Fast-acting Compression Research Articles

Amplitude Compression for Preventing Rollover at Above-Conversational Speech Levels.

Hearing aids provide nonlinear amplification to improve speech audibility and loudness perception. While more audibility typically increases speech intelligibility at low levels, the same is not true for above-conversational levels, where decreases in intelligibility ("rollover") can occur. In a previous study, we found rollover in speech intelligibility measurements made in quiet for 35 out of 74 test ears with a hearing loss. Furthermore, we found rollover occurrence in quiet to be associated with poorer speech intelligibility in noise as measured with linear amplification. Here, we retested 16 participants with rollover with three amplitude-compression settings. Two were designed to prevent rollover by applying slow- or fast-acting compression with a 5:1 compression ratio around the "sweet spot," that is, the area in an individual performance-intensity function with high intelligibility and listening comfort. The third, reference setting used gains and compression ratios prescribed by the "National Acoustic Laboratories Non-Linear 1" rule. Speech intelligibility was assessed in quiet and in noise. Pairwise preference judgments were also collected. For speech levels of 70 dB SPL and above, slow-acting sweet-spot compression gave better intelligibility in quiet and noise than the reference setting. Additionally, the participants clearly preferred slow-acting sweet-spot compression over the other settings. At lower levels, the three settings gave comparable speech intelligibility, and the participants preferred the reference setting over both sweet-spot settings. Overall, these results suggest that, for listeners with rollover, slow-acting sweet-spot compression is beneficial at 70 dB SPL and above, while at lower levels clinically established gain targets are more suited.

Automated Koala call detection using a cochlea model

Estimating koala populations is challenging, making surveys costly and often imprecise. This study examines the feasibility of recognising koala calls from audio recordings using a bio-inspired acoustic approach. In the work, we propose to use the Cascade of Asymmetric Resonators with Fast-Acting Compression (CAR-FAC) cochlea model to pre-process koala recordings and a support vector machine (SVM) for classification. The frequency-dependent gain and bandwidth properties of the CAR-FAC model in pre-processing noisy data are discussed and investigated. To evaluate the effectiveness of the proposed method, it is compared with an FFT-based system as a baseline method. The baseline methodology uses a Mel-spectrogram pre-processing and the same linear classifier as the proposed method as the back-end. In summary, we achieved 96.7% accuracy for the proposed system.

Using Auditory Characteristics to Select Hearing Aid Compression Speeds for Presbycusic Patients

ObjectivesThis study aimed to select the optimal hearing aid compression speeds (fast-acting and slow-acting) for presbycusic patients by using auditory characteristics including temporal modulation and speech-in-noise performance.MethodsIn total, 24 patients with unilateral or bilateral moderate sensorineural hearing loss who scored higher than 21 on the Montreal Cognitive Assessment (MoCA) test participated in this study. The electrocochleogram (ECochG) results, including summating potentials (SP) and action potentials (AP), were recorded. Subjects' temporal modulation thresholds and speech recognition at 4 individualized signal-to-noise ratios were measured under three conditions, namely, unaided, aided with fast-acting compression (FAC), and aided with slow-acting compression (SAC).ResultsThe results of this study showed that modulation discrimination thresholds in the unaided (−8.14 dB) and aided SAC (−8.19 dB) conditions were better than the modulation thresholds in the FAC (−4.67 dB) conditions. The speech recognition threshold (SRT75%) for FAC (5.21 dB) did not differ significantly from SAC (3.39 dB) (p = 0.12). A decision tree analysis showed that the inclusion of the AP, unaided modulation thresholds, and unaided SRT75% may correctly identify the optimal compression speeds (FAC vs. SAC) for individual presbycusic patients with up to 90% accuracy.ConclusionBoth modes of compression speeds improved a presbycusic patient's speech recognition ability in noise. The SAC hearing aids may better preserve the modulation thresholds than the FAC hearing aids. The measurement of AP, along with the unaided modulation thresholds and unaided SRT75%, may help guide the selection of optimal compression speeds for individual presbycusic patients.

Noise-robust text-dependent speaker identification using cochlear models.

