Cross-accent intelligibility of speech in noise: long-term familiarity and short-term familiarization.

Familiarization effects on word intelligibility in dysarthric speech.

BACKGROUND/AIMS This study investigated the effects of familiarization on naïve listeners' ability to perceive dysarthric speech produced by speakers with cerebral palsy and evaluated the degree of intelligibility improvement, both in the short and long term, as a function of (1) familiarization paradigms and (2) the number of familiarization phases. METHODS A total of 120 listeners (30 liste...

Transfer of Talker-Familiarity Effects

Familiarity with a talker facilitates perception for both heard speech (where speech from a familiar talker is better identified in noise) and for visual speech (where familiarity with a talker’s face assists visual speech recognition). Recently, it has even been shown that the talker familiarity effect can be produced cross modally, i.e., experience in speech-reading a talker facilitates perfo...

Short-term and Long-term Impact of Video-driven Metapragmatic Awareness Raising on Speech Act Production: A Case of Iranian Interme-diate EFL Learners

Comparing intelligibility of several non-native accent classes in noise

How does noise affect perception of foreign accented speech? English sentences spoken by non-native talkers with various language backgrounds were presented to native speakers. The Speech Reception Threshold (SRT) was found for 5 different accent classes, and contrasted with noise-free intelligibility. Modest trends towards global differences in intelligibility based on accent class were found,...

Long Short-Term Memory for Speaker Generalization in Supervised Speech Separation

Speech separation can be formulated as learning to estimate a time-frequency mask from acoustic features extracted from noisy speech. For supervised speech separation, generalization to unseen noises and unseen speakers is a critical issue. Although deep neural networks (DNNs) have been successful in noise-independent speech separation, DNNs are limited in modeling a large number of speakers. T...