Study Finds Shared Acoustic Patterns in Birdsong, Human Speech and Music

Overview

• The peer-reviewed paper published in Science on Thursday compared about 70,000 manually annotated bird vocal units from more than 350 species with human speech and song samples across 52 languages.
• Researchers identified three recurring features—small pitch steps between notes, regular syllable timing or rhythm, and phrase-final elongation where the last sound is lengthened—that appear in both bird vocalizations and human vocal arts.
• Those acoustic regularities occurred in vocal-learning groups such as songbirds, parrots and hummingbirds and in non-learning groups like suboscines and nightjars, leading authors to conclude learning is not required for the patterns.
• The team built the bird dataset from citizen-science recordings on xeno-canto and used manual annotation, and they say the results point toward shared motor, respiratory or sound-production constraints as plausible causes.
• Authors and an accompanying Perspective call for experimental, mechanistic work that tests breathing and vocal-production limits and for more study of non-learning bird groups to probe the biological roots of human communication.