One challenging issue in speaker identification (SID) is to achieve noise-robust performance. Humans can accurately identify speakers, even in noisy environments. We can leverage our knowledge of the function and anatomy of the human auditory pathway to design SID systems that achieve better noise-robust performance than conventional approaches. We propose a text-dependent SID system based on a real-time cochlear model called cascade of asymmetric resonators with fast-acting compression (CARFAC). We investigate the SID performance of CARFAC on signals corrupted by noise of various types and levels. We compare its performance with conventional auditory feature generators including mel-frequency cepstrum coefficients, frequency domain linear predictions, as well as another biologically inspired model called the auditory nerve model. We show that CARFAC outperforms other approaches when signals are corrupted by noise. Our results are consistent across datasets, types and levels of noise, different speaking speeds, and back-end classifiers. We show that the noise-robust SID performance of CARFAC is largely due to its nonlinear processing of auditory input signals. Presumably, the human auditory system achieves noise-robust performance via inherent nonlinearities as well.

Use of Emotional and Neutral Speech in Evaluating Compression Speeds.

Emotional speech differs from neutral speech in its envelope characteristics. Use of emotional speech materials may be more sensitive for evaluating signal processing algorithms that affect the temporal envelope. Subjective listener preference was compared between variable speed compression (VSC) and fast acting compression (FAC) amplitude compression algorithms using neutral and emotional speech. The study used a single-blinded, repeated measures design. Twenty hearing-impaired (HI) listeners with a bilaterally symmetrical, mild- to-moderately severe sensorineural hearing loss and 21 listeners with normal hearing (NH) participated. Speech was processed using FAC and VSC algorithms. A paired-comparison paradigm assessed subjective preference for FAC versus VSC using emotional and neutral speech materials. The significance of subjective preference for compression algorithm (FAC or VSC) was evaluated using a linear mixed effects model at each combination of stimulus type (emotional or neutral speech) and hearing group (NH or HI). HI listeners showed a preference for VSC over FAC when listening to emotional speech. The same listeners showed a nonsignificant, preference for VSC over FAC when listening to neutral speech. NH listeners showed preference for VSC over FAC for both neutral and emotional speech materials. These results suggest that the subjective sound quality of emotional speech is more susceptible than neutral speech to changes in the signal introduced by FAC. Clinicians should consider including emotional speech materials when evaluating listener preference for different compression speeds in the clinic.

A Biologically Inspired Sound Localisation System Using a Silicon Cochlea Pair

We present a biologically inspired sound localisation system for reverberant environments using the Cascade of Asymmetric Resonators with Fast-Acting Compression (CAR-FAC) cochlear model. The system exploits a CAR-FAC pair to pre-process binaural signals that travel through the inherent delay line of the cascade structures, as each filter acts as a delay unit. Following the filtering, each cochlear channel is cross-correlated with all the channels of the other cochlea using a quantised instantaneous correlation function to form a 2-D instantaneous correlation matrix (correlogram). The correlogram contains both interaural time difference and spectral information. The generated correlograms are analysed using a regression neural network for localisation. We investigate the effect of the CAR-FAC nonlinearity on the system performance by comparing it with a CAR only version. To verify that the CAR/CAR-FAC and the quantised instantaneous correlation provide a suitable basis with which to perform sound localisation tasks, a linear regression, an extreme learning machine, and a convolutional neural network are trained to learn the azimuthal angle of the sound source from the correlogram. The system is evaluated using speech data recorded in a reverberant environment. We compare the performance of the linear CAR and nonlinear CAR-FAC models with current sound localisation systems as well as with human performance.

Can Dual Compression Offer Better Mandarin Speech Intelligibility and Sound Quality Than Fast-Acting Compression?

The aim of this study was to evaluate the efficacy of dual compression for Mandarin-speaking hearing aid users. Dual compression combines fast and slow compressors operating simultaneously across all frequency channels. The study participants were 31 hearing aid users with symmetrical moderate-to-severe hearing loss, with a mean age of 67 years. A new pair of 20-channel behind-the-ear hearing aids (i.e., Phonak Bolero B90-P) was used during the testing. The results revealed a significant improvement in speech reception thresholds in noise when switching from fast-acting compression to dual compression. The sound quality ratings revealed that most listeners preferred dual compression to fast-acting compression for listening effort, listening comfort, speech clarity, and overall sound quality at +4 dB signal-to-noise ratio. These results are consistent with predictions based on the theoretical understanding of dual and fast-acting compression. However, whether these results can be generalized to other languages or other dual compression systems should be verified by future studies.

