Evaluation of a Transient Noise Reduction Strategy for Hearing Aids DOI: /jaaa

Transcription

1 J Am Acad Audiol 23: (2012) Evaluation of a Transient Noise Reduction Strategy for Hearing Aids DOI: /jaaa HaiHong Liu* Hua Zhang* Ruth A. Bentler Demin Han* Luo Zhang* Abstract Background: Transient noise can be disruptive for people wearing hearing aids. Ideally, the transient noise should be detected and controlled by the signal processor without disrupting speech and other intended input signals. A technology for detecting and controlling transient noises in hearing aids was evaluated in this study. Purpose: The purpose of this study was to evaluate the effectiveness of a transient noise reduction strategy on various transient noises and to determine whether the strategy has a negative impact on sound quality of intended speech inputs. Research Design: This was a quasi-experimental study. The study involved 24 hearing aid users. Each participant was asked to rate the parameters of speech clarity, transient noise loudness, and overall impression for speech stimuli under the algorithm-on and algorithm-off conditions. During the evaluation, three types of stimuli were used: transient noises, speech, and background noises. The transient noises included knife on a ceramic board, mug on a tabletop, office door slamming, car door slamming, and pen tapping on countertop. The speech sentences used for the test were presented by a male speaker in Mandarin. The background noises included party noise and traffic noise. All of these sounds were combined into five listening situations: (1) speech only, (2) transient noise only, (3) speech and transient noise, (4) background noise and transient noise, and (5) speech and background noise and transient noise. Results: There was no significant difference on the ratings of speech clarity between the algorithm-on and algorithm-off (t-test, p ). Further analysis revealed that speech clarity was significant better at 70 db SLP than 55 db SPL (p, 0.001). For transient noise loudness: under the algorithm-off condition, the percentages of subjects rating the transient noise to be somewhat soft, appropriate, somewhat loud, and too loud were 0.2, 47.1, 29.6, and 23.1%, respectively. The corresponding percentages under the algorithm-on were 3.0, 72.6, 22.9, and 1.4%, respectively. A significant difference on the ratings of the transient noise loudness was found between the algorithm-on and algorithm-off (t-test, p, 0.001). For overall impression for speech stimuli: under the algorithm-off condition, the percentage of subjects rating the algorithm to be not helpful at all, somewhat helpful, helpful, and very helpful for speech stimuli were 36.5, 20.8, 33.9, and 8.9%, respectively. Under the algorithm-on condition, the corresponding percentages were 35.0, 19.3, 30.7, and 15.0%, respectively. Statistical analysis revealed there was a significant difference on the ratings of overall impression on speech stimuli. The ratings under the algorithm-on condition were significantly more helpful for speech understanding than the ratings under algorithmoff (t-test, p, 0.001). Conclusions: The transient noise reduction strategy appropriately controlled the loudness for most of the transient noises and did not affect the sound quality, which could be beneficial to hearing aid wearers. Key Words: Digital noise reduction, hearing aid, hearing impairment, loudness *Beijing Tong Ren Hospital, Capital Medical University; Beijing Institute of Otolaryngology; Key Laboratory of Otolaryngology Head and Neck Surgery (Capital Medical University); Ministry of Education; Department of Communication Sciences and Disorders, University of Iowa Zhang Luo, Ph.D., No.17 Hou Gou Lane, Beijing Institute of Otolaryngology, Beijing, China, ; Phone: , ; Fax: ; dr.luozhang@gmail.com This study was supported by National Science Foundation of China ( ) and Projects in the National Science and Technology Pillar Program during the Eleventh Five-Year Plan Period (2008BAI50B01). 606

2 Impulse Noise Reduction Strategy/Liu et al Abbreviations: BTE 5 behind the ear; DNR 5 digital noise reduction; WDRC 5 wide dynamic range compression The majority of individuals with mild or moderate hearing impairment indicate that their primary problem is hearing in noise, with or without amplification provision (Kochkin, 2005). To address this problem, various signal-processing techniques have been introduced for implementation in hearing aids. For example, most hearing aids have some type of proprietary digital noise-reduction (DNR) algorithm that utilizes a slow-acting, modulation-based strategy and/ or a fast-acting filtering strategy (Bentler, 2005; Mueller and Ricketts, 2005; Bentler and Chiou, 2006). These technologies all share the general goal of reducing the level of signals identified as noise or other unwanted signal distortions. Whereas earlier algorithms were confined to noise classification