KEMAR Manikin Research Articles

Front-back bias and the 1000-Hz interaural intensity anomaly

Free field measurements were made of the level differences, as measured in an ear canal, for a noise source in front of the head compared to a source in back. The measurements were made on a Kemar manikin for 19 lateral angles over the complete 180-degree range for front-back comparisons. There were 7 octave bands, 125–8000 Hz, and two horizontal planes. Front-back level differences were generally positive except for the dramatic case of 1000 Hz, where the differences were negative over an azimuth range of 120 degrees. When the head-related elevation of the sources was increased from zero to 10 degrees, the range of the negative front-back differences grew to 140 degrees. These physical results are consistent with the strong tendency for listeners to identify a one-third octave noise band in a narrow range near 1000 Hz as a source in back. This tendency may be the origin of the often-observed Grantham effect, wherein the difference limen for the interaural level difference is larger near 1000 Hz than at higher or lower frequencies. The connection is that sources in back of the listener, whatever their frequencies, are not localized nearly as well as sources in front.

Acoustic feedback control in hearing aids with frequency warping

Acoustic Feedback Control continues to be a challenging problem due to the emerging form factors in advanced hearing aids and hearables, especially for severe and profound losses. In order to mitigate the effects of feedback in hearing aids, we present a novel use of frequency warping called “Freping,” for manipulating signals in the frequency domain. Freping is a non-linear transformation which breaks the similarity between the input and acoustic feedback signals of a hearing aid to prevent feedback. In this work, we implemented Freping in realtime for feedback management, as well as novel techniques to preserve the naturalness of sound. Our prior work has shown that Freping improves Added Stable Gain (ASG) in the presence of feedback while maintaining the same Hearing Aid Speech Quality Index (HASQI). We present new results for feedback management with Freping by performing “on-ear” measurements on an acoustic KEMAR manikin using standard hearing aid verification systems. We conducted feedback tests to measure the ASG with various configurations of Freping for various hearing loss levels using the NAL-NL2 prescriptions on a realtime open-source speech processing system for hearing aids. We also compare Freping feedback performance with that of commercial hearing aids.

Residual Hearing Does Not Influence the Effectiveness of Beamforming when Using a Cochlear Implant in Conjunction with Contralateral Routing of Signals

Introduction: Contralateral routing of signals (CROS) overcomes the head shadow effect by redirecting speech signals from the contralateral ear to the better-hearing cochlear implant (CI) ear. Here we tested the performance of an adaptive monaural beamformer (MB) and a fixed binaural beamformer (BB) using the CROS system of Advanced Bionics. Methods: In a group of 17 unilateral CI users, we evaluated the benefits of MB and BB for speech recognition by measuring speech reception threshold (SRT) with and without beamforming. MB and BB were additionally evaluated with signal-to-noise ratio (SNR) measurements using a KEMAR manikin. We also assessed the effect of residual hearing in the CROS ear on the benefits of MB and BB. Speech was delivered in front of the listener in a background of homogeneous 8-talker babble noise. Results: With CI-CROS in omnidirectional settings with the T-mic active on the CI as a reference, BB significantly improved SRT by 1.4 dB, whereas MB yielded no significant improvements. The difference in effects on SRT between the two beamformers was, however, not significant. SNR effects were substantially larger, at 2.1 dB for MB and 5.8 dB for BB. CI-CROS with default omnidirectional settings also improved SRT and SNR by 1 dB over CI alone. Residual hearing did not significantly affect beamformer performance. Discussion: We recommend the use of BB over MB for CI-CROS users. Residual hearing in the CROS ear is not a limiting factor for fitting a CROS device, although a bimodal option should be considered.

