Just back from a stimulating three days at the Second Conference on Computer Simulation of Musical Creativity in Milton Keynes organised by Robin Laney, where I gave a keynote on "Representing Creative Rhythms". In addition to the content described in the abstract below, I also presented previous work on the Mimi human-machine improvisation system (work with Isaac Schankler, Alexandre François, Dennis Thurmond) and the MorpheuS(*) music generation system (work with Dorien Herremans). I took part in two panels: Bob Sturm's tongue-in-cheek "Now that deep learning has solved music, what's next?" and Iris Yuping Ren's "Applying music patterns in generation". Two other panels discussed musical creativity and HCI, evaluation of creative music systems. This new community consists of representatives from academia and industry (Sibelius, Jukedeck, Melodrive), several of the young researchers were also game developers. The other keynote speaker, Anna Jordanous, is inventor of SPECS, the Standardised Procedure for Evaluating Creative Systems.
Keynote: Representing Creative Rhythms
Speaker: Elaine Chew
The creativity of music performance is no longer a disputed fact; however, the very nature of that creativity remains illusive, mainly because the performer’s creative input is hard to represent in written form for close inspection. I shall present some experiments in using conventional music notation to represent, in human-readable form, exact timings in music performance. The results, when compared to the original scores, makes tangible the gap between the actual and abstract rhythms, thereby offering a new means of representing performers' creative work.
Next, electrocardiographic recordings of cardiac arrhythmias are subject to the same process of transcription. The notation makes concrete the musical rhythms of the arrhythmias, patterns that are obscured in the predominant individual beat morphology- and frequency-based representations. To make visceral the experience of arrhythmia, the transcribed rhythms are matched to existing music based on rhythmic similarity and the retrieved scores disassembled and collaged, with minor adjustments, to create performable music that re-animates the abnormal heart rhythms.
The symbolic representation not only gives form to these temporal experiences, it also provides access to a host of new analytical approaches and acts as a conduit to novel applications in computer-simulated musical creativity.
* Acknowledgement
The MorpheuS project received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 658914.