Perceived Sound Quality Dimensions Influencing Frequency-Gain Shaping Preferences for Hearing Aid-Amplified Speech and Music.

Hearing aids are typically fitted using speech-based prescriptive formulae to make speech more intelligible. Individual preferences may vary from these prescriptions and may also vary with signal type. It is important to consider what motivates listener preferences and how those preferences can inform hearing aid processing so that assistive listening devices can best be tailored for hearing aid users. Therefore, this study explored preferred frequency-gain shaping relative to prescribed gain for speech and music samples. Preferred gain was determined for 22 listeners with mild sloping to moderately severe hearing loss relative to individually prescribed amplification while listening to samples of male speech, female speech, pop music, and classical music across low-, mid-, and high-frequency bands. Samples were amplified using a fast-acting compression hearing aid simulator. Preferences were determined using an adaptive paired comparison procedure. Listeners then rated speech and music samples processed using prescribed and preferred shaping across different sound quality descriptors. On average, low-frequency gain was significantly increased relative to the prescription for all stimuli and most substantially for pop and classical music. High-frequency gain was decreased significantly for pop music and male speech. Gain adjustments, particularly in the mid- and high-frequency bands, varied considerably between listeners. Music preferences were driven by changes in perceived fullness and sharpness, whereas speech preferences were driven by changes in perceived intelligibility and loudness. The results generally support the use of prescribed amplification to optimize speech intelligibility and alternative amplification for music listening for most listeners.

Effect of the number of amplitude-compression channels and compression speed on speech recognition by listeners with mild to moderate sensorineural hearing loss.

The use of a large number of amplitude-compression channels in hearing aids has potential advantages, such as the ability to compensate for variations in loudness recruitment across frequency and provide appropriate frequency-response shaping. However, sound quality and speech intelligibility could be adversely affected due to reduction of spectro-temporal contrast and distortion, especially when fast-acting compression is used. This study assessed the effect of the number of channels and compression speed on speech recognition when the multichannel processing was used solely to implement amplitude compression, and not for frequency-response shaping. Computer-simulated hearing aids were used. The frequency-dependent insertion gains for speech with a level of 65 dB sound pressure level were applied using a single filter before the signal was filtered into compression channels. Fast-acting (attack, 10 ms; release, 100 ms) or slow-acting (attack, 50 ms; release, 3000 ms) compression using 3, 6, 12, and 22 channels was applied subsequently. Using a sentence recognition task with speech in two- and eight-talker babble at three different signal-to-babble ratios (SBRs), 20 adults with sensorineural hearing loss were tested. The number of channels and compression speed had no significant effect on speech recognition, regardless of babble type or SBR.

Perceptual Evaluation of Signal-to-Noise-Ratio-Aware Dynamic Range Compression in Hearing Aids.

Dynamic range compression is a compensation strategy commonly used in modern hearing aids. Fast-acting systems respond relatively quickly to the fluctuations in the input level. This allows for more effective compression of the dynamic range of speech and hence enhanced the audibility of its low-intensity components. However, such processing also amplifies the background noise, distorts the modulation spectra of both the speech and the background, and can reduce the output signal-to-noise ratio (SNR). Recently, May et al. proposed a novel SNR-aware compression strategy, in which the compression speed is adapted depending on whether speech is present or absent. Fast-acting compression is applied to speech-dominated time–frequency (T-F) units, while noise-dominated T-F units are processed using slow-acting compression. It has been shown that this strategy provides a similar effective compression of the speech dynamic range as conventional fast-acting compression, while introducing fewer distortions of the modulation spectrum of the background and providing an improved output SNR. In this study, this SNR-aware compression strategy was compared with conventional fast- and slow-acting compression in terms of speech intelligibility and subjective preference in a group of 17 hearing-impaired listeners with varying degree of hearing loss. The results show a speech intelligibility benefit of the SNR-aware compression strategy over the conventional slow-acting system. Furthermore, the SNR-aware approach demonstrates an increased subjective preference compared with both conventional fast- and slow-acting systems.

Effects of Hearing Loss and Fast-Acting Compression on Amplitude Modulation Perception and Speech Intelligibility.

