A Practical Approach towards an Exploratory Framework for Physical Modeling

Much of the research on physical modeling deals with developing accurate simulations of different physical mechanisms (Keefe 1992; Bensa et al. 2002; Välimäki et al. 2006; Bilbao 2009). The research presented in this article focuses instead on the control of these simulations. Our work involved developing a framework for controlling physical models that focuses on creativity and exploration. Understanding the qualities of physical models, not only in regard to algorithmic and sonic properties but also in regard to intuitiveness, naturalness, interaction features, mapping possibilities, and potential gestural control can assist in the creative exploration of physical modeling.

When developing new musical interfaces, whether they are based on instrument-like, instrument-inspired, extended, or alternate controllers (to use the classification by Miranda and Wanderley 2006), a central goal seems to be to enhance the expressivity and intimacy of novel musical instruments (Poepel 2005; Dobrian and Koppelman 2006; Schlessinger and Smith 2009; Jones et al. 2009). Although we acknowledge that these are indeed important features, they are not the focus of this article. Rather, the article focuses on the ability of a novel instrument to be explored in order to produce interesting sonic output and to spark creative ideas in the mind of the user. In other words, we try to implement models based on existing techniques in ways that encourage the user to explore sonic properties in a creative manner.

The creative cognitive processes that occur when creative activities take place can be described by the Geneplore model (Finke, Ward, and Smith 1996)—a heuristic model that describes creative activities as a combination of generative and exploratory processes. There is a generation of ideas, which might not even be complete or accurate ideas but merely germs of an idea; they just bring with them signs of originality and appropriateness. These “preinventional” ideas are then explored, and the creative activity is thus the alternation back and forth between generation and exploration. Because the activity of generation and exploration is a continual, iterative process, it is not alway known whether a particular process can be categorized as either one or the other. However, for creativity to occur (according to the Geneplore model), both generation and exploration must take place.

For our purposes, we must examine the exploratory qualities of the sound-synthesis model, the physicality of the interaction, and how the integration of the two can encourage exploration. Exploration has to do with the manner in which a user interacts with the instrument. The sound-synthesis model in itself will of course affect what sort of sounds one will produce using a given instrument, but so too will the perceived affordances and constraints of the overall instrument, which exist on all levels—from the complexity/accessibility of the sound synthesizer, to the mapping of sensor data to sound, to integration with other musical systems. These guide the user, proposing/recommending a certain use of the instrument. The user acquires knowledge about how the instrument works—knowledge that is inherent in the instrument (which is what Magnusson 2009 calls an epistemic tool). It is the interplay between this inherent knowledge and the user’s knowledge that determines the interaction. By acknowledging that the different instruments can encourage different types of interaction, it is possible to design specifically for a certain type of interaction.

The rest of the article is composed of two parts. First is a discussion about control of and interaction with physical models leading to a set of design directives for exploratory control of physical models. The second part describes the implementation and evaluation of the PHYSMISM: a hardware interface we created to investigate these design directives.

Design Directives for Exploratory Control of Physical Models

To understand how an instrument can encourage exploration, we must examine the space in which the user navigates when using the instrument. Exploration is facilitated when the system responds to the user in an intuitive but not necessarily predictable manner. The response can be unpredictable as long as it brings with it clues to how one could reproduce such a response. The epistemic (knowledge-related), heuristic, and...