Richter Pál, Pávai István, Mórocz András
Folk Music Archives on the Way of Becoming Public
Revista de Etnografie şi Folclor – Journal of Ethnography and Folklore,New Series 1–2. , p. 49–52, Bucureşti, 2009
The Folk Music Archives at the Institute for Musicology (ZTI) preserves circa 18,000 hours sound recordings of folk music and audiovisual recordings of folk dance, including the phonograph cylinders of the Museum of Ethnography, the majority – about 70% (12,000 hours) – digitized. One of the most important tasks – beside the proper storage and retrievability – is to enter the data about field collections and sound recordings into a computerized database. This work began in the early nineties with the data processing of the closed and the so-called historical systems separated from the rest of the material, then after the unification of the earlier data input methods, in 1999 we created an interactive and dynamic Internet website-system that provides a database in which the main types of old Hungarian folksongs are searchable by their musical attributes and acquisitional and geographical details. The database system developed especially for folklore archives (Folklore Archiving System) is integrated with a thesaurus, a hierarchical structure of keywords which is maintained by the cataloguists themselves. The metadata system uses innovative features to allow high level, realistic cataloguing. As a result of recent developments, there are several interfaces to provide access to different subcollections of the database through the Internet or provide access points to external systems such as ethnoArc. The uploading of different part-units and part-databases has started with the Bartók-system, and we aim at making the whole collection available to researchers and for the general public as well. ZTI’s database system aims detailed, scientific level of metadata processing while providing convenient search tools for both scientific users and the greater public.
The Folk Music Archives at the Institute for Musicology (ZTI) was officially established on January 1, 1999. Naturally, however, the Folk Music Research Group of the Hungarian Academy of Sciences has had a technical department since the very beginnings (1951) and continued work under the auspices of the Institute for Musicology which was created through the merging of the Bartók Archives and the Folk Music Research Group. The archive housing sound recordings of folk music and audiovisual recordings of folk dance has been systematically set up and has won outstanding renown in Europe. The sound recordings of the Archive include 18,000 hours of music, 15,000 hours of which are the Archive’s own recordings, that is, the source materials can also be found here, and the remaining 3,000 hours are made up of copies of recordings stored in other archives. The latter include the phonograph cylinders of the Museum of Ethnography, which are available both in analogue and digital form in our Archive. The majority of the sound recordings, about 70% (12,000 hours) have already been digitized.
Naturally, proper storage and retrievability of data are an integral part of archiving. Thus, another important task is to enter the data about field collections and sound recordings into a computerized database. Following the initiative of Ferenc Sebő the computerized registry of the folk music collection of ZTI began in the early nineties under the guidance of László Dobszay. The electronic dataprocessing of the closed and the so-called historical systems separated from the rest of the material took place within the frames of this work. From the mid-‘90s István Pávai suggested the unification of the earlier data input methods and their conversion to an up-to-date database. These steps were taken soon. With the help of a competition in 1999 we created an interactive and dynamic Internet websitesystem that provides a database in which the main types of old Hungarian folksongs are searchable by their musical attributes and acquisition and geographical details. The capacity of data-processing systems and the possibilities of computerized immediate display of different media have grown remarkably in the last decade. This extensive technical development makes it possible that earlier independent databases of part-repertories be joined, and data be supplied with the playing of the media attached to them.
The software development that began in the first years of the new century on the basis of István Pávai’s plans served this purpose. As a result we now have a database management system at our disposal. The system was developed especially for folklore archives (Folklore Archiving System), and the uploading of different part-units and part-databases has started. Since the Bartók-system's was the most detailed upload that was conducted in the ‘90s and we succeeded in digitizing the record cards, the Folk Music Archive had the chance to first publish the Bartóksystem online from its collection (http://db.zti.hu/br). Similarly to the Bartók system we aim at making the whole collection available to researchers. The website will only differ from it in that the restricted parts of the collection will not be accessible from the web. We are able to provide filtered data from the subcollections and present the material in keeping with the different classification systems.
