Figure 1 - uploaded by Shuenn-Tsong Young
Content may be subject to copyright.
Example loudness-growth curves for Aescu HRL-1. (a). Original loudness-growth curve of a hearing-impaired individual. (b). Loudness-growth curve of a hearing-impaired individual based on the Mandarin speechmap. doi: 10.1371/journal.pone.0080831.g001
Source publication
The purpose of this study was to design and to verify a new hearing-aid fitting strategy (Aescu HRL-1) based on the acoustic features of Mandarin. The subjective and objective outcomes were compared to those fitted with NAL-NL1 (National Acoustic Laboratory Non-Linear, version1) in Mandarin-speaking hearing-aid users.
Fifteen subjects with sensorin...
Similar publications
The current study used the self-fitting algorithm to allow listeners to self-adjust hearing-aid gain or compression parameters to select gain for speech understanding in a variety of quiet and noise conditions. Thirty listeners with mild to moderate sensorineural hearing loss adjusted gain parameters in quiet and in several types of noise. Outcomes...
Citations
... The amplification is calibrated according to prescribed methods. In the calibration of adult hearing aids, there are two commonly recommended prescribed methods for fitting: NAL-NL2 and DSL v.5 [1]. These methods use nonlinear gain fitting strategies based on either loudness normalization or loudness equalization principles. ...
... In combination with the volume slider and the sliders in certain frequency ranges, a gain of ±20 dB is possible. δ(k) = ra · (sp volume + sp bass · nv bass (k) + sp mid · nv mid (k) + sp treble · nv treble (k)) (1) with: ...
Smartphones are increasingly being used to enable patients to play an active role in managing their own health through applications, also called apps. The latest generation of sound processors for cochlear implants offer Bluetooth connectivity that makes it possible to connect smartphones or tablets and thus enable patients to modify their hearing sensation or measure system parameters. However, to achieve a high adoption rate and secure operation of these applications, it is necessary to design intuitive user interfaces (UI) for end users. The main goal of the current study was to evaluate the usability of two different UIs. A second goal was to compare the hearing outcomes based on the patient’s adjustments. The two different UIs were explored in a group of adult and older adult bimodal cochlear-implant users, with adjustments possible for both the cochlear implant and the contralateral hearing aid. One of the UIs comprised a classical equalizer and volume-dial approach, while the second UI followed a 2D-Surface concept, to manipulate the corresponding sound parameters. The participants changed their fitting parameters using both UIs in seven different sound scenes. The self-adjusted settings for the different scenarios were stored and recalled at a later stage for direct comparison. To enable an assessment of reliability and reproducibility, the self-adaptation was also repeated for two of the seven sound scenes. Within minutes, the participants became accustomed to the concept of both UIs and generated their own parameter settings. Both UIs resulted in settings that could be considered similar in terms of spontaneous acceptance and sound quality. Furthermore, both UIs showed high reliability in the test–retest procedure. The time required for adjustment was significantly shorter with the 2D-Surface UI. A closer look at the bimodal aspect shows that participants were able to compensate for differences in loudness and frequencies between the cochlear implant and the hearing aid. The blind comparison test showed that self-adjustment led to a higher acceptance of the sound perception in more than 80% of the cases.
... They attributed such a Korean-specific difference to less use rate of high-frequency phonemes in Korean vis-à-vis English. Similar results have also been reported for Turkish (Yüksel & Gündüz, 2018) and Mandarin (Lai et al., 2013). Thus, there is a reasonable body of evidence suggesting that LTASS for certain languages may be significantly different than the universal LTASS, which is currently used in prescriptive formulae for fitting hearing aids. ...
Purpose
In this work, we have determined the long-term average speech spectra (LTASS) and dynamic ranges (DR) of 17 Indian languages. This work is important because LTASS and DR are language-dependent functions used to fit hearing aids, calculate the Speech Intelligibility Index, and recognize speech automatically. Currently, LTASS and DR functions for English are used to fit hearing aids in India. Our work may help improve the performance of hearing aids in the Indian context.
Method
Speech samples from native talkers were used as stimuli in this study. Each speech sample was initially cleaned for extraneous sounds and excessively long pauses. Next, LTASS and DR functions for each language were calculated for different frequency bands. Similar analysis was also performed for English for reference purposes. Two-way analysis of variance was also conducted to understand the effects of important parameters on LTASS and DR. Finally, a one-sample t test was conducted to assess the significance of important statistical attributes of our data.
Results
We showed that LTASS and DR for Indian languages are 5–10 dB and 11 dB less than those for English. These differences may be due to lesser use rate of high-frequency dominant phonemes and preponderance of vowel-ending words in Indian languages. We also showed that LTASS and DR do not differ significantly across Indian languages. Hence, we propose a common LTASS and DR for Indian languages.
