My Thoughts on Beat-Making Bots

Artificial intelligence (AI) has found its way into almost every industry remotely related to technology, and music is no exception. Recently, technology has been developed that allows artificial intelligence to create entirely original music from scratch. I would like to preface this article by establishing that–yes–I may have a slight bias against music-making AI simply because I’d prefer to not be replaced by bots in the future realm of music. I’m just kidding–these music production bots are incredibly fascinating and I feel they have remarkable potential across music.

A website I’ve experimented with is ecrettmusic.com, and I feel it is a stellar example of AI’s potential in music production. This website uses a user-provided genre, mood, and scene–the music’s use case–to generate a beat entirely from scratch. This site randomly generates the melodies onsite, so no two users will ever receive the same melody, showing the versatility of Ecrett. Ecrett’s user interface also mimics an actual DAW, allowing the user to toggle the stems in the piece, giving access solely to the parts they want. What surprised me the most about Ecrett was not its usability or even its existence–it was its quality. With no more than 4 clicks, the site was able to make beats with multiple layers and great arrangement. I honestly wouldn’t be able to tell the pieces were computer-generated had I not known previously.

Another website I had experimented with was boomy.com. This platform, however, impressed me far less than Ecrett did. Boomy also functions based on a user-provided genre, but I feel the final products I was given failed to compete with Ecrett. There was almost no melodic content to the pieces I was given, and as YouTuber Patrick CC put it in his video using the site, “Boomy is basically just making noise.” Although Boomy’s melodies were not too great, I was impressed by the quality of the percussion it generated.

A trait I noticed common to both Boomy and Ecrett was that the music was quite repetitive. Ecrett handled this slightly better because it had more intricate arrangement, but both sites generated music that, by modern standards, was a little stale. However, this leads me to believe that while beat-making bots may not be as applicable in the commercial record industry just yet, they may be greatly suited for producing music in settings like some scoring for video games or films, where music is meant to be a little more repetitive.

Although beat-making bots might not be able to replace the record industry’s top producers quite yet, their very existence is astonishing. The fact that any music at all can be computer-generated onsite from scratch shocks me because it reveals that quite literally anyone or anything can create music, given a little guidance.


Previous
Previous

The Value of Stem Separation

Next
Next

My Weekend at the Guthman Musical Instrument Competition