研究者詳細 - 水町　光徳

その他リンク： https://www.jstage.jst.go.jp/article/jsp/28/4/28_129/_pdf

Subband Optimization of Broadband Nonlinear Beamformer with Long Short-Term Memory 査読有り国際誌

M. Mizumachi, K. Nashiro

Proc. 30th International Congress on Sound and Vibration Paper ID: 550 2024年07月

Frequency dependency of inter-channel cross-correlation for stereo sources 査読有り国際誌

M. Mizumachi, T. Horiuchi

Proc. 156th AES Convention Express Paper 27 2024年06月

発声時の音環境操作ツールにおける実時間処理とインタフェースの実装について

河原英紀, 矢田部浩平, 松井淑恵, 程島奈緒, 水町光徳, 榊原健一

聴覚研究会資料 54 ( 3 ) 247 - 252 2024年05月

An Efficient Method of Frequency Bandwidth Limitation for Wireless-based Ultrasonic Data Collection System 査読有り国際誌

R. Shigyo, D. Nobayashi, K. Tsukamoto, M. Mizumachi, T. Ikenaga

Proc. 2024 RISP International Workshop on Nonlinear Circuits, Communications and Signal Processing (NCSP’24) 28PM2-2-2 2024年03月

Relationship between the Temporal Variation of Fundamental Frequency and Subjective Hoarseness 査読有り国際誌

S. Tomimatsu, M. Mizumachi

Proc. 2024 RISP International Workshop on Nonlinear Circuits, Communications and Signal Processing (NCSP’24) 28AM2-1-5 2024年03月

ヘッドフォン装着時に外耳道入口において収録される信号の内容について

河原英紀, 矢田部浩平, 松井淑恵, 程島奈緒, 榊原健一, 水町光徳

聴覚研究会資料 54 ( 2 ) 93 - 98 2024年02月

Contributions of temporal envelope and temporal fine structure to segregation of Lombard speech from background noise 査読有り

Tamura S., Mizumachi M.

Acoustical Science and Technology ( 一般社団法人日本音響学会 ) 45 ( 1 ) 33 - 44 2024年01月

担当区分：最終著者記述言語：英語掲載種別：研究論文（学術雑誌）

Lombard speech is known to make it easier for listeners to perceive speech signals in the presence of background noise. The aim of this study was to investigate how temporal characteristics such as the temporal envelope and temporal fine structure (TFS) in Lombard speech contribute to improve the segregation of Lombard speech from background noises. In the speech production task, we confirmed that Lombard speech had increased low-frequency amplitude modulations in the temporal envelope and increased frequency modulations in TFS compared to plain speech in frequency ranges covering first and second formants. In a speech perception task, ease of speech segregation was compared between Lombard and plain speech. We also compared ease of segregation of tone-vocoded speech created from Lombard and plain speech to examine the effect of eliminating TFS. We found that Lombard speech with both high amplitude and frequency modulations had improved speech segregation in both original and tone-vocoded speech conditions, but the difference in ease of speech segregation between Lombard and plain speech tended to be larger in original speech than tone-vocoded speech conditions. The present findings suggest that an enhanced envelope is the main contributor to improved segregation of Lombard speech, while increased frequency modulation plays a complementary role.

DOI： 10.1250/ast.e23.24

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85181966313&origin=inward

Frequency dependency of inter-channel cross-correlation for stereo sources 査読有り国際誌

Mizumachi M., Horiuchi T.

AES Madrid 2024 156th Audio Engineering Society Convention 2024年01月

担当区分：筆頭著者,　責任著者記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

Spatial audio has become popular in recent audio applications. Traditional 2-ch stereo sources also provide rich spatial representation. The authors have previously proposed amplitude-based and phase-based crosstalk generators for earphones and headphones reproduction. The parameters for those crosstalk generators were optimized to perceptually generate the suitable stereo stage width compared with the binaural rendering. Investigating the dominant frequency range, which perceptually determines the stereo stage width, is essential in spatial audio. This paper discusses the frequency dependency of inter-channel cross-correlation for 2-ch stereo sources. The frequency dependency of inter-channel cross-correlation, of course, depends on each music source. Those characteristics are investigated for typical classical and vocal music sources to discuss the perceptual aspects of spatial audio.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=105003269203&origin=inward

Proposal of Protocols for Speech Materials Acquisition and Presentation Assisted By Tools Based on Structured Test Signals 査読有り国際誌

Kawahara H., Sakakibara K.I., Mizumachi M., Yatabe K.

