Embedded Ultrasonics: A Microcontroller-Based Multichannel Ultrasound Recorder for Behavioural Field Studies

Abstract

Background: A variety of behavioural studies of echolocation and acoustic communication in free-flying bats depend on multichannel ultrasonic recordings. However, access to these methods remains geographically uneven: tropical and subtropical regions harbour much of the world's bat diversity yet remain underrepresented in experimental behavioural bioacoustics. Conventional systems often depend on costly audio interfaces, laptops, and specialist expertise. Although instrumentation is only one contributor to this disparity, a compact, affordable, openly documented platform can reduce a practical barrier to field research, ecological monitoring, and conservation efforts. Results: Here, I present BATSY4-PRO, an open-source four-channel recorder designed for behavioural and ecological field studies. Built around a Teensy 4.1 microcontroller and two WM8782 analogue-to-digital converters, it records synchronised 16-bit audio at 192 kHz directly to microSD storage. The recorder weighs under 150 g, operates from a 5 V supply, and, together with the four-microphone array, can be constructed for approximately 200 euros. It combines retrospective and buffer-prefilled forward recording with selectable real-time monitoring. While all four full-spectrum channels are recorded, users can monitor any individual microphone or a four-channel mix through tunable heterodyne conversion or direct passthrough. Field recordings showed consistently high signal-to-noise ratios comparable to recordings obtained with a commercially available multichannel soundcard and supported time-difference-of-arrival localisation of bat flight trajectories. Because the intended applications involve moving animals, I used grid-based Monte Carlo simulations to define the usable localisation field of the deployed array and evaluate how simulated source velocity affects localisation performance. The simulations identified a geometry-dependent near-array region of useful performance, followed by a sharp decline in accuracy with increasing range, and showed that localisation errors increased with simulated source velocity. Conclusions: BATSY4-PRO establishes an open, reproducible workflow from synchronised field recording and real-time monitoring to three-dimensional source localisation. Its configurable firmware logic and extensibility allow the same compact system to be adapted to distinct behavioural questions, species, and habitats without dependence on laboratory-bound or proprietary multichannel platforms. Wider adoption can expand the geographic and taxonomic scope of experimental bat bioacoustics, strengthen local capacity for ecological monitoring, and enable comparative investigation of echolocation behaviour under natural conditions.