Re: spurious partial in 2nd repeat abc->ly

[David Wright]

On Sat 20 Feb 2021 at 08:42:08 (+0000), Lilypond-User wrote:
> On Fri, 19 Feb 2021 at 19:45, Mark Stephen Mrotek wrote:
> 
> Thanks for the help Mark et al.
> >
> > The second repeated section starts with a partial measure.
> 
> But there is no \partial there in the source code, hence my problem.   With
> the source as shown in my original post the resulting PDF shows a partial
> and bar lines struck through beams, which is absurd in this context.
> 
> Given your hints, I have dotted the last notes in the alternatives as this:

No, please don't do that! This is dance music in strict time:
you'll have a hall-full of twisted ankles or worse.

Your new post shows this is a 48-bar jig, not 40. So my advice
for the time being would be to write out the last eight bars
and enclose them in a repeat, upbeat and all. Don't try to
nest the 4-bar repeat in the last eight bars: it's tricky,
and it doesn't add clarity.

> This compiles fine and the resulting PDF looks OK, sort of.  But this is
> cheating.
> 
> My understanding is that if a part starts with a \partial, then the last
> bar should consist of 5 quavers, not six.

Yes, and this applies to every part of the tune, every repeat.
So when a tune starts on the upbeat, like this, you have to
decide whether to enclose the upbeat within the repeat, or not.

In the first eight bars of this tune, the repeat uses the same note
for the upbeat, so you can include it as you have done. Similarly
for the second repeated section.

However, the last eight bars are different: the upbeat to the
dotted b (b4.) is c the first time, but f♯ the second time, so
that c (the last note in what is really bar 32) would *never* be
repeated in your OP source (of 40 bars).

By unfolding the 4-bar repeat, that problem goes away because the
f♯ is embedded in the 8-bar phrase. So the third repeat is just
like the first one.

I think I've marked all my changes with ←.

> Remember my original route to this ly was: python abc2xml.py file.abc >
> test.xml;musicxml2ly test.xml"

I'm not sure where the wrong notes came from …

> I have taken the liberty of attaching the PDF resulting from abcm2ps (which
> is what I am trying to emulate).  To me it seems more legal in that it
> features 5 quavers in the alternatives.

… because they're correct in the PDF.

As well as listening to the MIDI (which immediately pointed to the
first pitch error as well as the errors you introduced), I would
also unfold the score so that you can follow along. That way, the
rhythmic errors don't occur just when your eyes are scanning back
to the start of the repeat.

> In Part C, I have not yet worked out how to show the 1,2 and 3,4
> alternatives.  But that is another day's work.

I'm not even sure that the voltas are labelled correctly anyway.
They should be 1,3 and 2,4.

BTW I don't think you could get away with 144 bars of this one tune.
Or were those repeats just for your own proof-reading sake?

> BTW, I love Lilypond dearly and am so grateful for its tolerance of my
> musical ignorance.

Cheers,
David.

newmend.ly
Description: Text document

newmend.pdf
Description: Adobe PDF document

newmend.midi
Description: MIDI audio

newmend-unfold.pdf
Description: Adobe PDF document