sLowlife: Sonification of Plant Study Data

sLowlife is an exhibit about plant biology at the United States Botanic Garden [1]. I provided the sound. The premise of the exhibit is that changes in the lives of plants happen too slowly for us to notice. The exhibit features time-lapse videos of plant growth, some at the microscopic level. By speeding up the depiction of plant growth processes, we humans—distracted as we are by our fast-moving, multitasking world—can develop an appreciation for them.

I sought to provide a sonic environment that helps visitors ease into a contemplative frame of mind. The sound is soft and slow moving, coming across as an auditory analog of the slow changes in the lives of plants. Three stereo streams are deployed in the exhibit space in such a way that a visitor moving from one part of the exhibit to another experiences a variable mix between the streams. I designed the sound so that each stream loops over a different time span, and the initial combination never repeats.

There is a further dimension to the sound: All the textures are constructed, to varying degrees, using data from the experiments on which the exhibit is based. I wrote programs to extract data sets and map them to appropriate musical dimensions for realization within RTcmix, a sound synthesis and processing package that I develop with colleagues [2].

As I worked on the sound, I wrestled with the issue of the audibility of sonified data. This is not a new idea: think of Dufay's 15th-century motet "Nuper rosarum flores," which is based on the dimensions of Brunelleschi's dome [3]. It was probably enough for Dufay to know that his piece shared architectural details with the dome—confirming a Platonic ideal of harmony—without expecting anyone to hear this structure. Contrast Dufay with Steve Reich, who insists on audible process rather than conceptual justification [4]. To hear his Piano Phase well is to follow the phase-shifting process and to realize the importance of the moments when parts realign. Where does that leave sonification of plant study data? Must the data have an audible result?

I try several approaches. For one texture, I take the changing measurements of seedling curvature in response to light and map these to the rhythm of synthetic bell notes. This sound accompanies a montage of many time-lapse videos of seedlings bending toward light (Fig. 2). The individual videos have an unsteady appearance and they loop over varying time spans. The total image is a delightful interaction of many similar seedling "dances." Because the mapping of curvature data to rhythm is so straightforward, and because the sound itself is simple—only four pitches repeating throughout—it is easy to perceive a general correspondence between visual images and sound.

For another texture, I let the experiment data remain subordinate to my musical intuition, with the result that the effect of the data on the sound is not readily identifiable. The experiment treats plants with varying strengths of blue light and records the fluctuating amount of light transmitted through leaves. Each sustained note in this texture comprises multiple oscillators with variable tuning. As a leaf lets more light through, the oscillators diverge from matching each other in pitch and so sound more "out of tune." However, I control the depth of these tuning changes intuitively, by reacting to the sound while turning a knob. Moreover, I rely on my sense of harmonic quality and motion when selecting pitches. It turns out that the aspects of the sound under my intuitive control are far more noticeable than the ones under the influence of the experiment data. In this case, then, the link between the data and the sound is more conceptual than audible.

The text accompanying the exhibit informs the visitor that the sound has some basis in the experiment data. Can visitors hear this connection? Does simply knowing about the connection affect the way visitors perceive the sound—perhaps by assuring them that the sound moves beyond mere mood enhancement—even if they cannot hear the effects of the data directly? Questions like these...