based on envelope modulation rate and depth, current DNR algorithms may also include overall level, center frequency, spectral tilt, and other acoustic parameters that control the activation and strength of gain reduction. The goal of the sloweracting classic DNR algorithm is to improve listening ease in backgrounds of environmental noise. The faster-acting DNR (e.g., Wiener filtering) is intended to enhance speech perception in noise due to its capability of removing the noise during the pauses of speech. Several studies have investigated the effectiveness of DNR on transient sounds, such as doors slamming, dishes clattering, key jingling, car doors shutting, and so on (Bentler et al, 1993a; Bentler et al, 1993b; Boymans and Dreschler, 2000; Walden et al, 2000; Cord et al, 2002; Surr et al, 2002; Ricketts et al, 2003; Palmer et al, 2006a). The majority reported that DNR increased comfort or ease of listening while finding a lack of evidence for improvement in speech intelligibility (Bentler et al, 1993a; Bentler et al, 1993b; Walden et al, 2000). Many everyday environmental sounds cause listeners to report increased aversiveness with their hearing aids, even when both DNR strategies (fast or slow-acting) are being utilized (Palmer et al, 2006b; Bentler et al, 2008). To this end, transient- or impulsenoise reduction strategies are being developed (Keidser et al, 2007). Transient sounds consist of an intense pulse of acoustic energy rising well above the long-term average of the surrounding environment. They are typically broadband and always brief (transitory) in duration. Aside from the intensity and brevity of transients, they also are characterized by a leading waveform edge that rises rapidly (Voigt et al, 1980; Henderson and Hamernik, 1986). The more traditional approach for controlling transient noise in a hearing aid has been through the use of the automatic gain control (AGC) and/or peak clipping. However, traditional AGC circuits with attack times of 2 10 msec may be too slow to catch a transient noise that can rise to an intense peak in less than 1 msec (actually can be less than 0.5 msec) (Henderson and Hamernik, 1986). Peak clipping may be a better alternative because it is instantaneous; that is, clipping removes the peak of the transient noise, rather than reacting to a sudden increase in amplitude. However, it is well known that peak clipping can generate considerable distortion (Dillon, 2001). Furthermore, peak clipping only occurs when the hearing aid is at or near saturation. There are many aversive transient noises that are not intense enough to reach the peak clipping threshold, yet hearing aid wearers still report negative reaction to those less intense transient noises, such as dishes clattering, car doors shutting, or even pen tapping. In fact, the rapid onset time of the transient noises may not be reflected in the overall sound pressure level average over time, although the effect can be annoyance complaints (Hiramatsu et al, 1983; Fidell et al, 2002). In 2006, a transient noise detection and reduction algorithm called AntiShock TM was introduced in Unitron hearing aids. In brief, the principle of the algorithm is to apply time-domain detection and prediction with frequency-domain management. The acoustic transient sound is detected in the time-domain. The signal in the time-domain then is transformed into the frequency-domain for the signal processing and transient sound reduction management. Afterward, the frequency-domain signal is transformed back to timedomain. In general, the signal transformation from time-domain to frequency-domain and then back to time-domain is frame based by applying a certain window such as a Hamming. The frame size is typical 2 N samples such as 64 for 32-bin fast Fourier transform (FFT), which corresponds to a time length of 3.2 msec for a sampling rate of 20 khz. This process creates a certain time-delay t(1 10 msec according to the actual system implementation) between the signal input and signal output. The quick transient sound detectioninthetime-domainprovidesearlyprediction for the transient reduction processing in the frequency-domain. To date, there has been little evidence that either of the implemented strategies has a positive impact on transient sounds. The purposes of this study are (1) to evaluate the loudness control of AntiShock on various transient sounds (e.g., door slamming, knife on a ceramic board, pen tapping on a countertop) and (2) to determine whether AntiShock has a negative impact on sound quality of intended speech inputs. 607

