'Ficedula': an open-source MATLAB toolbox for cutting, segmenting and computer-aided clustering of bird song

摘要

Qualitative and quantitative assessments of bird song repertoires are important in studies related to song learning, sexual selection and cultural evolution. Despite methods for automatic analysis, it is still necessary to engage in manual cutting, segmenting and clustering of bird song elements in many cases. Here, we describe a program, the Ficedula Toolbox, which has been made available for free to the bird song research community and has recently come into extensive use. The main advantages of this package are the opportunity to conduct all processing steps in one framework and the option of carrying out computer-aided manual clustering. Output files are ready for further analyses, such as estimation of repertoire size, sequential analysis and repertoire overlap calculation. With this program, findings based on empirical data from the Collared Flycatcher (Ficedula albicollis) song show high inter-observer similarity, and thus, reproducible results. The toolbox may be especially applicable to the analysis of song in species with moderately high repertoires.