Paper linked above https://peerj.com/articles/16349/ links to .wav sound files in the Data Availability section https://datadryad.org/stash/dataset/doi:10.5061/dryad.ht76hd... and 20 minute .mp4 files https://zenodo.org/records/8247515 (and R code in a PDF) and a "Timeline of Interaction" PDF. Seems that's the closest to a "transcript".