[MEI-L] dots on chords and rests; chords with notes of differing duration

[Roland, Perry D. (pdr4h)]

Thanks for providing the Chopin example.  Staff 2 of measure 1 illustrates what I was trying to get at.  The 3rd 8th note of the left hand has half-note heads, indicating, as you say, the introduction of an additional voice.  But to the naïve encoder (OMR) it could appear to be an 8th note chord with "funny" note heads.  So, according to OMR software, the "proper" encoding could be:

<staff n="2">
  <layer>
    <beam>
      <note dur="8"/>
      <note dur="8"/>
      <chord dur="8">
        <note dur="2"/>
        <note dur="2"/>
      </chord>
      <note dur="8"/>
    </beam>
  </layer>
</staff>

and, according to a human:

<staff n="2">
  <layer>
    <beam>
      <note dur="8"/>
      <note dur="8"/>
      <space dur="8"/>
      <note dur="8"/>
    </beam>
  </layer>
  <layer>
    <space dur="4"/>
    <space dur="8"/>
    <chord dur="2" stem.dir="down">
      <note dur="2"/>
      <note dur="2"/>
    </chord>
  </layer>
</staff>

In the first encoding, the half notes do indeed cause the voice to "overlap" itself, but this is of no concern to the OMR program.

MEI is flexible enough to allow both encodings without privileging one or the other.  Both are "correct", depending on the purpose of the encoding.  I believe the second one could be derived from the first, permitting a "smartening up" of the OMR results.

--
p.


