Sony CSL Paris creates an AI that can make its own drum tracks
Promises it’s “not to replace musicians” 🙄
Sony CSL Paris, also known as Sony Computer Science Laboratories in Paris¹, has created a new AI (read: algorithm, because there’s no such thing as artificial intelligence yet), which, after scanning 665 songs, is now able to generate its own kick-drum tracks. AJ Dellinger from Engadget:
Sony is the latest company to dip its toes into AI-powered music. The company revealed this week that its researchers have created a machine learning model that can create kick-drum tracking. According to Sony, the artificial intelligence is capable of building “musically plausible” drum patterns based on existing instruments being used on the song.
In order to train the AI system, Sony’s researchers compiled data from 665 different songs from a wide range of genres including pop, rock and electronica. The songs all feature ryhthm instruments, bass, kick and snares that were available as separate 44.1kHZ audio tracks. With the contexual signals of those tracks, the researchers created drum kicks by setting a drum samples at all amplitude peaks. The AI system conditionally generates the kickdrum patterns based on the characteristics of the other material that it is placed around, regardless of the song’s tempo and changes in speed or duration.
Sony, of course, promises that there’s nothing sinister at work here and that everything they’re doing is meant in support of artists.
The goal is not to replace musicians, but to provide them with better tools to be more efficient in realizing their creative ideas. DrumNet is based on an artificial neural network which learns rhythmic relationships between different instruments and encodes these relationships in a 16-dimensional style space. A similar example is the Logic Pro X Drummer, allowing the user to specify the playing style by navigating a two-dimensional space. The difference of DrumNet to the Logic Pro X Drummer, however, is that it dynamically adapts to the existing music. In its current form, DrumNet can either autonomously generate kick drum tracks (following the statistics of the training data), be controlled by manually navigating the style space, or be used to extract a style from an existing piece.
As opposed to many other generative music technologies, we aim to directly use existing audio tracks as input to which we generate the kick drum track as audio output. Using audio input directly makes DrumNet more useful for music production than models based on MIDI input. We show the generality of the model, by providing many examples of full songs with different generated kick drum tracks on this website. For a proof-of-concept, we trained the model only on kick drum rhythms, but we are currently extending the model to generate a whole drum set.
As you may or may not know, after their purchase of EMI, Sony became the world’s largest record company, which in their most recent earnings reported a healthy profit.
Sales at Music improved 11.5% to ¥202.3 billion due to higher sales for Music Publishing resulting from the consolidation of EMI, and higher sales for Recorded Music. The segment’s operating income was ¥38.3 billion, which increased from ¥32.1 billion in the prior-year quarter.
The biggest contribution to those earning results? Streaming. So much of music today is already around today and gone tomorrow, but with the advent of streaming, we now have access to those millions of songs at any given time, resulting in record companies making a return on everything for infinity. Unfortunately, artists only see a fraction of those returns, but they do get something.
With money left on the table, rest assured that record companies, knowing how small of a player they are in the modern world compared to other entertainment mediums, are looking at every possible way to get even more money from music sales. The fewer artists they have to worry about, in this case drummers, the more money they can hold onto.
It’s further hard not to project that, in a decade from now, they’ll have songs that have been purely auto-generated by their ‘AI’ systems, thus removing any cost barrier, like paying an artist.
You can hear the sample tracks created here, but no matter how good, I don’t know about you, I like my art created by artists and not computers.
¹ I had no idea this division even existed
Thanks for being an SRN reader. If you liked this post and want to help me cover Sony even more with reviews, videos, and podcasts, please clap (you can 👏 more than once!), share on your favorite social sites, and subscribe to the SRN newsletter.
