Annotated audio of the loud nest calls of brooding female rooks in five colonies with individual-level identification

This data accompanies our article entitle 'The loud nest call of female rooks (Corvus frugilegus) shows a vocal signature at the individual level, but not at the nest cluster or colony levels'.

The dataset.zip archive contains audio and annotations obtained from recording five different colonies of rooks during the breeding season. The audio was compressed losslessly to .flac files from the original wav format to save storage space. The annotations are the corresponding .txt files containing timestamps and individual identity for each call used in the article, intended for viewing in Audacity (open the flac files, then import the annotation file with the same name). Finally, the tabelau_vocs.tsv file is essentially all the .txt files aggregated for the analysis. 

Note that to run the analysis, either the .flac audio files must be converted back to .wav files, or the tableau_vocs.tsv file should be edited to point at the .flac files instead of the original .wav files.

The intermediate_results.zip archive contains various intermediate results used in the analysis. These are meant to be used with the provided .R scripts in the same archive (up-to-date versions of these scripts may be found in the gitlab repo above), which reproduce the statistical analysis and the figures in the article and supplementary.

Finally, the supplementary.zip archive contains the supplementary materials provided with the article, including the supplementary.pdf file containing additional analyses, and the example_nest_calls folder containing example of nest calls from various females in the five colonies.

The example nest calls are provided at various stages of processing: raw audio extracted from the spectrogram (the .wav files without suffixes), high-pass filter above 100 Hz (.wav files with the 'filtered' suffix), audio with all the processing steps including the spectral gating-based denoising, pre-amplification, and centering (.wav files with the 'processed' suffix), and spectrograms used as inputs for the analysis after dB-scaling and Mel frequenct-scaling (.png files).

The rook_vocal_signatures-main.zip archive contains all the code as of the date of submission (August 21st, 2024) used in the article. This was added purely to anonymise the submission as much as possible. Up-to-date code should still be consulted from the link above (will be added after the paper is accepted.