Conclusions
We showed that differences in LTASS and DR for Indian languages vis-à-vis English are large and significant. Such differences may be attributed to phonetic and linguistic characteristics of Indian languages.
... A study by Lai et al. [19] , in which a new hearing-aid fitting strategy (Aescu HRL-1) based on the acoustic features of Mandarin was compared with NAL-NL1, showed the same tendencies as our results. Namely, the performance based on Aescu HRL-1 was as good as that of NAL-NL1 for Mandarinspeaking hearing aid users, and the participants responded that Aescu HRL-1 provides a more natural, better sound quality than does NAL-NL1. ...
Objective
Hearing aid amplification rationales have typically been developed by using global averages of the long-term average speech spectrum (LTASS) from Western European languages. However, there are few reports on hearing-aid amplification based on acoustic-phonetic characteristics of the Japanese language. This study’s objective is to investigate the LTASS for Japanese, and to compare a typical amplification rationale originally developed mainly for Western European languages with an amplification rationale specifically adjusted to the LTASS for Japanese.
Methods
LTASS for two speech materials provided by four Japanese talkers were analyzed using 1/3 octave bandwidth filters. The speech was recorded with different levels of vocal effort, yielding three LTASS for “soft”, “moderate” and “loud” speech. From these results, a gain offset of the hearing-aid amplification for Japanese was obtained as compared to ANSI S3.5. Speech intelligibility for an amplification rationale for Western European languages and the newly-developed Japanese version was obtained for presentation levels of 50 dB SPL, 65 dB SPL and 80 dB SPL. Nineteen people with mild to moderate hearing loss participated in the speech intelligibility experiment. Scores in% correct were arcsine-transformed and subjected to repeated measures ANOVA with pairwise comparisons of significant main effects using Bonferroni adjustments for multiple comparisons.
Results
The LTASS for Japanese was slightly different from the values of previous reports. A comparison of LTASS values to ANSI S3.5 with values for Japanese showed that the Japanese amplification rationale for “moderate” speech levels required more gain in the low-frequency area, and less gain in the high-frequency area. There was no significant difference in the speech intelligibility level between the amplification characteristics of Western European languages and Japanese language at each presentation level.
Conclusion
It was shown that for hearing-aid amplification for Japanese, adjustments based on LTASS differences for Western European Languages could be made. This preserved speech intelligibility at the same level as the original amplification rationale, suggesting that there was no need to consider differences in phonetics of Japanese to optimize speech understanding.
... Next, all the amplified speech passages are analyzed in 1/3 octave bands over several seconds to provide the long-term average speech spectrum (LTASS). Previous studies [17][18][19][20] indicated that the LTASS is a good predictor of the real speech output from compression-based HAs because it can represent the average energy of an HA during real speech conditions. In addition, the speech intelligibility index (SII) [21] can be obtained from the Speechmap method. ...
PurposeAlmost all smartphone-based personal sound amplification applications (PSAPs) do not provide amplification specifications as well as medical hearing aid (HA) products. Therefore, the main purpose of this study is to develop an HA-simulator and investigate the possible fitting range of this smartphone-based PSAP app equipped with a commercial headset, and to evaluate whether the aided outcome provides reasonable benefits for individuals with hearing loss by analyzing the Speechmap and speech intelligibility index (SII).Methods
An HA-simulator with a four-channel compression structure was developed and evaluated using a smartphone equipped with Apple EarPods. Measurement methods specified by the ANSI S3.22 were used to evaluate the electroacoustic performance of the proposed system. The Speechmap method and SII were used to check if the amplified speech for five different hearing loss configurations. These ANSI S3.22 and Speechmap measurement results were then compared with a digital HA under the same test conditions.ResultsThe results demonstrated that the proposed HA-simulator system provided an adequate output sound level for patients with mild-to-moderate hearing loss and a tolerable group delay for the user while providing an electroacoustic performance comparable to that of a digital HA based on ANSI S3.22 test criteria. The amplified speech signals presented by Speechmap and the corresponding SII values indicated that the proposed HA-simulator provided benefits comparable to those of medical HA.Conclusions
These results indicate that a smartphone-based HA simulator could provide suitable performance to that of a medical HA, and it potentially aid individuals with hearing loss.
... A standard sound quality question, ''I think that this method provides a high sound quality,'' was used to rate scores [46] based on mean opinion score (MOS) comparison [47]. The MOS rating is the most widely used measure for subjective 43292 VOLUME 7, 2019 quality tests, in which subjects rate the test speech on a scale from 5 (i.e., strongly agree) to 1 (i.e., strongly disagree). ...
