Adaptive Sonification of Time-Resolved Experimental Trajectories
Abstract
This software provides a Python-based platform for transforming time-resolved experimental data into sound. Initially developed to explore protein-folding dynamics measured using single-molecule optical tweezers, the algorithm is designed to be applicable to a broad range of one-dimensional experimental trajectories and time-series datasets. The program converts numerical measurements into musical pitches through normalization and logarithmic frequency mapping across 128 discrete pitch levels. Local maxima and minima are identified using Savitzky–Golay smoothing and peak detection. An adaptive repetition algorithm emphasizes these features while preserving the overall playback duration, allowing changes and transitions in the underlying data to be represented through variations in pitch and rhythmic density. Two audio-rendering modes are available: electronic synthesis and piano-like additive synthesis. Both use the same data-to-pitch mapping and event timing, providing alternative auditory representations of the same numerical trajectory. The software is intended as a flexible tool for scientific sonification, enabling researchers to explore patterns, fluctuations, and transitions in complex time-dependent signals through auditory perception. Potential applications extend beyond protein dynamics to other experimental measurements and numerical simulations that produce one-dimensional trajectories. Implementation and Usage The software is implemented in Python and requires NumPy and SciPy. The repository includes the source code (sonify.py), installation and usage instructions (README.md), dependency information (requirements.txt), and an MIT license. Users can provide their own one-dimensional numerical trajectories and generate WAV audio files using the sonify_trajectory function. Uniformly sampled data are recommended when preservation of relative experimental timing is important. No experimental datasets are included in this release.