> -----Original Message-----
> From: mei-l [mailto:mei-l-bounces at lists.uni-paderborn.de] On Behalf Of
> Byrd, Donald A.
> Sent: Thursday, October 15, 2015 4:07 PM
> To: Music Encoding Initiative
> Subject: Re: [MEI-L] dots on chords and rests; chords with notes of differing
> duration
> 
> Ah yes, we do need to think in terms of the various domains to be clear
> about all this.
> 
> Odd you didn't get the attachment; I checked, and this is one of the rare
> cases where I failed to forget it :-) . Anyway, it's now available at
> 
>     https://drive.google.com/file/d/0B2p91S3Iky-
> VZ1hkNXl3MzZncnM/view?usp=sharing
> 
> For case #2, you say "I could imagine a situation where the following chord is
> actually only a quarter note in duration and the half note heads indicate…"
> Surely you mean "the following chord is actually only a quarter note LATER"?
> If so, I've situations -- especially in Chopin -- that could be interpreted this
> way. In that case, a voice is overlapping with itself in the logical domain: an
> unpleasant situation, and one that could be avoided here by assuming the
> presence of an additional voice… But I've also seen (very rarely) situations
> where I think the most sensible interpretation really is a voice overlapping
> itself :-| .
> 
> --DAB
> 
> 
> On Oct 15, 2015, at 3:44 PM, "Roland, Perry D. (pdr4h)"
> <pdr4h at eservices.virginia.edu> wrote:
> 
> >
> > Hi Don,
> >
> > Thanks for the classification – I think it’s dead on.  The situation’s a little
> more complex, however, in that we must be mindful of  *which* duration
> we’re talking about – logical, visual, or gestural.
> >
> > In any case, MEI can handle all these cases:
> >
> > 1.  The differences in “dotted-ness” of the notes is in the visual domain.
> The key word here is “intended”, the scribe just didn’t get around to putting
> in all the dots.  In MEI, the omitted dots can be encoded using <supplied>.  In
> the visual and gestural domains, this is the same as class 2; that is, the
> chord’s duration is the longest of its (dotted) notes.
> >
> > <chord>
> >   <note dur="2" dots="1"/>
> >   <note dur="2" dots="1"/>
> >   <note dur="2">
> >     <supplied>
> >       <dot/>
> >     </supplied>
> >   </note>
> > </chord>
> >
> > 2.  As you say, the duration of the *chord* is often equal to that of the
> longest note, although I could imagine a situation where the following chord
> is actually only a quarter note in duration and the half note heads indicate
> some kind of laissez vibrer indication.  In this situation, it’s a good idea to
> state the duration of the chord explicitly.
> >
> > <chord dur="2"/>
> >   <note dur="2"/>
> >   <note dur="2"/>
> >   <note dur="4"/>
> >   <note dur="4"/>
> > </chord>
> >
> > 3. In this case it makes sense to split the “chord” into multiple layers.
> Unfortunately, the attachment didn’t make it through (or perhaps you forgot
> it), so I’ll just make up an example:
> >
> > <staff>
> >   <layer n="1">
> >     <chord>
> >       <note dur="2"/>
> >       <note dur="2"/>
> >     </chord>
> >   </layer>
> >   <layer n="2">
> >     <chord>
> >       <note dur="4"/>
> >       <note dur="4"/>
> >     </chord>
> >   </layer>
> > </staff>
> >
> > The visual appearance is / can be the same, but the difference between ex.
> 1 and ex. 3 is in the intention of the encoder – to explicitly say that there are
> 2 layers / voices here.  This has consequences for the encoding depending on
> what comes later in the shorter of the two “voices”; that is, whether/how the
> difference is “made up”.  If the shorter voice has any material following this
> chord, say starting on beat 3 in this example, then space has to be inserted to
> make the 2 layers the same length, or at least, account for beat 2 in layer 2.
> >
> > Jo, would you like to draft an explanation for the Guidelines?
> >
> > --
> > p.
> >
> >
> > From: mei-l [mailto:mei-l-bounces at lists.uni-paderborn.de] On Behalf Of
> Byrd, Donald A.
> > Sent: Thursday, October 15, 2015 2:14 PM
> > To: Music Encoding Initiative
> > Subject: Re: [MEI-L] dots on chords and rests; chords with notes of
> > differing duration
> >
> > I expanded the subject line because (IMO) the basic issue here has to do
> with _any_ differing durations; there may or may not not be dotted notes.
> >
> > Good question, Johannes; this is a bit tricky! After looking at a few
> examples (Mozart, Chopin, Brahms, Amy Beach, Ysaye), I'd say there are
> three situations. The first is the one you mention, involving dotted and
> undotted versions of the same duration, where there aren't enough dots for
> all the notes (generally because of 2nds) but all notes are intended to be
> played with the same duration.
> >
> > The second situation is where the notes obviously belong to one voice and
> some really are intended to be sustained longer than others. Every example
> I've seen in music for strings is like this and is a chord of at least three notes,
> the obvious reason for the notation being that modern instruments can
> sustain only two notes. Anyway, the duration of the chord is simply the
> duration of its longest note.
> >
> > But there's yet another situation, of which the first page of Chopin's
> "Raindrop" Prelude (in D-flat, Op. 28 no. 15) has many examples; the
> attached excerpt is from a Universal edition. Here there are clearly two
> voices, and the longer note(s) (halfs or dotted halfs) are in a different voice
> from the shorter one(s) (eighths or dotted eighths). Or it might be better to
> think of the longer notes as "really" hiding eighth-note heads at the same
> pitch, so there are two notes of the same pitched attacked at the same time.
> There are two examples of this situation in m. 1. For tbe one on the 2nd
> beat, the duration of the chord in one voice is a half, while in the other voice
> it's an eighth.
> >
> > --Don
> >
> >
> > On Oct 15, 2015, at 11:10 AM, Johannes Kepper <kepper at edirom.de>
> wrote:
> >
> >
> > Dear all,
> >
> > thanks for the responses. By no means I question that there are chords
> with insufficient dots for all contained notes. The point is: Are these lacking
> dots just lacking in the visual domains, but contained in the logical domain (~
> intended to be played), or do they really say that the notes should be played
> with different durations. Especially in manuscripts, I think it is often the first
> case. However, there are certainly cases when chords intentionally have
> notes of uneven durations, and they're probably not restricted to printed
> music. My question is if these chords should be modeled in MEI using the
> <chord> element, or if they should be split up into layers. I know this may
> result in problems with parallel vs. shared stems, but from a logical point of
> view, it seems more faithful to me. If you keep different durations inside a
