>1 }
>> }
>>>>
>>
>> These chords can be played with one finger! ;)
>
> So what do you recommend be done?
>
> I can make it so no fingerings are added, and then diagrams that are not
> completely fingered will show dots that have a finger of 0 (that's the way
> I started). It seemed to me that having a finger of 0 is confusing.
Agreed.
> Also, by far the most common barre is with the first finger. So allowing
> an open chord to be transposed to a barre chord with a first finger barre
> seemed useful.
Agreed.
>
> I don't have an algorithm to figure out fingering. The .ly file explains
> exactly how the implemented fingering algorithm works. It assumes that
> multiple notes played on the same fret with no fingering should be
> represented as a barre with the first finger. That's all it does.
Yes, again no lack of clarity in your .ly file and still I fell into the trap
when I tried to quickly cook up a minimal example for a bug report. Sorry for
my useless test case.
>
> If the user specifies fingerings that don't work right with the algorithm,
> they won't work right.
Sounds fair to me because I consider your algorithm to be very useful but it
would most certainly lead to a lot of complaints/bug reports (see above!)...
>
> Maybe I should back the code back to where a barre will be indicated only
> if the fingers are specified in the chord, and no barre will be generated
> when an open chord is transposed. That would be a straightforward fix.
> It would never generate any surprising results, because it would only
> generate barres that the user had entered by indicating that multiple
> notes should be played with the same finger.
>
> What do you think?
Sounds reasonable. Would something like this work?
\transpose f g { <f,-1 c-3 f-4 a-2 c'-1 f'-1>1 }
(At the moment it generates a false additional barre indicator.)
If yes, I’d say it’s a very good compromise.
Thanks for your work & patience!
patrick
_______________________________________________
lilypond-user mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/lilypond-user
