We're sorry but this page doesn't work properly without JavaScript enabled. Please enable it to continue.
Feedback

E-LAUTE: establishing a music edition platform using Semantic Web technologies

00:00

Formal Metadata

Title
E-LAUTE: establishing a music edition platform using Semantic Web technologies
Title of Series
Number of Parts
15
Author
Contributors
License
CC Attribution - ShareAlike 3.0 Germany:
You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor and the work or content is shared also in adapted form only under the conditions of this
Identifiers
Publisher
Release Date
Language

Content Metadata

Subject Area
Genre
Abstract
E-LAUTE (https://e-laute.info/) is creating a comprehensive edition of German renaissance lute tablatures (GLT), a historically widespread music notation that has been largely neglected by modern research. It interlinks music and textual encodings, notation images, audio, semantic annotations, and bibliographic metadata by using open data formats and Linked Data throughout the entire process. Additionally, it builds on research data and information architectures provided by the Technical University of Vienna (research workflow management) and the Austrian National Library (ÖNB; GAMS digital edition platform and triplestore). We are extending the ÖNB platform with facilities for incorporating multifaceted music information, and we are augmenting the Music Encoding Initiative’s (MEI) existing XML schema for the representation of GLT documents. We contextually enrich the MEI encodings through interconnection with textual encodings of contemporary lyrics and instructional material, IIIF facsimile images, audio recordings (produced both project-internally and externally), and additional metadata. To do so, we apply Linked Data ontologies, XML transclusions between encoding schemas, and Web Annotations for external contributions through decentralized Solid pods. We aim to create a central hub for managing the enriched data and for publishing the results in uniform and state-of-the-art formats (e.g. JSON-LD), providing open APIs (e.g. SPARQL) and contributing innovative approaches to music informatics and musicological research, thus serving the needs of music researchers, practitioners and enthusiasts alike.
Computer animationEngineering drawing
Computer animation
Computer animationEngineering drawing
Diagram
Computer animation
Computer animation
Meeting/Interview
Computer animation
Computer animation
Computer animation
Computer animationMeeting/Interview
Transcript: English(auto-generated)
So welcome to our presentation. Together with David, we're going to talk about a digital music edition that we are building at the Austrian National Library. But of course, this is a big project that involves many institutions, and we're going to explain the structure and
how Sematic Web plays a role in building up this digital edition and how we add some functionality to it. But because it's a little special topic, I think it would be good to just give
a short introduction what we mean by tablatures, so that you can follow also the presentation. Basically, a tablature is what you see on the right. It is a term that we also use nowadays, for example, for a guitar.
So it's a notation system that basically shows how a guitarist or anybody that plays a plucked instrument, how he can use this, where he can put his fingers on the neck of the instrument, on the different frets, as we say, and which strings he should pluck.
And this way he can produce the necessary sound. So it's basically a notation system. And the Ilaut edition is basically based on these tablatures that were created in the Renaissance, in the German speaking area.
And it is a project that basically gathers all tablatures in this German notation system from different libraries, from Germany, Switzerland, Austria, and also Poland. Before moving on to the next slide, an important thing also to mention, again, it's a little detailed, but it will help
maybe understand some aspects of the edition later, is that these tablatures, so these lute pieces, basically are based also on vocal works. They could be monophonic or polyphonic, and they're basically, most of them, they are transcriptions of these vocal works. So there's a
lot of interlinking between sources, and this is partially where linked data will take place, but also in other aspects of the project. So basically, I mentioned the tablatures, but basically in our digital edition, we have also a text, because we have
also instructions on how one can write a tablature, how one can transcribe the vocal work to a piece for lute. We also have sometimes the lyrics of some songs, because most of this music came from songs, or it would accompany some songs, and
sometimes we also have text on the margins, the marginalia, which also are some notes that the different scribes made and explain the music sources. This means for the people who maybe are related to digital humanities, we need to transcribe this information
into XML, so it's more accessible and semantically usable. For text, there is a TI, I guess it's known to many people, but when it comes to music, we have the so-called Music and Coding Initiative.
TI is Text and Coding Initiative, and MEI is the music one. And just giving a very also brief introduction to MEI, basically it's an XML schema for music documents representing the semantically intellectual structures behind the music notation. And of course it's also a good basis for digital, of course also analog rendering of
these sources. And you just get an idea for people who maybe somehow have some basic, at least music knowledge, you can see how notes are represented in MEI, with the different attributes about the position of the note, which octave, which pitch, duration, stem direction, and different things.
Going back now to the intellectual, let's say, basis for our digital edition, basically there's not only the German lute tablature, because our project is basically the German lute tablatures, which is one notation system.
There are different notation systems for lute tablatures of that era. And we have the French, the Italian, the Spanish one. And on the left side, you see an Italian, which can be read also by lutenists today, because it's very easy.
The different lines that you see are basically the strings and the numbers. So the exact position on the neck of the instrument, so the frets. With fret number one should put his finger on so that we produce the sound.
On the right, you see again a diagram basically of a German lute tablature, and there you can see how more complicated it becomes, because for its combination of fret and chord, we have a different symbol. And this makes the German lute tablature much more complicated. So coming back to the MEI, now you see a little more detailed version of an MEI document for a lute tablature. In
