Meme detail

Text detected in the image

We used a GLM to try to predict the distance in UMAP space from the differences in acoustic parameters.For this we chose a subset of ten syllables from every syllable type produced by the 49 birds plot- ted in Fig. 2c (228 syllable types × 10 versions of each syllable = 2,280 syllables). We then calculated the UMAP coordinates and eight core acoustic features for each syllable: duration, mean pitch, mean peak frequency, mean frequency modulation, mean entropy, mean good- ness of pitch, mean amplitude modulation and mean amplitude. We computed the pairwise UMAP distance between every pair of syllables and their pairwise feature space difference for each acoustic parameter.

Popularity