Ing. Václav Vencovský, Ph.D.

The basilar membrane in the cochlea can be modeled as an array of fluid coupled segments driven by stapes vibration and by the undamping nonlinear force simulating cochlear amplification. If stimulated with two tones, the model generates additional tones due to nonlinear distortion. These distortion products (DPs) can be transmitted into the ear canal and produce distortion-product otoacoustic emissions (DPOAEs) known to be generated in the healthy ear of various vertebrates. This study presents a solution for DPs in a two-dimensional nonlinear cochlear model with cochlear roughness-small irregularities in the impedance along the basilar membrane, which may produce additional DPs due to coherent reflection. The solution allows for decomposition of various sources of DPs in the model. In addition to the already described nonlinear-distortion and coherent-reflection mechanisms of DP generation, this study identifies a long-latency DPOAE component due to perturbation of nonlinear force. DP wavelets that are coherently reflected due to impedance irregularities travel toward the stapes across the primary generation region of DPs and there evoke perturbation of the nonlinear undamping force. The ensuing DP wavelets have opposite phase to the wavelets arising from coherent reflection, which results in partial cancellation of the coherent-reflection DP wavelets. (C) 2022 Author(s).

Middle-ear muscle reflex (MEMR) involving the stapedius was thought to be activated by loud sounds (approx. 80 dB SPL). However, recent findings indicate that the threshold may be much lower (around 15 dB less i.e. 65 dB SPL). This work presents analysis of MEMR effect on forward and reverse transmission, and pressure measured by a probe in the external auditory canal. MEMR is simulated in a middle ear model by increasing the stiffness of the stapedius.

This paper evaluates the ability of several algorithms to detect impulse distortions (clicks) in audio signals. The systems are evaluated against data from a listening test conducted using real audio signals provided by a vinyl manufacturer. Some of the signals contained clicks due to damage during the manufacturing process. An evaluation of click detection algorithms against listening test results focuses on the ability of the click-detection algorithms to detect perceptible clicks. The results presented in this paper show that an algorithm that employs a hearing model detected audible clicks with a lower false detection rate than the other algorithms in the test and that the wavelet transform-based algorithm with a dynamic threshold outperformed the other algorithms.

Stimulus-frequency otoacoustic emissions (SFOAEs) are generated by coherent reflection of forward traveling waves by perturbations along the basilar membrane. The strongest wavelets are backscattered near the place where the traveling wave reaches its maximal amplitude (tonotopic place). Therefore, the SFOAE group delay might be expected to be twice the group delay estimated in the cochlear filters. However, experimental data have yielded steady-state SFOAE components with near-zero latency. A cochlear model is used to show that short-latency SFOAE components can be generated due to nonlinear reflection of the compressor or suppressor tones used in SFOAE measurements. The simulations indicate that suppressors produce more pronounced short-latency components than compressors. The existence of nonlinear reflection components due to suppressors can also explain why SFOAEs can still be detected when suppressors are presented more than half an octave above the probe-tone frequency. Simulations of the SFOAE suppression tuning curves showed that phase changes in the SFOAE residual as the suppressor frequency increases are mostly determined by phase changes of the nonlinear reflection component.

The amplitudes of distortion-product otoacoustic emissions (DPOAEs) may abruptly decrease even though the stimulus level is relatively high. These notches observed in the DPOAE input/output functions or distortion-product grams have been hypothesized to be due to destructive interference between wavelets generated by distributed sources of the nonlinear-distortion component of DPOAEs. In this paper, simulations with a smooth cochlear model and its analytical solution support the hypothesis that destructive interference between individual wavelets may lead to the amplitude notches and explain the cause for onset and offset amplitude overshoots in the DPOAE signal measured for intensity pairs in the notches.

This paper presents a rate-code model of binaural interaction inspired by recent neurophysiological findings. The model consists of a peripheral part and a binaural part. The binaural part is composed of models of the medial superior olive (MSO) and the lateral superior olive (LSO), which are parts of the auditory brainstem. The MSO and LSO model outputs are preprocessed in the interaural time difference (ITD) and interaural level difference (ILD) central stages, respectively, which give absolute values of the predicted lateralization at their outputs, allowing a direct comparison with psychophysical data. The predictions obtained with the MSO and LSO models are compared with subjective data on the lateralization of pure tones and narrowband noises, discrimination of the ITD and ILD, and discrimination of the phase warp. The lateralization and discrimination experiments show good agreement with the subjective data. In the case of the phase-warp experiment, the models agree qualitatively with the subjective data. The results demonstrate that rate-code models of MSO and LSO can be used to explain psychophysical data considering lateralization and discrimination based on binaural cues.

