[mei-neumes-ig] MEI Neumes Module Draft, Version 0.2

Wed Sep 20 21:23:30 CEST 2017

Thank you, Elaine! This clarifies a few things.

"Shape" is a term I borrowed from the process of data modelling to mean two objects which share the same properties and values, but which may be completely different instances of something. Two houses with one door and 10 windows could be said to have the same 'data shape', and it doesn't matter if one is a bungalow and one a skinny three-story Dutch house, and one is called a 'house' and another a 'huis'. 

In this case, it simply means that two things that have a pitch contour of "up-then-down," for example, must be instances of the same class of things regardless of any particular name or identifier attached to them. The thing is defined by its properties, and not by a name that is given to it.

Perhaps you could clarify a few things about the functioning of your proposed elements. It seems that there are properties of these symbols that are non-exclusive, and that a single symbol can carry with it multiple dimensions of description. 

As an encoder, when would I choose to use a 'gravis' over a 'puncta' if they can both mean 'lower' pitch? I understand that only puncta can be 'same or lower' but it seems, on the face of it, that there might be some confusion about when to use each. It also seems like there are three signs available to encode the 'same' pitch, two for 'lower', and only one for 'higher.' Will this cover the range of pitch movements necessary in the repertoires?

Both 'puncta' and 'producta' also seem to have an impact on duration, which I think should be clearly separated from pitch. Is there any way to separate the duration implications out from them? If we remove duration from the meaning of the sign (and encode it in a separate attribute), are 'punctus' and 'productus' effectively the same thing? 

Is there a sign for 'same or higher with longer duration'? Is that necessary?

Is there a reason the 'neume' element has been replaced with 'ncg' in the supplied encoding?