Nonnegative matrix factorization (NMF) is a useful decomposition technique for multivariate data. More recently, NMF technology was used as a noise-estimation stage for a Wiener-filtering-based noise reduction (NR) method to improve the quality of noisy speech. Previous studies showed that this method provides better sound quality performance than conventional NMF-based approaches; however, there is still scope for improving the performance under noisy listening conditions. More specifically, the performance of an NMF noise estimator for calculating the noise level is considered sensitive to diverse noise environments and signal-to-noise ratio conditions. Therefore, we proposed an adaptive algorithm that derives an adaptive factor ($\alpha $ ) to adjust the weight between the estimated speech and noise levels on the basis of the signal-to-noise level for the gain function of the Wiener-filtering-based NR method to further improve the sound quality. Two objective evaluations and listening tests evaluated the benefits of the proposed method, and experimental results show that better output sound quality and competed for speech intelligibility performance can be achieved when compared with conventional unsupervised NR and NMF-based methods.
... Additionally, although the benefits of the NC + DDAE NR approach outweigh those of the conventional NR approaches, there is still room for improvement. Auxiliary features, such as fundamental frequency cues , pitch (Chen et al. 2010, band importance function (ANSI 1997), and incorporating the dynamic range of Mandarin speech (Lai et al. 2013) to achieve better intelligibility and sound quality for enhanced speech (Xu et al. 2015) could provide further benefits. More recently, researchers have started to work on multi-tiered acoustic noises using the deep learning-based models, such as Williamson and Wang (2017). ...
Objective:
We investigate the clinical effectiveness of a novel deep learning-based noise reduction (NR) approach under noisy conditions with challenging noise types at low signal to noise ratio (SNR) levels for Mandarin-speaking cochlear implant (CI) recipients.
Design:
The deep learning-based NR approach used in this study consists of two modules: noise classifier (NC) and deep denoising autoencoder (DDAE), thus termed (NC + DDAE). In a series of comprehensive experiments, we conduct qualitative and quantitative analyses on the NC module and the overall NC + DDAE approach. Moreover, we evaluate the speech recognition performance of the NC + DDAE NR and classical single-microphone NR approaches for Mandarin-speaking CI recipients under different noisy conditions. The testing set contains Mandarin sentences corrupted by two types of maskers, two-talker babble noise, and a construction jackhammer noise, at 0 and 5 dB SNR levels. Two conventional NR techniques and the proposed deep learning-based approach are used to process the noisy utterances. We qualitatively compare the NR approaches by the amplitude envelope and spectrogram plots of the processed utterances. Quantitative objective measures include (1) normalized covariance measure to test the intelligibility of the utterances processed by each of the NR approaches; and (2) speech recognition tests conducted by nine Mandarin-speaking CI recipients. These nine CI recipients use their own clinical speech processors during testing.
Results:
The experimental results of objective evaluation and listening test indicate that under challenging listening conditions, the proposed NC + DDAE NR approach yields higher intelligibility scores than the two compared classical NR techniques, under both matched and mismatched training-testing conditions.
Conclusions:
When compared to the two well-known conventional NR techniques under challenging listening condition, the proposed NC + DDAE NR approach has superior noise suppression capabilities and gives less distortion for the key speech envelope information, thus, improving speech recognition more effectively for Mandarin CI recipients. The results suggest that the proposed deep learning-based NR approach can potentially be integrated into existing CI signal processors to overcome the degradation of speech perception caused by noise.
... Participants were asked to fill in the IOI-HA and SSQ-HA questionnaires after each 4-week interval, with either the NLFC or EB-NLFC. 17 ...
Background/purpose:
The high frequency information of consonant messages is important for recognition of speech. Recently, the nonlinear frequency compression (NLFC) technique has been shown to improve the speech perception in patients with high frequency hearing loss. In Mandarin, seven consonants are located over 10-16 kHz. Extended-bandwidth (EB) NLFC may provide an additional benefit for recognition of Mandarin words. The purpose of this study was to explore the effects of NLFC and EB-NLFC on Mandarin word recognition in patients with high frequency hearing loss.
Methods:
Fourteen native Mandarin-speaking adult patients, aged 20-65 years with bilateral, moderate to severe, sensorineural hearing loss, specifically high frequency hearing loss were included in single-blind randomized study. The assessment tools included the Mandarin Monosyllable Recognition Test (MMRT), Mandarin Hearing in Noise Test (MHINT), and International Outcome Inventory for Hearing Aids (IOI-HA) and sound quality scale of the hearing aids. The patients were tested under unaided condition, after which they were randomly assigned to wear NLFC and EB-NLFC hearing aids, alternatively, in a crossover fashion. After each 4-week block, the patients were tested again to obtain the test outcomes.