Deteriorated speech comprehension is a common manifestation of the age-related decline of auditory functions (presbycusis). In this study, speech comprehension ability and other parameters of auditory function were assessed in elderly subjects with and without tinnitus. Apart from high-frequency audiograms in quiet and in background noise, speech recognition thresholds in silence or in competitive babble noise, and the ability to understand temporally gated speech, we measured also sensitivity to frequency modulation (FM) and interaural delay, gap detection thresholds (GDT), or the difference limens of intensity. The results show that in elderly participants, tinnitus per se has little influence on speech comprehension. The tinnitus patients also show similar GDT, sensitivity to interaural intensity difference, and sensitivity to FM as the NT subjects. Despite these similarities, nevertheless, significant differences in auditory processing have been found in the tinnitus participants: a worse ability to detect tones in noise, a higher sensitivity to intensity changes, and a higher sensitivity to interaural time differences. Additional correlation analyses further revealed that speech comprehension in the T subjects is dependent on the sensitivity to temporal modulation and interaural time delay, while these correlations are weak and non-significant in the NT subjects. Therefore, despite similarities in average speech comprehension and several other parameters of auditory function, elderly people with tinnitus exhibit different auditory processing, particularly at a suprathreshold level. The results also suggest that speech comprehension ability of elderly tinnitus patients relies more on temporal features of the sound stimuli, especially under difficult conditions, compared to elderly people without tinnitus.

Distortion product otoacoustic emissions (DPOAEs) are evoked by two stimulus tones with frequency f1 and f2 of ratio f2 / f1 in the range between approximately 1.05 and 1.4. This study theoretically and experimentally analyzes the cubic 2 f1 - f2 DPOAE for different stimulus levels of one of the tones while the other is constant. Simulations for f2 / f1 of 1.2 and moderate stimulus levels (30-70 dB sound pressure level) indicate that cubic distortion products are generated along a relatively large length of the basilar membrane, the extent of which increases with stimulus level. However, apical from the place of maximum nonlinear force, the wavelets generated by these distributed sources mutually cancel. Therefore, although the spatial extent of the primary DPOAE sources broadens with increasing stimulus level (up to 1.5 oct), the basilar-membrane region contributing to the DPOAE signal is relatively narrow (0.6 oct) and level independent. The observed dependence of DPOAE amplitude on stimulus level can be well-approximated by a point source at the basilar-membrane place where the largest distortion product (maximum of the nonlinear force) is generated. Onset and offset of the DPOAE signal may contain amplitude overshoots (complexities), which are in most cases asymmetrical. Two-tone suppression was identified as the main cause of these onset and offset complexities. DPOAE measurements in two normal-hearing subjects support the level dependence of the steady-state DPOAE amplitude and the asymmetry in the onset and offset responses predicted by the theoretical analysis.

A two-dimensional nonlinear cochlear model was used to study the dependence of the nonlinear-distortion component of cubic distortion product otoacoustic emissions (DPOAEs) on the levels of the primary tones: f(1), f(2.) DPOAE was simulated for a fixed frequency ratio between the primaries f(2)/f(1) = 1.2 and for two amplification gains at f(2) of 1.2, 2.4 and 4.8 kHz. The simulated optimal primary levels depend on frequency. Loss of the gain affects the optimal levels at the lowest intensities for which the DPOAE amplitude would fall below the noise level in real experiments. The nonlinear force acting in the model as a source of DPOAEs along the basilar membrane (BM) was calculated. Simulations showed that the nonlinear force spreads over a wide part of the length of BM, especially at large levels of the primaries and low f(2). Contributions along the source length may cancel each other out if the phase difference between them is half a cycle. This contributes to saturation of the DPOAE amplitude and causes notches. (C) 2018 The Author(s). Published by S. Hirzel Verlag . EAA. This is an open access article under the terms of the Creative Commons Attribution (CC BY 4.0) license.

