Croghan, N., Arehart, K. & Kates, J. (2014). Music preferences with hearing aids: effects of signal properties, compression settings and listener characteristics. Ear & Hearing, in press.
This editorial discusses the clinical implications of an independent research study and does not represent the opinions of the original authors.
For the hearing aid wearer, speech is arguably the most important sound but hearing aid satisfaction is affected by the way in which the devices process other environmental sounds, including music. Adoption of hearing aids by active, technology savvy users makes their ability to process music with optimal sound quality and minimal distortion is more important than ever. Though modern hearing aids do an effective job of processing speech, even in the presence of competing noise, many hearing aid users report that hearing aids either make no difference or make music less enjoyable (Leek et al., 2008).
Music and speech have different spectral and temporal characteristics, with music often being higher in intensity and more dynamic than speech (Chasin, 2003, 2006, 2010). Speech maintains somewhat similar and predictable acoustic characteristics across talkers; in contrast, the spectral and temporal characteristics of music vary widely from one instrument to another and one piece to another (Chason & Russo, 2004). Not surprisingly, some studies have indicated that the best hearing aid circuit characteristics and settings for speech recognition may not be optimal for music perception (Higgins et al, 2012; van Buuren et al., 1999). For instance, faster compression time constants may be helpful for restoring speech audibility and loudness perception (Moore, 2008) but listeners may prefer longer release times for listening to music (Hansen, 2002; Moore, 2011).
Recorded music heard by hearing-aid users is subject to two stages of compression; compression limiting during the studio recording and wide-dynamic range compression in the hearing aid. Processing at both of these stages could impact music sound quality and subsequent enjoyment by the listener. Croghan and colleagues investigated the acoustic and perceptual effects of compression on music processed through hearing aids. They examined the effect of compression limiting prior to hearing aid processing and compared slow versus fast hearing aid compression time constants as well as small versus large numbers of channels. In addition to these compression variables, they examined potential effects of suprathreshold processing and prior musical training.
Eighteen hearing aid users, ranging from 49 to 87 years of age, participated in the study. Subjects were divided into non-musician and musician groups. Two pieces of music, one classical and one rock, were used in the study. The pieces were selected to be relatively unfamiliar to the subjects, to reduce any effect of prior experience or expectations. To simulate studio processing, the music samples were recorded in three compression limiting conditions: no compression, mild compression limiting, and heavy compression limiting. These compression conditions were applied to the music samples prior to hearing aid processing.
Music was presented over binaural headphones, via a simulated hearing aid. Individual WDRC hearing aid simulations were programmed according to NAL-NL1 formulae for each subject. Stimuli were processed with two sets of compression release times and processing channels: fast (50msec) vs. slow (1000msec) release times and 3-channels vs. 18 channels. Two linear conditions were also included, using the NAL-R prescription with 3-channels and 18-channels for frequency shaping. The combination of 3 compression limiting, 4 WDRC and 2 linear conditions resulted in 18 processing conditions for each piece of music. Stimuli were presented at 65dB SPL and subjects made preference judgments in a 2-interval forced-choice paradigm.
To examine the effect of suprathreshold processing, three psychophysical tests were administered. Loudness perception was measured with the Contour Test of Loudness Perception (Cox et al., 1997), amplitude modulation depth discrimination was measured using speech-shaped noise modulated at 4Hz and frequency selectivity was measured with psychophysical tuning curves (Sek et al. 2005; Sek & Moore, 2011). The music stimuli were also analyzed with a modification of the Hearing Aid Speech Quality Index (HASQI; Kates & Arehart, 2010). Roughly stated, the HASQI provides an object sound quality rating by comparing time-frequency modulation and long-term spectrum of an unmodified signal to a modified one (the modified signal being one with the targeted signal processing applied). Not surprisingly, the lowest HASQI values, indicating the most difference between unprocessed and processed stimuli, were observed for fast WDRC combined with heavy compression limiting.
