Session: 16-01-01: Government Agency Student Poster Competition

Paper Number: 150490

150490 - Differentiating Dynamic System Outcomes Through Sonification 

Data sonification is a powerful tool to communicate information through sound in ways that cannot be easily conveyed visually. One of the strengths of sonification lies in its ability to help differentiate very small and large changes in the state of a system simultaneously. The combination of dynamic audio and visual elements in a computer-based presentation provides yet further opportunities to convey complex and dynamic data sets, reinforcing learning through interaction with the model directly. This benefit is especially useful for understanding nonlinear natural processes, modeling them to help uncover such systems’ emergent behaviors. One such behavior of interest is dynamic systems that involve chaotic outcomes. Using sonification to present data generated from nonlinear systems producing deterministic chaos (distinguished from systems presenting more generalized stochasticity) provides a more intuitive grasp of how the systems behave compared to more traditional of means of presenting this kind of data using mathematic procedures such as Fast Fourier Transformations. This is especially the case for teaching about these systems to general audiences.

 

In this project, photosynthesis was modeled using an existing model (“Photo3”) which can represent all three photosynthetic pathways, C3, C4 and Crassulacean acid metabolism (CAM). The system of interest for this investigation involves CAM photosynthetic processes. The specific plant that the model was parameterized for this project was Agave tequiliana, a CAM plant native to Jalisco, Mexico. The Photo3 model was adapted to run as the script agave on the norns platform, a small Linux-based sound computer using the Lua and SuperCollider programming languages. Rewriting the script in Lua for the norns platform allows a user to present visually and sonically the three outcomes possible in dynamic systems: a stable fixed point, periodic behavior and aperiodic behavior. Temperature conditions are the primary drivers of the state of the system and the script allows for interaction with the model as the user can move from outcome to outcome by changing the model’s temperature settings. The norns computer performs signal calculations and displays visual elements through a 128x64 pixel screen. 

Separation of state conditions via sonification is achieved through signal calculations performed using SuperCollider. The dynamic values associated with malic acid and circadian state in the photosynthetic model are sonified as the agave script iterates through the model. Signals processed for the utilization of malic acid by the plant for conversion into CO₂ sets the volume of the sonification as well as the center frequency of a bandpass filter. Circadian state sets a sequence of tones that play a note in a major or minor scale, depending on whether the circadian state value is greater than or less than 0.5.

As the ambient temperature increases, the leaf temperature increases as well, causing the cycle to shift through each of the system's three dynamic outcomes. A sonification profile was achieved in each of the three outcomes. The lowest temperature modeled was 285K and the sound created was quasi-rhythmic. In addition to the chaotic nature of the sound, the visual display also shows oscillations at the extremes of the loop. At temperatures between 293K and 301K the state of the system shifts from aperiodic to periodic. The sonification profile during the periodic outcome loses the oscillations that were seen in the visual display and caused distortions in the audio display. At temperatures 301K and 305K the system moves to a stable fixed point. The sonification profile at the fixed point goes to silence representing a lack of productivity. The uncovering of unique audio profiles for each of the possible outcomes in the dynamic system demonstrates that sonification is a viable solution to present emergent behavior in a dynamic system. 

Presenting Author: Duncan Turley Portland State University

Presenting Author Biography: Duncan Turley is a graduate student at Portland State University working on his master's in environmental engineering focused machine learning, remote sensing, and sonification projects with data from natural dynamic systems.

Authors:

Duncan Turley Portland State University
Jonathan Snyder Unaffiliated
Samantha Hartzell Portland State University