For a given type of headphones the sound pressure levels of pure tones at which an adequate number of young listeners without hearing loss just perceived the tones are called Reference Equivalent Threshold Sound Pressure Levels (RETSPLs). RETSPLs for the Sennheiser HD 650 circumaural headphones were measured in this paper for twenty five young listeners whose age was in the range of 19 to 28 years. The data were measured at frequencies between 125 Hz and 16 kHz. Since RETSPLs for this type of headphones are not currently available this paper may be useful for researchers who employ this type of headphones in experiments. The paper also compares RETSPLs for various types of supra-aural and circumaural headphones, which are available in standards.

It has been suggested that the stimulus frequency otoacoustic emissions (SFOAEs) are generated due to coherent reflections of forward-Traveling waves. In this paper, the emissions were simulated by a two-dimensional hydrodynamic cochlea model with parameters based on real dimensions of the human cochlea. Simulations were done using a low-level pure tone at frequencies between 0.5 and 1.3 kHz, and 3 and 6 kHz. In agreement with animal experimental data, the ratio between the simulated SFOAE group delays and group delays estimated from the basilar membrane transfer functions were between 1 and 2. In agreement with the coherent reflection filtering theory, the SFOAE delay was affected by the frequency selectivity of the cochlear filters. However, the simulated results gave different tuning ratios for two models differing only in tuning of their cochlear filters.

The distortion product otoacoustic emissions (DPOAEs) are generated when the cochlea is stimulated by two pure tones with different frequencies f1 and f2. Onset of the DPOAE amplitude may have a nonmonotonic complex shape when the f2 is pulsed during a stationary f1 input. Observed complexities have been explained as (1) due to the secondary source of the DPOAE at the distortion product (DP) characteristic site, and (2) due to the spatial distribution of DP sources with different phases. There is also a third possibility that the complexities are due to the suppression of the f1 basilar membrane (BM) response during the f2 onset. In this study, a hydrodynamic cochlea model is used to examine influence of f1 suppression on the time course of DPOAE onset. In particular, a set of simulations was performed for frequency ratio f2/f1 = 1.26 and various levels of the primary tones (L1 and L2=30-70 dB SPL) to determine the relationship between time dependencies of the DPOAE onset and the suppression of the f1 BM response. The model predicts that suppression of the f1 BM response can cause suppression of DPOAE amplitude during the onset period.

This study is focused on the perceived roughness of two simultaneous harmonic complex tones with ratios between their fundamental frequencies set to create intervals on just-tempered (JT) and equal-tempered (ET) scales. According to roughness theories, ET intervals should produce more roughness. However, previous studies have shown the opposite for intervals in which the lower fundamental frequency of the complex was equal to 261.6 Hz. The aim of this study is to verify and explain these results by using intervals composed of complexes whose spectral components were generated with either a sine starting phase or with a random starting phase. Results of the current study showed the same phenomenon as previous studies. To examine whether the explanation of the phenomenon lies in the function of the peripheral ear, three roughness models based upon this function were used: the Daniel and Weber (1997) model, the synchronization index (SI) model, and the model based on a hydrodynamic cochlear model. For most of the corresponding JT and ET intervals, only the Daniel and Weber (1997) model predicted less roughness in the ET intervals. In addition to this, the intervals were analyzed by a model simulating the auditory periphery. The results showed that a possible cause for the roughness differences may be in the frequencies of fluctuations of the signal in the peripheral ear. For JT intervals the fluctuations in the adjacent places on the simulated basilar membrane had either the same frequency or integer multiples of that frequency and were synchronized. Since a previous study showed that synchronized fluctuations in adjacent auditory filters lead to higher roughness than out of phase fluctuations (Terhardt, 1974), this may cause more roughness across JT and ET intervals.

This study presents a system which detects clicks in sound (audible degradations). The system is based on a computational model of the peripheral ear. In order to train and verify the system, a listening test was conducted using 89 short samples of analog (vinyl) records. The samples contained singing voice, music (rock’n’roll), or both. We randomly chose 30 samples from the set and used it to train the system; then we tested the system using the 59 remaining samples. The system performance expressed as a percentage of correct detections (78.1%) and false alarms (3,9%) is promising.

