Bone Conduction Microphone Research Articles

Improving bone-conducted speech quality: a novel approach based on speech production principles

Bone conduction microphone (BCM) which records speech that travels through at least one human tissue, demonstrates a significant advantage in capturing voice in noisy environments. The ergonomic design and versatile placement options of BCMs further enhance their utility, allowing integration with various head equipment such as earplugs, glasses, helmets, or NRBC masks. Despite these advantages, ensuring the quality and intelligibility of recorded speech remains a challenge. Past efforts have involved various post-treatments uniformly applied to all signals, to address these drawbacks, with limited success. In this paper, we introduce a novel approach aimed at enhancing BC speech quality by implementing a select number of filters grounded in speech production principles. In this study, treatments both dependent and independent of phonemes are applied. Subjective tests are conducted to assess the efficacy of these filters and compare their effects.

Comparison of Acoustic Performance of Different Sensors for the Purposes of On-Body Auscultation in Noisy Environments

Lung auscultation is a widely used clinical tool for quick physical assessment of the respiratory system. The ideal sensor preferable for lung auscultation would have high sensitivity to body-conducted sounds with a good resilience to airborne noise. This work envisions to benchmark different sensors starting from air coupled microphones, a bone conduction microphone and PVDF (Polyvinylidene difluoride) sensors on a small chest simulator in a controlled environment. The purpose of this work is to assess the suitability of different sensors, which could be used for lung auscultation both in silent and noisy environments. The proposed work evaluates the performance of different kinds of sensors in terms of their frequency response, noise resilience and a quality metric that expresses sensor’s robustness to airborne noise.

Enabling Real-Time On-Chip Audio Super Resolution for Bone-Conduction Microphones.

Voice communication using an air-conduction microphone in noisy environments suffers from the degradation of speech audibility. Bone-conduction microphones (BCM) are robust against ambient noises but suffer from limited effective bandwidth due to their sensing mechanism. Although existing audio super-resolution algorithms can recover the high-frequency loss to achieve high-fidelity audio, they require considerably more computational resources than is available in low-power hearable devices. This paper proposes the first-ever real-time on-chip speech audio super-resolution system for BCM. To accomplish this, we built and compared a series of lightweight audio super-resolution deep-learning models. Among all these models, ATS-UNet was the most cost-efficient because the proposed novel Audio Temporal Shift Module (ATSM) reduces the network's dimensionality while maintaining sufficient temporal features from speech audio. Then, we quantized and deployed the ATS-UNet to low-end ARM micro-controller units for a real-time embedded prototype. The evaluation results show that our system achieved real-time inference speed on Cortex-M7 and higher quality compared with the baseline audio super-resolution method. Finally, we conducted a user study with ten experts and ten amateur listeners to evaluate our method's effectiveness to human ears. Both groups perceived a significantly higher speech quality with our method when compared to the solutions with the original BCM or air-conduction microphone with cutting-edge noise-reduction algorithms.

Regional Language Speech Recognition from Bone-Conducted Speech Signals through Different Deep Learning Architectures.

Bone-conducted microphone (BCM) senses vibrations from bones in the skull during speech to electrical audio signal. When transmitting speech signals, bone-conduction microphones (BCMs) capture speech signals based on the vibrations of the speaker's skull and have better noise-resistance capabilities than standard air-conduction microphones (ACMs). BCMs have a different frequency response than ACMs because they only capture the low-frequency portion of speech signals. When we replace an ACM with a BCM, we may get satisfactory noise suppression results, but the speech quality and intelligibility may suffer due to the nature of the solid vibration. Mismatched BCM and ACM characteristics can also have an impact on ASR performance, and it is impossible to recreate a new ASR system using voice data from BCMs. The speech intelligibility of a BCM-conducted speech signal is determined by the location of the bone used to acquire the signal and accurately model phonemes of words. Deep learning techniques such as neural network have traditionally been used for speech recognition. However, neural networks have a high computational cost and are unable to model phonemes in signals. In this paper, the intelligibility of BCM signal speech was evaluated for different bone locations, namely the right ramus, larynx, and right mastoid. Listener and deep learning architectures such as CapsuleNet, UNet, and S-Net were used to acquire the BCM signal for Tamil words and evaluate speech intelligibility. As validated by the listener and deep learning architectures, the Larynx bone location improves speech intelligibility.

