The act and process of songwriting is vastly under-researched within popular music studies (Barber 2016 b), despite being such an integral part of not only the music industry but to the concept of music. Whilst Barber looks specifically at the process of songwriting in terms of passion and productivity (Long and Barber, 2014), it is important to look at the current developments within the field of songwriting.
Like many areas of the music industries, the current developments within the field of songwriting are primarily revolving around technology. This isn’t surprising; technology has dictated the actions of the music industries since music became an industry – Wikstrom even suggests that technological developments are the easiest way to map the chronology of the music industry through the framework he utilises in writing a music industrial-history (2009). In addition, Burkhart demonstrates how technology can disrupt and impact the musical distribution process with his celestial jukebox model (Burkart 2013). It is clear technology can disrupt industrial musical processes, but how has it affected songwriting?
The use of ‘machine learning’ has already been used by companies such as Youtube for content ID purposes to help combat copyright infringements, as well as to pull apart tracks from complete mixes. Recently however, there has been a new emphasis placed on ‘machine learning’ (or AI) in terms of music creation. In a less disruptive manner, the technology will be able to be used by artists and composers to give them more options when creating music – although this isn’t the primary concern for some songwriters. The primary concern is that the technology could feasibly replace some forms of composition completely, even if this isn’t the main purpose of the technology (Cooke 2016).
This raises many questions – based on technology’s effects on other music industry sectors, one can assume it is going to drastically change the songwriting process. However, it is certainly questionable if a machine or algorithm could replicate the aspects of the songwriting process as listed by Barber (2016 a), such as passion, inspiration, preparation, idea generation, routine and collaboration, amongst others. This isn’t to say AI has never been used to compose music – it has, with a song known as ‘Daddy’s Car’ being composed by an AI program called The Flow Machine, using a vast database of varied sheet music, a human style prompt and human-written lyrics, production and mixing (Lobenfeld 2016). In addition to this point, the discourse surrounding the idea of ‘manufactured pop’ is very supportive of AI-composed music – the name alone connotes ideas of music being created in a factory production line, not dissimilar to Adorno’s theory of standardisation (1990), in which every song at its core is the same.
The idea of songwriting becoming an automated process then, is not a new one. Since Adorno’s lamentations of the standardisation of popular music there has always been a discourse based around the idea of mass-produced music. However, it isn’t until recently that the act of songwriting has been challenged by a new technology, which could eliminate the need for some human composers in certain types of music. It could however also be used to essentially write new songs for legacy artists such as The Beatles or Motorhead – Daddy’s Car was programmed to be written in the style of The Beatles, theoretically an algorithm could be developed that could take The Beatles’ body of work and either create new compositions in a similar style or could even predict what the band would have written next.
- Adorno, T. (1990). On Popular Music. In: S. Frith and A. Goodwin, ed., On Record, 1st London: Routledge, pp.302-314
- Barber, S. (2016 a). Songwriting and the Recording Industry.
- Barber, S. (2016 b). The Brill Building and the creative labour of the professional songwriter. In: K. Williams and J. Williams, ed., The Cambridge Companion to the Singer-Songwriter, 1st Cambridge: Cambridge University Press.
- Burkart, P. (2013). Music in the Cloud and the Digital Sublime. Popular Music and Society, 37(4), pp.393-407.
- Cooke, C. (2016). AI may replace some forms of music composition, but it will enhance others | Complete Music Update. [online] Completemusicupdate.com. Available at: http://www.completemusicupdate.com/article/ai-may-replace-some-forms-of-music-composition-but-it-will-enhance-others/ [Accessed 7 Dec. 2016].
- Lobenfeld, C. (2016). Hear the first-ever full pop song composed by artificial intelligence. [online] FACT Magazine: Music News, New Music. Available at: http://www.factmag.com/2016/09/22/hear-first-complete-pop-song-composed-artificial-intelligence/ [Accessed 7 Dec. 2016].
- Long, P. and Barber, S. (2014). Voicing Passion: The Emotional Economy of Songwriting. European Journal of Cultural Studies, 18(2), pp.142-157.