The combined effect of source directivity and binaural listening on near-field binaural speech transmission index evaluation

The binaural speech transmission index (BSTI) is mainly affected by the effect of binaural listening, background noise, and acoustic conditions in a given space, as well as source characteristics such as directivity. However, little attention has been paid to the combined effect of source directivity and binaural listening on the BSTI in near-field regions. This study thus investigates this issue using BSTI measurements in an anechoic environment to exclude the influence of acoustic conditions. Three directional loudspeakers are adopted as the sources, and the impulse responses under different radiation directions in the horizontal plane are measured in an anechoic chamber. These impulse responses are then convolved with the near-field head-related transfer functions of a KEMAR manikin to obtain the binaural impulse responses, which are further used to calculate the BSTIs. The results show that the BSTI decreases considerably with an increasing deviation of the axial direction of the source from the listener (i.e., an increasing radiation angle) due to the directivity pattern of the source, which is almost independent of the azimuth and the distance between the source and listener. Finally, the BSTI correction functions of the source radiation angles are established with third-order polynomial regression for the three sources and can be integrated into the azimuth-dependent BSTI model to yield a comprehensive model that considers the source radiation angle as well as the azimuth and distance between the source and listener.

Effect of wearing personal protective equipment on acoustic characteristics and speech perception during COVID-19

With the COVID-19 pandemic, the usage of personal protective equipment (PPE) has become ‘the new normal’. Both surgical masks and N95 masks with a face shield are widely used in healthcare settings to reduce virus transmission, but the use of these masks has a negative impact on speech perception. Therefore, transparent masks are recommended to solve this dilemma. However, there is a lack of quantitative studies regarding the effect of PPE on speech perception. This study aims to compare the effect on speech perception of different types of PPE (surgical masks, N95 masks with face shield and transparent masks) in healthcare settings, for listeners with normal hearing in the audiovisual or auditory-only modality. The Bamford–Kowal–Bench (BKB)-like Mandarin speech stimuli were digitally recorded by a G.R.A.S KEMAR manikin without and with masks (surgical masks, N95 masks with face shield and transparent masks). Two variants of video display were created (with or without visual cues) and tagged to the corresponding audio recordings. The speech recording and video were presented to listeners simultaneously in each of four conditions: unattenuated speech with visual cues (no mask); surgical mask attenuated speech without visual cues; N95 mask with face shield attenuated speech without visual cues; and transparent mask attenuated speech with visual cues. The signal-to-noise ratio for 50 % correct scores (SNR50) threshold in noise was measured for each condition in the presence of four-talker babble. Twenty-four subjects completed the experiment. Acoustic spectra obtained from all types of masks were primarily attenuated at high frequencies, beyond 3 kHz, but to different extents. The mean SNR50 thresholds of the two auditory-only conditions (surgical mask and N95 mask with face shield) were higher than those of the audiovisual conditions (no mask and transparent mask). SNR50 thresholds in the surgical-mask conditions were significantly lower than those for the N95 masks with face shield. No significant difference was observed between the two audiovisual conditions. The results confirm that wearing a surgical mask or an N95 mask with face shield has a negative impact on speech perception. However, wearing a transparent mask improved speech perception to a similar level as unmasked condition for young normal-hearing listeners.

Azimuth-dependent model of binaural speech transmission index based on near-field head-related transfer functions

As well as the background noise and acoustic conditions of a given enclosed space, the binaural effect has a significant influence on the binaural speech transmission index (BSTI), especially for nearby sources. This effect can be quantitatively described by head-related transfer functions (HRTFs). As HRTFs are highly individual, the BSTIs from one person’s HRTF should differ from those of other people. However, this issue has rarely been studied. In the current work, we used the near-field HRTFs of 56 people to obtain the corresponding BSTIs, and analyzed the individual differences among them. The results show that the average standard deviation among the 56 individual BSTIs is almost within the just-noticeable-difference level of 0.03, and decreases with increasing sound source distance. We then acquired BSTIs from the KEMAR manikin and found that there were no significant differences with the median BSTIs across the 56 subjects. Thus, using the BSTIs of the KEMAR manikin, we developed regression BSTI models as a function of the azimuth based on seventh-order polynomial regression at different source distances.