3 Journal of the American Academy of Audiology/Volume 23, Number 8, 2012 AntiShock Algorithm METHODS During the transient noise detection, two thresholds, minimum transient noise contrast level and minimum transient noise index, are used for transient noise detection. Transient noise contrast level (dl) is defined as dl 5 L 2 Sn, where L is the sound pressure level in db and Sn is the fast averaging of a speech signal over a short duration of about 2 msec so that it can reflect the normal speech signal change over time. Transient noise index is defined as T index 5 a L Sn t1 t0, where a is the coefficient for transient noise index normalization, t1 is the time when the strength of the transient signal reaches the peak level, and t0 is the start point of the transient signal. These two thresholds can be determined through a self-learning process or predetermined measurement so that daily life nontransient signals (e.g., speech, music, normal acoustic sounds) are not detected as transient noise, and that a transient sound such as a door slamming will be detected as transient noise. That is, transient noise contrast level is determined by the difference between the peak level and the signal floor, and transient noise index is determined by use of a transient noise index normalization constant. The algorithm compares the transient noise contrast level and the transient noise index with respective thresholds and indicates an acoustic transient if one or both thresholds are exceeded. A stronger and sharper transient noise will generate a stronger transient noise index. The transient noise detection runs in real time using the thresholds of minimum transient noise contrast level and minimum transient noise index. According to the transient noise contrast level and transient noise index, the presence and strength of a transient noise can be determined. During the transient sound reduction management, the Gain Reduction [g(t)] can have three phases, named AntiShock attack phase, AntiShock holding phase, and AntiShock release phase, respectively. (1) AntiShock attack phase: g(t) 5 A 0 exp a(t2t1), t 2 (t1, t1 1 t), where A 0 is the full-on gain reduction as the Anti- Shock is fully engaged, which is to control the strength of reduction; a is the time constant for AntiShock attack speed; and t1 is the time when the strength of the transient sound reaches the peak level. (2) AntiShock holding phase: g(t) 5 (A 0 exp at ) exp 2b(t2t12t), t 2 (t1 1 t, t2 1 t), where b is the time constant for AntiShock release speed and t2 is the release point of the transient signal. (3) AntiShock release phase: g(t) 5 (A 0 exp at2b(t22t1) ). exp kðt t2 sþ, t. (t2 1 t), where l is the time constant for AntiShock final release. The A 0, a, b, andl can be predefined as constants or they can be adaptively updated according to the actual transient sound type and strength. In general, the higher the transient sound strength, the higher the A 0 and a.the shorter the system delay, the higher the a. A speech signal and transient noise (knife knocks on a plate) waveform with and without the AntiShock is shown in Figure 1. The upper panel shows the original transient noise (a) and speech signals (b and c); the lower panel shows the signals processed after the algorithm (a is the transient noise; b and c are speech signals, respectively). Participants Twenty-four new hearing aid users volunteered to participate in this study, including 11 females and 13males.Themedianageoftheparticipantswas 47.0 yr, ranging from 13 to 80 yr (SD of 24.7 yr). Prior to the evaluation, the purpose of the study was explained to the participants, and their informed consents were obtained. Pure tone audiometry showed that all participants had a sensorineural hearing loss with thresholds ranging from 25 to 90 db HL from 250 to 8000 Hz. The mean pure tone thresholds are shown in Table 1. To rule out possible cognitive difficulties, the Mini Mental State Examination (MMSE) was Figure 1. Speech signal and transient noise (knife knocks on a plate) waveform with and without the AntiShock. The upper panel shows the original transient noise (a) and speech signals (b and c); the lower panel shows the signals processed after the algorithm (a is the transient noise; b and c are speech signals, respectively). The horizontal axis represents the time (msec). The vertical axis represents the amplitude (sound pressure level). 608

4 Impulse Noise Reduction Strategy/Liu et al Table 1. Mean Pure-Tone Hearing Thresholds for All Subjects (n 5 24) Frequency (khz) Right (db HL) SD Left (db HL) SD administered. The MMSE scores for these participants ranged from 29 to 30, indicating that the participants cognitive performance was normal (Folstein et al, 1975). The use of human subjects was reviewed and approved by the Institutional Review Board of Beijing Tongren Hospital, Capital Medical University. Evaluation Environment and Evaluation Condition All the data collection was conducted in a calibrated IAC sound-isolating booth. The noise floor was below 24 dba. The participant was seated in the center of the room at 0 azimuth to a single speaker, located 1 m away. The speaker was placed at ear level to each participant, with an average height of 87 cm. Hearing Aids The Unitron Next 8 series