BIOACOUSTICS · ACOUSTIC INTELLIGENCE
We build machine-learning models that listen. We record biodiversity sound, annotate it, and train systems to detect and infer the patterns human ears miss, the sound that moves below and above what we can hear.
01 / CAPTURE
Active and passive recording systems and Autonomous Recording Units (ARUs) capture species sound continuously, at the sites and frequencies conventional surveys miss.
02 / PROCESS
Noise reduction, spectral cleaning, and multimodal annotation turn raw field audio into structured, research-grade datasets, every event timestamped and tagged.
03 / INTELLIGENCE
AI models detect and classify species, predict behaviour, count populations, and raise alerts, biodiversity surveillance that runs around the clock.
BIOACOUSTICS · ACOUSTIC INTELLIGENCE
We build machine-learning models that listen. We record biodiversity sound, annotate it, and train systems to detect and infer the patterns human ears miss, the sound that moves below and above what we can hear.
01 / CAPTURE
Active and passive recording systems and Autonomous Recording Units (ARUs) capture species sound continuously, at the sites and frequencies conventional surveys miss.
02 / PROCESS
Noise reduction, spectral cleaning, and multimodal annotation turn raw field audio into structured, research-grade datasets, every event timestamped and tagged.
03 / INTELLIGENCE
AI models detect and classify species, predict behaviour, count populations, and raise alerts, biodiversity surveillance that runs around the clock.
What we do
Acoustic intelligence is teaching machines to hear. We record sound, annotate it, and train machine-learning models to detect, classify, and infer the patterns inside it.
Most of the acoustic world sits outside human hearing, and outside manual surveys. We build the full pipeline, capture, processing, annotation, and model training, so sound becomes data, and data becomes a system that can listen on its own.
The unheard spectrum
0.04%
Your ears hear 0.04% of the animal kingdom. The rest moves below and above human hearing, and that is where we listen.
Between 17 and 20 Hz, elephants are speaking.
Between 120 and 200 kHz, bats are hunting.
Source: Fay, Hearing in Vertebrates, 1988 · Sinauer Associates
What we offer
Three ways to work with the lab on acoustic intelligence, whether you want to capture sound, label it, or build the models that learn from it.
We teach field recording, and we train local recordists across regions to capture sound where flying a team in is not practical. You record from your place; the dataset grows.
We turn raw field audio into labelled, model-ready data: spectrogram tagging, behaviour-linked annotation, and quality control, done to research standards.
We build and train acoustic-intelligence models that detect and classify species from sound, the brains that turn a sound archive into a system that can listen on its own.
We build capacity on the ground: we train recordists locally so remote and hard-to-reach sites can be captured by the people already there, rather than waiting on a visiting team.
How it works
Four stages, one continuous pipeline, from capture in the field to species-level intelligence.
The same four stages, seen as one funnel, each step concentrates raw field audio into something a model can act on.
The work behind the models
The intelligence we build supports conservation and biodiversity work, including collaboration with Antz Systems. These are the kinds of applications and deliverables that come out of it.
Where it is used
What you receive
Who it is for
Research infrastructure
We build structured, large-scale species sound databases for AI training, research, conservation, and sound design, with categorized recordings, behaviour-tagged datasets, environmental and geographic metadata, spectrogram indexing, and AI-ready formatting.
65 Hz · C2, the bioacoustics resonance
Whether you want to learn field recording, annotate datasets, or build acoustic-intelligence models with us, tell us where you are and what you want to capture.
Other divisions
