Rainer Beutelmann

Prediction of speech intelligibility in spatial noise and reverberation for normal-hearing and hearing-impaired listeners.

Binaural speech intelligibility of individual listeners under realistic conditions was predicted using a model consisting of a gammatone filter bank, an independent equalization-cancellation (EC) process in each frequency band, a gammatone resynthesis, and the speech intelligibility index (SII). Hearing loss was simulated by adding uncorrelated masking noises (according to the pure-tone audiogram) to the ear channels. Speech intelligibility measurements were carried out with 8 normal-hearing and 15 hearing-impaired listeners, collecting speech reception threshold (SRT) data for three different room acoustic conditions (anechoic, office room, cafeteria hall) and eight directions of a single noise source (speech in front). Artificial EC processing errors derived from binaural masking level difference data using pure tones were incorporated into the model. Except for an adjustment of the SII-to-intelligibility mapping function, no model parameter was fitted to the SRT data of this study. The overall correlation coefficient between predicted and observed SRTs was 0.95. The dependence of the SRT of an individual listener on the noise direction and on room acoustics was predicted with a median correlation coefficient of 0.91. The effect of individual hearing impairment was predicted with a median correlation coefficient of 0.95. However, for mild hearing losses the release from masking was overestimated.

Revision, extension, and evaluation of a binaural speech intelligibility model

This study presents revision, extension, and evaluation of a binaural speech intelligibility model (Beutelmann, R., and Brand, T. (2006). J. Acoust. Soc. Am. 120, 331-342) that yields accurate predictions of speech reception thresholds (SRTs) in the presence of a stationary noise source at arbitrary azimuths and in different rooms. The modified model is based on an analytical expression of binaural unmasking for arbitrary input signals and is computationally more efficient, while maintaining the prediction quality of the original model. An extension for nonstationary interferers was realized by applying the model to short time frames of the input signals and averaging over the predicted SRT results. Binaural SRTs from 8 normal-hearing and 12 hearing-impaired subjects, incorporating all combinations of four rooms, three source setups, and three noise types were measured and compared to the models predictions. Depending on the noise type, the parametric correlation coefficients between observed and predicted SRTs were 0.80-0.93 for normal-hearing subjects and 0.59-0.80 for hearing-impaired subjects. The mean absolute prediction error was 3 dB for the mean normal-hearing data and 4 dB for the individual hearing-impaired data. 70% of the variance of the SRTs of hearing-impaired subjects could be explained by the model, which is based only on the audiogram.

Stream Segregation in the Perception of Sinusoidally Amplitude-Modulated Tones

Amplitude modulation can serve as a cue for segregating streams of sounds from different sources. Here we evaluate stream segregation in humans using ABA- sequences of sinusoidally amplitude modulated (SAM) tones. A and B represent SAM tones with the same carrier frequency (1000, 4000 Hz) and modulation depth (30, 100%). The modulation frequency of the A signals (fmodA) was 30, 100 or 300 Hz, respectively. The modulation frequency of the B signals was up to four octaves higher (Δfmod). Three different ABA- tone patterns varying in tone duration and stimulus onset asynchrony were presented to evaluate the effect of forward suppression. Subjects indicated their 1- or 2-stream percept on a touch screen at the end of each ABA- sequence (presentation time 5 or 15 s). Tone pattern, fmodA, Δfmod, carrier frequency, modulation depth and presentation time significantly affected the percentage of a 2-stream percept. The human psychophysical results are compared to responses of avian forebrain neurons evoked by different ABA- SAM tone conditions [1] that were broadly overlapping those of the present study. The neurons also showed significant effects of tone pattern and Δfmod that were comparable to effects observed in the present psychophysical study. Depending on the carrier frequency, modulation frequency, modulation depth and the width of the auditory filters, SAM tones may provide mainly temporal cues (sidebands fall within the range of the filter), spectral cues (sidebands fall outside the range of the filter) or possibly both. A computational model based on excitation pattern differences was used to predict the 50% threshold of 2-stream responses. In conditions for which the model predicts a considerably larger 50% threshold of 2-stream responses (i.e., larger Δfmod at threshold) than was observed, it is unlikely that spectral cues can provide an explanation of stream segregation by SAM.