Comparison of laboratory and clinical computational metrics of acoustic measures of hearing-aid processed speech

Computational auditory metrics are used to characterize hearing-aid fittings in the laboratory and the clinic. The purpose of this study is to compare the speech intelligibility index (SII) as a clinical metric to the laboratory based SII and the hearing-aid speech perception index (HASPI) and hearing-aid speech quality index (HASQI) laboratory metrics. These comparisons are drawn from a comprehensive dataset of hearing-aid fittings for 120 hearing-aid recipients from a hospital-based audiology clinic. Hearing-aid devices are drawn from multiple manufacturers and multiple technology levels, and a total of nine hearing-aid devices are included. Acoustic recordings are made with the hearing-aids positioned on the KEMAR manikin and include multiple hearing-aid settings for each patient (manufacturer’s recommended fit, first fit by the audiologist, and the final fitting selected by the patient). Metric values are computed for each of the settings, and these metric comparisons highlight the advantages and challenges of each metric in characterizing hearing-aid signal processing. In addition, the results provide insight regarding hearing-aid fittings from the manufacturer’s fitting versus clinician/patient driven fittings. [Work supported by NIH R01 DC012289, NIH NRSA T32DC012280, and GN Hearing OCG6790B.]

Modeling sound-source localization of two sinusoidally amplitude-modulated noises in a sound field

Yost and Brown [JASA 133 (2013)] investigated the ability of listeners to localize two simultaneously presented, independent noises presented over loudspeakers from different locations. These experiments demonstrated that SAM noises that were out of phase at two spatially separated loudspeakers led to better localization performance than when the SAM noises were in phase at each loudspeaker. Performance was improved at SAM rates as high as almost 500 Hz as compared to non-SAM noise. Yost and Brown hypothesized that listeners’ behavior might be explained as a temporal-spectral (T-S) analysis, and showed that such an approach could, qualitatively, account for some of their behavioral data. This presentation will explore the degree to which a quantitative, biologically inspired peripheral/brainstem auditory model can predict some of the listener performance presented in Yost and Brown 2013. Specifically, the model structure includes simulated auditory nerve response using the Zilany et al. model (JASA, 2014) with rate-count-based estimation of interaural differences of time and intensity as they occur in the presented stimuli recorded with a KEMAR manikin.

New Evidence-based Fitting Paradigm for Over-the-Counter Hearing Aids

New Evidence-based Fitting Paradigm for Over-the-Counter Hearing Aids

Attenuation of airborne noise by wet and dry neoprene diving hoods

The insertion losses of five neoprene diving hoods of varying thicknesses (2 mm–9 mm) were measured in one-third octave bands using a Kemar manikin in a diffuse broadband noise field. The insertion losses were measured in air for both dry and wet hoods. The insertion loss was calculated as the sound level in each frequency band measured with the hood, minus the corresponding sound level measured without the hood. The insertion losses were similar for both ears of the manikin. Both wet and dry hoods neither attenuated nor amplified sound below 250 Hz. Between 315 Hz–1250 Hz, the insertion loss of each hood was negative, displaying a broad resonance with a gain of 6–8 dB. In this frequency range the hood acts as a mass-spring system, resonating like a drum skin when stretched over the ears. Above 1000 Hz, the insertion loss increased with frequency (10 dB per octave), reaching a maximum of 5000 Hz–6000 Hz. Wetting each hood did not significantly affect the insertion loss; the 'drum-skin' resonance frequency was marginally lower with a wet hood, and insertion losses may be marginally greater between 1000 Hz– 10 000 Hz. The resonance frequency decreased with increasing thicknesses of hood, and the insertion loss at frequencies above the resonance increased with hood thickness.

