Effects of noise and reverberation on speech recognition with variants of a multichannel adaptive dynamic range compression scheme

The Noise Reduction Algorithm May Not Compensate for the Degradation in Output Signal-to-Noise Ratio Caused by Wide Dynamic Range Compression

PURPOSE

Most modern hearing aids (HAs) employ wide dynamic range compression (WDRC) and noise reduction (NR) algorithms. It is known that the nonlinear effects of WDRC and NR cause changes to the output signal-to-noise ratio (SNR) of an HA. However, the relative contributions of WDRC and NR to the nonlinear effects are not fully understood. The current study investigated (a) whether WDRC or NR dominates the nonlinear effects measured at the output of a digital HA and (b) whether the electroacoustic effectiveness of NR depends on WDRC parameters while input SNR and background noise are systematically varied.

METHOD

Test stimuli were Connected Speech Test sentences in multitalker babble noise (2- or 20-talker), presented at input SNRs ranging from -10 to +10 dB. The HA was programmed using multiband WDRC set according to the National Acoustic Laboratories for Nonlinear HA fitting formula 2 prescriptive fits for four standard audiograms and two compression speeds. The NR algorithm of the HA was switched on or off in separate conditions. Nonlinear electroacoustic effects from the WDRC and NR algorithms were assessed by measuring the output SNR of the HA using a phase-inversion technique. To investigate whether there are other factors that may be important besides the output SNR, the Hearing Aid Speech Intelligibility Index and the Hearing Aid Speech Quality Index were applied to the recordings to generate inferences on aided speech intelligibility and perceived speech quality.

RESULTS

Results showed that WDRC dominated the net nonlinear effect at low-input SNRs, and the net nonlinear effect of WDRC and NR was reduced at high-input SNRs. Results also showed that the effectiveness of NR depended on compression parameters. The effectiveness of NR was partially explained by the trend of Hearing Aid Speech Intelligibility Index and Hearing Aid Speech Quality Index scores, potentially indicating that the Hearing Aid Speech Intelligibility Index and Hearing Aid Speech Quality Index scores may capture factors that cannot be captured by the output SNR metric.

CONCLUSIONS

Results suggest that the individual signal-processing stages in an HA should not be considered as independent. Electroacoustic evaluation of WDRC and NR algorithms in isolation is not sufficient to capture the combined nonlinear effect of the two algorithms.

SUPPLEMENTAL MATERIAL

https://doi.org/10.23641/asha.25962541.

The Effects of Signal to Noise Ratio, T60 , Wide-Dynamic Range Compression Speed, and Digital Noise Reduction in a Virtual Restaurant Setting

OBJECTIVES

Hearing aid processing in realistic listening environments is difficult to study effectively. Often the environment is unpredictable or unknown, such as in wearable aid trials with subjective report by the wearer. Some laboratory experiments create listening environments to exert tight experimental control, but those environments are often limited by physical space, a small number of sound sources, or room absorptive properties. Simulation techniques bridge this gap by providing greater experimental control over listening environments, effectively bringing aspects of the real-world into the laboratory. This project used simulation to study the effects of wide-dynamic range compression (WDRC) and digital noise reduction (DNR) on speech intelligibility in a reverberant environment with six spatialized competing talkers. The primary objective of this study was to determine the efficacy of WDRC and DNR in a complex listening environment using virtual auditory space techniques.

DESIGN

Participants of greatest interest were listeners with hearing impairment. A group of listeners with clinically normal hearing was included to assess the effects of the simulation absent the complex effects of hearing loss. Virtual auditory space techniques were used to simulate a small restaurant listening environment with two different reverberation times (0.8 and 1.8 sec) in a range of signal to noise ratios (SNRs) (-8.5 to 11.5 dB SNR). Six spatialized competing talkers were included to further enhance realism. A hearing aid simulation was used to examine the degree to which speech intelligibility was affected by slow and fast WDRC in conjunction with the presence or absence of DNR. The WDRC and DNR settings were chosen to be reasonable estimates of hearing aids currently available to consumers.

RESULTS

A WDRC × DNR × Hearing Status interaction was observed, such that DNR was beneficial for speech intelligibility when combined with fast WDRC speeds, but DNR was detrimental to speech intelligibility when WDRC speeds were slow. The pattern of the WDRC × DNR interaction was observed for both listener groups. Significant main effects of reverberation time and SNR were observed, indicating better performance with lower reverberation times and more positive SNR.

CONCLUSIONS

DNR reduced low-amplitude noise before WDRC-amplified the low-intensity portions of the signal, negating one potential downside of fast WDRC and leading to an improvement in speech intelligibility in this simulation. These data suggest that, in some real-world environments that include both reverberation and noise, older listeners with hearing impairment may find speech to be more intelligible if DNR is activated when the hearing aid has fast compression time constants. Additional research is needed to determine the appropriate DNR strength and to confirm results in wearable hearing aids and a wider range of listening environments.