Prediction of binaural speech intelligibility with frequency-dependent interaural phase differences

The aim of this study was to test the hypothesis of independent processing strategies in adjacent binaural frequency bands underlying current models for binaural speech intelligibility in complex configurations and to investigate the effective binaural auditory bandwidth in broad-band signals. Speech reception thresholds (SRTs) were measured for binaural conditions with frequency-dependent interaural phase differences (IPDs) of speech and noise. SRT predictions with the binaural speech intelligibility model by Beutelmann and Brand (2006, J. Acoust. Soc. Am. 120, 331-342) were compared with the observed data. The IPDs of speech and noise had a sinusoidal shape on a logarithmic frequency scale. The bandwidth between zeros of the IPD function was varied from 18 to 4 octaves. Speech and noise had either the same IPD function (reference condition) or opposite signs of the IPD function (binaural condition). Each condition had two subconditions with alternating and non-alternating signs, respectively, of the IPD function. The binaural unmasking with respect to the reference condition decreased from 6 dB to zero with decreasing IPD bandwidth for the alternating condition while it stayed significantly larger than zero for the non-alternating condition. The observed results were well predicted by the model with an analysis filter bandwidth of 2.3 equivalent rectangular bandwidths (ERBs).

A comparative study of seven human cochlear filter models

Auditory models have been developed for decades to simulate characteristics of the human auditory system, but it is often unknown how well auditory models compare to each other or perform in tasks they were not primarily designed for. This study systematically analyzes predictions of seven publicly-available cochlear filter models in response to a fixed set of stimuli to assess their capabilities of reproducing key aspects of human cochlear mechanics. The following features were assessed at frequencies of 0.5, 1, 2, 4, and 8 kHz: cochlear excitation patterns, nonlinear response growth, frequency selectivity, group delays, signal-in-noise processing, and amplitude modulation representation. For each task, the simulations were compared to available physiological data recorded in guinea pigs and gerbils as well as to human psychoacoustics data. The presented results provide application-oriented users with comprehensive information on the advantages, limitations and computation costs of these seven mainstream cochlear filter models.

Aging effects on the binaural interaction component of the auditory brainstem response in the Mongolian gerbil: Effects of interaural time and level differences

The effect of interaural time difference (ITD) and interaural level difference (ILD) on wave 4 of the binaural and summed monaural auditory brainstem responses (ABRs) as well as on the DN1 component of the binaural interaction component (BIC) of the ABR in young and old Mongolian gerbils (Meriones unguiculatus) was investigated. Measurements were made at a fixed sound pressure level (SPL) and a fixed level above visually detected ABR threshold to compensate for individual hearing threshold differences. In both stimulation modes (fixed SPL and fixed level above visually detected ABR threshold) an effect of ITD on the latency and the amplitude of wave 4 as well as of the BIC was observed. With increasing absolute ITD values BIC latencies were increased and amplitudes were decreased. ILD had a much smaller effect on these measures. Old animals showed a reduced amplitude of the DN1 component. This difference was due to a smaller wave 4 in the summed monaural ABRs of old animals compared to young animals whereas wave 4 in the binaural-evoked ABR showed no age-related difference. In old animals the small amplitude of the DN1 component was correlated with small binaural-evoked wave 1 and wave 3 amplitudes. This suggests that the reduced peripheral input affects central binaural processing which is reflected in the BIC.

Amplitude and phase equalization of stimuli for click evoked auditory brainstem responses

Although auditory brainstem responses (ABRs), the sound-evoked brain activity in response to transient sounds, are routinely measured in humans and animals there are often differences in ABR waveform morphology across studies. One possible reason may be the method of stimulus calibration. To explore this hypothesis, click-evoked ABRs were measured from seven ears in four Mongolian gerbils (Meriones unguiculatus) using three common spectrum calibration strategies: Minimum phase filter, linear phase filter, and no filter. The results show significantly higher ABR amplitude and signal-to-noise ratio, and better waveform resolution with the minimum phase filtered click than with the other strategies.