Influence of Breath Noise on Voice Detection with Bone Conduction Microphone in MRI-Scanner Noise

Influence of Breath Noise on Voice Detection with Bone Conduction Microphone in MRI-Scanner Noise

Study of occlusal acoustic parameters in assessing masticatory performance

BackgroundPrevious masticatory studies have focused on a variety of measurements of foods and boluses or kinematic parameters and sound during mastication. To date, the masticatory sound research of has been limited due to the difficulties of sound collection and accurate analysis. Therefore, significant progress in masticatory sound has not been made. Meanwhile, the correlation between acoustic parameters and mastication performance remains unclear. For the purpose of exploring the acoustic parameters in measuring mastication performance, the bone-conduction techniques and sound analysis were used, and a statistical analysis of acoustic and occlusal parameters were conducted.MethodsThe gnathosonic and chewing sounds of fifty-six volunteers with healthy dentate were recorded by a bone-conduction microphone and further analyzed by Praat 5.4.04 when intercuspally occluding natural foods (peanuts) were consumed. The granulometry of the expectorated boluses from the peanuts was characterized by the median particle size of the whole chewing sequence (D50a) and the median particle size during the fixed chewing strokes (D50b). The chewing time of the whole chewing sequence (CTa), the chewing time of the fixed chewing strokes (CTb), the chewing cycles (CC), and the chewing frequency (CF) were recorded and analyzed by the acoustic software. The acoustic parameters, including gnathosonic pitch, gnathosonic intensity, mastication sound pitch of the whole chewing sequence (MPa), mastication sound pitch of the fixed chewing strokes (MPb), mastication sound intensity of the whole chewing sequence (MIa) and mastication sound intensity of the fixed chewing strokes (MIb), were analyzed. Independent sample t-test, Spearman and Pearson correlation analyses were used where applicable.ResultsSignificant difference in parameters CC, MIa, CF and D50a were found by sex (t-test, p < 0.01). The masticatory degree of the test foods was higher in women (CC, 24.25 ± 5.23; CF, 1.70 ± 0.21 s−1; D50a, 1655.07 ± 346.21 μm) than in men (CC, 18.14 ± 6.38; CF, 1.48 ± 0.18 s−1; D50a, 2159.21 ± 441.26 μm). In the whole chewing sequence study, a highly negative correlation was found between MIa and D50a, and a highly positive correlation was found between MIa and CF (r = − 0.94, r = 0.82, respectively, p < 0.01). No significant correlation was found between the remaining acoustic parameters and mastication parameters. In the fixed chewing strokes study, a highly negative correlation was found between MIb and D50b (r = − 0.85, p < 0.01). There was no significant correlation between the rest of the acoustic parameters and the mastication parameters.ConclusionsMastication sound intensity may be a valuable indicator for assessing mastication. Acoustic analysis can provide a more convenient and quick method of assessing mastication performance.

Intelligibility of bone-conducted speech detected on the scalp

Bone-conduction microphones (BCMs) can detect speakers' voices with high signal-to-noise ratio even under extremely noisy environments like a machine factory or an engine room of a watercraft. BCMs are ordinarily attached to the front of the neck (larynx), therefore, it is sometimes accompanied by discomfort and esthetic problems. In order to solve such problems, we have been developing a novel BCM system built in a helmet, however, characteristics of bone-conducted speech detected on the scalp need to be clarified. In this study, mono-syllable articulations of bone-conducted speech detected at several locations on the head and neck were measured. Also, the speech transmission index (STI), objective measure of signal transmission quality, was calculated. The results indicated that the forehead and the vertex showed better articulation and STI than the mastoid process of the temporal bone, the mandibular condyle, and the occiput. In terms of the gender difference, the forehead and the vertex showed higher scores for the male voice, whereas the cheek showed the highest for the female voice. Additionally, the larynx showed lower scores than others. These results indicated that the attenuation of high-frequency components are smaller at the forehead and the vertex.

