Technically, metadata is the essence of a streaming service. It has an impact on all levels of the streaming experience since it defines the search results, discovery features and browsing experience. The size of the catalogue is equally important as the quality of the structured data. Both are needed so that for example the recommendation algorithms can use the data to select the album it thinks you are looking for. The prospects and possibilities are endless within the field of metadata.
Let’s zoom in to the classical music peculiarity and why it is so unique. First of all, metadata for popular music is relatively simple as it only covers fields such as the name of the artist, possibly a featuring artist, track name and the name of the song – that’s basically it. On the other hand, classical music structure is far more complex. A composition can be performed by multiple artists, each with their own interpretation. Some classical works have been recorded more than 100 times, but other streaming services are not designed to enable you to choose effectively among the hundreds of albums that have been recorded.
Second, one work in classical music can have several movements (parts). Most concertos have three movements, symphonies four and operas typically have over 10 parts. Other streaming services are designed to browse through single movement songs because that is the dominant structure in pop music. A classical music lover however does not typically want to start with the third movement but prefers to listen to the full work from a specific composer.
Third, classical music often has more than one artist involved in the recordings: there can be a soloist, but also a conductor and an orchestra. Sometimes there can be several soloists each with their own influence on how this particular recording sounds. For pop music fans it is typically less important who plays the instruments on the track, but for the true classical lover it matters whether Lang Lang is supported by the Berlin Philharmonic or the Boston Philharmonic Orchestra.
Fourth, there are the spelling variations. Tchaikovsky is not always Tchaikovsky. He is spelled differently on different albums and we need to understand that Peter Tchaikovsky, P. I. Tchaikovsky, Tchaichovski or Tchaikovsky all lead to the same composer. On the other hand, let’s not forget that families like Bach and Strauss had multiple famous composers whereas separating their works from each other requires editorial attention for identifying the Strauss I from Strauss II.
Fifth, there is other complexity to take in consideration. Think about for example cataloguing system like the opus numbers, nicknames like ‘the Moonlight Sonata’ or ‘Haffner’. Operas with acts and arias within acts, the works of anonymous composers, identical composition names in the repertoires, remastered albums of older recordings and albums with works of different composers. The list is endless.
As a conclusion the meta data of classical music is much more complex than the metadata of pop music albums. The algorithms the existing streaming services use are not treating classical music different from pop music. The result is obvious: a mess. Works do not appear in logical order, movements are presented separately and inconsistently and the names of artists and conductors are not displayed. Most classical music lovers undergo the so-called ‘search drama’ every time they search for a particular work or recording.
The solution is easy, in theory: building a database with all classical music works ever composed. In reality it is of course not that easy. There are more than 100,000 classical works composed and the number of metadata parameters is enormous. Name of the composer, opus number, work title, name of the movement, nicknames of works, names of the conductors, name of the lead artists and the name of the orchestra just to name a few. Due to the complexity of it, this kind of database does not exist. Or actually, did not exist, until we created it.
Our team of musicologists and data engineers have worked for more than two years on creating an encompassing classical music database - and the work is on-going. By leveraging existing databases, books, online sources, formulas and a lot of manual editorial work they have created the Primephonic classical music metadatabase. Classical music as a genre is alive and kicking, new works and albums are added every week. And although we believe we have more than 95% of all works in our database, we may have missed some. If you notice a missing work, please contact us at firstname.lastname@example.org. Together, we can address the need for a classical music metadatabase. Classical music and you all deserve it.