Virtual performance

random trip report

David P. Anderson
May 2022

Software synthesizers like Pianoteq produce reasonably good piano sounds (or you can get a MIDI grand). Computers can play notes with precise timing and dynamics. So one can imagine performing a piano piece using a computer.

How would this work? You (the "performer") would start with a digital score for the piece - perhaps a MusicXML file. Then you'd use software to add nuance: changes to the dynamics, timing, and articulation. You'd spend time - perhaps weeks or months - making it sound exactly how you want. You could then play it for a live audience, or put it on the Web. We call this a "virtual performance" (in contrast with a performance where you physically play an instrument).

When I describe the idea of virtual performance to my pianist friends, their initial reaction is usually dismissive. This is understandable: the act of physically playing a piece - especially for a live audience - can be magical: it combines emotion, physicality, listening, and spontaneity. It's hard to imagine that anything done sitting at a computer could produce a result as beautiful or meaningful as a live performance.

But I have a different viewpoint. When I work on a piece, it starts as a mental process. I hear the piece, perhaps from a recording or by reading through the score. It sits in my mind, gestating. I imagine different ways it might sound, and different moods or feelings it might evoke. This translates into choices in tempo, dynamics, articulation etc. - i.e., nuance. This exploration might take place at the keyboard or away from it.

Eventually I have a mental model of how I want the piece to sound. Then there's the (usually arduous) task of getting my fingers to play in a way that approximates the mental model. But:

  • My fingers have limited strength and coordination, and my technique is not great. So when I practice the piece, I have to find a compromise between the mental model and what my fingers can do. The mental model gets blurry.
  • The moment I stop practicing the piece, it starts to disappear from my fingers. Soon I can't play it well anymore. If I want to perform it again - or if I have a new idea that changes the mental model - I have to do the practicing all over.
  • There are people with fantastic musical minds who can't play well or at all.

These problems don't exist with virtual performance. If the software tools are good, you can realize your mental model precisely.

In addition:

  • If you want to revisit the piece years later, you can tinker with it immediately, no practice needed.
  • Virtual performance makes true collaboration possible. Two or more people can work together on a virtual performance. In contrast, a live performance may have input from teachers and coaches, but ultimately it's the creation of the performer.
  • Beyond speed and precision, there are things you can do with software that you can't do physically. For example, Numula provides "virtual" sustain and soft pedals that apply to an arbitrary subset of notes, sort of like an enhanced sostenuto pedal.

And I reject the notion that live performance is the sole domain of musical magic. For me, most of the magic happens as the mental model is forming, long before the performance.

By analogy: a painter forms a mental model of an image and spends weeks or months realizing it in their studio. The result can be beautiful and can express emotion. We don't demand that the painter create the painting in real time, in front of the viewer. Why should we make this demand of musicians?

In fact - on my soapbox now - it seems to me that today's live-performance aesthetic is a bit pompous and self-aggrandizing. The performer takes the composer's genius and presents it to the audience as if it were their own, and expects to be lavished with applause afterward.

Copyright 2024 © David P. Anderson