Design and implementation of dual pressure variation chambers for bone conduction microphone

This study presents a bone conduction microphone (BCM) to detect the variation of air pressure resulted from the skull vibration. The presented BCM consists of a commercial MEMS microphone, the sensing circuits, the bulk metal, the deformable polymer diaphragm, and two printed circuit board (PCB) covers. In this design, two chambers and a vibrating spring-mass structure are hermetic sealed by two PCB covers. As the spring-mass structure vibrates, the pressure of one chamber will increase and that of another will decrease. Thus, the vibration could introduce a higher pressure load as well as a larger sensing signal for the BCM. In application, the device with the dimensions of 3.5 × 2.65 × 1.48 mm3 is implemented. Measurements show the device has a sensitivity of −38.8 dBV, THD < 0.48% at 1 kHz with 1 g excitation, and ±5 dB bandwidth for 100 Hz∼6.7 kHz. Frequency responses of different samples show good repeatability. Furthermore, air leakage effect and crosstalk of the skull vibration sensing module have also been investigated in this paper. The results demonstrate the feasibility of the presented BCM.

Bone conduction microphone measures in the external auditory canal: An application for implementing cross-talk cancellation for bone-conducted sound

Bone conduction microphone measures in the external auditory canal: An application for implementing cross-talk cancellation for bone-conducted sound

Chewing Monitoring With Bone-Conduction Microphone for Body Area Network

The research described in the paper has developed a system for acquiring, processing, and analyzing the diet-related oral motors and the accompanying body movements, therefore realizes dietary activities monitor. It contains three parts, such as sensor node, mobile client, and server. The mentioned sound is defined as chewing, talking, coughing and silence, while the body movement states are classified as immobile or micro-move, walking, and running. The system is a wearable device that monitors the oral sound activities by recording of and the detection generated during the oral. It also includes an accelerometer to simultaneous records of human body. Bone-conduction microphone is adopted to collect the oral sounds for avoiding noise effects. Hardware system is developed based on special architecture, with respect to the balance between the computing cost, device size, power dissipation and network transmission load. The research proposes a new method to get MFCC coefficients. According to the order, it is assigned in the sensor and server to protect the private information and guarantee parallelism of the data collection and calculation. The oral sound bone-conduction sound database related diet activity is constructed as the training data for the system. Client software uses statistical methods to analyze the oral and body movements derived from it, therefore obtains the user’s dietary laws. Through experiments, the optimal configuration parameters of the system are determined. The recognition rate is up to more than 90%. According to the modern medical researches on diet health, a corresponding system is designed and developed to give a quantitative assessment to the personal dietary laws, and visualize the corresponding information to provide users reasonable guidances.

Portable Obstructive Sleep Apnea Detection Device Based on Bone Conduction Microphone

Snoring is a common physiological phenomenon but it is also an important sign of a life-threatening sleep disorder called Obstructive sleep apnea (OSA). Studies show that OSA affects 14% of men and 5% of women while many patients remain undiagnosed because of the high cost of professional hospital examination. We designed a wearable, portable device based on bone conduction microphone which allows users to perform OSA preliminary screening by themselves at home. Bone conduction microphone is adopted to achieve high quality signal recording while keeping the device convenient to use. Snoring sound recording won’t be affected when people turn over during sleep. After snoring signal recording, apnea or hypopnea part is recognized with Hidden Markov models (HMMs) as the basis for diagnosis of OSA.

