Search

This paper proposes a polar coding mechanism for the underwater acoustic (UWA) channel, which improves the performance in the bit error rate (BER) compared with the original code construction algorithms and decoding methods invented by Arikan. It studied the performance of polar coding in the UWA communication system with the Orthogonal Frequency Division Multiplexing (OFDM) technique in theory and simulation. The simulation results show that the BER of the polar code with the code rate 1/2 in the underwater time-varying channel can reach 10−4∼10−5 when the signal-to-noise ratio (SNR) is at 4 dB, which is about 0.5∼1 dB better than LDPC and Turbo codes. Moreover, the BER performance of polar codes under different UWA channel models and parameters, code lengths, and code rates were also compared. The results demonstrate that the proposed mechanism matches the UWA channel well, thus making polar codes suitable and reliable for the UWA communication systems.

Underwater Acoustic Target Recognition (UATR) remains one of the most challenging tasks in underwater signal processing due to the lack of labeled data acquisition, the impact of the time-space varying intrinsic characteristics, and the interference from other noise sources. Although some deep learning methods have been proven to achieve state-of-the-art accuracy, the accuracy of the recognition task can be improved by designing a Residual Network and optimizing feature extraction. To give a more comprehensive representation of the underwater acoustic signal, we first propose the three-dimensional fusion features along with the data augment strategy of SpecAugment. Afterward, an 18-layer Residual Network (ResNet18), which contains the center loss function with the embedding layer, is designed to train the aggregated features with an adaptable learning rate. The recognition experiments are conducted on the ship-radiated noise dataset from a real environment, and the accuracy results of 94.3% indicate that the proposed method is appropriate for underwater acoustic recognition problems and sufficiently surpasses other classification methods.

© 2015 Taylor & Francis Conclusion: This study described objective and subjective evaluations of the Nurotron® Venus™ Cochlear Implant System and indicated that this system produced a satisfactory performance. Objective: To observe the performance of the Nurotron® Venus™ cochlear implant (CI) system via electrophysiological and psychophysical evaluations. Methods: A 26-electrode CI system was specially designed. The performance of MRI in animal and cadaveric head experiments, EABR in cats experiment, the correlation between ESRT and C level, and psychophysics evaluations in clinical trials were observed. Results: In the animal and cadaveric head experiments, magnet dislocation could not be prevented in the 1.5 T MRI without removal of the internal magnet. The EABR was clearly elicited in cat experiment. In the clinical trial, the ESRT was strongly correlated with C level (p

In order to reduce the computational burden of the algorithm for two-sided correlation transformation, this paper proposes a fast algorithm based on the orthogonal propagator. Compared with the two-sided correlation transformation algorithm, the new method uses the propagation operator instead of the eigenvalue decomposition to estimate the feature vector of the noise subspace. There is no need to build the focusing matrix, therefore the computation complexity is significantly reduced for wide-band processing. Furthermore, theoretical analysis shows that there is no loss in the estimation performance since it implicitly concentrates the energy of each frequency bin in the process of solving the coherent process. The simulation results verify the benefit of the new method under the white Gaussian noise assumption.

Objectives: The main goal of this study was to investigate the effects of acoustic characteristics, including timbre and fundamental frequency (F0), on the musical pitch discrimination of cochlear implant users. Methods: Eight postlingually deafened cochlear implant users were recruited, along with 8 control subjects with normal hearing. Pitch discrimination tests were carried out using test stimuli from 4 musical instruments plus synthetic complex stimuli. Three reference tones with different F0s were used. Results: The mean difference limens were 1.8 to 10.7 semitones in the just-noticeable difference task and 2.1 to 13.6 semitones in the pitch-direction discrimination task for different timbre and F0 combinations. Three-way analysis of variance showed that the acoustic characteristics of the musical stimuli, such as timbre and F0, significantly influenced pitch discrimination performance. Conclusions: Acoustic characteristics determine the complexity of the electrical stimulation pattern, which directly affects performance in pitch discrimination. A place pattern with a clear and regular low-order harmonic structure is most important for good pitch discrimination. A clear F0-related temporal pattern is also useful when the F0 is low. Pitch perception performance will worsen when there is interference in the high-frequency channels.

