Auditory Numerosity Judgement Is Driven by Non-Spatial Factors Across Azimuth, Elevation, and Distance.
Jakab Pilaszanovich, Max Schulz, Lukas Lange and 4 others
Listeners consistently undercount how many voices are in a scene, and this undercount is the same whether voices come from left-right, up-down, or near-far — even though spatial hearing differs enormously across those directions.
What drives the estimate is how much frequency-time space the mixture fills.
Key findings
1Numerosity judgement was highly similar across azimuth, elevation, and distance (slopes 0.51, 0.54, 0.46), even though localisation accuracy and spatial release from masking differed strongly across those same dimensions.
2Neither individual localisation accuracy nor spatial release from masking predicted a person's numerosity slope; adding spatial-hearing predictors did not improve the model beyond spatial dimension alone.
3A 10-percentage-point increase in spectrotemporal coverage was associated with 0.13 additional reported sources overall, with the strongest effects between three and five presented sources across all three spatial dimensions.
Still to come
How it was doneWhat they found
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
All participants were healthy young adults (mean age 25.1) with self-reported normal hearing; no formal audiometric screening was performed, so subtle high-frequency losses cannot be ruled out. Four of the 21 participants were excluded on outlier criteria, reducing the analysed sample. Speech stimuli were generated by a neural text-to-speech system rather than recorded from real talkers, which may limit ecological validity for natural multi-talker scenes. The tested range was limited to two through six concurrent sources; behaviour at higher source counts is unknown. In the distance condition, reverberation served both as a spatial cue and as a source of spectrotemporal smearing, so the two cannot be separated. Single laboratory, single setup per dimension; generalisability to other rooms, speaker arrays, or populations is untested.
Declared interests
Funded by the German Research Foundation (DFG) within the Priority Programme AUDICTIVE (grant SCHO1079/2-1). The authors declared no potential conflicts of interest.
The easy way to misread this
Do not read the null spatial-hearing result as evidence that spatial hearing is irrelevant to multi-talker listening in general. This study measured numerosity estimation — a rough 'how many' guess — in healthy young adults using synthetic speech. It does not test speech understanding, source tracking, or communication performance, and it says nothing about patients with hearing loss, auditory processing disorder, or real-world communication difficulty.
Summarised by AI from the full paper, without a clinician reviewing it. Check it against the source before it changes what you do. Read it on PubMed →