Automated pure tone audiometry with true wireless stereos earbuds

Amounts of automated pure tone audiometer applications have been developed for personal terminals in recent years. However, most of them require specifically designed headphones, which are usually expensive and not accessible to most people. The commercially available true wireless stereos earbuds (e.g., Honor Flypods and Apple Airpods) got a prevalence in recent years. In this study, Flypods were used in a specifically developed automated audiometer whose algorithms and calibrations were modified for wireless earphones. The calibration was established by using both a KEMAR manikin and a loudness comparison method. Twenty subjects with mild to moderate hearing loss were recruited in a clinical verification experiment. The average deviation of automated audiometry was 3.1 dB, when referring to the thresholds measured by a conventional manual audiometer with a pair of TDH 39 headphones. Most (63.2%) of the deviations were under 5 dB. The sensitivity and specificity across frequencies from 125 to 8000 Hz were 95.4% and 62.6%, respectively. The results demonstrate that it is practical to utilize this sort of affordable wireless earphones for preliminary hearing level screening.

Measured localization performance at low stimulus levels with an adaptive tracking task

Sabin et al. (2005) measured localization accuracy for 250-ms broadband noises that varied in location along a spherical surface, and varied in stimulus level, from 0 to 60 dB above the detection threshold. They found that, for lower stimulus levels, responses tended to be biased towards lower elevations and to the front, and were more accurate at higher stimulus levels. Whereas Sabin et al. (2005) randomized level and location between trials, the current study fixed location and increased level until a correct response was given (or the 80 dB SPL limit was reached). Our initial level was 12.5 dB SPL, and our step size 2.5 dB SPL. Responses tended to be biased towards lower vertical elevations and to the front at low stimulus levels, and be more accurate at higher levels, consistent with Sabin et al. The percentage of front/back errors was generally greater for locations in the rear hemifield relative to those in the frontal hemifield, and for locations at higher elevations relative to locations at lower elevations. Audibility analysis, performed using a KEMAR manikin, showed that localization errors tended to decrease the most when audible bandwidth increased at the acoustically better ear.

Computing interaural differences using idealized head models

The spherical model of the human head, attributable to Lord Rayleigh, accounts for important features of observed interaural time differences (ITD) and interaural level differences (ILD), but it also fails to capture many details. To gain an intuitive understanding of the failures, we computed ITDs and ILDs for a succession of idealized shapes approximating the human head: sphere, ellipsoid, ellipsoid plus cylindrical neck, ellipsoid plus cylindrical neck plus disk torso. Calculations were done as a function of frequency (100–2500 Hz) and for source azimuths from 10 to 90 degrees using finite-element models. The computations were compared to free-field measurements on a KEMAR manikin. The spherical head model approximated many measured interaural features, but the frequency dependence tended to be too flat in both ITD and ILD. The ellipsoidal head produced greater variation with frequency and therefore agreed better with the measurements, reducing the RMS discrepancies in both ITD and ILD by 35%. Adding a neck further increased the frequency variation. Adding the disk torso further improved the agreement, especially below 1000 Hz, decreasing the ITD discrepancy by another 21\%. The evolution of models enabled us to associate details of interaural differences with overall anatomical features. [Work supported by the AFOSR grant 11NL002.]

A Free Database of Head Related Impulse Response Measurements in the Horizontal Plane with Multiple Distances