Bone-conducted speech enhancement using deep denoising autoencoder

Bone-conduction microphones (BCMs) capture speech signals based on the vibrations of the speaker's skull and exhibit better noise-resistance capabilities than normal air-conduction microphones (ACMs) when transmitting speech signals. Because BCMs only capture the low-frequency portion of speech signals, their frequency response is quite different from that of ACMs. When replacing an ACM with a BCM, we may obtain satisfactory results with respect to noise suppression, but the speech quality and intelligibility may be degraded due to the nature of the solid vibration. The mismatched characteristics of BCM and ACM can also impact the automatic speech recognition (ASR) performance, and it is infeasible to recreate a new ASR system using the voice data from BCMs. In this study, we propose a novel deep-denoising autoencoder (DDAE) approach to bridge BCM and ACM in order to improve speech quality and intelligibility, and the current ASR could be employed directly without recreating a new system. Experimental results first demonstrated that the DDAE approach can effectively improve speech quality and intelligibility based on standardized evaluation metrics. Moreover, our proposed system can significantly improve the ASR performance by a notable 48.28% relative character error rate (CER) reduction (from 14.50% to 7.50%) under quiet conditions. In an actual noisy environment (sound pressure from 61.7 dBA to 73.9 dBA), our proposed system with a BCM outperforms an ACM, yielding an 84.46% reduction in the relative CER (proposed system: 9.13% and ACM: 58.75%).

A REVIEW ON MULTIMODAL SPEAKER RECOGNITION

A review on multimodal speaker recognition (SR) is being presented. For many decades the speaker recognition has been studied and still it has grabbed the interest of many researchers. Speaker recognition includes of two levels –system training and system testing. The robustness of the speaker recognition system depends on the training environment and testing environment as well as the quality of speech .Air conducted (AC) Speech is a source from which speaker is recognized by extracting the features. The performance of the speaker recognition system depends on AC speech. further to improve the robustness and accuracy of the SR system various other sources(Modals) like Throat Microphone ,Bone Conduction Microphone, array of microphones,Non Audible murmur, non auditory information like video are used in complementary with standard AC microphone. This paper is purely a review on SR and various complimentary modals.

E8-5 MRI駆動音による振動と遮蔽した骨伝導マイクロホン音声のSN比について

E8-5 MRI駆動音による振動と遮蔽した骨伝導マイクロホン音声のSN比について

Combination method air and bone conducted speech for speaker recognition in i-vector space

Recently, some new sensors, such as bone-conductive microphones, throat microphones, and non-audible murmur (NAM) microphones, besides conventional condenser microphones have been developed for collecting speech data. Accordingly, some researchers began to study speaker and speech recognition using speech data collected by these new sensors. From these new sensors, we focus on a bone-conductive microphone. Recently, the speaker recognition method based on i-vector will be state-of-the-art. Hence, in this paper, first, we report the speaker recognition performance using bone-conducted speech based on i-vector-based speaker recognition system. In addition, we propose a speaker recognition method combined bone-conducted speech with air-conducted speech. In this paper, we investigate three combination methods, which are a distance combination method, an i-vector combination method, and a feature combination method. To evaluate the proposed methods, we conducted the speaker identification experiments. From experimental results, performance of the bone-conducted speech is almost same as that of the air-conducted speech under the condition of the enrolment and the evaluation speech collected on same session. In addition, the experimental results show that all proposed methods are able to improve the speaker recognition performance of air- and bone-conducted speech.

Relation between area of shielding bone-conduction microphone and quality of speech communication in noisy environment

The doctor does not receive the transmitted voice of the patient when the patient becomes unwell in MRI examination, because the voice is faint and MRI acoustical noise is loud. Our purpose is to improve the quality of speech communication of bone-conduction microphone to transmit the patient's faint voice to the doctor. We propose the theoretical formula using the characteristic of vibration propagation for calculating the quality in the case of the shielding microphone. In this report, the improvement level of speech communication quality in the case of the shielding microphone is estimated by the proposed formula in order to discuss the accuracy of the proposed formula. Here, the improvement level of shielding was calculated with a phantom. The phantom consists of a ceramic plate and a silicon rubber representing bones and muscles, respectively. The vibration of vertical direction on the phantom was measured with accelerometers to calculate the improvement level. The results showed that the quality of ...

