Probing AM detection in noise with reverse correlation – a pilot study - Archive ouverte HAL Accéder directement au contenu
Poster De Conférence Année : 2019

Probing AM detection in noise with reverse correlation – a pilot study

Résumé

Evidence from psycholinguistic studies suggests that the impact of steady noise on envelope perception has two main components: (1) a masking effect: at low SNR, crucial envelope cues in the signal become masked, compromising recognition accuracy; and (2) a confounding effect: intrinsic envelope modulations arising from the filtering of noise into critical bands can be confused with useful modulations in the signal. Reverse correlation techniques are particularly suitable for exploring the confounding effect of noise on perception. However, the only attempt at applying this technique to an AM detection task yielded mixed results [Ardoint et al., 2007]. This project aims at estimating psychophysical kernels in the envelope domain for a 4-Hz AM detection task. Here we will present some promising pilot data on 2 participants (5.000 and 2.000 trials), together with simulated data from the Modulation Filterbank (MFB) model. The results reveal that both real and simulated listeners are able to track the modulation peaks in the 4-Hz target. However, unlike the MFB model, human psychophysical kernels are in phase opposition with the ideal template, suggesting that they may represent fluctuations in the target and masker as a short-term signal-to-noise ratio in the envelope power domain (SNRenv).
Fichier principal
Vignette du fichier
Poster_ARCHES_2019.pdf (647.03 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03016571 , version 1 (29-03-2024)

Identifiants

  • HAL Id : hal-03016571 , version 1

Citer

Léo Varnet, Christian Lorenzi. Probing AM detection in noise with reverse correlation – a pilot study. ARCHES, 2019, Paris, France. ⟨hal-03016571⟩
33 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More