A freely available collection of Head-Related Impulse Response (HRIR) measurements is introduced. The impulse responses were acquired in an anechoic chamber using a KEMAR manikin at four different loudspeaker distances – 3 m, 2 m, 1 m and 0.5 m – reaching from the far field to the near field. The loudspeaker was positioned at ear height and the manikin was rotated with a high-precision stepper motor in one degree increments. Besides the raw HRIRs also datasets are available which have been compensated for the use with specific headphone models. Introduction In order to localize a sound source, the human auditory system takes advantage of the time lag and level difference between the two ear signals as well as cues added by reflection and diffraction by torso, head and external ears [1]. A head-related impulse response (HRIR), acquired under free-field conditions, is able to describe these features. Several databases of HRIRs are freely available [2, 3, 4], but their angular resolution in the horizontal plane is limited and only [4] contains measurements for more than one source distance (0.8 m and 3 m). For distances larger than 1 m the distance between the manikin and the source is assumed to have no influence on the binaural cues, only the amplitude decreases for larger distances [5]. If the source is positioned closer to the listener, changes in the binaural cues have to be considered. This study will provide HRIR databases measured in the horizontal plane with a resolution of 1◦ for sources positioned at a distance of 0.5 m, 1 m, 2 m and 3 m. φ r = 0.5 m, 1 m, 2 m, 3 m Figure: Coordinate system used in this paper. The azimuth angle φ describes the placement of the loudspeaker counter-clockwise around the head. Here φ = −60◦. HRIR Databases The HRIR databases described here are freely available under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 license and can be downloaded at http://dev.qu.tu-berlin.de/projects/measurements/wiki The datasets are available in different formats suitable for a variety of applications: .mat files for use in GNU octave and Matlab, .daff files in the Open Directional Audio File Format (OpenDAFF), an emerging open standard for spatial audio data, and .wav files for use in the SoundScape Renderer (SSR).

Preferred sound levels of portable music players and listening habits among adults: A field study

The main purpose of this descriptive field study was to explore music listening habits and preferred listening levels with portable music players (PMPs). We were also interested in seeing whether any exposure differences could be observed between the sexes. Data were collected during 12 hours at Stockholm Central Station, where people passing by were invited to measure their preferred PMP listening level by using a KEMAR manikin. People were also asked to answer a questionnaire about their listening habits. In all, 60 persons (41 men and 19 women) took part in the questionnaire study and 61 preferred PMP levels to be measured. Forty-one of these sound level measurements were valid to be reported after consideration was taken to acceptable measuring conditions. The women (31 years) and the men (33 years) started to use PMPs on a regular basis in their early 20s. Ear canal headphones/ear buds were the preferred headphone types. Fifty-seven percent of the whole study population used their PMP on a daily basis. The measured LAeq60 sec levels corrected for free field ranged between 73 and 102 dB, with a mean value of 83 dB. Sound levels for different types of headphones are also presented. The results of this study indicate that there are two groups of listeners: people who listen less frequently and at lower, safer sound levels, and people with excessive listening habits that may indeed damage their hearing sensory organ in time.

Effect of spatial separation, extended bandwidth, and compression speed on intelligibility in a competing-speech task

The benefit for speech intelligibility of extending the bandwidth of hearing aids was assessed when the target speech (sentences) and background (two talkers) were co-located or spatially separated. Also, the relative benefits of slow and fast compression were assessed. Sixteen hearing-impaired (HI) subjects with mild-to-moderate high-frequency hearing loss and eight normal-hearing (NH) subjects were tested. The target and interfering sounds were recorded using a KEMAR manikin and were located at +/-60 degrees azimuth, either co-located or spatially separated. Simulated binaural hearing-aid processing using five-channel slow or fast compression was performed offline, with gains set individually for each HI subject. Upper cutoff frequencies were 5, 7.5, or 10 kHz. Processed stimuli were presented via headphones. For both NH (unaided) and HI subjects, there was no significant effect of cutoff frequency for the co-located condition, but a small but significant benefit from increasing the cutoff frequency from 5 to 7.5 kHz for the spatially separated condition. For the HI subjects, slow compression gave slightly but significantly higher scores than fast compression for the spatially separated but not for the co-located condition. There were marked individual differences both in the benefit from extended bandwidth and in the relative benefit of slow and fast compression.