A Priori SNR Estimation Using Air- and Bone-Conduction Microphones

This paper proposes an a priori signal-to-noise ratio (SNR) estimator using an air-conduction (AC) and a bone-conduction (BC) microphone. Among various ways of combining AC and BC microphones for speech enhancement, it is shown that the total enhancement performance can be maximized if the BC microphone is utilized for estimating the power spectral density (PSD) of the desired speech signal. Considering the fact that a small deviation in the speech PSD estimation process brings severe spectral distortion, this paper focuses on controlling weighting factors while estimating the a priori SNR with the decision-directed approach framework. The time--frequency varying weighting factor that is determined by taking a minimum mean square error criterion improves the capability of eliminating residual noise and minimizing speech distortion. Since the weighting factors are also adjusted by measuring the usefulness of the AC and BC microphones, the proposed approach is suitable for tracking the parameter even if the characteristic of environment changes rapidly. The simulation results confirm the superiority of the proposed algorithm to conventional algorithms in high noise environments.

The effects of delayed auditory feedback revealed by bone conduction microphone in adult zebra finches.

Vocal control and learning are critically dependent on auditory feedback in songbirds and humans. Continuous delayed auditory feedback (cDAF) robustly disrupts speech fluency in normal humans and has ameliorative effects in some stutterers; however, evaluations of the effects of cDAF on songbirds are rare. We exposed singing young (141–151 days old) adult zebra finch males to high-amplitude cDAF. cDAF exposure was achieved by the recording of bone-conducted sounds using a piezoelectric accelerometer, which resulted in high-quality song recordings that were relatively uncontaminated by airborne sounds. Under this condition of cDAF, birds rapidly (2–6 days) changed their song syllable timing. The one bird for which we were able to maintain the accelerometer recordings over a long period of time recovered slowly over more than a month after cDAF was discontinued. These results demonstrate that cDAF can cause substantial changes in the motor program for syllable timing generation over short intervals of time in adult zebra finches.

The effect of bone conduction microphone placement on intensity and spectrum of transmitted speech items

Speech signals can be converted into electrical audio signals using either conventional air conduction (AC) microphone or a contact bone conduction (BC) microphone. The goal of this study was to investigate the effects of the location of a BC microphone on the intensity and frequency spectrum of the recorded speech. Twelve locations, 11 on the talker's head and 1 on the collar bone, were investigated. The speech sounds were three vowels (/u/, /a/, /i/) and two consonants (/m/, /∫/). The sounds were produced by 12 talkers. Each sound was recorded simultaneously with two BC microphones and an AC microphone. Analyzed spectral data showed that the BC recordings made at the forehead of the talker were the most similar to the AC recordings, whereas the collar bone recordings were most different. Comparison of the spectral data with speech intelligibility data collected in another study revealed a strong negative relationship between BC speech intelligibility and the degree of deviation of the BC speech spectrum from the AC spectrum. In addition, the head locations that resulted in the highest speech intelligibility were associated with the lowest output signals among all tested locations. Implications of these findings for BC communication are discussed.

Model-based speech enhancement using a bone-conducted signal

Codebook-based single-microphone noise suppressors, which exploit prior knowledge about speech and noise statistics, provide better performance in nonstationary noise. However, as the enhancement involves a joint optimization over speech and noise codebooks, this results in high computational complexity. A codebook-based method is proposed that uses a reference signal observed by a bone-conduction microphone, and a mapping between air- and bone-conduction codebook entries generated during an offline training phase. A smaller subset of air-conducted speech codebook entries that accurately models the clean speech signal is selected using this reference signal. Experiments support the expected improvement in performance at low computational complexity.