Cheers!
-Andrew

> On 20 Sep 2017, at 16:55, Elaine Stratton Hild <elaine.stratton_hild at uni-wuerzburg.de> wrote:
> 
> Dear All,
>  
> Thank you for your work on this.
>  
> I agree completely, Andrew, that the neume names should be able to be included in the encoding, but the encoding should not depend on them. A quick thought concerning your suggestion below… “neume name” (rather than “neume type” would probably be more specific and clearer to musicologists (like me!) using the module.
>  
> Just to clarify… the proposal Thomas Weber and I have put forward does not rely upon neume names. The designations we recommend (“acuta” / “gravis” \ “punctus” . and “producta” _ ) are the earliest historical designations for notational signs, and—most importantly—they are not limited to one notational type. (Perhaps this is along the lines of what you meant with "data shapes", Andrew?)
> 
> They designate basic shapes that themselves convey basic information (acuta = one note with the same or higher pitch; gravis = one note with lower pitch; punctus = one note with same or lower pitch and normal duration; producta = one note with same or lower pitch and longer duration).
>  
> They are especially convenient for encoding because they can be combined to create all the notational signs of all types medieval notation.
>  
> (Small historical detour... For the writer of the 10th century treatise where these terms are first used for musical notation, the signs formed binary oppositions to each other: “Acuta” and “gravis” were drawn from the study of grammar, with the acuta placed over vowels where the voice rose in pitch and the gravis placed with syllables where the voice fell in pitch. “Punctus” and “producta” are symbols drawn from verse analysis: short syllables in Latin were marked with punctus; long syllables were marked with producta.)
>  
> We suggest using these designations (as basic "data shapes"), rather than the tilt attribute, because the basic meaning conveyed by an acuta (the logical domain) remains consistent across all medieval notations, but the nuanced tilt of the sign (the visual domain, whether it’s written “north” or “northeast”, for example) varies with individual scribes.
>  
> Many thanks and all best wishes,
> Elaine   
>  
> 
> 
> 
> 
> 
> Dr. Elaine Stratton Hild
> Corpus monodicum: Die einstimmige Musik des lateinischen Mittelalters    
> Julius-Maximilians-Universität Würzburg
> Institut für Musikforschung
> Domerschulstraße 13
> D-97070 Würzburg
> 
> 
> >>> Andrew Hankinson <andrew.hankinson at mail.mcgill.ca> 09/20/17 4:58 PM >>>
> Hi Thomas,
> 
> Thanks for this, and spending time putting forward some concrete suggestions. 
> 
> I'm also of the opinion that the current proposal has a significant bias towards the visual component. I believe this has emerged largely because the exact musical content of some repertoires is unknown, making a system built on a "logical" domain approach to musical description difficult to define. (that is, when all you have a symbol that is angled, and you don't even know what pitch it is, how do you describe it musically?) That said, I think it's worthwhile to try and tug the existing proposal in the direction of addressing and describing the musical content directly.
> 
> I have been under the impression that building a system built on explicit naming of neumes as elements would not be a good idea, based on critiques I have heard on the existing neume module. The same neume can have many names (across styles or languages) and building a system on neume names makes it more difficult to accurately capture the longer non-standard (i.e., "compound") neumes. Thus the reason for building a system based on shape contour, rather than shape name.
> 
> That isn't to say that names are not a valuable thing to encode, but the idea is that these names would be included in a separate taxonomy within the file (or even a separate controlled vocabulary) that might be referenced within an encoding. I think this will go further towards making a general encoding system for all neumed repertoires, instead of narrowing it down to a particular style. 
> 
> You might imagine something along the lines of:
> 
> <neume type="#clivis">
> ...
> </neume>
> 
> or
> 
> <neume type="http://example.org/neumevocabulary#clivis">
> ...
> </neume>
> 
> This would let you use your own controlled vocabulary to name things for your own encoding, while maintaining a more general 'data shape' for the musical structures, allowing them to be compared across repertoires.
> 
> -Andrew
> 
> > On 20 Sep 2017, at 15:26, Thomas Weber <thomas.weber at notengrafik.com> wrote:
> > 
> > Dear group,
> > 
> > 
> > Elaine and me looked at both the Hildegard example as well as the "standard" neumes table by Cardine and made an attempt at encoding the signs' properties with better separation of domains. We tried to add more logical domain information to the previously mostly graphical domain info.
> > 
> > Our thesis was that the main distinguishing factor between different use cases of the Neumes module is in how much detail the notation should be described. Many projects right now use the Volpiano font to capture only the melodic contour and potentially connection information. For Corpus monodicum, we want to encode more properties of single staff-bound components, mainly whether a component is an oriscus, quilisma or strophicus. Our understanding is that the Old Hispanic project as well as the Optical Neumes Recognition project need even more detailed information to be encodable.
> > 