Results:
Patients with hearing aids with EB-NLFC had a significantly better word and consonant recognition using the MMRT (p<0.05). The MHINT was better for the EB-NLFC group without significant differences. The EB-NLFC group had better scores in both the IOI-HA and sound quality scale but not statistically significant.
Conclusion:
Patients with high-frequency hearing loss may benefit more from using EB-NLFC for word and consonant recognition; however, the improvement was small under a noisy listening environment. The subjective questionnaires did not show significant benefit of EB-NLFC either.
... In other words, the high frequency information of speech will not be easy to hear for the subjects. Therefore, a lot of consonant information (high frequency parts) will be lost, thus decreasing the performance of speech perception, especially for Mandarin listening [23]. Compared with the clean speech in Fig. 3, the clean_FAME speech provided a sturdier high frequency speech signal. ...
... For the objective evaluations we compared the speech distortion index (SDI) (Chen et al., 2006), perceptual estimation of speech quality (PESQ) (ITU-T, 2001;Rix et al., 2001;Hu and Loizou, 2008), and segmental SNR improvement (SSNRI) (Chen, 2008) using utterances from the Aurora-4 task (Hirsch and Pearce, 2000;Parihar et al., 2004). The subjective listening test was conducted with a single-blind design using a standardized Mandarin speech database (Lai et al., 2013b). The experimental results confirm the effectiveness of the proposed GMAPA algorithm relative to several well-known speech enhancement algorithms. ...
... The purpose of this test was to investigate the subjective sound quality performance in each speech enhancement algorithm. A standard sound quality questionnaire in clinical trials was used to rate scores (Lai et al., 2013b). The questionnaire included the following five statements: (Q1) I think this method provides high sound quality, (Q2) I think this method provides a natural sound, (Q3) I can hear very clearly when I use this method, (Q4) I feel comfortable when I use this method, and (Q5) I can't hear noise when I use this method; the subject was asked to score these statements for each of the speech signals using a mean opinion score (MOS) comparison. ...
Spectral restoration methods for speech enhancement aim to remove noise components in noisy speech signals by using a gain function in the spectral domain. How to design the gain function is one of the most important parts for obtaining enhanced speech with good quality. In most studies, the gain function is designed by optimizing a criterion based on some assumptions of the noise and speech distributions, such as minimum mean square error (MMSE), maximum likelihood (ML), and maximum a posteriori (MAP) criteria. The MAP criterion shows advantage in obtaining a more reliable gain function by incorporating a suitable prior density. However, it has a problem as several studies showed: although MAP based estimator effectively reduces noise components when the signal-to-noise ratio (SNR) is low, it brings large speech distortion when the SNR is high. For solving this problem, we have proposed a generalized maximum a posteriori spectral amplitude (GMAPA) algorithm in designing a gain function for speech enhancement. The proposed GMAPA algorithm dynamically specifies the weight of prior density of speech spectra according to the SNR of the testing speech signals to calculate the optimal gain function. When the SNR is high, GMAPA adopts a small weight to prevent overcompensations that may result in speech distortions. On the other hand, when the SNR is low, GMAPA uses a large weight to avoid disturbance of the restoration caused by measurement noises. In our previous study, it has been proven that the weight of the prior density plays a crucial role to the GMAPA performance, and the weight is determined based on the SNR in an utterance-level. In this paper, we propose to compute the weight with the consideration of time–frequency correlations that result in a more accurate estimation of the gain function. Experiments were carried out to evaluate the proposed algorithm on both objective tests and subjective tests. The experimental results obtained from objective tests indicate that GMAPA is promising compared to several well-known algorithms at both high and low SNRs. The results of subjective listening tests indicate that GMAPA provides significantly higher sound quality than other speech enhancement algorithms.
Background noise is a critical issue for hearing aid device users; a common solution to address this problem is speech enhancement (SE). In recent times, a novel SE approach based on deep learning technology, called deep denoising autoencoder (DDAE), has been proposed. Previous studies show that the DDAE SE approach provides superior noise suppression capabilities and produces less distortion than any of the classical SE approaches in the case of processed speech. Motivated by the improved results using DDAE shown in previous studies, we propose the multi-objective learning-based DDAE (M-DDAE) SE approach in this study; in addition, we evaluated its speech quality and intelligibility improvements using seven typical hearing loss audiograms. The experimental results of our objective evaluations show that our M-DDAE approach achieved significantly better results than the DDAE approach in most test conditions. Considering this, the proposed M-DDAE SE approach can be potentially used to further improve the listening performance of hearing aid devices in noisy conditions.
