Loudness Compensation Research Articles

A Design Method for Gammachirp Filterbank for Loudness Compensation in Hearing Aids

Because the hearing impaired often experience different degrees of hearing loss along with the loss of frequencies, the loudness compensation algorithm in hearing aids decomposes the speech signal and compensates with different frequency bands based on their audiograms. However, the speech quality of the compensated signal is unsatisfactory because the traditional filterbanks fail to fully consider the characteristics of human hearing and personalized hearing loss. In this study, an effective design for the gammachirp filterbank for the loudness compensation algorithm was proposed to improve the speech quality of hearing aids. Firstly, a multichannel gammachirp filterbank was employed to decompose the signals. Then, the adjacent bands were merged into one channel, guided by the proposed combination method. After obtaining the personalized filterbank, each band conducted a loudness compensation to match the requirements of the audiograms. The excellent advantage of the gammachirp filterbank is that it can simulate the characteristics of the basilar membrane. Furthermore, the novel channel combination method considers the information from the audiograms and the relationship between frequency ranges and speech intelligibility. The experimental results showed that the proposed multichannel gammachirp filterbank achieves better speech signal decomposition and synthesis, and good performance can be gained with fewer channels. The loudness compensation algorithm based on the gammachirp filterbank effectively improves sentence intelligibility. The sentence recognition rate of the proposed method is higher than that of a system with a gammatone filterbank by approximately 13%.

RETRACTED: Multi-channel adaptive loudness compensation algorithm based on noise tracking in digital hearing aids

RETRACTED: Multi-channel adaptive loudness compensation algorithm based on noise tracking in digital hearing aids

Restoring Perceived Loudness for Listeners With Hearing Loss.

Normalizing perceived loudness is an important rationale for gain adjustments in hearing aids. It has been demonstrated that gains required for restoring normal loudness perception for monaural narrowband signals can lead to higher-than-normal loudness in listeners with hearing loss, particularly for binaural broadband presentation. The present study presents a binaural bandwidth-adaptive dynamic compressor (BBDC) that can apply different gains for narrow- and broadband signals. It was hypothesized that normal perceived loudness for a broad variety of signals could be restored for listeners with mild to moderate high-frequency hearing loss by applying individual signal-dependent gain corrections. Gains to normalize perceived loudness for narrowband stimuli were assessed in 15 listeners with mild to moderate high-frequency hearing loss using categorical loudness scaling. Gains for narrowband loudness compensation were calculated and applied in a standard compressor. Aided loudness functions for signals with different bandwidths were assessed. The deviation from the average normal-hearing loudness functions was used for gain correction in the BBDC. Aided loudness functions for narrow- and broadband signals with BBDC were then assessed. Gains for a 65 dB SPL speech-shaped noise of BBDC were compared with gains based on National Acoustic Laboratories' nonlinear fitting procedure version 2 (NAL-NL2). The perceived loudness for 20 real signals was compared to the average normal-hearing rating. The suggested BBDC showed close-to-normal loudness functions for binaural narrow- and broadband signals for the listeners with hearing loss. Normal loudness ratings were observed for the real-world test signals. The proposed gain reduction method resulted on average in similar gains as prescribed by NAL-NL2. However, substantial gain variations compared to NAL-NL2 were observed in the data for individual listeners. Gain corrections after narrowband loudness compensation showed large interindividual differences for binaural broadband signals. Some listeners required no further gain reduction for broadband signals; for others, gains in decibels were more than halved for binaural broadband signals. The interindividual differences of the binaural broadband gain corrections indicate that relevant information for normalizing perceived loudness of binaural broadband signals cannot be inferred from monaural narrowband loudness functions. Over-amplification can be avoided if binaural broadband measurements are included in the fitting procedure. For listeners with a high binaural broadband gain correction factor, loudness compensation for narrowband and broadband stimuli cannot be achieved by compression algorithms that disregard the bandwidth of the input signals. The suggested BBDC includes individual binaural broadband corrections in a more appropriate way than threshold-based procedures.