> > We came to the conclusion that the problem we still need to solve is better separation of visual and logical domains. Our impression is that the visual domain is of primary importance for both of the named projects. That might be the reason why the current draft seems to mainly focus on the visual domain.
> > 
> > We prepared the following table based on Cardine that outlines our suggested encoding.
> > 
> >  http://cm.notengrafik.com/2017-Graz-Cantus-Newtork/cardine-table/cardine_web.xml
> > 
> > The most obvious addition are sub-elements describing each <nc> as one of the seven basic single pitch signs (four basic signs based on Anonymus Vaticanus: acuta, gravis, punctus, producta; three special signs: quilisma, oriscus, strophicus). We chose to make them elements so that each of those can get its specific set of attributes describing them in more detail, like "extended" for acuta/gravis (understood to indicate larger intervals) or "waves" for quilisma. The idea behind that was to make it clearer which attributes are usable in which contexts, but this could also be done by means of Schematron rules when using attributes instead of elements - which we did in our pre-Graz state of ideas:
> > 
> >  http://cm.notengrafik.com/2017-Graz-Cantus-Network/presentation/presentation.html#variations
> >  http://cm.notengrafik.com/2017-Graz-Cantus-Network/presentation/presentation.html#code-sign
> > 
> > We suggest to represent the "angled" property using the @con attribute, as "angularity" is a property of the connection. Depending on the sign, the logical meaning of this would be to lengthen the component before or after the angled connection. This logical domain info could be recorded using @dur on the proper <nc>.
> > 
> > We picked up Ich's encoding of "O splendidissima" and made a modified version here:
> > 
> >  http://cm.notengrafik.com/2017-Graz-Cantus-Network/O_splindidissima.mei
> > 
> > Here we also tried to illustrate how more detailed visual domain information could added, namely the suggested @tilt. As the tilt is in most cases a consistent scribal idiosyncrasy, this might however better be solved using the tables of signs that Stefan Morent reminded us of recently. There, these idiosyncrasies could be described once for an entire file, source or hand. Johannes (hi to Johannes in CC) suggested that these "standard" patterns of writing might be best put in <scoreDef> and to encode deviations where they occur. These tables could also be useful to document variation among tokens.
> > 
> > We suggest to keep the table problem and how to link between the content and the tables (by semi-standardized names, by project specific naming/numbering schemes or by simple XML IDs, or maybe implicitly by characteristic sequence of single pitch signs) for later discussion, together with other topics like how to represent mi signs and rubrics.
> > 
> > We wish you a fruitful workshop in Halifax, and we are looking forward to hearing about any outcome concerning the neumes module!
> > Thomas
> > 
> > 
> > Am 07.09.2017 um 14:57 schrieb Ichiro Fujinaga, Prof.:
> >> Here’s what I have for the draft of the proposal for the new MEI Neumes Module schema.
> >> I’ve arbitrarily named it Version 0.2, so that we can refer back to it.
> >> 
> >> For ease of discussion and efficiency, please discuss one issue at a time in separate email threads, e.g., Re: @oriscus.
> >> 
> >> Thanks,
> >> 
> >> Ichiro
> >> 
> >> ==============================
> >> MEI Neumes Module Draft, Version 0.2
> >> 
> >> A <neume> element consists of one or more <nc> element(s).
> >> 
> >> A <nc> (neume component) is a single pitched event, although the exact pitch may not be known.
> >> 
> >> Attributes for <neume>:
> >> @intm (interval melodic; relative to the previous <neume>) {u|d|s|n|sh|sl} (u = up/high, d = down/low, s = same, n = neutral/unknown, sh = same or higher (but not lower), sl = same or lower (but not higher))
> >> @significative letters
> >> @hispanicTick (type1, type2)
> >> @hook
> >> 
> >> Attributes for <nc>: 
> >> @pname
> >> @oct
> >> @intm (interval melodic; relative to the previous <neume>) {u|d|s|n|sh|sl} (u = up, d = down, s = same, n = neutral/unknown, sh = same or higher (but not lower), sl = same or lower (but not higher))
> >> @wavy 
> >> @significative letters
> >> @curved (anticlockwise, clockwise)
> >> @angular (plain, v-shaped, staircase) (was @angled) 
> >> @liquescence
> >> @episema
> >> @extended
> >> @tilt {n|ne|e|se|s|sw|w|ne} (north, northeast, etc.)
> >> @quilisma (2, 3, or 4 curves)
> >> @oriscus (curved, flat, jagged)
> >> @con (gapped, looped) (connection to the previous <nc> within the same <neume>)
> >> 
> >> Removed from the previous version: @jagged, @flat, @long, @diagonalright, and @hispanicLoop from the <neume> level.
> >> 
> >> _______________________________________________
> >> mei-neumes-ig mailing list
> >> 
> >> mei-neumes-ig at lists.uni-paderborn.de
> >> https://lists.uni-paderborn.de/mailman/listinfo/mei-neumes-ig
> > 
> > -- 
> > 
> > Notengrafik Berlin GmbH
> > HRB 15007
> > 
> > UstID: DE 289234097
> > Geschäftsführer:
> > Thomas Weber und Werner J. Wolff
> > 
> > fon: +49 30 220661685
> > 
> > Leuschnerdamm 13
> > 10999 Berlin
> > 
> > notengrafik.com
> > 
> > _______________________________________________
> > mei-neumes-ig mailing list
> > mei-neumes-ig at lists.uni-paderborn.de
> > https://lists.uni-paderborn.de/mailman/listinfo/mei-neumes-ig
> 
> _______________________________________________
> mei-neumes-ig mailing list
> mei-neumes-ig at lists.uni-paderborn.de
> https://lists.uni-paderborn.de/mailman/listinfo/mei-neumes-ig
> _______________________________________________
> mei-neumes-ig mailing list
> mei-neumes-ig at lists.uni-paderborn.de
> https://lists.uni-paderborn.de/mailman/listinfo/mei-neumes-ig