To test the effect of different configuration settings we have used many Audiomoths (AMs) sited next to each other recording at the same time, and configured the exactly the same, apart from varying the one configuration setting that is currently under test.
Initially we started with a bank of 4 AMs, however it quickly became clear more Audiomoths running in parallel were needed to test the effects of things like amplitude threshold. We found 8 AMs allowed enough flexibility in testing, while allowing enough time for the recordings on the 8 mSD cards to be analysed within the day time, to allow for redeployment the following night.
Test location
We tested in a quiet rural Somerset garden with the AM’s set up overlooking a gravel driveway. There are no trees, and little vegetation nearby, and hence this is nearly the “ideal” quiet conditions.
The bats in this location have been watched for over 25 years and hence their behavior can be predicted.
- One or two Serotine roost in the roof of a cottage to the side of the drive.
- A Lesser horseshoe is resident in a bat loft in a separate outbuilding with the exit “bat chute” facing the AMs.
- Several Common and Soprano pipistrelle roost in the roof structures of the outbuilding.
The presence of two buildings can result in “echos” bouncing off the walls, however overall this location provides ideal conditions for consistently testing and analysing bat recordings across this range of species
analysis
After a nights recording the mSD cards were extracted and the number and size of files recorded.
Bat AutoId was then performed with the free Batclasify, and any manual identification and screenshots prepared using Audacity. We tested how accurate the AutoID of Batclasify was on AudioMoth recordings, and published the results here.