[MEI-L] FRBR in MEI

Tim Crawford t.crawford at gold.ac.uk
Wed Nov 21 10:11:55 CET 2012 

Dear All,

While I have absolutely no wish to get further involved in this  
fascinating discussion, it strikes me that yes, indeed, much of what  
Benjamin is talking about may be handled better by a music ontology.

And such a thing - the Music Ontology - has now reached an advanced  
state of development by researchers at the BBC and elsewhere over a  
number of years. For details and specification, see:

     http://musicontology.com/

I'm not sure this alone (or even the entire apparatus of the Semantic  
Web) will solve all the problem cases you might encounter or devise,  
but at least it might relieve some of the pressure caused by a desire  
to encode *everything* about a musical work within MEI ...

Keep up the great work!

Tim Crawford, London

On 21 Nov 2012, at 08:24, Benjamin Wolff Bohl wrote:

> Hi there,
> first thanks to Axel for sorting out that FRBR is for bibliographic  
> items and thus performances, that we have nothing more of than the  
> knowledge it happened are out of scope.
> I've somehow been thinking too much towards somthing like a music  
> ontology.
> Moreover the idea of having a hierarchy in FRBR might have mislead  
> me, prooving Peter's earlier mentioned concerns regarding this true.
>
> Nevertheless we still got the recordings to deal with!
> So Johannes, let's continue to disagree ;-)
>
>> Hi Benni,
>>
>> I hope I got one of your last mails wrong (in this regard), but  
>> just in case I didn't: By no means I wanted to keep you from  
>> commenting on this (or other) thread(s), as your comments are  
>> extremely valuable and helpful – even if I sometimes disagree. If I  
>> offended you somehow, that wasn't my intention, and I want to  
>> apologize for it.
>>
>> That being said, I may continue to disagree ;-) Actually, I don't  
>> think we're that far away. The one thing you seem to get wrong  
>> though is the process from expression to manifestation, which is in  
>> no case trivial and a mere technological step without artistic  
>> contribution. When you consider the efforts necessary to engrave a  
>> piece of music, or the work on the preparation of the WeGA scores  
>> we see every day, you will agree that even in the graphical domain,  
>> this step is indeed highly artistic and involves a whole bunch of  
>> people with different expertise. I agree that the workflows for  
>> making recordings are different, but both things seem to be  
>> comparable from this perspective, don't you think?
> I neither think that we are too far away from each other now. And I  
> never wanted to say that the transiton from expression to  
> manifestation was a mere technical, but maybe I should have  
> explained a little more what my initial graphic was all about with  
> the recordings, as by no means it would involve a mere technical  
> step. Beginning from the way the recording engineer set up his  
> microphones and what he did on his audio desk, across quite a couple  
> of steps involving editing (cutting, rather technical but  
> nevertheless with artistic implications), mixing (very artistic) and  
> mastering(as artistic as technical), that all would result in  
> archive material quite a lot of intellectual/artistic work is  
> involved in a record(ing).
>
> I'll have a try on this:
> WORK - examination -> edition (e1) ------------- engraving -> print  
> run (m1) -printing -> print copy (i1)
>
> If you have the above, and try to get a parallel idea on the way to  
> the copy of a record on your shelf (i2):
> (1) What will be expression?
> (2) What will be manifestation?
> (3) Is one stream of e-m-i this sufficient?
>
> WORK - examination -> artist's inpterpretation (e2) - - 
> >              ?             -> record copy (i2)
>
> or
>
> WORK - recording   ->          ?          - mastering -> label's  
> press run -pressing (m2) -> record copy (i2)
>
> Maybe let's try to fill this with one of Don's "Greatful" examples:
> The Song "Truckin" has been released Nov 1 1970 on the album  
> "American Beauty" and as a single. The album was recorded in AUG-  
> SEP 1970, although it might be the single version was recorded in  
> SEP or maybe this specific song was recorded in SEP.
>
> Truckin (w1) --> 1970-08 to 1970-09 Session Tapes (e2) --> 1970-11-1  
> Warner bros. release of "American Beauty" album (m2) --> record copy  
> (i2)
> Truckin (w1) --> 1970-08 to 1970-09 Session Tapes (e3) --> 1970-11-1  
> Warner bros. release of "Truckin" single (m3) --> record copy (i3)
>
> But the single version and album version differ quite a lot, album  
> length 5:09 and single 3:13 so we should specify a little more.
>
> Truckin (w1) --> 1970-08 to 1970-09 Session Tape album-verison (e2)  
