Work

Making sense of periodicity glimpses in a prediction-update-loop - a computational model of attentive voice tracking

Year: 2020

Type: article

Abstract: Humans are able to follow a given speaker even in challenging acoustic conditions. The perceptual mechanisms underlying this ability remain unclear. In this study, we present a computational model of attentive voice tracking, consisting of four main computational blocks: A) glimpsed feature extraction, B) foreground-background segregation, C) state estimation and D) top-down knowledge. Conceptually, the model brings together ideas related to auditory glimpses, vowel segregation and Bayesian inference. Algorithmically, it combines sparse periodicity feature extraction, sequential Monte Carlo sampling and probabilistic voice models. We evaluate the modelby comparing it with the data obtained in the psychoacoustic task, which measured the ability to track of one of two competing voices with time-varying parameters (fundamental frequency (F0) and first two formants (F1, F2)). We test three model versions, which differ in the type of information used in the segregation stage: version 1 uses oracle F0, version 2 uses estimated F 0 and version 3 uses estimated F0 and oracle F1 and F2. Version 1 outperforms human listeners, version 2 is not sufficient and version 3 is closest to explaining human performance.

Source: Zenodo (CERN European Organization for Nuclear Research)

DOI: 10.5281/zenodo.4091713

Authors: Luberadzka, Joanna, Kayser, Hendrik, Hohmann, Volker

Institution: Carl von Ossietzky Universität Oldenburg

Language: English

Topic: Neuroscience and Music Perception

Subfield: Cognitive Neuroscience

Field: Neuroscience

Domain: Life Sciences

Open Access status: green

Funders: Deutsche Forschungsgemeinschaft, National Institutes of Health, National Institute on Deafness and Other Communication Disorders

Awards: Open community platform for hearing aid research, SFB 1330: Hörakustik: Perzeptive Prinzipien, Algorithmen und Anwendungen, G4309561208