Effect of Harmonicity on the Detection of a Signal in a Complex Masker and on Spatial Release from Masking

The amount of masking of sounds from one source (signals) by sounds from a competing source (maskers) heavily depends on the sound characteristics of the masker and the signal and on their relative spatial location. Numerous studies investigated the ability to detect a signal in a speech or a noise masker or the effect of spatial separation of signal and masker on the amount of masking, but there is a lack of studies investigating the combined effects of many cues on the masking as is typical for natural listening situations. The current study using free-field listening systematically evaluates the combined effects of harmonicity and inharmonicity cues in multi-tone maskers and cues resulting from spatial separation of target signal and masker on the detection of a pure tone in a multi-tone or a noise masker. A linear binaural processing model was implemented to predict the masked thresholds in order to estimate whether the observed thresholds can be accounted for by energetic masking in the auditory periphery or whether other effects are involved. Thresholds were determined for combinations of two target frequencies (1 and 8 kHz), two spatial configurations (masker and target either co-located or spatially separated by 90 degrees azimuth), and five different masker types (four complex multi-tone stimuli, one noise masker). A spatial separation of target and masker resulted in a release from masking for all masker types. The amount of masking significantly depended on the masker type and frequency range. The various harmonic and inharmonic relations between target and masker or between components of the masker resulted in a complex pattern of increased or decreased masked thresholds in comparison to the predicted energetic masking. The results indicate that harmonicity cues affect the detectability of a tonal target in a complex masker.

Exploring binaural hearing in gerbils (Meriones unguiculatus) using virtual headphones

The Mongolian gerbil (Meriones unguiculatus) has become a key species in investigations of the neural processing of sound localization cues in mammals. While its sound localization has been tested extensively under free-field stimulation, many neurophysiological studies use headphones to present signals with binaural localization cues. The gerbils behavioral sensitivity to binaural cues, however, is unknown for the lack of appropriate stimulation paradigms in awake behaving gerbils. We close this gap in knowledge by mimicking a headphone stimulation; we use free-field loudspeakers and apply cross-talk cancellation techniques to present pure tones with binaural cues via “virtual headphones” to gerbils trained in a sound localization task. All gerbils were able to lateralize sounds depending on the interaural time or level difference (ITD and ILD, respectively). For ITD stimuli, reliable responses were seen for frequencies ≤2.9 kHz, the highest frequency tested with ITD stimuli. ITD sensitivity was frequency-dependent with the highest sensitivity observed at 1 kHz. For stimuli with ITD outside the gerbils physiological range, responses were cyclic indicating the use of phase information when lateralizing narrow-band sounds. For ILD stimuli, reliable responses were obtained for frequencies ≥2 kHz. The comparison of ITD and ILD thresholds with ITD and ILD thresholds derived from gerbils’ free-field performance suggests that ongoing ITD information is the main cue for sound localization at frequencies <2 kHz. At 2 kHz, ITD and ILD cues are likely used in a complementary way. Verification of the use of the virtual headphones suggests that they can serve as a suitable substitute for conventional headphones particularly at frequencies ≤2 kHz.

Effect of preceding stimulation on sound localization and its representation in the auditory midbrain

Prior stimulation can influence the perception of sound source location. Some psychophysical sound localization procedures differ in the amount of prior stimulation, which may affect measures of localization accuracy. If and how particularly the number of preceding stimuli affects sound localization and the neural representation of sound source position has not been investigated so far and will be the focus of the present report. We trained Mongolian gerbils in a left/right discrimination task where the target stimulus was preceded by silence or followed a number of reference stimuli. Localization thresholds decreased with the number of references presented before the target stimulus. The smallest thresholds were found after the presentation of a train of 5 reference stimuli and after silence. We recorded from units in the inferior colliculus (IC) of anaesthetized gerbils using virtual‐acoustic space stimuli mimicking the ones used in the behavioural task and applied signal detection theory to compare behavioural and neurometric thresholds. We found that neurometric thresholds based on spike rate information of single units covered a wide range of threshold values but only neurometric thresholds that were based on responses of small populations of IC units reached consistently thresholds we also observed in the behavioural experiment. Unlike behavioural thresholds, however, neurometric thresholds were independent of the number of reference stimuli suggesting that processing stages downstream from the IC might better reflect the effect of prior stimulation.