© 2014 The Authors. Although the cochlear implant has been widely acknowledged as the most successful neural prosthesis, only a fraction of hearing-impaired people who can potentially benefit from a cochlear implant have actually received one due to its limited awareness, accessibility, and affordability. To help overcome these limitations, a 26-electrode cochlear implant has been developed to receive China's Food and Drug Administration (CFDA) approval in 2011 and Conformité Européenne (CE) Marking in 2012. The present article describes design philosophy, system specification, and technical verification of the Nurotron device, which includes advanced digital signal processing and 4 current sources with multiple amplitude resolutions that not only are compatible with perceptual capability but also allow interleaved or simultaneous stimulation. The article also presents 3-year longitudinal evaluation data from 60 human subjects who have received the Nurotron device. The objective measures show that electrode impedance decreased within the first month of device use, but was stable until a slight increase at the end of two years. The subjective loudness measures show that electric stimulation threshold was stable while the maximal comfort level increased over the 3 years. Mandarin sentence recognition increased from the pre-surgical 0%-correct score to a plateau of about 80% correct with 6-month use of the device. Both indirect and direct comparisons indicate indistinguishable performance differences between the Nurotron system and other commercially available devices. The present 26-electrode cochlear implant has already helped to lower the price of cochlear implantation in China and will likely contribute to increased cochlear implant access and success in the rest of the world.This article is part of a Special Issue entitled .

Envelope is the vital part of one-dimensional data. The estimation of envelope can be treated as a demodulation problem. However, the definition of envelope is ambiguous and lack of an exact mathematical definition. It is commonly agreed that envelope varies slowly and in some empirical view it should pass the prominent peaks of the data smoothly. In this study, we propose an algorithm to directly use the prominent peaks and finally get the envelope by interpolation. We term it as empirical envelope estimation algorithm (EEEA). Envelope derived by EEEA has steerable smoothness and is adaptive to the data. It has clear physical meaning and has great potential in some off-line signal analysis applications, such as musical sound analysis and heart sound analysis.

This paper discusses a single-channel speech enhancement method for cochlear implant listeners. It is assumed that the Fourier Transform coefficients of speech and background noise have different statistical distributions. A statistical-model-based method is adopted to update the signal-to-noise ratio and estimate the background noise so that the musical noise and speech distortion induced by traditional spectral subtraction method can be effectively reduced. This enhancement method was evaluated on seven postlingually deaf Chinese cochlear implant listeners in comparison with other two speech enhancement methods. Test materials were Mandarin sentences corrupted by three different types of background noise. Experimental results showed that the proposed speech enhancement method could benefit the speech intelligibility of Chinese cochlear implant listeners. The results suggest that different noise types may affect the performance of different speech enhancement algorithms.

The present study investigated the effect of voice gender on Mandarin tone perception by cochlear implant users who speak Chinese. Six cochlear implant users were recruited in this study with their own devices. Psychoacoustic experiment on Mandarin tone perception was carried out with the use of 160 monosyllabic words uttered by a male speaker and a female speaker. Clean speech materials and their noisy counterpart at 10 dB signal-to-noise ratio were tested. Experimental result showed that cochlear implant users could achieve relatively good Mandarin tone perception of 72% correct at clean condition. However, the tone perception performance significantly degraded in noise environment. No significant differences were found between male voice and female voice. Electrodogram analysis indicated that tone-related information was severely masked by noise on the electrode output of cochlear implant.

Principal Component Analysis (PCA) is a common method for feature selection. In order to enhance the effect of selection, a Principal Component Analysis based on Mutual Information (PCAMI) algorithm is proposed. PCAMI introduces the category information, and uses the sum of mutual information matrix between features under different acoustic environments instead of covariance matrix. The eigenvectors of the matrix represent the transformation coefficients. The eigenvalues of the matrix are used to calculate the cumulative contribution rate to determine the number of dimension. The experiment on acoustic environment classification shows that PCAMI has better dimensionality reduction results and higher classification accuracy using neuron network than PCA.

In auditory science, Hilbert transform (HT) was used for decomposition of temporal envelope (E) and fine structure (TFS) [1]. Sometimes the envelope is called ‘Hilbert envelope’. Although HT has rigorous mathematical definition, it has little intrinsic physical relationship with E or TFS. Aims to investigate the rationality of that method for E or TFS extraction, this study introduced Empirical Mode Decomposition (EMD). Experiment results showed that HT was usually rational for envelope extraction of narrow-band signals, but not for that of non-narrow-band signals. However, please keep in mind that this conclusion is not a definite proof but a experimental implication under the theory of EMD.

This paper describes a periodicity-enhancement strategy which can be used to improve lexical tone perception for cochlear implant (CI) users. We focus on Cantonese, a major Chinese dialect with complex tone system. Acoustical simulations were tested with normal-hearing subjects. Temporal envelope and periodicity components (TEPC) below 500 Hz were extracted from four frequency bands. Stimuli were the combination of TEPC-modulated noises from sub-bands. For enhancement, the original temporal periodicity cues between 20 and 500 Hz were simplified by a sinusoidal wave whose frequency was equal to the exact fundamental frequency (F0) of the original speech or a linearly approximated F0. Tone identification tests were conducted using Cantonese disyllabic words. Results showed that tone scores with periodicity-enhanced TEPCs were significantly higher than the original ones. The approximated F0 contour provides similar tone perception performance to the exact F0 contour, indicating that exact F0 is unnecessary for the periodicity-enhancement strategy in CIs.