ANTECEDENTS AND ASPECTS OF THE FOLKLORE ARCHIVING SYSTEM
When István Pávai started to work for ZTI (1994) he continued work on the cataloguing of the Bartók System and the System of Folk Song Types with Ferenc Sebő and Pál Richter. Both systems are so called logical (music) file systems. The computer database for the sound archives has been developed by a team led by István Németh. Realizing that the two tasks needed to be coordinated Pávai suggested that the work of the two teams be harmonized, however, this took several years of preparation due to the abundance of data, inaccurate data entry, the incompatibility of fonts from various periods and an incongruence among the fields
of individual databases.
In the meantime, under the leadership of István Németh digitisation of the sound archives and saving the material onto CDs began. When deputy director of ZTI, Tibor Tallián made an invaluable contribution by applying and winning grants to fund the project, and later as director, he established the Folk Music Archives and asked Pávai to work for ZTI full-time and be the head of the archives.
In parallel, in the Hungarian Heritage House (HH) the László Lajtha Folklore Documentation Centre was established to make recordings of folk music and folk dance accessible to the public, a duty that had earlier been performed by the Resource Centre for Folk Dancers. The two institutions (ZTI and HH) were to have two distinct functions; the Folk Music Archives of ZTI, the central archives of Hungarian folk music, acquired all the existing records, whether original or digital copies, whereas the Folklore Documentation Centre of HH was to make a limited number of copies accessible to the general public for educational and cultural purposes. The legal predecessors of the two institutions had a long history of cooperation, thus, it is expected that collaboration between the legal successors will be equally fruitful.
Recordings were archived in a form that was durable and ensures data integrity granted by the technical and financial resources in a given period. At the same time, separate copies were made for researchers and for the general public, which may be used without corrupting the archive copy.
The earlier, off-line procedure of retrieving the material was to find the call number of the recording in the card catalogue, find it in the archives shelves and play the selection on a sound system. In contrast, users are rewarded by a video tape with a two-hour movie for the time spent on searching in the rental shop.
Whether examining folk music from a scientific, educational or cultural aspect, users are dealing with a set of melodies scattered among several different types of media, therefore, the time spent on trying to find them greatly increases.
Computer databases, on the other hand, offer a time efficient alternative, however, they necessitate a digitalization of the entire collection, creating archive copies, good quality compressed copies, a thematic structuring of the database, a thematic segmentation of the media, and defining the algorithms for indexes and retrieval. Naturally, a central server is also required which can be accessed by client systems.
The development of information technology, especially the possibility of media processing on low cost hardware called for the creation of a multimedia database system, which serves as a well organised and detailed database for the metadata and provides instant access to the multimedia content as well.
ZTI’s database system aims detailed, scientific level of metadata processing while providing convenient search tools for both scientific users and the greater public.
The system is integrated with a thesaurus, a hierarchical structure of keywords which is maintained by the cataloguists themselves. The thesaurus also serves as a structured view of parts of the folklore science applicable for the media content. Each branch has at least one descriptor (the “official” keyword) and can have multilingual synonyms (alternative notations). In addition, horizontal relations can be set up between branches, allowing search results to be extended. Due to the multi-hierarchical features, the same term can appear at different branches, therefore cataloguing and search functions can be performed in different thematic ways.
The metadata system uses innovative features to allow high level, realistic cataloguing. Thesaurus fields hold references to terms in the thesaurus, thus providing additional search features such as search on synonyms or collective terms. The date format allows usage of intervals and incomplete dates (i.e. year only, month only, seasons, etc.). In many cases multiple keywords are appropriate for describing document properties. Thesaurus fields and wordlist fields allow multiple values. As the thesaurus, wordlists are also maintained in multiple languages that allows not only search for language alternatives but display of data in user defined language.
Audio and video material is processed at media-part (segment) level. Metadata sheets are attached to each segment defined. With the help of the segmenter module, parts of the material can be selected at frame (HH:MM:SS:FF) accuracy. Navigation within the media is helped by various tools including audio envelope (generated by background process).
Media (video, audio, image) processing exploits Microsoft Windows standard subsystems (DirectX, GDI) providing compatibility and easy handling. Export and import functions are based on the usual Office tools.
As a result of recent developments, there are several interfaces to provide access to different subcollections of the database through the Internet or provide access points to external systems such as ethnoArc. Interfaces ensure adequate security and regulation of access rights regarding the legal constraints of publishing the material.