The purpose was to investigate the effects of hearing-loss and fast-acting compression on speech intelligibility and two measures of temporal modulation sensitivity. Twelve adults with normal hearing (NH) and 16 adults with mild to moderately severe sensorineural hearing loss were tested. Amplitude modulation detection and modulation-depth discrimination (MDD) thresholds with sinusoidal carriers of 1 or 5 kHz and modulators in the range from 8 to 256 Hz were used as measures of temporal modulation sensitivity. Speech intelligibility was assessed by obtaining speech reception thresholds in stationary and fluctuating background noise. All thresholds were obtained with and without compression (using a fixed compression ratio of 2:1). For modulation detection, the thresholds were similar or lower for the group with hearing loss than for the group with NH. In contrast, the MDD thresholds were higher for the group with hearing loss than for the group with NH. Fast-acting compression increased the modulation detection thresholds, while no effect of compression on the MDD thresholds was observed. The speech reception thresholds obtained in stationary noise were slightly increased in the compression condition relative to the linear processing condition, whereas no difference in the speech reception thresholds obtained in fluctuating noise was observed. For the group with NH, individual differences in the MDD thresholds could account for 72% of the variability in the speech reception thresholds obtained in stationary noise, whereas the correlation was insignificant for the hearing-loss group. Fast-acting compression can restore modulation detection thresholds for listeners with hearing loss to the values observed for listeners with NH. Despite this normalization of the modulation detection thresholds, compression does not seem to provide a benefit for speech intelligibility. Furthermore, fast-acting compression may not be able to restore MDD thresholds to the values observed for listeners with NH, suggesting that the two measures of amplitude modulation sensitivity represent different aspects of temporal processing. For listeners with NH, the ability to discriminate modulation depth was highly correlated with speech intelligibility in stationary noise.

Selected Cognitive Factors Associated with Individual Variability in Clinical Measures of Speech Recognition in Noise Amplified by Fast-Acting Compression Among Hearing Aid Users.

Objective:Previous work examining speech recognition in more challenging listening environments has revealed a large variability in both persons with normal and hearing impairments. Although this is clinically very important, up to now, no consensus has been reached about which factors may provide better explanation for the existing individual variability in speech recognition ability among hearing aid users, when speech signal is degraded. This study aimed to examine hearing-sensitivity skills and cognitive ability differences between listeners with good and poor speech recognition abilities.Materials and Methods:A total of 195 experienced hearing aid users (33–80 years) were grouped by higher or lower speech recognition ability based on their performance on the Hagerman sentences task in multi-talker babble using fast-acting compression algorithm. They completed a battery of cognitive abilities tests, hearing-in-noise and the auditory thresholds test.Results:The results showed that the two groups did differ significantly overall on cognitive abilities tests like working memory, cognitive processing speed and attentional shifting, but not on the attentional inhibitory test and non-verbal intelligence test.Conclusions:Listeners with poor compared to those with better speech recognition abilities exhibit poorer cognitive abilities, which place them in a disadvantaged position, and /or more susceptible to signal modifications (as a result of fast-acting compression signal processing), resulting in limited benefits from hearing aids strategies. The findings may have implications for hearing aid signal processing strategies selection in rehabilitations.

Quadratic distortion in a nonlinear cascade model of the human cochlea.

The cascade of asymmetric resonators with fast-acting compression (CARFAC) is a cascade filterbank model that performed well in a comparative study of cochlear models, but exhibited two anomalies in its frequency response and excitation pattern. It is shown here that the underlying reason is CARFAC's inclusion of quadratic distortion, which generates DC and low-frequency components that in a real cochlea would be canceled by reflections at the helicotrema, but since cascade filterbanks lack the reflection mechanism, these low-frequency components cause the observed anomalies. The simulations demonstrate that the anomalies disappear when the model's quadratic distortion parameter is zeroed, while other successful features of the model remain intact.

A FPGA Implementation of the CAR-FAC Cochlear Model

This paper presents a digital implementation of the Cascade of Asymmetric Resonators with Fast-Acting Compression (CAR-FAC) cochlear model. The CAR part simulates the basilar membrane's (BM) response to sound. The FAC part models the outer hair cell (OHC), the inner hair cell (IHC), and the medial olivocochlear efferent system functions. The FAC feeds back to the CAR by moving the poles and zeros of the CAR resonators automatically. We have implemented a 70-section, 44.1 kHz sampling rate CAR-FAC system on an Altera Cyclone V Field Programmable Gate Array (FPGA) with 18% ALM utilization by using time-multiplexing and pipeline parallelizing techniques and present measurement results here. The fully digital reconfigurable CAR-FAC system is stable, scalable, easy to use, and provides an excellent input stage to more complex machine hearing tasks such as sound localization, sound segregation, speech recognition, and so on.

