Interview Joachim Veit

It’s called TEXT-Grid, after all. As a musicologist, do you often work with text?

Veit: Our focus is of course on the music, or the notation. To that extent, in a sense we work with “special texts”, but texts consisting of words also play a big role when we are using sources such as the letters or diaries of composers, reviews etc. It gets complicated when we try to link word texts and notes to each other.

Living in the computer era – is the creation of note texts still that complicated?

Veit: There is a lot of good software, but editing musical notation is still very costly and full of errors, especially since modern programs with graphical user interface contain many automated processes that we need to disable in order to use the tool for scholarly purposes. For example, we place additional information in brackets or use ties that are interrupted in the middle to provide the user with information about the sources of the works. The standard programs were not designed for such subtleties, which is why at the Edirom project we are working on the development of tools for digital music editions. For several years now, we have been able to offer the user synchronized retrievable facsimiles of the music as graphic images. This represents significant progress for our work. When encountering different versions of texts in reviewing the manuscripts and earlier editions for the creation of a music edition, we can now visualize these variations. In the score, however, which is used to generate the individual instrument parts for practical use, these variants are not yet digitally usable. One of our main goals is therefore the semantic encoding of music, enabling us to view the different versions side by side and to really make use of them. We are participating in a project initiated by the University of Virginia at Charlottesville which is called the "Music Encoding Initiative (MEI)" in reference to the Text Encoding Initiative (TEI). Our goal is to develop encoding for the content of musical notation for research purposes. In the TextGrid project, we want to use this encoding and develop an input tool and a relatively simple viewer.

Are you developing your own music notation program?

Veit: No, that is not possible in such a short time with so few people. The first step is to provide a basic way to enter and display this code in a graphic user interface. The amount of code that represents a page of music is much larger than the code representing a printed page of text. If music texts are to be encoded in a machine-readable format, at present several thousand lines of code need to be entered for a single page of the score, either through the additional step of a rudimentary import from another format or by hand. This is not feasible for the long term. Right now our team is developing a graphic interface where you can click on certain types of notes – such as those with stems, beams and dynamics – just like you normally click on special characters, and then all data will be stored automatically in the MEI format. One click can thus generate 30 or 40 lines of code.

That will be a huge number of special characters…

Veit: In some respects, yes, but one of our problems is that there are so many different display options for the same musical content. Progress in the digital domain of musicology has been very slow for this reason. Our preferred method of data entry works reasonably well for content in the field of older music where the notes usually represent little more than pitch and duration, perhaps with an underlying text, which is easy to process. With classical music or even music from the Romantic era, the number of characters increases drastically and many of them are ambiguous. All these details must be taken into account when entering musical content. Think of a tie, for example: it can indicate that the notes are connected to a musical phrase, but it can also mean that a wind instrument should play this passage legato or that a string player should switch from up-bow to down-bow. Then there are these small dashes (strokes), which have caused much debate among musicologists, that are perhaps meant to signify a type of staccato. On closer examination, however, they can sometimes mean very different things, such as an accent, or a certain style of playing.

Can you identify more subtle details than before by using digital methods?

Veit: We have developed a type of synopsis tool in the Edirom project that we can use to display five or six sources side by side and compare them directly with each other. In this way, details and differences between manuscripts that had previously gone unnoticed are coming to the fore. Unfortunately we can not automatically search for these variations. This is of course our long-term goal, because only then will it be possible to operationalize the whole thing.

Is collaborative work an issue for you?

Veit: We are cooperating on the Weber edition with the Berlin State Library. All of us are working on the same sources, most of which are stored as facsimile files on our servers in Detmold. Unfortunately, our colleagues in Berlin are having difficulty setting up a steady connection to these servers. TextGrid would be ideal for this purpose because then there would be no problem making large amounts of data available and working together. Since everyone uses different programs, it has often been the case that not all of the special characters they inserted were always displayed. It is a giant step forward for us that our team has already integrated the musical Unicode characters fully compatibly into all of the editors in the TextGridLab. Instead of having to insert individual notes as graphics into the text, for instance, they can finally be entered as text characters. There is also a special field containing the standard combination of musical characters, so users can see them on the screen directly and easily copy them from the TextGrid program into other programs via drag and drop.

So you are already actively using TextGrid, although it is still in the beta phase?

Veit: Yes, and I also regularly use the dictionary network project when working with historical texts, such as looking up unfamiliar terms when preparing an edition of letters. It not only works much faster than poring through thick encyclopedia volumes, but it is also much more productive because there is the incredible opportunity to compare the various dictionaries. I also expect to benefit from the acquisition of the text corpus from the Digital Library. For instance, the biography of Weber written by his son is of immediate interest to us. It is still the standard biography, although there are some issues with it. So far, however, we lack detailed metadata and markup that would allow us to use it effectively. As soon as we have the possibility to check and enrich these data with comments, we will jump at the chance to make use of it!

Will conductors and musicians read music from screens someday?

Veit: At the Expo 2000 they experimented with this, but I doubt that playing from a screen will ever become the rule. The advantage for the conductors is much greater. When they work out a score in advance, they have to make a series of decisions, especially with recent scholarly editions that don’t restrict the range of possible interpretations that are always present. Imagine if conductors could click on the selected variants on the screen in an operationalized musical text to create a score that they deem consistent. From this new score they could then extract the individual instrumental parts they want to use with the orchestra. Against the background of such prospects, my current work sometimes seems almost like punishment: I compile my volumes with the aid of Edirom tools, and I would really like to pass on these new opportunities to users. But in the end we still print normal volumes on paper. For the printed version, I must always choose one variant over the others, and I can only offer the other variations to readers in difficult-to-use texts that are mostly without images.

Won’t musicians have a lot of reservations about using digital editions?

Veit: In our edition of Weber’s Clarinet Quintet, we enclosed a DVD with a comprehensive report and all the important sources for the edition in facsimile, giving everyone a clear picture of the variations as well as of the editors’ work. Musicians can choose to access this knowledge on different levels, or they can decide to view only the section that relates to their individual instrumental part. Thus an expensive scholarly edition can be utilized by a wider audience. The reaction from musicians was, “we want more of this”!

Interview by Esther Lauer.

Back to the press kit …