Lukas-Fabian Moser <[email protected]> writes:

> Thanks for the explanations!
>> Your idea is good.  The infrastructure is not there.  We probably should
>> have something like make-translator for stuff that just translates
>> events and/or meddles with context properties.
>>
>> So you probably need to work on the music expressions for now similar to
>> how articulate.ly does things.
>
> Just so I understand correctly: I shouldn't use an engraver at all but
> write a music function?

It's more "can't" rather than "shouldn't".

> And if I understood everything correctly to this point, this is more
> difficult since at the time a music function is interpreted, the
> actual position in time of the music in question is not yet known.

Yes, this is "this sucks" territory.  MIDI is a bit underrepresented.
We don't even have a Scheme representation of the grob equivalent for
MIDI (mobs ?).

> So, it might be a bit involved here since what I'm doing is inherently
> non-local (in contrast to, as far as I can see, \articulate): I think
> I would have to keep track of time signature changes, do some
> book-keeping on note lengths to know when a measure is complete, and
> so on. This seems certainly possible, and ... well ... let's say, it
> certainly is another opportunity to increase my understanding of
> Scheme and Lilypond. :-)

Best fix would be probably to create generic Scheme _translators_
(agnostic to being in a particular kind of output).  At first, this
might just entail not skipping over their creators in Midi context
creation.

-- 
David Kastrup

_______________________________________________
lilypond-user mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/lilypond-user

Reply via email to