> --> 1970-11-1 Warner bros. release of "American Beauty" album (m2) -- 
> > record copy (i2)
> Truckin (w1) --> 1970-08 to 1970-09 Session Tape single-version (e3)  
> --> 1970-11-1 Warner bros. release of "Truckin" single (m3) -->  
> record copy (i3)
>
> A little complication: the single version was not recorded but taken  
> from the album version and edited down from 5 to 3 minutes  
> nevertheless it is a own expression, but it hints us twords some  
> items that might reside in an archives shelf, namely:
> - session tapes : the tapes from the recording session (potentially  
> multi-track)
> - edit tapes : the tapes where all the nice parts from the session  
> tapes were cut together to make up the material for the work  
> (potentially multi-track)
> - mix tapes : a stereo mix version including lots of additional  
> features like for example delay effects etc. resembling the final  
> version
> - master tapes : an acoustically slightly reshaped version of the  
> mix tape version in order to fit the technical limitations of a  
> certain target medium like vinyl and some intellectual work to  
> smoothen the mix (e.g. making all songs on a record sound similar)
>
> If they are relevant for my MEI file, they should go into <source>,  
> but where should these go in FRBR?
> Maybe they should all be separate expressions with strong relations  
> to each other?
> So actually the only one in the direct same "FRBR hierarchy" would  
> be the master tape?
>
> Truckin (w1) --> 1970-09-XX "Truckin" Master Tape (e3) --> 1970-11-1  
> Warner bros. release of "Truckin" single (m3) --> record copy (i3)
>
> If I don't know about all the tapes I might just put?
> Truckin (w1) --> 1970-08 to 1970-09 recordings --> 1970-11-1 Warner  
> bros. release of "Truckin" single (m3) --> record copy (i3)
>
> or?
>
> Truckin (w1) --> 1970 Version (e3)--> 1970-11-1 Warner bros. release  
> of "Truckin" single (m3) --> record copy (i3)
>
> What do you think, is there still a problem?
> Is there anything interesting for you in the above?
>
>> Besides that, I totally agree that FRBR is not extremely  
>> prescriptive regarding how to model certain situations, but after  
>> thinking about it for some time, I (now) think that this is  
>> actually a benefit, as it doesn't enforce a specific setup, but  
>> allows projects to implement it as they see fit. So in the end, I'm  
>> not against your approach in general, I'm just against enforcing  
>> your approach. The current implementation of FRBR in MEI tries to  
>> keep this openness of FRBR, which I regard as a good thing. In the  
>> end, all of us could be wrong ;-)
> I never wanted to enforce anything only to show up possibilities to  
> be considered when implementing FRBR or test the current  
> implementation against. And I think you're absolutely right that an  
> openness could be a benefit as we certainly will miss possible  
> complicated situations.
>
> <salutation>benjamin</salutation>
>>
>> Best,
>> Johannes
>>
>> Am 20.11.2012 um 12:33 schrieb Axel Teich Geertinger:
>>
>>> Hi Benni
>>>
>>> Perhaps we should remember that FRBR is intended for  
>>> _bibliographic records_, not for descriptions of a work's  
>>> reception history. Thus, the premise for using FRBR is that in the  
>>> end we want to describe bibliographic items. Since a performance  
>>> itself isn't a bibliographic item, perhaps it does not have to fit  
>>> in? Only if it results in such an item (via manifestation), i.e. a  
>>> recording, it becomes truly relevant to use FRBR. The performance  
>>> in that case is not the primary thing we want to describe, it is  
>>> just the context that resulted in the recording manifestation.
>>>
>>> Just another 2 cents,
>>> Axel
>>>
>>> -----Oprindelig meddelelse-----
>>> Fra: mei-l-bounces at lists.uni-paderborn.de [mailto:mei-l-bounces at lists.uni-paderborn.de 
>>> ] På vegne af Benjamin Wolff Bohl
>>> Sendt: 20. november 2012 11:57
>>> Til: Music Encoding Initiative
>>> Emne: Re: [MEI-L] FRBR in MEI
>>>
>>> Hi Perry,
>>> thanks for some clarifying approaches
>>> further statements inline
>>>
>>> Am 16.11.2012 22:25, schrieb Roland, Perry (pdr4h):
>>>> Random comments on the discussion so far.  Sorry if this gets long.
>>>>
>>>> When contemplating performances and recordings, it seems to me  
>>>> that people often have trouble reaching agreement on the term  
>>>> "sound recording".  Andrew's slides label the *expression* as  
>>>> "the sound recording", but others might label the *manifestation*  
>>>> as "the sound recording".  You might say the expression is the  
>>>> "act of making a recording" and the manifestation is the  
>>>> "recording that results".
>>>>
>>>> To disentangle the different uses of the term "recording", it  