Verification of Estimated Output Signal-to-Noise Ratios From a Phase Inversion Technique Using a Simulated Hearing Aid

PURPOSE

The signal-to-noise ratio (SNR) for speech presented in background noise may vary after being processed by digital hearing aids with nonlinear signal processing algorithms, such as wide dynamic range compression (WDRC). A phase inversion technique has been previously developed to assess the output SNR of hearing aids. However, systematic validations of this technique have not been conducted. This study aims to validate the phase inversion technique.

METHOD

A simulated hearing aid with multichannel WDRC was implemented, from which the output SNRs, computed via shadow filtering, for connected speech in background noise were directly computed. The agreement between the shadow filter output SNRs and those estimated using the phase inversion technique for the same stimuli was utilized to validate the phase inversion technique. The background noise was 2- or 20-talker babble noise, and the speech stimuli were presented at SNRs of -10 to +10 dB at the input of the simulated hearing aid. The simulated hearing aid was configured to provide amplification for four representative audiograms, and the WDRC was set to be fast or slow acting. To investigate the effects of additive noise, independent of the presented noise stimulus, on the phase inversion estimated output SNR, the same simulated hearing aid was implemented with an additive Gaussian noise at its input (45 and 60 dB SPL).

RESULTS

Results showed that the phase inversion technique could either overestimate or underestimate output SNR, depending on the test condition; the estimation errors tended to coincide with temporal landmarks, such as natural pauses between consecutive sentences or fricatives; and increasing the simulated noise led to poorer estimates of output SNR.

CONCLUSIONS

Results imply that the accuracy of the phase inversion technique is dependent on the test conditions. Thus, the phase inversion technique should be used with caution, and its validity should be evaluated further.

A Data-driven Distance Metric for Evaluating the effects of Dynamic Range Compression in Adverse Conditions

Dynamic range compression is a crucial component in hearing aids, aiming to restore audibility for hearing-impaired listeners. However, determining suitable compression parameters, such as the time constants for the level estimation stage, remains a topic of debate, as the perceptual benefit of different parameter configurations varies depending on the acoustic conditions. In this study, a data-driven distance metric based on physical metrics was developed to evaluate and compare the performance of various compression systems. This analysis encompassed fast-acting and slow-acting compression, as well as a "scene-aware" compression that dynamically adjusted the release time constant based on the presence of the target. A reference system called "source-independent compression" was also considered, which had access to individual speech and noise signals. Multiple physical metrics were employed to assess the effects of these different compression systems under diverse acoustic conditions, including varying levels of interfering noise and degrees of room reverberation. Factor analysis was applied to derive a concise set of interpretable features representing the impact of compression, expressed as linear combinations of carefully selected objective metrics. The reduced dimensional representation enabled the use of the Manhattan distance to measure the similarity between the compression systems. Results demonstrated that the scene-aware compressor generally exhibited a smaller distance to the reference system compared to both the fast-acting and slow-acting compressors. This finding suggests that an adaptive compression system holds potential benefits across a range of acoustic conditions.

The Effect of Multichannel and Channel-Free Hearing Aids on Spectral-Temporal Resolution and Speech Understanding in Noise

BACKGROUND

Identifying and understanding speech is difficult for individuals with sensorineural hearing loss, especially in noisy environments. Possible causes include less audibility of the signal, impaired temporal resolution, and low selectivity of frequency. The hearing aid is the most common option used to minimize the problems faced by individuals with sensorineural hearing loss.

PURPOSE

This article investigates the effects of multichannel and channel-free hearing aid signal processing techniques on spectral-temporal resolution and speech understanding in noise.

RESEARCH DESIGN

An experimental study was used in which the determined tests were applied to the participants.

STUDY SAMPLE

Thirty-four individuals with bilateral symmetrical sensorineural hearing loss between the ages of 18 and 70 were included in our study.

DATA COLLECTION AND ANALYSIS

Spectral-temporally modulated ripple test, random gap detection test (RGDT), and Turkish matrix test were applied to the participants using multichannel and channel-free hearing aids. All the data obtained were compared statistically in terms of the performances of the hearing aids.

RESULTS

There was no significant difference between multichannel and channel-free hearing aids for spectral resolution and speech understanding in noise tests (p > 0.05). While there was no significant difference between the two hearing aids for 500 and 4,000 Hz RGDT in temporal resolution measurement (p > 0.05), for 1,000 Hz (p = 0.045), 2,000 Hz (p = 0.046), and composite RGDT (p = 0.001), statistically significant better performances were obtained with the channel-free hearing aids.

CONCLUSION

It is thought that faster processing of the incoming signal in the channel-free hearing aids improves the temporal resolution performance. It is predicted that our study findings might help to determine the signal processing technique that will maximize the communication skills of the patients in various conditions.

