First, an explanation from the creators themselves…
“We use a modified SampleRNN architecture to generate music in modern genres such as black metal and math rock. Unlike MIDI and symbolic models, SampleRNN generates raw audio in the time domain. This requirement becomes increasingly important in modern music styles where timbre and space are used compositionally. Long developmental compositions with rapid transitions between sections are possible by increasing the depth of the network beyond the number used for speech datasets. We are delighted by the unique characteristic artifacts of neural synthesis.”
Using this method, Dadabots have released music based off analysis of Meshuggah, The Dillinger Escape Plan, Krallice and more! You can check these out below along with each description and see what else has been machine made by heading over to their Bandcamp page!
This is some very interesting material to say the least…
This neural metal album is part of a submission to NIPS 2017 Workshop for Machine Learning, Creativity and Design: “Generating Black Metal and Math Rock”
Music was generated autoregressively with a sample recurrent neural network* trained on raw audio from the album Nothing by Meshuggah. The machine listened to Nothing 30 times over several days. The machine generated 1300 minutes of audio. After listening to Nothing 100 times over 15 years, a human listened to the machine audio, chose 18 sections from varied evolution points, and taped them together into a 10 minute album. Titles were generated by a Markov chain. The album cover was generated by neural style transferring the Nothing cover with itself.
Neural network attempts to play [hallucinate] the most spastic, mathematical, chaotic and contradicting metal [mathcore!] album ever released, Calculating Infinity by Dillinger Escape Plan.
This album is part of a submission to NIPS 2017 Workshop for Machine Learning, Creativity and Design: “Generating Black Metal and Math Rock”.
The music was generated with a recurrent neural network* trained on raw audio taken from the album. Titles were generated by a Markov chain. The album cover was created with neural style transfer.
This album is part of a submission to NIPS 2017 Workshop for Machine Learning, Creativity and Design: “Generating Black Metal and Math Rock”.
This album was generated with a recurrent neural network* trained on raw audio from the album “Diotima” by Krallice. All titles were generated by a Markov chain. The album cover was created with neural style transfer.
Links: Official Website // Bandcamp