this case, we have a French one, but this is also where linked data also starts to play a role, because as you can see with the digital rendering we have of that document, and this is made by a rendering engine, which is called Verovio.
Basically, the output is SVG, and in that case, the end user can select any portion of the music. In that case, you can see the upper left corner, this letter A is in blue, which means it is selected.
And basically, any selection, every individual note, every individual symbol basically can be a web resource, so it can be identified and we can reference it in any document. So even outside the digital edition, we can pinpoint to a specific portion of music, and this is where MEI is also very important.
But as I said, we are trying to build a multimodal edition, a critical digital music edition, and coping here are some words from a digital musicologist, Franz Wiering, already from 2009.
Basically, what we need to take into account is to have the different copies of the sources, so the facsimiles, but also so connections between music sources in a music archive, provide some annotations, and also give the space to some analytical tools.
And MEI also enables that as a symbolic representation of music. So in general, digital critical editions and also in the Austrian National Library,
they have a synoptic view, which means that we have two main pillars. On one side, we saw the diplomatic transcription of music, and then on the other side, we have an edited transcription. In our case, it becomes a little more complicated, so for each one of these two, we have different notation systems.
So we have German lute tablature, the Italian one, and the French one, because for many, as I mentioned before, for many instrumentalists, reading tablature in Italian or French tablature is much easier than in the German one. Then we also want to provide the same notation in common music notation, so in the regular notes that we know with a pentagram.
And as I said, because we have also text in our edition, then we have also literal and edited transcriptions of any text portions in our collections. Then on top of that, we provide also the facsimiles of the resources, and then we offer the possibility
also for the end users to provide audio recordings alongside the audio that we produce also in the project. And this can be made with the use of web annotations, but David is going to talk about it later on. So you see already, it's a kind of complex project. In any case,
interdisciplinarity plays a specific, very important role, and we have representatives from different areas. The main ones being digital humanities and musicology, but of course, we need the assistance of music informatics and German studies because of the text that is provided.
Going a little to the data infrastructure behind the project, I don't need to explain now the whole architecture. I will just concentrate on the down left part, so this yellow and reddish part.
We have a bibliographic database, basically it's a relational database which existed before the official beginning of the project. So we took this as our basis, and the musicologists and the German literature specialists,
they input their data, their metadata and any research data about the description of these sources. And then our role in the national library is to take this information and map it into RDF. And this then will become part of a triple store, which is going to be stored in GAMS.
GAMS is a digital asset management system that we have in the Austrian library. Then we're going to also pull different manuscripts through IIIF from all participating libraries, as I said in the German speaking area.
And then also with adding some other components, we're going to create the digital edition. And going a little farther to this bibliographic information, here you see some screenshots of this relational database, how it is for the researchers.
So we have different information, we may have text, insipids and titles of specific pieces and then the common uniform titles that are associated with them. We can describe the different sources, them being either manuscripts or prints. And then, because its source has many items within, then one can go to the level of
the item and also describe it in terms of the different genres, mediums of performances, notation types. Mediums of performances, I mean, because there are different kinds of luts, different, how we say, they were steamed.
So they used different ways to prepare the instrument and also some concordances with other instruments. And I'm taking at the moment, we are at the beginning basically where we are translating
this relational database or an SQL database basically to RDF by using Python and RDF Leap. And here you just see a part of it, it's still at the beginning. So for now we're just using a local namespace, so Elauter, but the goal is of course
to use other music ontologies and bibliographic ontologies so that we can have more interoperability with other systems. I'm gonna pause my presentation now and David is gonna take over and do his own part.
Maybe you could mute. Yes, thank you, that's good. Let me see if I can share my screen quickly. Just a moment please.
Yes, perhaps you could share your slides for me if you don't mind and then we can work through it that way. Thank you very much.
So maybe one more word to this while we're using a very basic or very localized
namespace here to get the data out of this relational database and into a graph form. As Elias has said, we are of course aiming to make this interoperable and to reuse other existing solutions as far as possible. Why aren't we just using something like the music ontology? You may have heard of this if you're interested in semantic web and music.
Other ontologies exist as well, for example the Torre Mus project from France. And the answer is that none of these quite sufficiently cover this sort of medieval or renaissance rather sources that we're working with.
So our sources are just older than the targets of these existing ontologies and thus we're working with problems that are more concrete for us. In fact the previous speaker had this very nice statement, making concrete statements more ambiguous.
This is something we have to accomplish because of things like contested attribution of sources. Even the whole notion of a work is problematic with the music that we are working with. So we are going to be adopting what we can from existing ontologies and then bridging using our own constructs.
So to give you an idea of what this is going to look like, at the moment we can show you an existing digital edition of the UNB using a similar infrastructure.