Effects of Slow- and Fast-Acting Compression on Hearing-Impaired Listeners’ Consonant–Vowel Identification in Interrupted Noise

There is conflicting evidence about the relative benefit of slow- and fast-acting compression for speech intelligibility. It has been hypothesized that fast-acting compression improves audibility at low signal-to-noise ratios (SNRs) but may distort the speech envelope at higher SNRs. The present study investigated the effects of compression with a nearly instantaneous attack time but either fast (10 ms) or slow (500 ms) release times on consonant identification in hearing-impaired listeners. Consonant–vowel speech tokens were presented at a range of presentation levels in two conditions: in the presence of interrupted noise and in quiet (with the compressor “shadow-controlled” by the corresponding mixture of speech and noise). These conditions were chosen to disentangle the effects of consonant audibility and noise-induced forward masking on speech intelligibility. A small but systematic intelligibility benefit of fast-acting compression was found in both the quiet and the noisy conditions for the lower speech levels. No detrimental effects of fast-acting compression were observed when the speech level exceeded the level of the noise. These findings suggest that fast-acting compression provides an audibility benefit in fluctuating interferers when compared with slow-acting compression while not substantially affecting the perception of consonants at higher SNRs.

Signal-to-Noise-Ratio-Aware Dynamic Range Compression in Hearing Aids.

Fast-acting dynamic range compression is a level-dependent amplification scheme which aims to restore audibility for hearing-impaired listeners. However, when being applied to noisy speech at positive signal-to-noise ratios (SNRs), the gain function typically changes rapidly over time as it is driven by the short-term fluctuations of the speech signal. This leads to an amplification of the noise components in the speech gaps, which reduces the output SNR and distorts the acoustic properties of the background noise. An adaptive compression scheme is proposed here which utilizes information about the SNR in different frequency channels to adaptively change the characteristics of the compressor. Specifically, fast-acting compression is applied to speech-dominated time-frequency (T-F) units where the SNR is high, while slow-acting compression is used to effectively linearize the processing for noise-dominated T-F units where the SNR is low. A systematic evaluation of this SNR-aware compression scheme showed that the effective compression of speech components embedded in noise was similar to that of a conventional fast-acting system, whereas natural fluctuations in the background noise were preserved in a similar way as when a slow-acting compressor was applied.

Cognitive Processing Speed, Working Memory, and the Intelligibility of Hearing Aid-Processed Speech in Persons with Hearing Impairment.

Previous studies have demonstrated that successful listening with advanced signal processing in digital hearing aids is associated with individual cognitive capacity, particularly working memory capacity (WMC). This study aimed to examine the relationship between cognitive abilities (cognitive processing speed and WMC) and individual listeners’ responses to digital signal processing settings in adverse listening conditions. A total of 194 native Swedish speakers (83 women and 111 men), aged 33–80 years (mean = 60.75 years, SD = 8.89), with bilateral, symmetrical mild to moderate sensorineural hearing loss who had completed a lexical decision speed test (measuring cognitive processing speed) and semantic word-pair span test (SWPST, capturing WMC) participated in this study. The Hagerman test (capturing speech recognition in noise) was conducted using an experimental hearing aid with three digital signal processing settings: (1) linear amplification without noise reduction (NoP), (2) linear amplification with noise reduction (NR), and (3) non-linear amplification without NR (“fast-acting compression”). The results showed that cognitive processing speed was a better predictor of speech intelligibility in noise, regardless of the types of signal processing algorithms used. That is, there was a stronger association between cognitive processing speed and NR outcomes and fast-acting compression outcomes (in steady state noise). We observed a weaker relationship between working memory and NR, but WMC did not relate to fast-acting compression. WMC was a relatively weaker predictor of speech intelligibility in noise. These findings might have been different if the participants had been provided with training and or allowed to acclimatize to binary masking noise reduction or fast-acting compression.

