Re: Counterpoint

[Urs Liska]

Am 08.11.2016 um 14:30 schrieb bart deruyter:

hey,

seems like my thoughts did spark some interest :-) . 

I had no knowledge of Humdrum at all,

This is pretty much a niche technology. But LilyPond is also a niche technology, so that comment isn't a judgement ;-)
Humdrum is a *data format* with a large set of analytical tools written for it. The original idea of creating Humdrum was to create a format usable for computer-assisted analysis.
Chances are that there will be Humdrum-to-LilyPond integration in the not-too-distant future BTW.

but I've read something about music21 before. If I'm reading it right Music21 is a python library to analyse music. I'm probably thinking way to simplistic, but frescobaldi is written in python too. Maybe there are possibilities to make the one work with the other.

In principle, yes.

I see in the music21 reference manual a converter from music21 to lilypond. If there would be a converter from lilypond to music21, it would be the easiest solution

One should ask the music21 people about that.

(maybe ly to musicxml export would be a good option too, but I don't think the musicxml exporter of frescobaldi is up to the task yet, or would I be wrong?).

You can alway try it out, but I wouldn't recommend relying on it professionally.

Embedded as an 'analysis' menu in frescobaldi (similare to page check rendering), frescobaldi could pass the score to music21, which analyses it, then sends the results back to frescobaldi, which renders it with the code generated by music21.lily.translate.

I think this will not be happening in this way, as I'm sure Wilbert Berendsen will not want to add such a heavyweight library to Frescobaldi.

What should be possible is calling/including music21 from within a "Snippet" that you can create/add to Frescobaldi. That should work already.

Another option is the plugin infrastructure that has been on the wish list for quite some time now.

Having analysis options and counterpoint checks, voiceleading checks etc. added to OpenLilyLib would on the other hand be very interesting because of less dependencies on external technologies and no conversion issues.

Indeed. But if these external technologies have already solved the actual problems one tries to solve that is a strong point.
I think it should still be possible to use openLilyLib as a wrapper, i.e. the interface to integrate to LilyPond documents and call external tools through Scheme. Just as another option.

Urs

Just thinking out loud here, I'm already glad there is some interest in it :-)

grtz,

Bart

http://www.bartart3d.be/
On Twitter
On Identi.ca
On Google+

2016-11-08 12:38 GMT+01:00 Frauke Jurgensen <address@hidden>:

In short, yes, such things exist, though not in Lilypond. I am a computational musicologist that collaborates in developing tools for analysing counterpoint. We've got tools like this to use in Humdrum or Music21. I think the Lilypond implementation would not be trivial (as Urs says), but I'll talk to one of my colleagues who is a much better programmer and see what he says.

On Tue, Nov 8, 2016 at 9:57 AM, Urs Liska <address@hidden> wrote:

Am 07.11.2016 um 13:21 schrieb bart deruyter:
> On a sidenote (perhaps for a different topic), in Musescore there is
> the possibility to create plugins which provide harmony checks,
> someone also did a plugin for a previous Musescore version which
> checked only first species counterpoint.
>
> I know lilypond's first purpose is creating sheet music, not composing
> music, but are there snippets of scheme or libraries around which
> could do the same?
>
> I think, for people who study counterpoint and voice leading, or any
> other rule-set in music, it would be very interesting to have a an
> option to check if they've followed the rules. In my case I have no
> teacher, can't afford private lessons, so I have to figure it out on
> my own without any way to check if I'm actually correct in
> interpreting the rules and executing the exercises.
>

I don't know if any code for this or similar purposes is already around
(I suspect not, otherwise you'd have got a reply), but I think from the
organizational POV it should be pretty easy to write something like
that. Basically it would work similar to the part combiner: take two (or
more) music expressions, perform the calculation and produce some
output. I don't immediately see how the actual content checks would have
to be implemented, but the infrastructure should be striaghtforward.

I can see different ways to approach it: one could have a function that
simply performs the checks and prints out the results to the console, or
it could actually modify the music expressions in a way that the results
are printed directly in the score (e.g. coloring or other visible hints).
In a similar way one could also write functions for harmonic analysis.

Probably the actual implementation is not all that trivial, and I
wouldn't start working on it. But I think it would make a good
openLilyLib package, and if someone is interested in the topic and has
the necessary Scheme skills I'd be happy to help with the openLilyLib
part of things.

Urs

_______________________________________________
lilypond-user mailing list
address@hidden
https://lists.gnu.org/mailman/listinfo/lilypond-user

_______________________________________________
lilypond-user mailing list
address@hidden
https://lists.gnu.org/mailman/listinfo/lilypond-user

_______________________________________________
lilypond-user mailing list
address@hidden
https://lists.gnu.org/mailman/listinfo/lilypond-user