A psychoacoustic environmental monitoring toolkit — applying sound quality metrics traditionally used in vehicle and appliance design to the analysis of ecological soundscapes.
About EcoSonic
EcoSonic applies psychoacoustic metrics — including loudness, brightness, and roughness — to the analysis of environmental soundscapes. These are the same metrics used by acoustic engineers to assess the sound quality of vehicles, appliances, and machinery. AELab applies them to a very different problem: measuring the acoustic health of ecosystems.
The research examines how changes in acoustic ecology — driven by climate change, development, invasive species, and other forces — affect both human and animal experience of sonic environments. By quantifying the psychoacoustic qualities of natural soundscapes, EcoSonic provides a new class of environmental indicator that complements traditional biodiversity surveys and ecological assessments.
The project has produced a mathematical model correlating Mel-frequency cepstral coefficients (acoustic features derived from sound recordings) with local weather data — enabling researchers to track how the soundscape responds to environmental change over time.
Methodology
High-quality acoustic recordings are made at monitoring sites across the American Southwest, capturing biophony, geophony, and anthrophony over extended periods.
Recordings are analyzed using psychoacoustic metrics — loudness, brightness, roughness, and articulation entropy — to quantify the perceptual character of the soundscape at each site and time point.
Machine learning techniques are applied to identify correlations between acoustic features (Mel-frequency cepstral coefficients) and environmental data including local weather, season, and land use.
The resulting models are used to assess acoustic ecology conditions, identify anomalies, and support environmental management strategy planning — accelerating analysis that would otherwise require intensive manual survey.
Applications
EcoSonic's researchers hypothesize that acoustic ecology monitoring tools can significantly accelerate environmental analysis and support management strategy planning at scales that conventional monitoring methods cannot reach. Machine learning techniques are applicable to identifying broad event categories — seasonal change, extreme weather events, long-term ecological shift — as well as fine-grained acoustic events like the arrival or departure of particular species.
The EcoSonic toolkit represents a new approach to environmental monitoring: one that treats the soundscape as an integrated signal carrying information about the health and dynamics of the whole ecosystem, rather than focusing on individual species or isolated indicators.
This approach connects directly to AELab's broader mission: sound is a critical environmental signifier. EcoSonic makes that signifier legible through science.