The 18 stimulus conditions were examined for the effect of compression on the overall dynamic range, amplitude by frequency and modulation of the music samples. Generally any increase in processing – increasing compression limiting, increasing the number of channels, going from linear to slow WDRC or slow to fast WDRC – reduced the dynamic range of the classical and rock music samples. WDRC caused more dynamic range reduction in the high frequencies. Compression limiting affected classical music similarly across frequencies, whereas rock music was affected more in the high and low-frequency regions than in the mid-frequencies. Compression – either WDRC or limiting – reduced the magnitude of modulation, likely making the rhythmic structure of the music less distinguishable.
Listener preference results for compression limiting and WDRC indicated some differences based on the type of music that was presented. For classical music, there was no significant difference between slow WDRC and linear processing, but both of these were preferred over fast WDRC. Mild or no compression was significantly preferred over heavy compression limiting. There was no effect for the number of channels on classical music preferences. Slightly different results were obtained for rock music: linear processing was preferred over both WDRC conditions and slow WDRC was significantly preferred over fast WDRC. There was no significant effect of compression limiting but the 3-channel condition was rated significantly better than the 18-channel condition.
The following listener-related factors were examined for their effects on preference: gender, musician vs. non-musician, PTA, dynamic range, tuning curve bandwidth and modulation depth discrimination threshold. Because PTA and dynamic range were strongly correlated to each other, these factors were excluded from the analysis. For classical music, the only significant findings were interactions among tuning curve bandwidth, WDRC condition and number of processing channels. Listeners with broader tuning curves showed a slight preference for linear amplification over WDRC and 3-channel over 18-channel processing. In contrast, the group with narrower tuning curves had a slight preference for slow WDRC and 18-channel processing. There were no other significant findings for listener-related factors. The authors posit that listeners with better frequency resolution may have preferred slow WDRC and 18-channel processing because they were able to resolve the harmonics and benefit from greater audibility. Conversely, listeners with poorer frequency resolution may have responded to reduced distortion in the linear and 3-channel conditions, despite potentially reduced audibility.
In this study, Croghan and her colleagues found that for music stimuli, compression limiting and WDRC reduced temporal envelope contrasts. These results are in agreement with previous studies using speech stimuli (Bor et al., 2008; Jenstad & Souza, 2005). They also found that compression limiting was more likely than WDRC to reduce the peaks of the modulation spectrum. This is somewhat in agreement with a previous report by Souza & Gallun (2010) on consonant discrimination, in which hearing aid compression limiting had an adverse effect but multi-channel, fast WDRC was beneficial. However, the authors point out that hearing aid compression limiting is different from music industry compression limiting in that the former compresses only high-level sounds and does not affect average (RMS) sound level.
The results of this study indicate that music was adversely affected by compression limiting and WDRC and that in general, listeners preferred listening to music with little or no compression. Listeners with broad psychophysical tuning curves showed a preference for 3-channel processing, whereas those with narrower tuning curves preferred 18-channel processing. This may be related to the ability of those with narrower tuning curves to perceive harmonics, especially in the classical piece, which is related to the perceived quality of stringed instruments (Chasin & Russo, 2004). This result is similar to a report using speech stimuli by Souza et al. (2012) in which listeners with better frequency resolution were more able to benefit from multi-channel compression. More research is needed to illuminate the relationship between suprathreshold processing and music perception. Traditional measures of psychophysical tuning curves is an unwieldy proposition for clinicians, but hearing aid users with impaired frequency resolution may require a modified treatment approach.
In contrast to previous studies in which musicians outperformed non-musicians on tests of frequency discrimination, speech discrimination and working memory (Parbery-Clark et al., 2009; 2012), Croghan and her colleagues found no significant difference in the psychophysical tests or preference ratings for musicians versus non-musicians. They point out, however, that their study used recorded music samples and preferences for live music cannot be extrapolated from their results. Clinicians should expect musicians to be analytical about the sound quality of their hearing aids and be prepared to offer a separate, manually accessible programs for music listening. Similarly, many non-musicians are music aficionados who would also appreciate an alternate program for music. In many hearing instruments, alternate music programs can be added using defaults available in manufacturer software, or can be individually customized. In music listening programs, special features like automatic directionality and noise reduction should be disabled and based on Croghan’s report, should probably have more linear processing and less compression than primary, everyday listening programs.