behind-the-ear (BTE) hearing aids were fitted bilaterally to each participant. This model of hearing aid utilizes wide dynamic range compression (WDRC) in eight individual channels, and themicrophoneresponsecanbesettoomnidirectional, fixed directional, or adaptive directional. The hearing aids were programmed using the manufacturer-implemented National Acoustic Laboratories Non-linear version 1 (NAL-NL1) fitting formula. Real-ear measurement was conducted (Aurical Plus TM ) to adjust each listener s frequency gain response to the prescribed NAL-NL1 target. The AntiShock can be set to mild, moderate, or maximum level. During the evaluation, for the Anti- Shock on condition, the hearing aid was programmed to its moderate level because the moderate level was the default setting and would be the most common setting for most hearing aid users. The other parameters (DNR, feedback management, volume control) were set to the default setting. The microphone response was set to omnidirectional. Unvented earmolds were ordered to prevent transient noise releasing from a vent. The earmolds were coupled to the BTE hearing aids using standard size 13 tubing. Stimuli Three types of stimuli were used: transient noise, speech, and background noise. The transient noises included knife on a ceramic board, mug on a tabletop, office door slamming, car door slamming, and pen tapping on countertop. The speech sentences used for the test were presented by a male speaker in Mandarin. The background noises included party noise and traffic noise. All of these sounds were combined into five listening situations: (1) speech only: sentences were presented at 55 and 70 db SPL; (2) transient noise only: knife on a ceramic board, mug on a tabletop, office door slamming, car door slamming, and pen tapping on countertop; (3) speech and transient noise: male speech at 55/70 db SPL mixed with office door slamming/car door slamming/knife on ceramic board/pen tapping on countertop; (4) background noise and transient noise: party noise mixed with office door slamming/mug on a tabletop, traffic noise mixed with car door slamming; (5) speech and background noise and transient noise: male speech at 55/70 db SPL mixed with party noise/traffic noise and office door slamming/ mug on a tabletop/car door slamming. The stimuli used in the evaluation and their sound pressure level are shown in Table 2. The stimuli were recorded by Liaison Standard BTE Recording Satellite with Digigram VX Pocket 440 sound card. The spectral information of the transient sounds is shown in Figure 2. Procedure During the evaluation, the order of presentations and algorithm settings were randomized across subjects. A single-blinded paradigm was used during the evaluation, wherein the participants were blinded to the status of the algorithm, when the algorithm-on or algorithm-off was switched by a clinician. Each participant was asked to rate the stimuli between the algorithm-on and algorithm-off conditions respectively. Prior to the data gathering, the participants were instructed that they would listen to each test signal twice using two different settings in the hearing aids. After the stimulus, the participants were asked to rate the following aspects of the stimulus: clarity, loudness, and overall impression of the sound on Table 2. Stimuli and the Sound Pressure Level Used in the Evaluation Type of stimuli Item of stimuli Sound pressure level (db SPL) Transient noise Knife on a ceramic board 85 Mug on a tabletop 94 Office door slamming 115 Car door slamming 91 Pen tapping on countertop 79 Speech sentence Male speech in Mandarin 55/70 Background noise Party noise 72 Traffic noise

5 Journal of the American Academy of Audiology/Volume 23, Number 8, 2012 Figure 2. Spectral information of the transient sounds. The horizontal axis represents the time (hms refers to decimal time). The vertical axis represents the frequency (Hz). 610

6 Impulse Noise Reduction Strategy/Liu et al Table 3. Rating Parameters and Listening Conditions Evaluation domain Rating item Listening condition Clarity of speech Lots of distortion Speech only Some distortion Speech 1 transient noise Understandable Speech 1 background noise 1 transient noise Clear Very clear Loudness of transient noise Too soft Transient noise only Somewhat soft Speech 1 transient noise Appropriate Background noise 1 transient noise Somewhat loud Speech 1 background noise 1 transient noise Too loud Overall impression for speech stimuli Not helpful at all Speech 1 transient noise Somewhat helpful Speech 1 background noise 1 transient noise Helpful Very helpful provided scales. The rating aspects vary according to stimulus conditions. (Refer to Table 3 for rating aspects and listening conditions. For example, under the transient noise only condition, the subject would hear a specific transient sound twice under different settings of the algorithm. After the presentation, the subject