Re: How to connect Midi keyboard to Lilypond?

[David Kastrup]

James Harkins <address@hidden> writes:

> David Kastrup <dak <at> gnu.org> writes:
>
>> It's been some time since I last tried, but the basic answer I arrived
>> at for myself was "don't bother".  The tools are not good enough right
>> now to save time.
>
> I have to agree with David here.
>
> When I first started looking at LilyPond, one of my first questions was
> about MIDI input. But then I realized:
>
> - I would have used MIDI for step input (hold a note, press a key for
> the rhythmic value). So, which is faster? Reaching for another
> keyboard to hold down, say, F and type 2, or just to type "f2" on one
> keyboard? Seemed to me that it would be faster to stick with one
> keyboard (the computer keyboard).

Well, add to that reasonably good rhythm detection so that you basically
just need to put in the bar checks and your input tool corrects its
conceptions accordingly.  Or make a completely separate input pass just
for entering the durations.  Or combine them, and update the guesses
based on the specified durations.

There are a number of ways in which one can imagine an actually helpful
way of working with a separate Midi input, or even with abusing the
computer keyboard itself as a Midi keyboard approximation.

And then there is the question of how convenient your editing tools make
it to pull apart something like a Midi performance of a piano concerto
into the kind of voicing you need for making LilyPond happy with the
music.  That's not the ordinary cut&paste support.  If I have something
like wrongly chorded expressions, how to cut out selected notes in
chords and then paste them out into a separate voice?

That's something that Emacs' LilyPond mode could conceivably be extended
to do with a reasonable degree of comfort, and of course it's a nice
challenge for something like Frescobaldi as well.  It's not strictly
related to Midi, but this sort of editing task is more likely to occur
with Midi-based workflows.

Anway, my point is: the currently available tools are not good enough
right now to save time.

I'm not saying that this means the idea is doomed.  I think that would
be a sour grapes stance.  But at the current point of time, the only
convincing reason I see for working with Midi input is if you plan on
improving the available tools, and in that case, full speed ahead!

-- 
David Kastrup