> chord, it seems unclear to me when following notes should start – after the
> shortest note in the chord, or after the longest. In essence, this introduces
> ambiguity in the encoding which normally isn't found in the original. I kind of
> like Perry's approach to say that the chord's attributes specify the logical
> duration, while the child notes provide information about the graphical
> appearance. That way, one can still "calculate" with the duration, while
> tracking the visuals independently from that. This would even allow to have
> different durations in chords (even though I still don't like them).
> >
> > My original concern was that querying the duration from the chord
> > element and grandchildren <dot> elements seems impractical. If we'd
> > clarify the guidelines in the direction above, I'm happy :-)
> >
> > Johannes
> >
> >
> > Am 15.10.2015 um 06:51 schrieb Christine Siegert
> <Christine.Siegert at beethoven-haus-bonn.de>:
> >
> >
> > Dear Johannes and all,
> > looking at sources from the time of about 1800, I recognize that there are
> many chords with notes of different durations. I think mainly of conventions
> of writing music. When a composer is writing a chord, he perhaps has notes
> of the same duration in mind (not always), but writes perhaps only one or
> two dots instead of three. As far as I see, these notes have to belong to the
> same layer. So it seems to me that the dots really are part of the notes.
> > All the best,
> > Christine
> >
> > Prof. Dr. Christine Siegert
> > Leitung Beethoven-Archiv und Verlag
> > Beethoven-Haus Bonn
> > Bonngasse 24-26
> > D-53111 Bonn
> > Tel.: +49(0)228-98175-22
> >
> > ________________________________________
> > Von: mei-l [mei-l-bounces at lists.uni-paderborn.de]" im Auftrag von
> > "Johannes Kepper [kepper at edirom.de]
> > Gesendet: Mittwoch, 14. Oktober 2015 19:13
> > An: Music Encoding Initiative
> > Betreff: Re: [MEI-L] dots on chords and rests
> >
> > I know this is and endless debate, and eventually, it's just a matter
> > of words, or analytical concepts, but I would always argue that chords
> > in MEI are always constituted by notes of the same duration.
> > Everything else is a matter of splitting things up into layers. This
> > can become quite ugly, as it easily results in a bunch of <space>
> > elements, but I strongly prefer this very strict approach over a loose
> > concept of a chord constituted by arbitrary notes just sharing a stem.
> > Again, I'm talking about how to encode things in MEI. I don't care
> > much how people call these things on paper…
> >
> > Just my 2c,
> > jo
> >
> >
> > Am 14.10.2015 um 18:42 schrieb Laurent Pugin <lxpugin at gmail.com>:
> >
> >
> > Quick question: how would you encode a chord where only some notes
> have a dot?
> >
> > On Oct 14, 2015 5:17 PM, "Byrd, Donald A." <donbyrd at indiana.edu>
> wrote:
> > I agree: chords themselves are never dotted. Chords containing a mixture
> of dotted and undotted notes have been observed in the music of Bartok and
> Brahms; they surely occur in other composers. The essential point is that
> chords can contain notes with any mixture of durations.
> >
> > --Don
> >
> >
> > On Oct 14, 2015, at 10:30 AM, "Roland, Perry D. (pdr4h)"
> > <pdr4h at eservices.virginia.edu>
> > wrote:
> >
> >
> >
> > The lack of <dot> within <rest> is an oversight that can be easily remedied.
> It will be there in the next release.
> >
> > <dot> within <chord>, however, is a completely different problem.  It
> seems to me that chords are never dotted, only the notes within them.  It is
> true that @dots is allowed on <chord>, but it was placed there as a
> convenience.  However, one cannot extrapolate from its existence the need
> to allow <dot> within <chord>.  As Jo has already noted, allowing <dot>
> within <chord> introduces unnecessary complexity and ambiguity, so I advise
> against that approach.
> >
> > I’m not trying to re-ignite the recent discussion re: so-called
> > “invisible accidentals” and its focus on redundancy, but since the
> > following encoding
> >
> > <chord dur=”2” dots=”1”>
> > <note />
> > <note />
> > <note />
> > </chord>
> >
> > is short hand for
> >
> > <chord>
> > <note dur=”2” dots=”1” />
> > <note dur=”2” dots=”1” />
> > <note dur=”2” dots=”1” />
> > </chord>
> >
> > and because these attributes are all optional, then there’s nothing
> > inherently wrong with
> >
> > <chord dur=”2” dots=”1”>
> > <note dur=”2” dots=”1” />
> > <note dur=”2” dots=”1” />
> > <note dur=”2” dots=”1” />
> > </chord>
> >
> > The @dur and @dots attributes on <chord> are redundant, but allow the
> kind of “one stop” rhythmic query Jo is seeking.
> >
> > Similarly, there is nothing wrong with the following:
> >
> > <chord dur=”2” dots=”1”>
> > <note dur=”2” dots=”1”>
> >   <dot /> <!—points to SVG shape -->
> > </note>
> > <note dur=”2” dots=”1”>
> >   <dot /> <!—points to SVG shape -->
> > </note>
> > <note dur=”2” dots=”1”>
> >   <dot /> <!—points to SVG shape -->
> > </note>
> > </chord>
> >
> > Again, there is redundancy between the @dots attribute on <note> and the
> <dot> element, but they serve different purposes -- @dots exists in the
> logical domain and <dot> exists in the visual one.  The <dot> element is
> providing additional information not communicated by the @dots attribute,
> not supplanting it.
> >
> > It’s often difficult to draw a bright line between the logical and
> > visual domains.  As a general rule, however, when there is only one
> > “source of information”, such as note/@dots, then it functions in both
> > domains.  The addition of another “source”, such as note/dot,
> > separates the domains.  In other words, the visual domain can often be
> > surmised from the logical, but making it explicit requires extra
> > information; that is, usually elements.  (BTW, the gestural domain can
> > also be inferred from the logical/visual, but making it explicit also
> > requires additional information, usually attributes.)
> >
> > <dot> is permitted inside <layer> in order to capture the ambiguous nature
> of dots in the mensural repertoire and probably should be disallowed by the
> CMN customization.  Using layer/dot in this case would be abuse in my
> opinion.
> >
> > --
> > p.
> >
> >
> >
> >
> >
> > <image002.png>
> > _______________________________________________
> > mei-l mailing list
> > mei-l at lists.uni-paderborn.de
> > https://lists.uni-paderborn.de/mailman/listinfo/mei-l
> 
> 
> ---
> Donald Byrd
> Woodrow Wilson Indiana Teaching Fellow
> Adjunct Associate Professor of Informatics Visiting Scientist, Research
> Technologies Indiana University Bloomington
> 
> 
> 
> 
> 
> 
> 
> _______________________________________________
> mei-l mailing list
> mei-l at lists.uni-paderborn.de
> https://lists.uni-paderborn.de/mailman/listinfo/mei-l