would rate the loudness of the transient sound. Each level of the rating was assigned a number, i.e., for the rating of the speech clarity, lots of distortion 5 1, some distortion 5 2, understandable 5 3, clear 5 4, and very clear 5 5.) Three ratings were obtained in the study: Clarity of speech (1 5 lots of distortion, 2 5 some distortion, 3 5 understandable, 4 5 clear and 5 5 very clear); Loudness of the transient noise (1 5 too soft, 2 5 somewhat soft, 3 5 appropriate, 4 5 somewhat loud and 5 5 too loud); Overall impression for speech stimuli (1 5 not helpful at all, 2 5 somewhat helpful, 3 5 helpful and 4 5 very helpful). As noted in Table 3, the clarity of speech was evaluated under the conditions of speech only, speech and transient noise, and speech and background noise and transient noise. The loudness of transient noise was evaluated under the conditions of transient noise only, speech and transient noise, background noise and transient noise, and speech and background noise and transient noise. The overall impression for speech stimuli was evaluated under the conditions of speech and transient noise and speech and background noise and transient noise. Clarity of Speech RESULTS The result for speech clarity showed that, under the algorithm-off condition, the average percentages of subjects rating the speech sounds some distortion, understandable, clear, and very clear were 0.5, 29.6, 31.5, and 38.4%. Under the algorithm-on condition, the corresponding average percentages were 2.3, 29.2, 33.6, and 35.0%, respectively. A paired samples t-test (based on the rating levels as numbers) was employed to compare the ratings between the algorithm-on and algorithm-off conditions. Statistical analysis revealed that there was no significant difference on the ratings of speech clarity ( p. 0.05). Furthermore, we performed a two-way ANOVA to examine the effect of speech intensity (55 versus 70 db SPL) and listening conditions (speech only, speech and transient noise, speech and background noise and transient noise) on the algorithm. This analysis revealed that the ratings are significantly different between speech intensity levels and across listening conditions (F and p, for speech intensity; F and p, for listening conditions). We found Table 4. Rating for Speech Clarity Listening condition Speech intensity (algorithm-on) (algorithm-off) Comparison between algorithm-on and -off (p-value) Speech only Speech 1 transient noise , Speech 1 background noise 1 transient noise

7 Journal of the American Academy of Audiology/Volume 23, Number 8, 2012 Figure 3. Distributions of the loudness ratings of the transient noises. The x-axis represents the loudness score rating; the y-axis represents the percentage of responses correspondence to each loudness score. Loudness score: 1 too soft; 2 somewhat soft; 3 appropriate; 4 somewhat loud; 5 too loud. better speech clarity at 70 db SPL than 55 db SPL, and the ratings of the speech clarity decreased as the difficulty of listening conditions increased. Ratings for speech clarity under the algorithm-on and algorithm-off for each listening condition are shown in Table 4. Loudness of Transient Noise Judgments of the transient noises loudness revealed that AntiShock effectively controlled most of the transient noises. Under the algorithm-off condition, the average percentages of participants rating of the transient noise sounded somewhat soft, appropriate, somewhat loud, and too loud were 0.2, 47.1, 29.6, and 23.1%, respectively. The corresponding average percentages under the algorithm-on condition were 3.0, 72.6, 22.9, and 1.4%, respectively. A paired samples t-test was employed to compare the ratings between the algorithm-on and algorithm-off conditions. Statistics showed that there was significant difference in the ratings of the loudness of the transient noise (t 5 612

8 Impulse Noise Reduction Strategy/Liu et al Table 5. Rating for Transient Noise Loudness Listening condition Transient noise (algorithm-on) (algorithm-off) Comparison between algorithm-on and off (p-value) Transient only , , , , ,0.001 Speech 1 transient noise , , , ,0.001 Transient 1 background noise ,0.001 Speech 1 background noise 1 transient noise , , ,0.001 Note: 1 5 office door slamming; 2 5 car door slamming; 3 5 knife on a ceramic board; 4 5 mug on a tabletop; 5 5 pen tapping on countertop , p, 0.001). The comparison of the loudness ratings for every transient sound between the algorithmon and algorithm-off is shown in Figure 3. Ratings for transient noise loudness under the algorithm-on and algorithm-off for each listening condition are shown in Table 5. Overall Impression for Speech Stimuli The overall impression for speech stimuli revealed that, under the algorithm-off condition, the average percentages of subjects rating not helpful at all, somewhat helpful, helpful, and very helpful for speech stimuli were 36.5, 20.8, 33.9, and 8.9%, respectively. Under