Multichannel loudness compensation method based on segmented sound pressure level for digital hearing aids

To improve the performance of gain compensation based on three-segment sound pressure level (SPL) in hearing aids, an improved multichannel loudness compensation method based on eight-segment SPL was proposed. Firstly, the uniform cosine modulated filter bank was designed. Then, the adjacent channels which have low or gradual slopes were adaptively merged to obtain the corresponding non-uniform cosine modulated filter according to the audiogram of hearing impaired persons. Secondly, the input speech was decomposed into sub-band signals and the SPL of every sub-band signal was computed. Meanwhile, the audible SPL range from 0 dB SPL to 120 dB SPL was equally divided into eight segments. Based on these segments, a different prescription formula was designed to compute more detailed gain to compensate according to the audiogram and the computed SPL. Finally, the enhanced signal was synthesized. Objective experiments showed the decomposed signals after cosine modulated filter bank have little distortion. Objective experiments showed that the hearing aids speech perception index (HASPI) and hearing aids speech quality index (HASQI) increased 0.083 and 0.082 on average, respectively. Subjective experiments showed the proposed algorithm can effectively improve the speech recognition of six hearing impaired persons.

A Comprehensive Method to Improve Loudness Compensation and High-Frequency Speech Intelligibility for Digital Hearing Aids

A Comprehensive Method to Improve Loudness Compensation and High-Frequency Speech Intelligibility for Digital Hearing Aids

Spectral and binaural loudness summation for hearing-impaired listeners

Sensorineural hearing loss typically results in a steepened loudness function and a reduced dynamic range from elevated thresholds to uncomfortably loud levels for narrowband and broadband signals. Restoring narrowband loudness perception for hearing-impaired (HI) listeners can lead to overly loud perception of broadband signals and it is unclear how binaural presentation affects loudness perception in this case. Here, loudness perception quantified by categorical loudness scaling for nine normal-hearing (NH) and ten HI listeners was compared for signals with different bandwidth and different spectral shape in monaural and in binaural conditions. For the HI listeners, frequency- and level-dependent amplification was used to match the narrowband monaural loudness functions of the NH listeners. The average loudness functions for NH and HI listeners showed good agreement for monaural broadband signals. However, HI listeners showed substantially greater loudness for binaural broadband signals than NH listeners: on average a 14.1 dB lower level was required to reach “very loud” (range 30.8 to −3.7 dB). Overall, with narrowband loudness compensation, a given binaural loudness for broadband signals above “medium loud” was reached at systematically lower levels for HI than for NH listeners. Such increased binaural loudness summation was not found for loudness categories below “medium loud” or for narrowband signals. Large individual variations in the increased loudness summation were observed and could not be explained by the audiogram or the narrowband loudness functions.

A Method of Wide Dynamic Range Compression Based on Curve Compression Applying on Hearing Aids

Recently, there is a growing trend to investigate wide dynamic range compression technique with the development of hearing aids. However, existing methods used to enhance expected speech as well as the background noise and classical linear compression can’t achieve the requirements of hearing impaired individuals’ reaction to the loudness compensation flexibly. To cope with the problems, this paper proposes a new approach to suppress the enhanced background noise and to dismiss abnormal sound caused by severe jittering from high frequency signal. What’s more, a scheme of curve compression is introduced to improve the loudness amplifying flexibility instead of conventional linear compression. Both the theoretical simulation and testing on cellphone hearing aid APP based on android system prove the proposed method can improve the performance of speech enhancement and provide more choices to strength sound with the curve compression ratio.

Single- and multi-channel modulation detection in cochlear implant users.

Single-channel modulation detection thresholds (MDTs) have been shown to predict cochlear implant (CI) users’ speech performance. However, little is known about multi-channel modulation sensitivity. Two factors likely contribute to multichannel modulation sensitivity: multichannel loudness summation and the across-site variance in single-channel MDTs. In this study, single- and multi-channel MDTs were measured in 9 CI users at relatively low and high presentation levels and modulation frequencies. Single-channel MDTs were measured at widely spaced electrode locations, and these same channels were used for the multichannel stimuli. Multichannel MDTs were measured twice, with and without adjustment for multichannel loudness summation (i.e., at the same loudness as for the single-channel MDTs or louder). Results showed that the effect of presentation level and modulation frequency were similar for single- and multi-channel MDTs. Multichannel MDTs were significantly poorer than single-channel MDTs when the current levels of the multichannel stimuli were reduced to match the loudness of the single-channel stimuli. This suggests that, at equal loudness, single-channel measures may over-estimate CI users’ multichannel modulation sensitivity. At equal loudness, there was no significant correlation between the amount of multichannel loudness summation and the deficit in multichannel MDTs, relative to the average single-channel MDT. With no loudness compensation, multichannel MDTs were significantly better than the best single-channel MDT. The across-site variance in single-channel MDTs varied substantially across subjects. However, the across-site variance was not correlated with the multichannel advantage over the best single channel. This suggests that CI listeners combined envelope information across channels instead of attending to the best channel.

