Characterizing structural variation in the notes of Adelaide's warbler (Setophaga adelaidae) songs

Acoustic structure varies at multiple levels in birdsong. Note-level analyses are beneficial because they can enable scientists to characterize song structures, define song types, and more precisely measure vocal performance. By-eye analysis of spectrograms is sometimes adequate for note classification. However, Adelaide’s warbler (Setophaga adelaidae) notes are difficult to classify visually because they contain a blend of discrete and continuous variation. We ran a novel, automated, image-based analysis called latent visualization to determine if discrete note types exist and to explore variation among notes. Notes differed primarily in general frequency characteristics, and note clusters (i.e., types) were not discrete. Contrary to our expectations, cluster labels disregarded note contour, which is the most important aspect of note structure for by-eye classification. Overall, latent visualization may not be ideal for species that produce frequency-modulated notes. We recommend running future image-based analyses on note contours alongside acoustic feature analysis for validation.

Authors

• Huang, Samantha Yeing-Sih
• University of Lethbridge. Faculty of Arts and Science

• Adelaide's warbler
• birdsong note variation
• birdsong acoustic structure
• latent visualization
• frequency-modulated notes
• birdsong acoustic variation

Identifier

hdl:10133/6601

OAI identifier oai:identifier

oai:opus.uleth.ca:10133/6601