ISCA Archive Interspeech 2025
ISCA Archive Interspeech 2025

Prediction of listening effort ratings for habitual and clear-Lombard speech presented in noise

Esther Janse, Chen Shen, Martin Cooke

Earlier research developed the high-energy glimpse proportion metric (HEGP) to capture those speech-dominant spectro-temporal regions, or glimpses, that survive energetic masking when speech is presented in noise. Whereas the HEGP metric has been shown to correlate with intelligibility of speech presented in noise, this study investigated whether HEGP predicts listening effort (LE) ratings, and if so, whether it does so differentially for habitual and clear-Lombard speech. Secondly, the study investigated whether acoustic measures of spectral balance, median F0 and F0 range, and articulation rate, explained additional variance in LE. Results showed that HEGP predicted LE ratings equally strongly across speaking styles. Beyond HEGP, wider F0 range and lower articulation rate were also associated with decreased LE, particularly in the habitual speech condition. Thus, the HEGP release-from-masking metric, complemented by speaking-style aspects, can help explain LE ratings in noise.