>>>> helps me to remember that an expression is not a physical entity,  
>>>> but a manifestation is.  Therefore, I prefer to think of the  
>>>> expression as "the performance" (the non-physical thing being  
>>>> recorded) and the manifestation as "the recording" (the physical  
>>>> thing).  This fits with the way libraries have traditionally  
>>>> cataloged recordings, i.e., CDs, LPs, cassettes, wax cylinders, ...
>>> I completely agree on that, being the reason why I used both the  
>>> terms recording and record with record being on the manifestation/ 
>>> item-level and recording being rather on the expression- 
>>> manifestation-level. Why so? Recording has to be subordinate to  
>>> work after all and a recording is not just a simple physical  
>>> manifestation but a multistep process involving conceptual and  
>>> creative work done by producers and engineers.
>>> So talking about a recording as only being a manifestation becomes  
>>> problematic as it is a intellectual process resulting in a  
>>> physical manifestation. That's the way I was looking on it (owed  
>>> to my audio engineering past) and of course it can be seen  
>>> differently.
>>>> In any case, the FRBR document, which Axel cites, says a  
>>>> *performance is an expression* and a *recording is a  
>>>> manifestation*.
>>> This is perfectly plausible when disregarding the intellectual  
>>> endeavour entangled with the "act of making a recording", as  
>>> mentioned before.
>>>> The usual "waterfall" kind of diagram is explained by saying the  
>>>> term
>>>> "work" applies to conceptual content; "expression" applies to the
>>>> languages/media/versions in which the work occurs; "manifestation"
>>>> applies to the formats in which each expression is available; and
>>>> "item" applies to individual copies of a single format.  (Here  
>>>> "media"
>>>> means "medium of expression", say written language as opposed to  
>>>> film,
>>>> and "format" means physical format, as in printed book as opposed  
>>>> to
>>>> audio CD.)
>>>>
>>>> Taking another tack, though, often it is easier for me to think  
>>>> of FRBR "from the bottom up", rather than start from the work and  
>>>> proceed "down" the waterfall diagram.  Using the recording  
>>>> example, the item is the exemplar I hold in my hand, the  
>>>> manifestation is all of the copies of that exemplar (or better  
>>>> yet, all the information shared by all those copies), the  
>>>> expression is the version of the work that is represented by the  
>>>> manifestation (e.g., Jo's nose flute + harpsichord version and  
>>>> the orchestral version are different expressions), and the work  
>>>> is an intellectual creation/idea (e.g., Bohl's op. 1, the one  
>>>> that goes da, da, da, daaaaaa, reeep! reeep! reeep!).
>>>>
>>>> Using this "bottom up" thinking helps avoid mental contortions  
>>>> regarding what the work is -- the work is simply the thing at the  
>>>> end of this mental process.  From there on, there are work-to- 
>>>> work relationships, so we don't have to think about whether  
>>>> "Romeo and Juliet", "Westside Story", and every other story about  
>>>> star-crossed lovers are expressions of an ur-work with its own  
>>>> manifestations and so on, which lead us to a different  
>>>> "waterfall" conclusion each time we discover a new work or  
>>>> expression.
>>> The idea of approaching the FRBR model "from the bottom" is great.  
>>> And to be honest was something I did when drawing my model,  
>>> especially concerning the record and recording portion of it. I  
>>> started out from work on the top right and from the individual  
>>> record bottom right and tried to fill in as many steps as  
>>> possible, always wondering whether it be physical or conceptual.  
>>> Actually I had the recording in between expression and  
>>> manifestation in the first place, as I had the audio tape or  
>>> digital audio in between manifestation and item.
>>> The parallel processes from a wok to an item (regardless of  
>>> whichever form this may have) are owed to perspective and goal.  
>>> When talking about graphical sources I completely agree with the  
>>> idea of a certain instrumentation version or the like being an  
>>> expression, a print run being a manifestation an individual copy  
>>> of which would be an item.
>>>> Instead of creating separate expression-level markup for each  
>>>> performance, Axel treats some expressions (performances) as  
>>>> events related to another expression of a work (the orchestral  
>>>> version vs. the nose flute version).  This is fine.  As Johannes  
>>>> already pointed out, separate <expression> elements for the  
>>>> performances can be generated from the <eventList> markup, if  
>>>> necessary.  Conversely, there's nothing wrong with creating  
