It seems like the “Rise of the Machines” has been a recurrent theme in the news lately, with even the New York Times running stories about how robots are taking our jobs and artificial intelligence in various forms is being used more and more in the workplace. But what about machines writing music?
Will robots and AI (artificial intelligence) become more involved in the music world as well, writing music for us or each other or at least acting as our intelligent studio assistants?
I don’t think electronic musicians are at risk of being replaced by robots quite yet, but I have noticed a couple interesting plug-in releases lately that offer to help you write music by generating musical notes and patterns based on your MIDI input. Composers and technologists like Tod Machover at the MIT Media Lab have been working on projects like this since the 80s, but now it seems that more and more of this research is filtering out of the lab and being put into a form that we can all use.
Plug-ins and software that actually generate new musical material like this are called ‘generative.’ One example of generative music software that some readers of this blog may be familiar with is the Newscool drum sequencer in Native Instruments’ Reaktor:
The Newscool sequencer uses a version of Conway’s Game of Life to generate new musical material. In Newscool, once you set up the initial parameters in the sequencer and press play, it will start running and continuously spin out new variations and beats based on its initial state, according to a set of simple mathematical rules.
But why use generative plug-ins at all? Is it cheating? Shouldn’t you write all the music yourself if you truly want to be original?
I don’t think so at all; using generative plug-ins like this can be a fascinating way to come up with new ideas and directions in your music. Of course, you always need to apply your musical taste to the final output and decide what sounds best out of the generated bits and pieces, but as we all know, it is often the ‘happy accidents’ and semi-random collisions that make for interesting music anyway, so why not use generative software to help come up with more of them?
Jnana is a Sanskrit word that means knowledge, and the new Jnana plug-ins from Collin Sullivan at Stanford’s CCRMA (the Center for Computer Research in Music and Acoustics) work by actually analyzing and understanding any MIDI material you’ve entered in Live, then generating variations and responses to it in a similar style.
There are actually two related Jnana plug-ins: “Jnana Live” will listen to real-time MIDI input and generate its own melodic phrases in response, and “Jnana Clips” will analyze preexisting clips in Live and generate variations on them.
If this description sounds a bit baffling, check out this video posted by the creator to get a better idea of what Jnana can do: a pianist plays phrases that go into Live and then Jnana comes up with musical responses in a similar style:
The Jnana plugs are open source and free to download from the creator’s website at Stanford, but note that they run exclusively in Ableton in Max for Live; they won’t be recognized at all without Max for Live installed.
I downloaded and installed the Jnana plug-ins myself without much trouble on OS X following the setup guide Sullivan has helpfully provided. There is no setup guide for Windows, but Sullivan suggests that the install process should be similar, basically just copying a few files into the correct locations on your computer.
H.P. Lovecraft’s Cthulhu is one of my favorite evil deities from literature and I’ve always been a fan of Chaosium’s Call of Cthulhu game and its various spin-offs, so when I saw an announcement about this monstrous new plug-in from Xfer Records’ Steve Duda I checked it out right away. (Duda is well-known as one of Deadmau5’s production partners and a collaborator with many other major artists as well as being a software developer.)
Cthulhu is a set of two MIDI effects that help you to generate new chord progressions and arpeggio patterns that make musical sense. These two effects work similarly: Cthulhu is able to analyze the MIDI notes and chords it receives, identify what chords are being used, and help you to create new chords or arpeggiated sequences based on what it hears.
It also incorporates a step sequencer that allows you to completely reshape the incoming MIDI, making unique and rhythmic riffs out of the notes you play.
There are many other arpeggio plug-ins on the market, but Cthulhu takes it to the next level with features such as chord-arpeggio mode, ties, duration / velocity sequencing, intelligent transposition and harmonization, polymetric sequencing and more.
To hear more of what Cthulhu can do, check out this intro video:
Dubspot blog editor John von Seggern has played techno at massive underground parties in China, remixed Western pop artists for the Indian music market (and vice versa), designed orchestral electronic sounds and effects for the Pixar film Wall-E, and presented his anthropological research on music technology at ethnomusicology conferences. He has authored two instructional books about computer music production and performance as well as the manual for Native Instruments’ popular software synthesizer Massive.