Influence of Audibility and Distortion on Recognition of Reverberant Speech for Children and Adults with Hearing Aid Amplification

BACKGROUND

Adults and children with sensorineural hearing loss (SNHL) have trouble understanding speech in rooms with reverberation when using hearing aid amplification. While the use of amplitude compression signal processing in hearing aids may contribute to this difficulty, there is conflicting evidence on the effects of amplitude compression settings on speech recognition. Less clear is the effect of a fast release time for adults and children with SNHL when using compression ratios derived from a prescriptive procedure.

PURPOSE

The aim of the study is to determine whether release time impacts speech recognition in reverberation for children and adults with SNHL and to determine if these effects of release time and reverberation can be predicted using indices of audibility or temporal and spectral distortion.

RESEARCH DESIGN

This is a quasi-experimental cohort study. Participants used a hearing aid simulator set to the Desired Sensation Level algorithm m[i/o] for three different amplitude compression release times. Reverberation was simulated using three different reverberation times.

PARTICIPANTS

Participants were 20 children and 16 adults with SNHL.

DATA COLLECTION AND ANALYSES

Participants were seated in a sound-attenuating booth and then nonsense syllable recognition was measured. Predictions of speech recognition were made using indices of audibility, temporal distortion, and spectral distortion and the effects of release time and reverberation were analyzed using linear mixed models.

RESULTS

While nonsense syllable recognition decreased in reverberation release time did not significantly affect nonsense syllable recognition. Participants with lower audibility were more susceptible to the negative effect of reverberation on nonsense syllable recognition.

CONCLUSION

We have extended previous work on the effects of reverberation on aided speech recognition to children with SNHL. Variations in release time did not impact the understanding of speech. An index of audibility best predicted nonsense syllable recognition in reverberation and, clinically, these results suggest that patients with less audibility are more susceptible to nonsense syllable recognition in reverberation.

Modeling the effects of dynamic range compression on signals in noise

Hearing aids use dynamic range compression (DRC), a form of automatic gain control, to make quiet sounds louder and loud sounds quieter. Compression can improve listening comfort, but it can also cause unwanted distortion in noisy environments. It has been widely reported that DRC performs poorly in noise, but there has been little mathematical analysis of these noise-induced distortion effects. This work introduces a mathematical model to study the behavior of DRC in noise. By making simplifying assumptions about the signal envelopes, we define an effective compression function that models the compression applied to one signal in the presence of another. Using the properties of concave functions, we prove results about DRC that have been previously observed experimentally: that the effective compression applied to each sound in a mixture is weaker than it would have been for the signal alone; that uncorrelated signal envelopes become negatively correlated when compressed as a mixture; and that compression can reduce the long-term signal-to-noise ratio in certain conditions. These theoretical results are supported by software experiments using recorded speech signals.

Compression and amplification algorithms in hearing aids impair the selectivity of neural responses to speech

In quiet environments, hearing aids improve the perception of low-intensity sounds. However, for high-intensity sounds in background noise, the aids often fail to provide a benefit to the wearer. Here, by using large-scale single-neuron recordings from hearing-impaired gerbils — an established animal model of human hearing — we show that hearing aids restore the sensitivity of neural responses to speech, but not their selectivity. Rather than reflecting a deficit in supra-threshold auditory processing, the low selectivity is a consequence of hearing-aid compression (which decreases the spectral and temporal contrasts of incoming sound) and of amplification (which distorts neural responses, regardless of whether hearing is impaired). Processing strategies that avoid the trade-off between neural sensitivity and selectivity should improve the performance of hearing aids.

The Type of Noise Influences Quality Ratings for Noisy Speech in Hearing Aid Users

Purpose The overall goal of the current study was to determine whether noise type plays a role in perceptual quality ratings. We compared quality ratings using various noise types and signal-to-noise ratio (SNR) ranges using hearing aid simulations to consider the effects of hearing aid processing features. Method Ten older adults with bilateral mild to moderately severe sensorineural hearing loss rated the sound quality of sentences processed through a hearing aid simulation and presented in the presence of five different noise types (six-talker babble, three-talker conversation, street traffic, kitchen, and fast-food restaurant) at four SNRs (3, 8, 12, and 20 dB). Results Everyday noise types differentially affected sound quality ratings even when presented at the same SNR: Kitchen and three-talker noises were rated significantly higher than restaurant, traffic, and multitalker babble, which were not different from each other. The effects of noise type were most pronounced at poorer SNRs. Conclusions The findings of this study showed that noise types differentially affected sound quality ratings. The differences we observed were consistent with the acoustic characteristics of the noise types. Noise types having lower envelope fluctuations yielded lower quality ratings than noise types characterized by sporadic high-intensity events at the same SNR.