the algorithm-on condition, the corresponding average percentages were 35.0, 19.3, 30.7, and 15.0%, respectively. A paired samples t-test was employed to compare the ratings between the algorithm on and off conditions. Statistics showed that there was significant difference for the ratings of overallimpressiononspeechstimuli(t , p, 0.001). That is, the ratings under the algorithm-on were significantly more helpful for speech understanding than the ratings under the algorithm-off. Further analysis revealed a significant difference across conditions for speech intensity at 70 db SPL (p, 0.001). However, for speech intensity at 55 db SPL, there was no significant difference between the algorithm-on and algorithm-off conditions (p ). The response distribution of overall impression for speech stimuli ratings between the algorithm-on and algorithm-off is shown in Figure 4. Ratings for overall impression for speech stimuli under the algorithm-on and algorithm-off for each listening condition are shown in Table 6. Clarity of Speech DISCUSSION Transient noise can be disruptive for people wearing hearing aids. Ideally, the transient noise should be detected and controlled by the signal processor without disrupting speech and other intended input signals. Our results showed that there were no significant differences on the ratings of speech clarity ( p. 0.05) between the algorithm-on and algorithm-off conditions. Under the algorithm-on condition, the percentage of subjects rating the speech sounds clear or very clear was 68.6%, and the corresponding percentages under the algorithm-off were 69.9%. Furthermore, there were no ratings of lots of distortion under the algorithm-on condition, and the ratings of some distortion only accounted for 2.9%. These results showed that speech clarity did not deteriorate while the AntiShock algorithm was actively detecting and controlling the transient noises. As mentioned before, peak clipping may be an option for the transient noise management because it is fast enough to catch the transient noise. However, it is well known that peak clipping can produce considerable distortion. An effectiveness algorithm should detect and control the noise without disrupting speech and other intended input signals. Loudness of Transient Noise Results showed that under the condition of Anti- Shock on, some transient noises still sounded somewhat loud or too loud with the percentages of 22.9 and 1.4%, respectively. We found that the rating of somewhat loud and too loud concentrated on office 613

9 Journal of the American Academy of Audiology/Volume 23, Number 8, 2012 certain loudness level differences rather than making all kinds of transient noises the same comfortable level. A recent study conducted by Johnson et al (2010) provided a comparison of subjective performance and benefit measured with the Abbreviated Profile of Hearing Aid Benefit (APHAB) questionnaire for two groups. One group used 1990s-era linear processing hearing aids, whereas the other group used more current WDRC-capable hearing aids. They found that the latter reduced the aversiveness of amplified sound, which also suggests that new noise reduction technology may ameliorate (to some extent) the common complaint that hearing aids cause many everyday sounds to be objectionably loud. Overall Impression for Speech Stimuli Figure 4. Distributions of the overall impression for speech stimuli. The x-axis represents the overall impression for speech stimuli score rating; the y-axis represents the percentage of responses correspondence to each score. Overall impression for speech stimuli score: 1 not helpful at all; 2 somewhat helpful; 3 helpful; 4 very helpful; A: speech intensity 5 55 db SPL; B: speech intensity 5 70 db SPL. door slamming, which is the most intensive transient stimulus (refer to the sound pressure level of the stimuli in Table 2). During the evaluation, unvented earmolds were used to prevent the transient noise escaping from a vent. In clinical practice of hearing aid fitting, for the patient with a mild hearing loss or the patient with a mild hearing loss in the low frequencies, venting could be used for the relief of loudness and occlusion effect. Furthermore, transient noise reduction should keep Results of the study revealed that the majority of the inexperienced hearing aid users found the algorithm to be helpful while listening to variety of speech, transient, and background noise stimuli. We found that there was no significant difference between the algorithm-on and algorithm-off conditions for speech intensity at 55 db SPL. However, the ratings under the algorithm-on were significantly higher than the ratings under the algorithm-off for speech intensity at 70 db SPL. This outcome suggests that the signal-tonoise ratio is important for the subject in considering the helpfulness of the algorithm. Compared with speech intensity at 55 db SPL, hearing aid users