Learning-Based Auditory Encoding for Robust Speech Recognition

This paper describes an approach to the optimization of the nonlinear component of a physiologically motivated feature extraction system for automatic speech recognition. Most computational models of the peripheral auditory system include a sigmoidal nonlinear function that relates the log of signal intensity to output level, which we represent by a set of frequency dependent logistic functions. The parameters of these rate-level functions are estimated to maximize the a posteriori probability of the correct class in training data. The performance of this approach was verified by the results of a series of experiments conducted with the CMU S phinx-III speech recognition system on the DARPA Resource Management, Wall Street Journal databases, and on the AURORA 2 database. In general, it was shown that feature extraction that incorporates the learned rate-nonlinearity, combined with a complementary loudness compensation function, results in better recognition accuracy in the presence of background noise than traditional MFCC feature extraction without the optimized nonlinearity when the system is trained on clean speech and tested in noise. We also describe the use of lattice structure that constraints the training process, enabling training with much more complicated acoustic models.

The just audible tonality of short exponential and Gaussian pure tone bursts

This paper reports on listening tests performed to investigate the just audible tonality (JAT) of decaying pure tone bursts. Both exponential and Gaussian functions are used to shape the envelopes of the tone bursts and critical band center frequencies between 150 and 7000 Hz are studied. Loudness compensation is implemented to compensate for the reduced loudness of short tone bursts and attack functions are used for minimizing clicks. By using the method of limits, a sequence of tone bursts with increasing decay times and constant frequencies is presented to the listeners at 0.9 s intervals. The first burst in the sequence which is perceived as being tonal is indicated by the listeners. When that happens, the decay times of the tone bursts decrease and the listeners are asked to select the first tone burst with no audible tonality. The listeners are allowed to freely define tonality. No reference is given. For frequencies above 3.4 kHz, the results indicate that tonality is just audible for tone burst lengths of approximately 2.6-3.0 ms. For the lowest stimuli frequencies, the corresponding burst length is approximately 20-23 ms.

System and method for varying low audio frequencies inversely with audio signal level

The present invention relates to an auto loudness circuit for performing loudness compensation automatically depending on the signal level. When the signal level decreases, loudness compensation is slowly introduced and as the signal level increases, loudness compensation is quickly removed. To do so, the auto loudness circuit utilizes a filter circuit with the characteristic of a first order bass boost. The filter circuit maintains a corner frequency which is proportional to the inverse of audio level in order to mimic the Fletcher-Munson curves. Because the circuit employs a capacitance-multiplier with a first order resistance capacitance filter, the bass boost is inversely proportional to the signal level. Thus, bass boost is achieved automatically as the program content changes so that the listener is unaware of significant changes in program material as signal levels change either through increase or decrease in volume, crescendo or new material.

Volume-responsive loudness compensation circuits, systems, and methods

An audio compensation system (20) for producing a sound compensated output signal (So). The system comprises a volume control circuit (24) for producing a volume-adjusted signal (Sic) by applying a volume adjustment to an audio signal in response to a volume setting, wherein the sound compensated output signal is responsive to the volume-adjusted signal. The system also comprises circuitry (26, 30, 32) for producing an amplified signal (Sifa) by amplifying a selected bandwith of signals in response to the volume setting. The sound compensated output signal is also responsive to the amplified signal.