Preserving spatial perception in rooms using direct-sound driven dynamic range compression.

Fast-acting hearing-aid compression systems typically distort the auditory cues involved in the spatial perception of sounds in rooms by enhancing low-level reverberant energy portions of the sound relative to the direct sound. The present study investigated the benefit of a direct-sound driven compression system that adaptively selects appropriate time constants to preserve the listener's spatial impression. Specifically, fast-acting compression was maintained for time-frequency units dominated by the direct sound while the processing of the compressor was linearized for time-frequency units dominated by reverberation. This compression scheme was evaluated with normal-hearing listeners who indicated their perceived location and distribution of sound images in the horizontal plane for virtualized speech. The experimental results confirmed that both independent compression at each ear and linked compression across ears resulted in broader, sometimes internalized, sound images as well as image splits. In contrast, the linked direct-sound driven compression system provided the listeners with a spatial perception similar to that obtained with linear processing that served as the reference condition. The independent direct-sound driven compressor created a sense of movement of the sound between the two ears, suggesting that preserving the interaural level differences via linked compression is advantageous with the proposed direct-sound driven compression scheme.

Preserving spatial perception in reverberant environments using direct-sound driven dynamic range compression

Fast-acting hearing-aid compression typically distort the auditory cues involved in the spatial perception of sounds in rooms, due to the enhancement of low-level reverberant energy portions of the sound relative to the direct sound. The present study investigated the benefit of a novel direct-sound driven compression scheme that adaptively selects appropriate time constants to preserve the listener’s spatial impression. Specifically, fast-acting compression was maintained for time-frequency (T-F) units dominated by the direct sound while the compressor was linearized via longer time constants for T-F units dominated by reverberation. This novel compression scheme was evaluated with normal-hearing listeners who indicated their perceived location and distribution of virtualized speech in the horizontal plane. The results confirmed that both independent compression at each ear and linked compression across ears resulted in more diffuse and broader, sometimes internalized, sound images as well as image split...

Effects of Modified Hearing Aid Fittings on Loudness and Tone Quality for Different Acoustic Scenes.

To compare loudness and tone-quality ratings for sounds processed via a simulated five-channel compression hearing aid fitted using NAL-NL2 or using a modification of the fitting designed to be appropriate for the type of listening situation: speech in quiet, speech in noise, music, and noise alone. Ratings of loudness and tone quality were obtained for stimuli presented via a loudspeaker in front of the participant. For normal-hearing participants, levels of 50, 65, and 80 dB SPL were used. For hearing-impaired participants, the stimuli were processed via a simulated hearing aid with five-channel fast-acting compression fitted using NAL-NL2 or using a modified fitting. Input levels to the simulated hearing aid were 50, 65, and 80 dB SPL. All participants listened with one ear plugged. For speech in quiet, the modified fitting was based on the CAM2B method. For speech in noise, the modified fitting used slightly (0 to 2 dB) decreased gains at low frequencies. For music, the modified fitting used increased gains (by 5 to 14 dB) at low frequencies. For noise alone, the modified fitting used decreased gains at all frequencies (by a mean of 1 dB at low frequencies increasing to 8 dB at high frequencies). For speech in quiet, ratings of loudness with the NAL-NL2 fitting were slightly lower than the mean ratings for normal-hearing participants for all levels, while ratings with CAM2B were close to normal for the two lower levels, and slightly greater than normal for the highest level. Ratings of tone quality were close to the optimum value ("just right") for both fittings, except that the CAM2B fitting was rated as very slightly boomy for the 80-dB SPL level. For speech in noise, the ratings of loudness were very close to the normal values and the ratings of tone quality were close to the optimal value for both fittings and for all levels. For music, the ratings of loudness were close to the normal values for NAL-NL2 and slightly above normal for the modified fitting. The tone quality was rated as very slightly tinny for NAL-NL2 and very slightly boomy for the modified fitting. For noise alone, the NAL-NL2 fitting was rated as slightly louder than normal for all levels, while the modified fitting was rated as close to normal. Tone quality was rated as slightly sharper for the NAL-NL2 fitting than for the modified fitting. Loudness and tone quality can sometimes be made slightly closer to "normal" by modifying gains for different listening situations. The modification for music required to achieve "normal" tone quality appears to be less than used in this study.