The aim of this study was to determine how melodic contour identification with cochlear implant was affected by the frequency range, timbre and intonation. Noise-excited vocoder was used to simulate the signal processing for cochlear implant system. Melodic contour identification tests were carried out with eight normal-hearing subjects. Stimuli from four instruments (piano, trumpet, clarinet and violin) were utilized with different frequency ranges and intonations. Our experimental results showed that (i) temporal envelope information was important for melodic contour identification when the spectral information was limited; (ii) frequency range, timbre and intonation could influence melodic contour identification performance of the normal-hearing subjects when listening to the acoustic simulation of cochlear implant.

The present study investigated musical pitch discrimination with acoustic simulation of cochlear implant. A 4-channel noise-excited vocoder was used to simulate the signal processing of cochlear implants. Eight normal-hearing subjects participated in this study. Psychoacoustic experiments on pitch-direction discrimination test were carried out to examine the ability of detecting pitch-change direction. Stimuli from four instruments (clarinet, trumpet, piano and violin) were utilized. Experimental results showed that temporal envelope and periodicity information was important for musical pitch discrimination when the spectral information was limited.

A new method for detecting and classifying loudspeaker faults is presented in this paper. Total response of high-order harmonics groups is measured and used as defect features of loudspeaker. Based on support vector machine (SVM), we built a classification system combined with one-class SVM and Directed Acyclic Graphic SVM (DAGSVM). Comparing with K-nearest neighbor (k-NN) classifier, the accuracy of the method is higher in the experiment.

As the most successful neural prosthesis, cochlear implants have provided partial hearing to more than 120,000 persons worldwide; half of which being pediatric users who are able to develop nearly normal language. Biomedical engineers have played a central role in the design, integration and evaluation of the cochlear implant system, but the overall success is a result of collaborative work with physiologists, psychologists, physicians, educators, and entrepreneurs. This review presents broad yet in-depth academic and industrial perspectives on the underlying research and ongoing development of cochlear implants. The introduction accounts for major events and advances in cochlear implants, including dynamic interplays among engineers, scientists, physicians, and policy makers. The review takes a system approach to address critical issues from design and specifications to integration and evaluation. First, the cochlear implant system design and specifications are laid out. Second, the design goals, principles, and methods of the subsystem components are identified from the external speech processor and radio frequency transmission link to the internal receiver, stimulator and electrode arrays. Third, system integration and functional evaluation are presented with respect to safety, reliability, and challenges facing the present and future cochlear implant designers and users. Finally, issues beyond cochlear implants are discussed to address treatment options for the entire spectrum of hearing impairment as well as to use the cochlear implant as a model to design and evaluate other similar neural prostheses such as vestibular and retinal implants.

Current experimental Cochlear Implant (CI) sound processing strategies like F0mod [Laneau, et.al., 2006], which aim to improve the pitch perception of CI subjects, amplitude-modulate the envelope of each frequency channel by the fundamental frequency (F0) of the speech sound. Due to the hardware limitation of CI devices, typical F0 estimation algorithms are too complex to be implemented. Low-complexity and accurate F0 estimator will be necessary for such manipulation. This study uses the Summary Auto-Correlation Function (SACF) to extract F0. The new pitch estimator was simulated and compared with typical algorithms that estimate F0 before the sub-band filtering. By the use of SACF, the new estimator reduces the computation load significantly and preserves good F0 estimation performance. # This work is supported by National Natural Science Foundation of China (11104316), and Shanghai Natural Science Foundation (11ZR1446000).

Pitch perception researches showed that temporal modulation plays an important role in pitch perception from unresolved harmonics. Fundamental frequency (F0) enhancement through temporal modulation of envelope was utilized in some cochlear implant (CI) signal processing strategies [e.g., Laneau, et.al., 2006] to improve the pitch perception of CI. In this study, Temporal Modulation Transfer Function (TMTF) and the relationship between Amplitude Modulation Rate (AMR) and pitch perception at different stimulation sites (apical, middle, and basal) were investigated. The stimuli are sinusoidal-amplitude–modulation pulse trains. Two psychophysical experiments were carried out: Two-Alternative-Forced Choice on modulated/un-modulated stimuli and AMR-Pitch Ranking. Experimental results showed that some of the CI subjects exhibited distinct performance at different stimulation sites, which implies that issues of stimulation sites and the subject's individual performance should be considered for F0 enhancement stra...

