ABSTRACT
This paper traces the infrastructural politics of automated music mastering to reveal how contemporary iterations of artificial intelligence (AI) shape cultural production. The paper examines the emergence of LANDR, an online platform that offers automated music mastering, built on top of supervised machine learning branded as artificial intelligence. Increasingly, machine learning will become an integral part of signal processing for sounds and images, shaping the way media cultures sound, look, and feel. While LANDR is a product of the so-called ‘big bang’ in machine learning, it could not exist without specific conditions: specific kinds of commensurable data, as well as specific aesthetic and industrial conditions. Mastering, in turn, has become an indispensable but understudied part of music circulation as an infrastructural practice. Here we analyze the intersecting histories of machine learning and mastering, as well as LANDR’s failure at automating other domains of audio engineering. By doing so, we critique the discourse of AI’s inevitability and show the ways in which machine learning must frame or reframe cultural and aesthetic practices in order to automate them, in service of digital distribution, recognition, and recommendation infrastructures.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1 From Jonathan’s informal conversations with colleagues in these organizations, it appears no other major changes occurred apart from rebranding.
2 We have chosen LANDR over competitors for reasons of approach. Izotope uses its machine learning to design software plugins to be installed on users’ computer, not in a dynamic ‘real time’ arrangement and therefore is not exactly comparable. Their algorithms are more ‘hand crafted’ with a machine learning supplement. Cloudbounce is less heavily capitalized, but also less accessible to authors.
3 There are numerous internalist accounts of the recent history of machine learning by computer scientists eg., http://www.andreykurenkov.com/writing/ai/a-brief-history-of-neural-nets-and-deep-learning/. As of yet, we have not found a history of this period of the field that treats the science as itself a cultural and social phenomenon.
4 For more on LANDR’s connection to the local scene, as well as how its algorithm works, see our companion essay (Sterne and Razlogova Citation2019).
5 Spoken language in interviews is lightly edited to read better as written language.
6 Crane did point out that there are services which provide both mixing and mastering, and some artists do their own mastering, but these are an exception. No mastering engineer we spoke with seemed threatened by LANDR; Crane reported the same impression.
7 This may be in part because of the contexts of DIY music production: a home studio with significant acoustic irregularities and consumer grade speakers would create more problems for a mastering engineer to solve than a recording done by professional engineers in a commercial studio.
8 Based on a technique called Music Information Retrieval, music recognition analyzes parts of a recording to compare it to an available database of recordings, and then makes a ‘guess’ as to a match.
9 We discuss this further in our companion essay (Sterne and Razlogova Citation2019).
Additional information
Funding
Notes on contributors
Jonathan Sterne
Jonathan Sterne teaches at McGill University and is, most recently, author of Diminished Faculties: A Political Phenomenology of Impairment (2021). Visit his website at https://sterneworks.org
Elena Razlogova
Elena Razlogova is an Associate Professor of History at Concordia University. She is the author of The Listener's Voice: Early Radio and the American Public (2011).