Dissertation
Acoustic complexity in real-world noise and effects on speech perception for listeners with normal hearing and hearing loss
University of Iowa
Doctor of Philosophy (PhD), University of Iowa
Summer 2022
DOI: 10.25820/etd.006434
Abstract
OBJECTIVE. Hearing loss is a major public health epidemic, and the primary treatment for hearing loss is hearing aids. Hearing aid uptake among people would could potentially benefit from hearing aid use, however, is very low. One primary reason for the low rate of hearing aid uptake is a lack of perceived benefit, particularly in noisy situations where people with hearing loss struggle to hear in the most. Many hearing aid technologies that are efficacious in laboratory tests of speech in noise perception are not effective in real-world noisy environments. One reason might be that the types of tests and noise stimuli typically used in audiology and hearing science research bare little resemblance to real-world noisy soundscapes. Improving our understanding of factors that affect listening in real-world noisy environments and using that knowledge to improve the ecological validity of laboratory and clinical assessments in audiology will improve hearing aid technologies, clinical decision making, and outcome assessment. To that end, the broad purpose of this study was to improve our understanding of real-world noise and its effects on speech perception for listeners with normal hearing and listeners with hearing loss. To do this, we measured the complexity of different real-world soundscapes using entropy, a way to quantify the amount of information in a system based on the probability structure of a system parameter. Using simulated real-world soundscapes in the lab, we characterized the effects of time-domain and frequency-domain entropy in real-world noise on speech perception among listeners with normal hearing and listeners with hearing loss in aided and unaided conditions. Then, participants collected data in real-world noisy environments they encountered in their daily lives by recording the environments and completing surveys while in the environment. We then investigated the effects of entropy in real-world soundscapes on speech intelligibility, listening effort, and perceived soundscape complexity.
DESIGN. Laboratory experiments: the complexity of noise recordings from eight real-world soundscapes were systematically quantified by their entropy in the time and frequency domains. Then, 50 segments the length of IEEE sentences were extracted from each environment. Twenty-five segments were extracted that varied from low to high entropy in the time domain and 25 segments were extracted that varied from low to high entropy in the frequency domain. Then, using a trial-by-trial design, participants repeated back IEEE sentences embedded in each segment in two blocks (time-domain entropy and frequency-domain entropy). For each trial, a random IEEE sentence was drawn, convolved with the room impulse response of the noise environment, and then combined with the noise at a fixed signal-to-noise ratio. Then, all trials were randomized and presented to the participant. Twenty-five participants with normal hearing and 25 participants with hearing loss participated. Participants with hearing loss completed the study twice: once without hearing aids and once with hearing aids, using the the Portable Hearing Aid Lab running the Open Master Hearing Aid software. Using linear mixed effects models, we characterized the effects of entropy on number of keywords correct. Real-world experiments: participants were asked to seek out 5-10 complex listening environments they encounter in their daily lives. Once in the environment, participants wore the Portable Hearing Hid lab and completed two ecological momentary assessments on a smartphone. The assessments asked participants questions about their environment as well as perception (e.g., speech intelligibility, listening effort, perceived complexity). The portable hearing aid lab also recorded their environment. For participants with normal hearing, the Portable Hearing Aid Lab provided no gain. For participants with hearing loss, the hearing aid switched from an unaided (acoustically transparent) to aided condition between the two momentary assessments. Using linear mixed effects models, we characterized the effects of entropy in the real-world on self-reported speech intelligibility, listening effort, and perceived soundscape complexity.
RESULTS. Laboratory experiments: 15,000 trials for the energy entropy block and 15,000 trials for the spectral entropy block were collected and analyzed. Entropy had significant effects on speech perception, with higher entropy resulting in poorer speech perception. Effects were larger for entropy in the time domain than the frequency domain. Effects were larger for participants with normal hearing than participants with hearing loss, and larger for participants with hearing loss in the aided than unaided condition. Speech perception differed significantly between most environments for all participants. Real-world experiments: effects of entropy in the real-world were more equivocal, but generally mirrored the findings of the laboratory experiments, with clearer effects of entropy in the time domain than frequency domain, for participants with normal hearing than hearing loss, and for participants with hearing loss in the aided than unaided condition. For speech listening, greater time-domain entropy generally yielded greater listening effort and higher perceived complexity for speech listening. Greater frequency-domain entropy yielded greater listening effort for participants with hearing loss in the aided condition. For overall environment perception, greater time-domain entropy was associated with higher overall ratings of environment complexity and listening effort, but only for participants with normal hearing and participants with hearing loss in the aided condition. For both laboratory and real-world experiments, speech perception was not different in the aided and unaided conditions for participants with hearing loss. Implications for ecological validity in hearing science and audiology research as well as clinical implications are discussed.
CONCLUSION. Entropy affected speech perception for listeners in a controlled laboratory experiment, with greater entropy resulting in poorer speech perception. Effects were larger for time-domain than frequency-domain entropy, and larger for listeners with normal hearing than listeners with hearing loss. Effects of entropy on listening performance in the real-world were more equivocal, with clearest effects for listening effort and complexity perception for listeners with normal hearing. Entropy should be considered when designing ecologically valid tests for speech perception and hearing aid benefit.
Details
- Title: Subtitle
- Acoustic complexity in real-world noise and effects on speech perception for listeners with normal hearing and hearing loss
- Creators
- Erik Jorgen Jorgensen
- Contributors
- Yu-Hsiang Wu (Advisor)Carolyn Brown (Committee Member)Octav Chipara (Committee Member)Inyong Choi (Committee Member)Shawn Goodman (Committee Member)Elizabeth Walker (Committee Member)
- Resource Type
- Dissertation
- Degree Awarded
- Doctor of Philosophy (PhD), University of Iowa
- Degree in
- Speech and Hearing Science
- Date degree season
- Summer 2022
- DOI
- 10.25820/etd.006434
- Publisher
- University of Iowa
- Number of pages
- xvii, 223 pages
- Copyright
- Copyright 2022 Erik Jorgen Jorgensen
- Comment
This thesis has been optimized for improved web viewing. If you require the original version, contact the University Archives at the University of Iowa: https://www.lib.uiowa.edu/sc/contact/.
- Language
- English
- Description illustrations
- illustrations (some color)
- Description bibliographic
- Includes bibliographical references (pages 155-187).
- Public Abstract (ETD)
Hearing loss is a major epidemic, and the primary treatment for hearing loss is hearing aid use. However, hearing aids do not work very well in the environments people with hearing loss struggle in the most–noisy environments. One reason for this is that little is known about real-world noise and how different types of noise affect our ability to hear speech in the real world. Historically, audiology and hearing science research has used artificial stimuli that does not resemble the types of noise we encounter in the real world. The purpose of this study was to measure complexity in noise from different real-world environments and investigate how it affects speech listening for people with normal hearing and people with hearing loss with and without hearing aids in both the laboratory and in the real world. Our laboratory experiments showed that complexity, or the degree of randomness, in real-world noise affects speech listening for all listeners, with larger effects for people with normal hearing and people with hearing loss wearing hearing aids. Our real-world experiments suggested that this finding holds in the real world, but the effects are less clear, particularly for listeners with hearing loss without hearing aids. Other factors that affect listening in the real world likely moderate the effects of noise complexity. Our findings have implications for the design of ecologically-valid tests and experiments in audiology and hearing science research, hearing aid design, and our general understanding of listening in noisy soundscapes.
- Academic Unit
- Communication Sciences and Disorders
- Record Identifier
- 9984285154002771
Metrics
8 File views/ downloads
167 Record Views