In sound classification problem, a feature selection method based on improved maximum mutual information is proposed. Better sound classification performance is expected with smaller computational effort by evaluating the “information content” of classification features, only selecting the relevant features of a classifying system, and excluding redundant ones. The proposed method is based on a “greedy” selection of the classification features. The mutual information is measured based on both the output class and the selected features. Simulation experiment was carried out to evaluate performance of sound classification with the proposed feature selection method. The selected features were feed into neural network for sound classification. The classification accuracy was compared with the mutual information feature selector (MIFS). Experimental results showed that the proposed method can produce better sound classification than the classical method. This work has been sponsored by Jiaxing Engineering Cent...

The Weighted Overlap and Add (WOLA) method is an optimization of frequency modulated DFT filter banks. However, the decrement in frequency resolution of speech signal caused by WOLA leads to noise estimation error. In this study, a speech power compression and gain compensation method was proposed to solve the low-resolution problem from the WOLA process. Finally, a multi-band spectral subtraction method was developed by combining the WOLA and the spectral subtraction algorithm in frequency domain. The noise reduction degrees were determined by the individual Signal-to-Noise Ratios (SNRs) in each frequency bands. Objective evaluation of the proposed noise reduction method was performed based on Itakura-Saito Distortion Measure under different types of noise. Statistical results showed that this proposed method could improve noisy speech by 5 dB SNR with little distortion. This noise reduction method has an advantage on low computation load, real-time processing and good performance on noise reduction. So this method can be implemented in embedded devices, e.g. hearing aid and cochlear implants. # This work is supported by National Natural Science Foundation of China (11104316). and Shanghai Natural Science Foundation (11ZR1446000).

Cochlea, known as `fourier analyzer' in auditory system of mammals, codes the mechanical wave to electrical signals tonotopically. Meanwhile, the auditory coder also works in temporal mode, where envelope (E, i.e. amplitude modulation (AM)) and temporal fine structure (TFS, i.e. frequency modulation (FM)) are both significant but not exactly defined. In several recent researches (e.g. [Smith, et. al, 2002]), hilbert transform (HT) was utilized to extract the E and TFS cues. Though HT is mathematically rigorous, it is lack of clear physical meaning on AM-FM. This research introduces the empirical mode decomposition (EMD) to auditory research field. EMD was designed for analysis of non-linear and non-stationary data including natural sound. It is observed that the outputs of a auditory filter (e.g. Gammachirp), which was born as narrow-band filters, are approximately intrinsic mode functions (IMF) that admit well-behaved hilbert transforms (implying having physical significance). However, non-narrow-band filter (NNBF) output can be decomposed into two obvious IMFs or more, implying that HT is not suited to NNBF. EMD with HT provides auditory researchers one more perspective on AM-FM decomposition of auditory signals.# This work is supported by National Natural Science Foundation of China (11104316) and Shanghai Natural Science Foundation (11ZR1446000).

Cochlear implants (CIs) were designed primarily to enhance the speech perception of western persons with bilateral sensorineural deafness. The CI technology used in the current commercial devices, in terms of both hardware and software, didn't focus on the delivery of tone or pitch related information. It's known that lexical tone is important to Mandarin-speaking people. To better encode pitch-related information, the current study will investigate the effect of temporal envelope periodicity to Mandarin tone perception of CI recipients. Temporal periodicity from each frequency band is manipulated corresponding to the fundamental frequency (F0) of the speech signal. Psychophysical experiment was carried out to evaluate the Mandarin tone perception performance of CI recipients in quiet and noisy conditions. Test materials were monosyllabic and disyllabic Chinese words with four different tones. Experimental results showed that F0-related coding on temporal periodicity can improve the Mandarin tone percepti...

The cochlear implant (CI) stimulation parameters such as pulse width, phase gap, pulse amplitude, and stimulation rate are the key and fundamental features in the objective evaluation of a CI device. In this study, an electronic evaluation system was developed to measure the above CI parameters for fully understanding the performance of a CI device. This system is multi-functional, powerful and flexible, which contains several modules, e.g. source signal generation, I/O synchronization, pulse sequence acquisition, data processing, and pulse/waveform display. A user-friendly graphic interface was developed including many useful functions and options to make different measurement and evaluation. The measurement results and original data can be easily exported from the evaluation system for further analysis. The accuracy and limitation of the system will also be discussed in this work. This system is suitable for CI parameter verification, speech processing research strategy evaluation, multi-electrode strea...