Wavelet-based nonlinear AGC method for hearing aid loudness compensation

Hearing impairment can often be corrected by medical or surgical treatment, provided the loss is not due to cochlear pathology (sensorineural loss). For sensorineural loss, the only corrective action is to wear a hearing aid. Many perceptual differences between normal and sensorineural hearing-impaired listeners are due to differences in the dB levels at which sound is detected by the ear (auditory threshold) and the associated dynamic range over which it is comfortable to listen to (loudness sensation). To adequately compensate for this, the processing of the auditory signal should be nonlinear and time-varying. Two wavelet-based compression algorithms have been developed: automatic gain control (AGC) with linear amplification and nonlinear compression AGC. The nonlinear AGC is a compression algorithm, which models loudness sensation. The wavelet transform separates the input into seven frequency bands corresponding to the critical bands of the human auditory system. For each frequency band, multiplying the wavelet coefficients by the gain can amplify or compress nonlinearly and smoothly, depending on the signal level, time and spectrum. Results suggest that the nonlinear approach, while maintaining the general spectral structure of the signal, is perceptually superior to linear AGC and compensates better for audibility loss.

Development of digital hearing AIDS.

This issue of Trends has attempted to introduce the reader to many of the historical and developmental issues surrounding the digitization of hearing aid devices. The enormous scope of digital processing issues and their interface to otopathologic listeners can only be marginally presented within the page limitations of this paper. Nevertheless, it should be clear that a great deal of effort and expense have preceded the introduction of the first sustainable true digital products which have only recently reached consumer distribution. Loudness compensation strategies have already evolved to immediate implementation within the current generation of digital products. Field studies in progress will presumably help revise and refine those of future iterations. New processing strategies undoubtedly will evolve in close connection to expanded diagnostic techniques and elaborated understanding of individual processing requirements. Hence, continued research in speech science and auditory pathology will be important components of expanded industry successes. Equally important will be further reductions in the power requirements of processing circuits complemented by simultaneous increases in computational capacity. Those manufacturers who have already brought digital devices to the hearing impaired community are, of course, to be congratulated. Meanwhile, those professionals involved in the rehabilitative application of new processing technology and hearing-impaired consumers who have long awaited more effective technologies welcome the continued development of this exciting trend in audio enginearing. Genuine progress at the level of user benefits will almost certainly require vigorous efforts in advanced diagnostics and evaluation techniques (Cox, 1993; Byrne, 1996). M.I.T. professor, Nicholas Negroponte (1995), states categorically, “Being digital is the license to grow.” It is hoped that the reader of this issue of Trends will sense that the economic growth of the hearing aid industry, the growth of the professions that interface with it, and the expansion of improved rehabilitative assistance to the hearing impaired are all inexorably linked to the digital transformation that is now well under way. Many heroic and creative efforts by scientists all over the world formed the developmental foundations of current and future successes. Considerably more research and development will undoubtedly continue to enlarge the benefits and usefulness of DSP hearing aids, but to some extent each new device will contain within them a subtle portion of the many earlier efforts and discarded attempts to grow hearing aid operations into the powerful versatility of “being digital.”

Wavelet analysis in recruitment of loudness compensation

Wavelet-based multiband dynamic range compression is developed to compensate for a common hearing impairment known as recruitment of loudness. The algorithm combines standard compression with intensity-level dependent gain calculation. Complexity and performance are similar to traditional techniques. Further, the methods established can be applied to the more adaptive wavelet packets and local cosine bases which model the speech signal more closely. >

