Are you a fan of death metal? If yes, then you would be glad to know that Dadabots – an AI death metal band – is playing non-stop on YouTube and you can enjoy their music in real time. The video is a non-stop, the algorithmically-generated flow of technical death metal. It sure is an amazing feat of creativity from machines whether you are a fan of death metal or not.
Dadabots have been created by CJ Carr and Zack Zukowski, a duo of musicians that have a keen interest in algorithmically created music. The duo has been busy for the past few years working on creating a neural network that is recurring and capable of creating original composition after being trained on particular datasets from a singular genre of music.
The early experiments by the duo included a variety of genres up until they realized that punk and metal was far better suited to the random nature of the algorithm. Carr and Zukowksi write in the recent paper, ‘We observed that electronic music and hip-hop instrumentals did not seem to train as well as organic, electro-acoustic ensembles of musicians. Music genres like metal and punk seem to work better, perhaps because the strange artifacts of neural synthesis (noise, chaos, grotesque mutations of voice) are aesthetically pleasing in these styles. Furthermore, their fast tempos and creative use of loose performance techniques translate well to SampleRNN’s rhythmic distortions.’
The final product has been named Dadabots. It has produced 10 albums till date and is inspired by artists such as Dillinger Escape Plan, Meshuggah, and NOFX. Algorithms were not only created for each album but also for the album art and generating song titles. The recent project of Dadabots is a non-stop, live stream on YouTube known as Relentless Doppelganger. Dadabots were trained on the music of Archspire – a Canadian band.
Carr, speaking about the recent project, said, ‘Most nets we trained made shitty music. Music soup. The songs would destabilize and fall apart. This one was special though.’ The output has been amazingly consistent and stable thus leading the duo to create a live-stream that is autonomously playing everything that the neural network is generating in real time. Check it out below and let us know what you think of it!