Croghan’s study provides insight into the ways in which hearing aid signal processing affects music acoustics and perception. Our current knowledge of music acoustics and hearing aid signal processing may be more meaningfully applied to the technical design of hearing aids than to routine clinical practice. While opportunities remain for meaningful advancement in the processing of music through hearing aids, some clinical advice can be offered:
- Musicians or individuals with strong musical interests may benefit from a dedicated memory, optimized for music listening.
- Optimization of a dedicated music listening memory is best attempted following a patient’s initial adaptation period to new hearing aids.
- Follow-up visits addressing music perception and sound quality should include multiple music samples of the patient’s own selection.
- Using default settings for music listening, the patient should be prompted to set the playback loudspeakers to a level they find pleasing for a given music sample.
Although the preferences of each patient are different, these suggestions are a solid foundation for providing patients with a high-quality music listening experience.
Chasin, M. (2003). Music and hearing aids. Hearing Journal 56, 36-41.
Chasin, M. (2006). Hearing aids for musicians. Hearing Review 13, 11-16.
Chasin, M. (2010). Amplification fit for music lovers. Hearing Journal 63, 27-30.
Chason, M. & Russo, F. (2004). Hearing aids and music. Trends in Amplification 8, 35-47.
Cox, R., Alexander, G. & Taylor, I. (1997). The contour test of loudness perception. Ear and Hearing 18, 388-400.
Croghan, N., Arehart, K. & Kates, J. (2014). Music preferences with hearing aids:
effects of signal properties, compression settings and listener characteristics. Ear & Hearing, in press.
Hansen, M. (2002). Effects of multi-channel compression time constants on subjectively perceived sound quality and speech intelligibility. Ear and Hearing 23, 369-380.
Higgins, P., Searchfield, G. & Coad, G. (2012). A comparison between the first-fit settings of two multichannel digital signal-processing strategies: music quality ratings and speech-in-noise scores. American Journal of Audiology 21, 13-21.
Leek, M., Molis, M. & Kubli, L. (2008). Enjoyment of music by elderly hearing-impaired listeners. Journal of the American Academy of Audiology 19, 519-526.
Moore, B. (2008) . The choice of compression speed in hearing aids: Theoretical and practical considerations and the role of individual differences. Trends in Amplification 12, 103-112.
Moore, B., Fullgrabe, C. & Stone, M. (2011). Determination of preferred parameters for multichannel compression using individually fitted simulated hearing aids and paired comparisons. Ear and Hearing 32, 556-568.
Moore, B. & Glasberg, B. (1997) A model of loudness perception applied to cochlear hearing loss. Auditory Neuroscience 3, 289-311.
Neumann, A., Bakke, M. & Hellman, S. (1995a). Preferred listening levels for linear and slow-acting compression hearing aids. Ear and Hearing 16, 407-416.
Parbery-Clark, A., Skoe, E. & Lam, C. (2009). Musician enhancement for speech-in-noise. Ear and Hearing 30, 653-661.
Parbery-Clark, A., Tierney, A. & Strait, D. (2012). Musicians have fine-tuned neural distinction of speech syllables. Neuroscience 219, 111-119.
Sek, A., Alcantara, J. & Moore, B. ( 2005). Development of a fast method for determining psychophysical tuning curves. International Journal of Audiology 44, 408-420.
Sek, A. & Moore, B. (2011). Implementation of a fast method for measuring psychophysical tuning curves. International Journal of Audiology 50, 237-242.
van Buuren, R., Festen, J. & Houtgast, T. (1999). Compression and expansion of the temporal envelope: Evaluation of speech intelligibility and sound quality. Journal of the Acoustical Society of America 105, 2903-2913.