Micro power A/D converter for audio signals

The increasing computing power provided by modern signal processors makes the application of digital signal processing algorithms more and more attractive in various fields, including hearing aids. This trend is evident in recent publications, e.g., [l-4], covering multichannel compression and loudness compensation techniques among others. Power consumption is a parameter of major concern in hearing aid equipment. This of course is not in favor of digital signal processing approaches due to the additional need for signal conversion from analog to digital and vice versa. Hearing aid equipment requires an audio signal with a bandwidth of about 6 kHz and a dynamic range of about 80 dB. The power consumption for an adequately specified analog-to-digital (A/D) converter including a suitable anti-aliasing filter should be considerably less than 1 mW. A/D converters with such a low power consumption and a 14-bit resolution at above 10 kHz sampling rate are clearly not standard components. Instrumentation-type A/D converters are understood to provide a resolution according to the number of specified bits; i.e., the upper bound for quantization noise is expected to be half the least significant bit. Furthermore, anti-aliasing filters in such applications are required to keep aliasing below that same bound. Fortunately, in audio applications the requirements may be loosened without loss of perceived signal quality. One aspect is resolution. Numerous masking experiments [5] indicate that quantization noise at a level of 25 dB below signal components does not cause any perceptual degradation. Consequently, an adaptive 6-bit quantizer [6] will do as shown in the front end section of Fig. 1, i.e., left of the codeword memory labeled M, to M5. Another aspect is aliasing. An attractive design approach includes oversampling and realizing antialiasing by a digital filter. In the case of audio signals a sampling rate of 32 kHz and subsequent low-pass filtering and downsampling to 16 kHz will do. The resulting fold-over frequency at 8 kHz allows rather loose constraints on the decimation filter. There are two reasons for this. Most sounds reveal a spectrum that falls off at a considerable rate above 8 kHz. Exceptions to this mainly comprise fricative sounds. However, moderate aliasing is perceptually not relevant in this case. Consequently, a simple nonrecursive digital filter with favorable implementation properties is successfully used in the proposed design. It shows 6 dB attenuation at 8 kHz fold-over frequency, 30 dB at 11 kHz, and side lobes below 40 dB. The perceptual requirements are very well met in this way. This digital lowpass filter is indicated in Fig. 1 on the right side of the codeword memory. The remaining parts of this paper are organized as follows: details of front end processing are described in Section 2. Section 3 describes how the codeword memory is organized. In particular, this section shows how 14-bit-wide codewords are generated by combining the information supplied by quantizer and gain control. The decimation filter is described in detail in Section 4 and conclusions are summarized in Section 5.

Constraints on nonlinear amplification for recruitment of loudness compensation

A nonlinear amplification technique has been developed to compensate for recruitment of loudness in sensorineural hearing losses [J. C. Rutledge, “Time‐Varying, Frequency‐Dependent Compensation for Recruitment of Loudness,” Ph.D. thesis, Georgia Inst. of Technol., 1989]. This technique uses a sinusoidal speech model to incorporate a model of the psychoacoustic masking of sinusoids in normal hearing and in hearing‐impaired persons. It operates on both a time‐varying and frequency‐dependent basis. The drawback is that the model processing may distort the vowel formant structure and thus cause confusions between vowels. Therefore constraints on the relative amplitude levels of the sinusoidal components at different frequency regions should be incorporated into the model. These constraints have been studied and will be discussed.

A digital hearing aid that compensates loudness for impaired listeners and its performance

A method of designing a digital hearing aid for speech signal suitable for sensorineural hearing loss with a narrow audible range is proposed. First, the input signal is divided into short time blocks (8 ms) and its spectrum is analyzed using short‐time FFT. Then the frequency‐gain characteristic of the digital filter for the block is determined using loudness compensation functions that describe the relation between the loudness of normal listeners and that of impaired listeners, so that the input signal within the block is projected on to the dynamic range of the subject. This algorithm is realized using a digital signal processor (DSP). The results of monosyllabic intelligibility tests to evaluate the performance of the system for six sensorineural impaired listeners are also shown. [Work supported by Grant‐in‐Aid for Scientific Research from the Ministry of Education and the Ministry of Public Welfare.]

Speech Enhancement by Filtering in the Loudness Domain.

Several attempts that employed dynamic compression to compensate for cochlear impairment in combination with recruitment have failed to improve speech intelligibility. One reason is the attenuation of dynamic intensity modulations in the range of 1 to 50 Hz which contribute significantly to speech discrimination. In the approach presented here, these modulations were enhanced rather than attenuated by an envelope highpass filter operating in the loudness domain as follows: the output of a filterbank was lowpass filtered to extract the envelope, converted to loudness in accordance with normal hearing, highpass-filtered and converted back to the 'desired' envelope by the loudness characteristic of each individual hearing-impaired subject. The output signal was finally obtained by summing the outputs of the filterbank channels after correcting their amplitudes. The algorithm was implemented by a computer to simulate a digital hearing aid fitted to the loudness-scaling function of each individual. Loudness scaling and speech-discrimination data in quiet and in noisy conditions are presented for 2 impaired subjects with and without highpass filtering in the loudness domain. The results indicate that loudness compensation without highpass filtering improves speech intelligibility for a wide range of input levels by achieving an optimum gain characteristic at all input levels. The highpass filter applied to the envelope has no noticeable effect on speech discrimination either in quiet or in noise.