This is not a musical project, this is the Peter Handke Notitzbücher project. Peter Handke is an Austrian Nobel laureate and these are his notebooks in digital form. So this is generated by this digital infrastructure at the UNB which we are
now extending as the first project to use music to work with these music encodings. So here for example you have a facsimile image served via IIIF on the left, you have a textual encoded representation of the right. And now we can see...
So we saw a number of different slides showing facsimiles, encodings and also transcriptions, both diplomatic transcriptions and edited transcriptions for reading.
Here is our goal for eLauter, these are CSS and HTML mockups that we are currently working to fill with the data from the project. I don't know if you can still hear me, I seem to have fallen off grid.
Sorry, I seem to have lost connection briefly but I think I'm back. So here you can see different views that will be possible, Italian lutablature both in diplomatic and edited transcription. And then here an example of IIIF image on the left of the tablature and the vocal model that Ilya has mentioned on the right.
And then here we have an example of some French lutablature rendered and the MEI code that is backing that up. So as I mentioned these are all generated using the digital infrastructure at UrnB.
This is built on top of GAMS, the asset management system for humanities data developed in Graz in Austria. On the one hand we have the encoding files, MEI and TEI files as well as IIIF. These encodings are translated to RDF using some XSLT transforms.
The encodings are also rendered for use in the browser. The music now using Verovio, this is an extension developed as part of our Elauter project. And then on the other hand we have the graph that is being generated from the relational database entered by our humanists.
And this is stored in a BlazeGraph triple store that is connected to this GAMS database. This GAMS asset management system is used to then fill in individual components on the HTML view that is generated.
But we are in a session about end users and this project is very much about serving this somewhat forgotten data to people who are excited and interested in it. So that includes musicologists of course but also very much the community of lute enthusiasts and lute players.
This is a relatively small but very engaged community who often have a strong digital affinity as well as can be seen with various predecessor lute projects that are available electronically. And this German lute aperture is thus far very very much underexplored also in the digital realm.
So we want to be able to serve these people, we want to allow them to annotate, share annotations, share their own performances in using our resource here. And we have a number of predecessor projects that we've been involved in on annotations in the context of linked data and music encoding.
These have traditionally been web annotations that target fragments of an MEI file, fragments, media fragments of audio recordings or regions of IIIF images.
But a disadvantage of that is that we're always annotating on the surface of our evidence objects of these web resources which doesn't really fulfill the semantic role that we want. We're wanting to make statements about musical phrases for example, not about some XML elements.
And so this music annotation ontology which was developed by our colleagues at the University of Oxford and Paderborn comes to our rescue here. It injects this middle layer of musical objects which allow us to describe
different facets of information as being part of the same, for example, musical phrase. And this can then be annotated using our web annotations as musicological objects rather than the musical objects that are being annotated. Right, thank you.
We have built an editor for music encodings that runs in the web browser. It's called MEIFriend. It has been taken up quite enthusiastically by the music encoding community. But relevant for this talk is that we also have an annotation panel built into this editor.
And this is able to generate web annotations, RDF annotations of the musical objects you saw on the previous slide. So it's possible to select parts of the score and identify them as a musical object. And then also to use these other annotation functions to make statements about these objects.
Thus the musicological objects of the music annotation ontology. And all of this is powered by the Solid platform, if you're unaware of it. It's a project for decentralized personal data on the web. It is both a storage place for linked data and other types of data as well as an identity provider.
So it's possible, for example, with MEIFriend to log into a Solid pot. Filius, thank you. To write annotations into Solid Read from there. To share them with other users' Solid pots.
So for example I can share with my colleagues or my friends. Or also to publish and to the wider web of data. Other applications can then be built that also use the same resources integrating with the other applications on the level of the data within the Solid pot.
And further applications can then have the same role sort of interacting on a data level. And so we are planning to make use of this in a pilot project called LÜT.digital. In which people will be able to annotate the resources served through the ELATI projects with their own textual annotations but also with audio examples.
So that concludes the talk for today. Here are some references and I thank our funders and the colleagues in our project across various different institutions. Thank you.
But thank you for your talk. It was really impressive. As you said it was complex but it's impressive really. Maybe I can ask you one tiny question because we are actually out of time. But did you have contact to other projects that are using EI for example? So yes.
So EI is a community effort that's been developed over the last 20 years. And we as a project are very closely involved with the activities that are being undertaken as part of this initiative. And specifically for this project we actually contributed an extension for German LÜT tablatures.
So there was already support for other kinds of tablatures within EI. But this slightly idiosyncratic German tablature format wasn't supported before. So now this schema has been extended and will be featured in the next release of MEI.
And there will also be a release of the Verovio renderer coming soon which will be able to render these tablatures graphically. That sounds really good. Thank you. I really like that you share your knowledge in that way and really contribute also to the MEI in the big scale.
So let's see how many other LÜT projects will come up in other countries maybe now that they heard your talk. Okay but then I will conclude this session.