Handcrafted perfection? Jonathan Sterne’s talk on the music mastering service LANDR

von Neela Janssen

Websites, programs, apps - talking about IT and online services surrounding us everyday in an academic context, we are often lost for words: their use and presence is too familiar for academic distance, their structure and programming too complex to understand, their impact feels diffuse and diverse. And still, as Jonathan Sterne’s talk on the music mastering service LANDR shows, much can be gained if one is willing to struggle through the complex task of approaching contemporary online phenomena. Sterne most significantly shows how important it is to understand not only how an online service is used, but also it’s underlying programming and data structure. Only through taking this basis seriously we can get an insight in the dynamics of these services and their use.

Sterne consequently starts with LANDR as an example of artificial intelligence and machine learning: LANDR is mastering music, not with the help of a human mastering engineer, but through sound analysis and its access to training data.

Why is mastering a task that could be done by artificial intelligence, Sterne asked the people behind LANDR. The answer he presents in his talk is then both motivated by social aspects as well as technical factors: On one hand, mastering is usually done without the artist present. In contrast to mixing, where the artists usually give their opinion and the songs are edited multiple times, mastering is the last and final step, done solely by mastering engineers in their studios. On the other hand, mastering, again in contrast to mixing, is much less genre-specific. Which means, that even though a program has to handle different genres, the task of mixing these songs is quite the same. This is important because, as Sterne’s interviewees confirm: ‟We can do mastering with data.”

Sterne is then mostly concerned about the impact on sound standardisation and the profession of mastering engineers. As the program behind LANDR has no concept of genre and relies on the training data as well as on the data it is presented for mastering by the customers using it, the question of its standardising effect is indeed a pressing one. There are always cultural politics involved in the process of creating such a program with its categorisations and stylistic choices and these politics should be looked at, as Sterne emphasises multiple times.

Still, he also notes LANDR’s democratic factor: it provides cheap and fast service to people in the music industry as well as for those who want to get in or are just starting to produce music or sound.

With LANDR, they get access to a “professional sound” that would otherwise exclude DIY artists from the mainstream scene. LANDR is able to fabricate tracks that can be uploaded to streaming services as Spotify or Soundcloud and listened to without distinction, at least for the consumer’s ear.

In this fact lies something that connects LANDR to many other services found online: the access to (at least so-called) professional services to non-professionals. It doesn’t matter if it’s about compiling a photo book from the latest family camping holiday, turning a wedding picture into a high definition wallpaper or baking a cake not only with exquisite ingredients found online, but refined with a coloured sugar coating made with professional supplies shipped to you. For every creative process, there seems to be a professional finish available if you’re willing to pay some money. In the case of LANDR as in the case of all the other examples listed above, questions arise: Why is it so appealing to make your own creative product look, sound or taste like the ones that can be bought in stores? And what does that mean for the professionals and their alignment? What is the impact on our definition of hobby and creative recess in contrast to profession and art?