listened at more favorable signal-to-noise ratios at 70 db SPL, which may have contributed to a better rating for overall impression for speech stimuli. Kochkin (2002) reported that the most important improvement hearing aid users want to see is improvement in understanding in noise. In an earlier study, the benefit of noise reduction was found to be able to improve subjective comfort for the listeners. That is, Mueller et al (2006) evaluated the aided annoyance levels of 22 participants using the Acceptable Noise Level (ANL) test. Results indicated that the mean aided ANL score improved approximately 4 db with the use of the manufacturer s algorithm, indicating that noise-reduction algorithms can reduce annoyance from noise, at least in laboratory conditions. Table 6. Rating for Overall Impression for Speech Stimuli Listening condition Speech intensity (algorithm-on) (algorithm-off) Comparison between algorithm-on and off (p-value) Speech 1 transient noise ,0.001 Speech 1 background noise 1 transient noise ,

10 Impulse Noise Reduction Strategy/Liu et al CONCLUSION This study evaluated a transient noise reduction strategy (AntiShock) on first-time hearing aids users. Three main conclusions can be drawn: (1) the algorithm appropriately controlled the loudness for most of the transient noises; (2) the algorithm did not negatively affect the clarity of speech; and (3) the majority of the inexperienced hearing aid users noted the algorithm to be helpful while listening to the variety of speech, noise, and transient stimuli. Dispensing clinicians should consider transient noise reduction in their selection of optimal hearing aid features for patients. Acknowledgments. The authors acknowledge Zhang Xuyang for his help with the analysis and Henry Luo (Unitron) for his help in design and technical clarification. REFERENCES Bentler RA. (2005) Effectiveness of directional microphones and noise reduction schemes in hearing aids: a systematic review of the evidence. J Am Acad Audiol 16: Bentler RA, Anderson CV, Niebuhr D, Getta J. (1993a) A longitudinal study of noise reduction circuits, I: objective measures. J Speech Hear Res 36: Bentler RA, Anderson CV, Niebuhr D, Getta J. (1993b) A longitudinal study of noise reduction circuits, II: subjective measures. J Speech Hear Res 36: Bentler RA, Chiou LK. (2006) Digital noise reduction: an overview. Trends Amplif 10: Bentler RA, Wu YH, Kettel J, Hurtig R. (2008) Digital noise reduction: outcomes from laboratory and field studies. Int J Audiol 47: Boymans M, Dreschler WA. (2000) Field trials using a digital hearing aid with active noise reduction and dual-microphone directionality. Audiology 39: Cord MT, Surr RK, Walden BE, Olsen L. (2002) Performance of directional microphone hearing aids in everyday life. J Am Acad Audiol 13: Dillon H. (2001) Hearing Aids. New York: Thieme Medical Publishers. Fidell S, Silvati L, Pearsons K. (2002) Relative rates of growth of annoyance of impulsive and non-impulsive noises. J Acoust Soc Am 111: Folstein MF, Folstein SE, McHugh PR. (1975) Mini-mental state: a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12: Henderson D, Hamernik RP. (1986) Impulse noise: critical review. J Acoust Soc Am 80: Hiramatsu K, Takagi K, Yamamoto T. (1983) Experimental investigation on the effect of some temporal factors of nonsteady noise on annoyance. J Acoust Soc Am 74: Johnson JA, Cox RM, Alexander GC. (2010) Development of APHAB norms for WDRC hearing aids and comparisons with original norms. Ear Hear 31: Keidser G, O Brien A, Latzel M, Convery E. (2007) Evaluation of a noise-reduction algorithm that targets non-speech transient sounds. Hear J 60:29, 32, 34, Kochkin S. (2002) Consumers rate improvements sought in hearing instruments. Hear Rev 9: Kochkin S. (2005) Customer satisfaction with hearing instruments in the digital age. Hear J 58: Mueller HG, Ricketts TA. (2005) Digital noise reduction. Hear J 58: Mueller HG, Weber J, Hornsby BWY. (2006) The effects of digital noise reduction on the acceptance of background noise. Trends Amplif 10: Palmer CV, Bentler RA, Mueller HG. (2006a) Evaluation of a second-order directional microphone hearing aid: II. Self-report outcomes. J Am Acad Audiol 17: Palmer CV, Bentler RA, Mueller HG. (2006b) Amplification with digital noise reduction and the perception of annoying and aversive sound. Trends Amplif 10: Ricketts T, Henry P, Gnewikow D. (2003) Full time directional versus user selectable microphone modes in hearing aids. Ear Hear 24: Surr RK, Walden BE, Cord MT, Olsen L. (2002) Influence of environmental factors on hearing aid microphone preference. JAm Acad Audiol 13: Voigt P, Godenhielm B, Ostlund E. (1980) Impulse noise measurement and assessment of the risk of noise induced hearing loss. Scand Audiol Suppl (Suppl. 12): Walden B, Surr R, Cord M, Edwards B, Olson L. (2000) Comparison of benefits provided by different hearing aid technologies. J Am Acad Audiol 11: