Humans perceive bird choruses as individuals up to about eight; larger groups become ‘summary’ sounds

# What the study did

# Why this matters Most research on auditory scene analysis focuses on speech and music. Natural environments contain many non-human sound sources that overlap in time. Understanding whether humans parse these natural mixtures as distinct signals or as a single texture affects how we think about perception in the field, biodiversity monitoring by ear, and the limits of human auditory scene analysis outside urban settings.

# Key findings

• Species segregation: When mixtures contained different species, listeners tended to segregate the components when the species' summary statistics were sufficiently different. In other words, perceptual separation between species depended on measurable disparities in their sound statistics.

• Shared mechanisms: The results indicate that the same segregation mechanisms used for human-produced sounds and geophysical noises also apply to natural bird vocalizations. The auditory system balances detailed representation for small numbers of sources and statistical summaries for dense acoustic scenes.

# How they tested it

# Practical implications

• Field listening and citizen science: People can reliably detect and discriminate individual birds in modestly sized choruses, but dense choruses will be perceived as an acoustic background rather than separable individuals.

• Acoustic monitoring: Automated and human-based monitoring strategies should account for a perceptual threshold where abundance information becomes encoded in summary statistics rather than in discrete detections.

• Modeling natural soundscapes: Auditory summary-statistic approaches capture important aspects of human perception for natural acoustic scenes and can guide development of tools that aim to mirror human listening.

# Open data and reproducibility The paper links the human psychophysics data, the sound database catalog IDs, and the code for sound generation, acoustic analysis, and auditory models. Data and code are available on GitHub and an OSF repository provided by the authors.

# Who did the work and where it appeared The study was conducted by Richard McWalter and Christian Lorenzi and published in PLOS Computational Biology on September 21, 2026.

# Bottom line Human listeners separate individual bird vocalizations reliably up to about eight same-species singers. Larger, denser choruses are encoded as time-averaged auditory textures, while mixtures of different species can remain perceptually segregable if their sound statistics differ enough. The pattern aligns with summary-statistic models of auditory scene analysis used for other sound categories.