The term roughness is used to describe a specific sound sensation which may occur when listening to stimuli with more than one spectral component within the same critical band. It is believed that the spectral components interact inside the cochlea, which leads to fluctuations in the neural signal and, in turn, to a sensation of roughness. This study presents a roughness model composed of two successive stages: peripheral and central. The peripheral stage models the function of the peripheral ear. The central stage predicts roughness from the temporal envelope of the signal processed by the peripheral stage. The roughness model was shown to account for the perceived roughness of various types of acoustic stimuli, including the stimuli with temporal envelopes that are not sinusoidal. It thus accounted for effects of the phase and the shape of the temporal envelope on roughness. The model performance was poor for unmodulated bandpass noise stimuli.

It has been shown that masked thresholds for complex tone maskers may depend on the relative phase between the spectral components of said maskers. Since these masker phase effects are less pronounced in hearing-impaired listeners, it indicates a possible role of peripheral compression. In order to study this phenomenon, we used a previously published physical model of the cochlea. We implemented the model into a version of the temporal-window model and used it to predict masked thresholds in harmonic complex tone maskers. The predicted thresholds were qualitatively similar with behavioral data (reproduced from the literature) of normal-hearing and hearing-impaired listeners: differences between the maximal and minimal masked thresholds decreased with increasing bandwidth of auditory filters (i.e., with increasing loss of peripheral compression). This effect was independent on the duration of the temporal window. In the predicted data, correlations between the maximum masking difference and the bandwidth of the cochlear filters were significant. However in the behavioral data, the correlations were significant but less pronounced. A loss of peripheral compression affects the isointensity responses of the simulated cochlear filters – broadens the response magnitudes and changes the curvature of the response phases. In the cochlear model, the broadening of the cochlear filters did not compensate for the decrease of the maximum masking difference caused by the changed phase curvature. If the relation between the filter bandwidth and compression loss is the same in the cochlear model and in the real cochlea, the less pronounced correlations observed in the behavioral data may indicate changed temporal resolution or other impairment in the auditory system.

Vencovský [ISMA2014, p. 483-488] introduced a new roughness model and showed its performance in comparison with listening tests for three types of complex acoustics stimuli: amplitude-modulated harmonic complex tones, real samples of pathological voices (sustained vowel /a/), harmonic intervals of the chromatic scale composed of two harmonic complex tones (dyads). This study extends his results. It adds a new type of stimuli (low-frequency harmonic complex tones), slightly changes the method used to estimate roughness, and, for a comparison, depicts results for the Daniel and Weber roughness model, and the Leman synchronization index (SI) roughness model. The Vencovský model performed well for all of the stimuli. The Daniel and Weber model performed well for the first type of stimuli, but its results were poor for the rest of the stimuli. The SI model performed well for the first three types of the stimuli but poorly for the low-frequency harmonic complex tones.

The implementation of binaural auditory model able to reflect the lateral position of tones with interaural time differences is presented. The model is composed of two parts, monaural processing model adapted from Dau and the binaural processing model designed by authors. The binaural processing model is simulating medial superior olive (MSO), part of human brain stem which is claimed to be responsible for coding of the temporal differences between signals in two ears. Designed model is not using most widely used framework for binaural models, Jeffress' delay line, instead of it own designed approach inspired by Grothe's paper is implemented. The output of the model was compared with the subjective listening tests taken from literature. The results show that presented model is able to reliably reflect the subjective data.

Human hearing system allows us to localize the sound source in the space. Superior olivary complex placed in the human brain stem is believed to be responsible for this ability. Physiological model simulating the function of this complex was presented in the paper (Ville Pulkki, Toni Hirvonen, Acta Acoustica united with Acoustica, 2009, vol95. pp. 883 - 899). The model was implemented into Matlab and verified by authors of the presented paper. It is planned to use this model in the system allowing to detect the position of the sound source in the space.

Estimation of pure tone hearing threshold is a part of hearing tests conducted in order to know the sensitivity of the human auditory system. This paper presents manual audiometer implemented in Matlab which can be used to conduct such measurements. Audiometer was calibrated by means of an artificial ear and pistonphone. Modified Hughson-Westlake method usually used in manual audiometry is described.

