Parselmouth for bioacoustics: Integrating Praat into the Python scientific ecosystem

Abstract

As collected datasets become larger and computational analyses become ever more complex, the efficient processing of bioacoustical data is a crucial problem to tackle. Often, during data exploration and analysis, different research software packages need to be flexibly combined in a script. A typical example of such a multi-faceted workflow is the extraction of acoustic parameters from a recording, which are then plotted and tested for statistical significance. Parselmouth is an open-source Python library for Praat, a widely used acoustics and phonetics software package implementing acoustic algorithms and analyses regularly adopted in bioacoustics research. Parselmouth's goal is to provide a full-fledged Python library that integrates efficiently into the larger Python ecosystem. This way, it not only simplifies the application of Praat’s functionality within a typical data analysis workflow but also enables the creation of new experimental tools. Parselmouth’s contribution to bioacoustics research can be highlighted through concrete examples of studies we have conducted, e.g., on vocal flexibility in seals. Moreover, the integration of Praat’s functionality into a general-purpose programming language allows for novel, more complex experimental setups: for example, the integration of Parselmouth into a custom-created software tool permits live-monitoring and instantly evaluating the vocal development during animal training.