Interaural coherence for noise bands: Waveforms and envelopes

This paper reports the results of experiments performed in an effort to find a formulaic relationship between the interaural waveform coherence of a band of noise gamma(W) and the interaural envelope coherence of the noise band gamma(E). An interdependence described by gamma(E)=pi/4+(1-pi/4)(gamma(W))(2.1) is found. This relationship holds true both in a computer experiment and for binaural measurements made in two rooms using a KEMAR manikin. Room measurements are used to derive a measure of reliability for the formula. Ultimately, a user who knows the waveform coherence can predict the envelope coherence with a small degree of uncertainty.

Angle dependent effects for impulse noise reduction for hearing protectors.

The proposed U.S. Environmental Protection Agency regulation for labeling hearing protection devices (HPDs) includes an impulsive noise reduction rating . In 2009, the American National Standards Institute Subcommittee for noise approved a revised standard for measuring the impulsive insertion loss of HPDs, ANSI/ASA S12.42-2009. The exposure at the ear in response to a forward-propagating wave depends strongly on the orientation of the head with respect to the direction of propagation. Furthermore, the insertion loss varies with the peak sound pressure level. This paper reports the results of tests performed using an acoustic shock tube to produce peak impulses of approximately 160-dB peak sound pressure level. Two manikins were evaluated: the GRAS KEMAR manikin equipped with 1/2 and 1/4 in. microphone in a GRAS 711 IEC coupler and the Institute de Saint Louis manikin equipped with a Bruel & Kjaer IEC 711 coupler equipped with a 1/4 in. microphone. The manikin heads were rotated through ±90 deg relative to the direction of the oncoming wavefront and impulsive peak insertion loss was measured according to S12.42-2009. [Portions of the research were supported by U.S. EPA Interagency Agreement No. 75921973-01-0.]

Computation of the head-related transfer function via the fast multipole accelerated boundary element method and its spherical harmonic representation

The head-related transfer function (HRTF) is computed using the fast multipole accelerated boundary element method. For efficiency, the HRTF is computed using the reciprocity principle by placing a source at the ear and computing its field. Analysis is presented to modify the boundary value problem accordingly. To compute the HRTF corresponding to different ranges via a single computation, a compact and accurate representation of the HRTF, termed the spherical spectrum, is developed. Computations are reduced to a two stage process, the computation of the spherical spectrum and a subsequent evaluation of the HRTF. This representation allows easy interpolation and range extrapolation of HRTFs. HRTF computations are performed for the range of audible frequencies up to 20 kHz for several models including a sphere, human head models [the Neumann KU-100 ("Fritz") and the Knowles KEMAR ("Kemar") manikins], and head-and-torso model (the Kemar manikin). Comparisons between the different cases are provided. Comparisons with the computational data of other authors and available experimental data are conducted and show satisfactory agreement for the frequencies for which reliable experimental data are available. Results show that, given a good mesh, it is feasible to compute the HRTF over the full audible range on a regular personal computer.

Angle dependent effects for impulse noise reduction for hearing protectors.

The proposed U.S. Environmental Protection Agency regulation for labeling hearing protection devices (HPDs) includes an impulsive noise reduction rating. In 2009, the American National Standards Institute Subcommittee for noise approved a revised standard for measuring the impulsive insertion loss of HPDs, ANSI/ASA S12.42-2009. The exposure at the ear in response to a forward-propagating wave depends strongly on the orientation of the head with respect to the direction of propagation. Furthermore, the insertion loss varies with the peak sound pressure level. This paper reports the results of tests performed using an acoustic shock tube to produce peak impulses of approximately 160-dB peak sound pressure level. Two manikins were evaluated: the GRAS KEMAR manikin equipped with 1/2- and 1/4-in. microphone in a GRAS 711 IEC coupler and the Institute de Saint Louis manikin equipped with a Bruel & Kjaer IEC 711 coupler equipped with a 1/4 in. microphone. The manikin heads were rotated through ±90 deg relative to the direction of the oncoming wavefront and impulsive peak insertion loss was measured according to S12.42-2009. [Portions of the research were supported by U.S. EPA Interagency Agreement No. 75921973-01-0.]