Meme detail

Text detected in the image

parameters). We next generated spectrograms for each syllable and embedded them using Euclidean UMAP, as described above.We then used multiple linear regression to examine correlations between each acoustic parameter and the positions of song-syllable spectrograms plotted in UMAP1 and UMAP2 latent feature space. To further analyse correlations, we examined the Pearson correlation coefficients between UMAP1 and UMAP2, as well as between different acoustic metrics. The

Popularity