Tethys

Tethys is a scientific workbench that allows the integration of acoustic detections, classifications, and localizations of aquatic animals with environmental data from a wide variety of souces. We and our collaborators built Tethys as we saw the opportunity to build systems that will let us analyze long-term data sets. Frequently, people analyze a few seasons worth of data and publish the results. While the summary results are archived in the publication, detailed results that could be used in subsequent analyses are lost. Tethys lets researchers preserve detailed information and provides simple acces to environmental data. Learn more at the Tethys web site.

Silbido

Our group has worked in the area of automated annotation for toothed-whale whistles. We have createad a MATLAB padckage silbido that enables bioacousticians to use some of the algorithms that we have published. In particular, our initial graph search algorithm and our first deep-learning annotator are availble on GitHub in the silbido repository. Silbido contains components written in MATLAB, Java, and C++. When cloning the repository, you must compile the Java and C++ code, instructions on how to do this are in the documentation directory. Windows users can download one of the silbido releases in which the code has already been compiled.

The methods used are described in:

P. Conant. P. Li, X. Liu, H. Klinck, E. Fleishman, D. Gillespie, E.-M. Nosal, and M. A. Roch (2022). “Silbido profundo: An open source package for the use of deep learning to detect odontocete whistles,” J. Acoustical Soc. Am., 152(6), pp. 3800-3808. doi:10.1121/10.0016631.

Li, P., Liu, X., Palmer, K. J., Fleishman, E., Gillespie, D., Nosal, E.-M., Shiu , Y., Klinck, H., Cholewiak, D., Helble, T., and Roch, M. A. (2020). “Learning Deep Models from Synthetic Data for Extracting Dolphin Whistle Contours,” in Intl. Joint Conf. Neural Net. (Glasgow, Scotland, July 19-24), pp. 10. DOI: 10.1109/IJCNN48605.2020.9206992

M. A. Roch, T.S. Brandes, B. Patel, Y. Barkley, S. Baumann-Pickering, M.S. Soldevilla, “Automated extraction of odontocete whistle contours,” J. Acous. Soc. Am., Vol. 130(4), pp. 2212-2223, 2011. doi:10.1121/1.3624821.

Requirements: Matlab 2021b or later (might work with 2019), Signal and Image processing toolboxes, free ONNX toolbox to use the deep learning annotator.

The data used to deveop silbido consists of towed array, dipping hydrophone, and stationary platform data recorded in the Southern California Bight and Palmyra Atoll. A subset of these data were annotated by trained analysts using a cubic spline annotation tool that is part of silbido. These data were released as the DCLDE 2011 conference data set. They have recently been added to the National Centers for Environmental Information data archive:

M. A. Roch, Y. Barkley, X. Zhang, M. S. Soldevilla, S. Baumann-Pickering, and J. A. Hildebrand, “DCLDE 2011 Conference Data,” NOAA National Centers for Environmental Information, 2025. doi:10.25921/wsnc-js16.

The dataset also includes code in Python, Java, and MATLAB for reading and writing the tonal annotation files.

(Note: These data were submitted shortly before the US Government shutdown and will not be avialable until the US government resumes normal operations.)