Binaural auditory model able to reflect the lateral position of tones with interaural time differences is presented. Overall model is composed of two parts, the first part is monaural model presented in the literature and the second part is the binaural model designed by the authors. The binaural model is in contrast to the most often used models not using a delay line concept. The output of the model represents signal in the medial superior olive (MSO) in auditory brainstem which is claimed to be responsible for coding of temporal differences between signals in two ears. The model is also extended by a cognitive part which allows to calculate signal lateralization comparable with data obtained by means of subjective listening tests taken from the literature. The results shows that model is able to reliably reflect the subjective data.

Hearing system allows us to localize the sound sources in the open space. Sound is entering auditory system by ears placed on contralateral sides of the head. This generally causes that the sounds coming from the different angles in a space are not exactly the same in both ears. Interaural time and level differences can occure. These differences allow us to localize the sound in space. Binaural model allowing to detect interaural level differences between left and right ear was developed by Ville Pulkki and Toni Hirvonen. Implementation of the LSO part of model in MATLAB is desribed in this paper.

Time-domain computational auditory model composed of consecutive outer- and middle-ear, cochlear frequency selectivity, inner hair cells, squaring expansion, adaptation, low pass modulation filter and modulation filterbank and constant variance noise stage is described. The model ability to simulate experimental data, such as intensity discrimination and simultaneous masking is evaluated by means of an optimal detector.

Auditory model which was designed in order to simulate the real function of the peripheral parts of the human auditory system, specifically outer- and middle-ear and the response of basilar membrane (BM) and inner hair cells in the cochlea according to physiological and psychophysical observations, and thus it can account for a variety of phenomena affecting the perceived sound signals, for example nonlinear distortions and intermodulation products. Capablitity of the model to simulate psychophysically observed phenomena, such as masking patterns is verified.

Pinna, head and body causes distortion of the perceived sound and it, in turn, help to sound source localization in medial plane, where the interaural time and level difffferences between the signals coming to both ears are the same. Auditory model is used in this paper to analyse which spectral parts of transient sound signal are mostly affffected by the pinna, head and body in dependence on the sound source position in space. The presented results show that mainly infuenced bands are between 5 and 12 kHz, as was already shown in literature, and additionally to these data, some other bands located in lower frequencies (between 1 and 4 kHz). Subjective measurements need to be done in order to evaluate these results that lower frequencies are also important for the sound source localization in the medial plane.

A computational auditory model in the objective audio quality assessment system PEMO-Q was replaced by nonlinear auditory model which was recently evaluated in comparison to the psychophysical data (simultaneous and time masking, detection of amplitude modulation etc). The modified system was used to assess the audio quality of various sound samples. Measured objective difference grade (ODG) data were compared with the subjective difference grade (SDG) data. The ODG data were also compared with the ODG data obtained by means of the original PEMO-Q system. The correlation coefficients between the SDG and the ODG data were calculated. Although the used auditory model can better simulate psychophysical data than auditory model used in the original PEMO-Q system, correlation between SDG and ODG data measured in this paper is higher for the original PEMO-Q system.

A physiological auditory model is described. The model simulates a processing of a sound by an outer, middle and inner ear. The nonlinear inner ear model comprises the cochlear frequency selectivity model and the inner hair cells model proposed according to mammalian physiological data. A capability of the auditory model to simulate human psychophysical masking data is verified.

A physiological auditory model is described. The model simulates a processing of a sound by an outer, middle and inner ear. The nonlinear inner ear model comprises the cochlear frequency selectivity model and the inner hair cells model proposed according to mammalian physiological data. A capability of the auditory model to simulate just noticeable level difference data and masking data is verified.

Analyzer system intended to visualize the sound signal in the time and frequency domain is presented. The system was realized in Matlab and it can process sound signals of wav files. Spectral analyzer takes into account the perception of the sound by humans and thus the magnitude spectrum can be visualized with frequency axis in Barks end ERBs. Spectrum can be weighted by the transfer function of A-curve.

Ear model which consists of two main stages is implemented. The first stage is outer/middle ear filter and the second stage is computational algorithm of cochlear frequency selectivity based on dual resonance nonlinear type of filters. Influence of settings the parameters which affect the nonlinearity of cochlear filterbank on simulated tone-on-tone masking patterns is evaluated.