Background/Objectives: Adult hearing-impaired patients qualifying for cochlear implants typically exhibit less than 60% sentence recognition under the best hearing aid conditions, either in quiet or noisy environments, with speech and noise presented through a single speaker. This study examines the influence of deep neural network-based (DNN-based) noise reduction on cochlear implant evaluation. Methods: Speech perception was assessed using AzBio sentences in both quiet and noisy conditions (multi-talker babble) at 5 and 10 dB signal-to-noise ratios (SNRs) through one loudspeaker. Sentence recognition scores were measured for 10 hearing-impaired patients using three hearing aid programs: calm situation, speech in noise, and spheric speech in loud noise (DNN-based noise reduction). Speech perception results were compared to bench analyses comprising the phase inversion technique, employed to predict SNR improvement, and the Hearing-Aid Speech Perception Index (HASPI v2), utilized to predict speech intelligibility. Results: The spheric speech in loud noise program improved speech perception by 20 to 32% points as compared to the calm situation program. Thus, DNN-based noise reduction can improve speech perception in noisy environments, potentially reducing the need for cochlear implants in some cases. The phase inversion method showed a 4–5 dB SNR improvement for the DNN-based noise reduction program compared to the other two programs. HASPI v2 predicted slightly better speech intelligibility than was measured in this study. Conclusions: DNN-based noise reduction might make it difficult for some patients with significant residual hearing to qualify for cochlear implantation, potentially delaying its adoption or eliminating the need for it entirely.

Hearing aids have traditionally been designed to facilitate speech perception. With regards to music perception, previous work indicates that hearing aid users frequently complain about music sound quality. Yet, the effects of hearing aid amplification on musical perception abilities are largely unknown. This study aimed to investigate the effects of hearing aid amplification and dynamic range compression (DRC) settings on musical scene analysis (MSA) abilities and sound quality ratings (SQR) using polyphonic music recordings. Additionally, speech reception thresholds in noise (SRT) were measured. Thirty-three hearing aid users with moderate to severe hearing loss participated in three conditions: unaided, and aided with either slow or fast DRC settings. Overall, MSA abilities, SQR and SRT significantly improved with the use of hearing aids compared to the unaided condition. Yet, differences were observed regarding the choice of compression settings. Fast DRC led to better MSA performance, reflecting enhanced selective listening in musical mixtures, while slow DRC elicited more favorable SQR. Despite these improvements, variability in amplification benefit across DRC settings and tasks remained considerable, with some individuals showing limited or no improvement. These findings highlight a trade-off between scene transparency (indexed by MSA) and perceived sound quality, with individual differences emerging as a key factor in shaping amplification outcomes. Our results underscore the potential benefits of hearing aids for music perception and indicate the need for personalized fitting strategies tailored to task-specific demands.

This study evaluated a deep-neural-network denoising system using model-based design, comparing it with adaptive filtering and beamforming across various noise types, SNRs, and hearing-aid fittings. A KEMAR manikin fitted with five audiograms was recorded in reverberant and non-reverberant rooms, yielding 1,152 recordings. Speech intelligibility was estimated using the HASPI from 1,152 KEMAR manikin recordings. Effects of processing strategy and acoustic factors were tested with model-based within-device design that account for repeated recordings per device/program and fitting. Linear mixed model results showed that the DNN with beamforming outperformed conventional processing, with strongest gains at 0 and +5 dB SNR, moderate benefits at −5 dB in low reverberation, and none in medium reverberation. Across SNRs and noise types, the DNN combined with beamforming yielded the highest predicted intelligibility, with benefits attenuated under moderate reverberation. Azimuth effects varied; because estimates were derived from a better-ear metric on manikin recordings. Additionally, this paper reports comparisons using metrics of sound quality, for an intrusive metric (HASQI) and the pMOS non-intrusive metric. Results indicated that model type interacted with processing and acoustic factors. HASQI and pMOS scores increased with SNR and were moderately correlated (r² ≈ 0.479), supporting the use of non-intrusive metrics for large-scale assessment. However, pMOS showed greater variability across hearing aid programs and environments, suggesting non-intrusive models capture processing effects differently than intrusive metrics. These findings highlight the promise and limits of non-intrusive evaluation while emphasizing the benefit of combining deep learning with beamforming to improve intelligibility and quality.

Music is an important signal class for hearing aids, and musical genre is often used as a descriptor for stimulus selection. However, little research has systematically investigated the acoustical properties of musical genres with respect to hearing aid amplification. Here, extracts from a combination of two comprehensive music databases were acoustically analyzed. Considerable overlap in acoustic descriptor space between genres emerged. By simulating hearing aid processing, it was shown that effects of amplification regarding dynamic range compression and spectral weighting differed across musical genres, underlining the critical role of systematic stimulus selection for research on music and hearing aids.

Background/Objectives Adult hearing-impaired patients qualifying for cochlear implants typically exhibit less than 60% sentence recognition in their best hearing aid condition, either in quiet or noisy environments, with speech and noise presented through a single speaker. This study examines the influence of deep neural network-based (DNN-based) based noise reduction on cochlear implant evaluation. Methods Speech perception was assessed using AzBio sentences in both quiet and noisy conditions (multi-talker babble) at 5 and 10 dB signal-to-noise ratio (SNR) through one loudspeaker. Sentence recognition scores were measured for 10 (5-bilateral, 5-unilateral) hearing-impaired patients using three hearing aid programs: calm situation, speech in noise, and spheric speech in loud noise (DNN-based noise reduction). Speech perception results were compared to bench analyses comprising the phase inversion technique, employed to predict SNR improvement, and the Hearing-Aid Speech Perception Index (HASPI v2), utilized to predict speech intelligibility. Results The spheric speech in loud noise program improved speech perception by 20 to 32% points as compared to the calm situation program. This improvement might make it difficult for some patients to achieve the