>>>> separate <expression> elements for each performance and relating  
>>>> them to other appropriate expressions and/or relating them  
>>>> directly to the work.  If necessary, given accurate place and  
>>>> date information, the <eventList> kind of markup could be created  
>>>> from the separate <expression> elements.  So, six of one ...
>>> I can agree here, too. I only wondered if the sound wave resulting  
>>> from the performance was the physical item (specific performers on  
>>> a specific date), then consequently a series of performances by  
>>> conductor and orchestra would make up for the manifestation, the  
>>> expression then would be the concept that the conductor developed  
>>> studying his "source material" and making up the way he wanted the  
>>> composition to be realized ergo his "personal version" of the  
>>> piece, somewhat of a personal edition.
>>> The performance material of course being an item of a certain  
>>> print run
>>> (manifestation) of a certain edition (expression), having strong  
>>> relationships to all of the above.
>>>> Johannes said "If there is a manuscript of the nose flute  
>>>> version, the information about it would be spread between the  
>>>> manifestation (source) and the item."  Well, maybe.  But, I think  
>>>> in this case it would be fine to describe the manifestation and  
>>>> the item in a single place (within <source> in MEI) because  
>>>> there's only one manifestation and one (and only one) item  
>>>> associated with that manifestation.  This is the traditional way  
>>>> manuscripts have been described, pre-FRBR.  Practically speaking,  
>>>> the manifesation and the item are the same thing.  But, as soon  
>>>> as you want to say something special about a particular *part*  
>>>> (as in "chunk", not performer part) of the manifestation, you  
>>>> have to split these up again, for example, when one section of a  
>>>> manuscript is located in Prague and another is in Manitoba.
>>> This was the idea behind me marking/stretching the autograph from  
>>> expression to item.
>>>
>>> /benjamin
>>>> This is not the case with printed material where there is  
>>>> *always* more than one item created from a manifestation, but it  
>>>> is still traditional to describe the manifestation and item as  
>>>> though they are the same thing.  For example, it is common to  
>>>> follow the manifestation's author, title, place of publication,  
>>>> etc. with information about the location where one can obtain an  
>>>> examplar of the manifestation, say, UVa Library M 296.C57 1987.
>>>>
>>>> Johannes also said "So if you have two more measures in a source,  
>>>> this
>>>> source establishes a new expression in FRBR."  Again, maybe.  The  
>>>> FRBR
>>>> report (1997, amended and corrected through 2009) says
>>>>
>>>> "Variations within substantially the same expression (e.g.,  
>>>> slight variations that can be noticed between two states of the  
>>>> same edition in the case of hand press production) would normally  
>>>> be ignored or, in specialized catalogues, be reflected as a note  
>>>> within the bibliographic record for the manifestation. However,  
>>>> for some applications of the model (e.g., early texts of rare  
>>>> manuscripts), each variation may be viewed as a different  
>>>> expression."
>>>>
>>>> The issue is in the determination of whether 2 things are  
>>>> "substantially the same expression".  As with many things, this  
>>>> depends on the person making the determination, there is no  
>>>> single correct answer.  We intend that MEI will provide the tools  
>>>> for accurate description using either approach.
>>>>
>>>> Just my 2 cents,
>>>>
>>>> --
>>>> p.
>>>>
>>>> __________________________
>>>> Perry Roland
>>>> Music Library
>>>> University of Virginia
>>>> P. O. Box 400175
>>>> Charlottesville, VA 22904
>>>> 434-982-2702 (w)
>>>> pdr4h (at) virginia (dot) edu
>>>> _______________________________________________
>>>> mei-l mailing list
>>>> mei-l at lists.uni-paderborn.de
>>>> https://lists.uni-paderborn.de/mailman/listinfo/mei-l
>>>
>>>
>>> _______________________________________________
>>> mei-l mailing list
>>> mei-l at lists.uni-paderborn.de
>>> https://lists.uni-paderborn.de/mailman/listinfo/mei-l
>>>
>>> _______________________________________________
>>> mei-l mailing list
>>> mei-l at lists.uni-paderborn.de
>>> https://lists.uni-paderborn.de/mailman/listinfo/mei-l
>>
>> _______________________________________________
>> mei-l mailing list
>> mei-l at lists.uni-paderborn.de
>> https://lists.uni-paderborn.de/mailman/listinfo/mei-l
>
>
> _______________________________________________
> mei-l mailing list
> mei-l at lists.uni-paderborn.de
> https://lists.uni-paderborn.de/mailman/listinfo/mei-l
>

More information about the mei-l mailing list