Skip to main content

43. Developing a Bioacoustics Analysis for Anti-Bat Sounds

Alyssa Culver, Dr. Jesse Barber

Culver final poster - view content on post page
Select to view full poster image

Watch Intro Video


Bats and moths have been in an evolutionary arms race for at least 50 million years. To get a better understanding our lab has gone out around the world, recording thousands of different moths ultra sound anti-bat clicks (Barber & Conner 2007).


  1. Collected acoustic data from moths around the world
  2. Created a file name pipeline
  3. Clipped moth acoustic files for each individual moth
    1. Three analyzed cycles
    2. One 100 millisecond cycle
    3. One 350 millisecond cycle
  4. Create spectrograms of clipped files
  5. Analyzed spectrograms using a machine-learning approach to Python.


Work in progress as still clipping files and creating spectrograms


file name examples, d

Keep file names relative to the data that was collected, but in such a way that each file name is unique.

Decipher strong clicking patterns amongst weak patterns or bat calls

sound file

spectograpm of moth's modulation

Developing a pipeline for a meta database will make future data collection more readily available for analysis.

Process pipeline diagram
1. Process of collecting moth ultrasound response data 2. Pulling files based off spreadsheets that has usable data, finding the files, clipping, naming, and archiving 3. Most files has more empty space than usable signals; or they are embedded within bat echolocation calls 4. Finding the area in the file with the highest sound-noise ratio, clipping the best representation to 100 and 350 milliseconds long files 5. Spectrogram of a moth’s modulation cycle that will be used for analysis


Work in progress


We speculate that moths with similar clicking patterns will be clustered together. However, data archiving is still in progress.

Additional Information

For questions or comments about this research, contact Alyssa Culver at