Challenges in predicting the lyric intelligibility of musical segments for older individuals with hearing loss

Whitmer WM, McShefferty D, Akeroyd MA, Bannister SC, Barker JP, Cox TJ, Fazenda BM, Firth J, Graetzer SN, Greasley AE, et al. Challenges in predicting the lyric intelligibility of musical segments for older individuals with hearing loss. In: 2025 AES International Conference on Artificial Intelligence and Machine Learning for Audio. Audio Engineering Society; 2025. Paper 14. Available from: https://aes.org/publications/elibrary-page/?id=23003

@inproceedings{Whitmer2025_23003,
  author = {Whitmer, William M and McShefferty, David and Akeroyd, Michael A and Bannister, Scott C and Barker, Jon P and Cox, Trevor J and Fazenda, Bruno M and Firth, Jennifer and Graetzer, Simone N and Greasley, Alinka E and Dabike, Gerardo Roa and Vos, Rebecca},
  title = {{Challenges in predicting the lyric intelligibility of musical segments for older individuals with hearing loss}},
  booktitle = {2025 AES International Conference on Artificial Intelligence and Machine Learning for Audio},
  note = {Paper 14},
  year = {2025},
  month = sep,
  publisher = {Audio Engineering Society},
  url = {https://aes.org/publications/elibrary-page/?id=23003}
}

TY  - CPAPER
TI  - Challenges in predicting the lyric intelligibility of musical segments for older individuals with hearing loss
AU  - Whitmer, William M
AU  - McShefferty, David
AU  - Akeroyd, Michael A
AU  - Bannister, Scott C
AU  - Barker, Jon P
AU  - Cox, Trevor J
AU  - Fazenda, Bruno M
AU  - Firth, Jennifer
AU  - Graetzer, Simone N
AU  - Greasley, Alinka E
AU  - Dabike, Gerardo Roa
AU  - Vos, Rebecca
T2  - 2025 AES International Conference on Artificial Intelligence and Machine Learning for Audio
M1  - Paper 14
PY  - 2025
DA  - 2025/09/02
UR  - https://aes.org/publications/elibrary-page/?id=23003
PB  - Audio Engineering Society
LA  - en
AB  - Difficulty understanding lyrics is a major barrier to enjoying music for people with hearing loss. Novel production techniques using machine learning could aid lyric understanding, but objective metrics for training are needed, and these must be informed by the experiences of the target population. Currently, there are no data on the lyric-recall ability of older individuals with hearing loss. Thirteen older participants with mostly mild-sloping hearing loss listened to and recalled 100 vocal segments of popular music that varied in genre, duration and word count. In each trial, participants heard a randomly chosen segment twice with a 5-s interstimulus interval. Reproduction was over headphones at 65 dB(A) using individualised frequency-dependent nonlinear gain to compensate for hearing loss. The proportion of words recalled correctly varied greatly across samples, from 0-100%, and varied as a function of genre, similar to past studies with different populations. Individual intelligibility across segments was correlated with age; segment intelligibility across individuals, however, was not correlated with word count or rate. An objective metric for lyric intelligibility for listeners with hearing loss was developed based on Whisper automatic speech recognition model and individualised hearing-loss simulation. At best, this metric was only able to account for 56% of the variance in listening test results and often underpredicting intelligibility. Improving sung-lyric intelligibility is a different challenge from spoken-speech enhancement for those with hearing loss. Not only does the enhancement need to be considered within the overall enjoyment of the music, but the variation in results as seen in the current study reflects variations in genre, orchestration and vocal quality in sung music.
ER  -