AI Study on Birdsong: Scientist Discovered 8 Basic Sound Patterns Across 3000 Bird Species

Artificial Intelligence (AI) study published in Science has analysed more than 116,000 recordings of 3000 bird species. The study highlights all global birdsong is built from 8 acoustic building blocks which are three trills, three whistles and two complex notes. 

The study was led by Quentin Bacqule, a PhD student at the University of Saint Etienne. The report highlights the evolutionary history of birds' song which shows 82% variation. 

Eight Universal Motifs of Bird Species

The researchers utilised crowd sourced recordings gathered from across the world via citizen science databases. They evaluated them using an automated sound analysis method modulation power spectrum. This technique measures how sound fluctuates across different frequencies and time. 

How birds arrange their repertoires using a combination of just 8 recurring sound motifs. Three types of trills were executed at various frequencies ranging from slow rhythms to ultrafast sequences.

Three types of whistles were characterised by distinct modifications such as flat and clean tones. 

Two complex acrostic which are categorised as atonal chaotic notes and harmonic stacks.. 

According to data a bird’s vocal is governed by their evolution and local habitat such as the impact of evolutionary history when comparing song against family lineage and body characteristics and other behaviors. 

The researchers discovered evolutionary history of birds  accounts for a staggering 82% of variation found in song motifs. 

The environment surrounding birds habitat plays an equal role in fine tuning sounds. Birds living in dense hick vegetation and high humidity easily. Birds in temperate regions find open spaces and often complex rapid trills that carry rich data over shorter distances.