2024 27th Conference on the Oriental Cocosda International Committee for the Co Ordination and Standardisation of Speech Databases and Assessment Techniques O Cocosda 2024 Proceedings 2024年01月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

We propose protocols for acquiring speech materials, making them reusable for future investigations, and presenting them for subjective experiments. We also provide means to evaluate existing speech materials' compatibility with target applications. We built these protocols and tools based on structured test signals and analysis methods, including a new family of the Time-Stretched Pulse (TSP). Over a billion times more powerful computational (including software development) resources than a half-century ago enabled these protocols and tools to be accessible to under-resourced environments.

DOI： 10.1109/O-COCOSDA64382.2024.10800149

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85215666447&origin=inward

SUBBAND OPTIMIZATION OF BROADBAND NONLINEAR BEAMFORMER WITH LONG SHORT-TERM MEMORY 査読有り国際誌

Mizumachi M., Nashiro K.

Proceedings of the International Congress on Sound and Vibration 2024年01月

担当区分：筆頭著者,　責任著者記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

Nonlinear beamformers with neural networks have become popular with the remarkable development of machine learning. The concept of the neural network-based beamformer has been proposed in the early 1990s. The nonlinear beamformer trained with a single sinusoidal signal succeeds in sharpening its beampattern at the specific frequency. In other words, narrowband beamforming was achieved using traditional neural networks. Broadband beamforming is, however, indispensable in dealing with acoustic signals such as speech. Mizumachi and Arase have proposed a wideband nonlinear beamformer to sharpen the mainlobe in the speech band. It was challenging to achieve a sharp mainlobe in a lower frequency range below 500 Hz with an 8-ch small-scale microphone array. Then, the authors investigated subband optimization of the nonlinear beamformer in the lower frequencies at 500 Hz. In this study, the authors expand the frequency range of the subband optimization of the broadband nonlinear beamformer. It is confirmed that the nonlinear beamformer with long short-term memory could optimize with the different parameter settings, such as a suitable cell size and unit number, even in the low frequency of 250 Hz and high frequency of 8 kHz.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85205379908&origin=inward

様々な音響測定用信号を離散フーリエ変換に基づく測定法に用いる際の性能比較

河原英紀, 矢田部浩平, 水町光徳, 榊原健一, 北村達也

信学技報 EA2023-57 2023年12月

Acoustic measurement framework for audio systems based on structured periodic test signals 査読有り国際誌

Kawahara H., Yatabe K., Sakakibara K.I., Mizumachi M., Kitamura T.

GCCE 2023 - 2023 IEEE 12th Global Conference on Consumer Electronics 227 - 228 2023年10月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

We introduce a framework for acoustic measurement that allows everyday sounds into acoustic measurement test signals and standard test signals such as swept-sine and MLS (Maximum Length Sequence). The framework extends our extended Time-Stretched-Pulse generation method and safeguarding method for any signals to make them appropriate for acoustic measurement. The extension uses the periodic nature of discrete Fourier transform combined with orthogonalization based on the Walsh-Hadamerd matrix. This framework enables simultaneous measurement of acoustic attributes such as linear time-invariant response, signal-dependent non-linear and time-invariant components, and random time-varying components. We also discuss and demonstrate practical scenarios in classes and laboratory lessons using tools developed based on the proposed framework.

DOI： 10.1109/GCCE59613.2023.10315633

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85179760475&origin=inward

Simultaneous Measurement of Multiple Acoustic Attributes Using Structured Periodic Test Signals Including Music and Other Sound Materials 査読有り国際誌

Kawahara H., Yatabe K., Sakakibara K.I., Mizumachi M., Kitamura T.

2023 Asia Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2023 173 - 180 2023年10月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

We introduce a general framework for measuring acoustic properties such as liner time-invariant (LTI) response, signal-dependent time-invariant (SDTI) component, and random and time-varying (RTV) component simultaneously using structured periodic test signals. The framework also enables music pieces and other sound materials as test signals by “safeguarding” them by adding slight deterministic “noise.” Measurement using swept-sin, MLS (Maxim Length Sequence), and their variants are special cases of the proposed framework. We implemented interactive and real-time measuring tools based on this framework and made them open-source. Furthermore, we applied this framework to assess pitch extractors objectively.

DOI： 10.1109/APSIPAASC58517.2023.10317411

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85180006417&origin=inward

使える音声資料の作り方を再考する

河原英紀, 榊原健一, 水町光徳, 矢田部浩平, 北村達也, 森勢将雅

音声コミュニケーション研究会資料 ( 音声コミュニケーション研究委員会 ) 3 ( 4 ) n/a 2023年09月

記述言語：日本語掲載種別：研究論文（研究会，シンポジウム資料等）

記録する音声資料をできるだけ価値があるものとする方法を検討している。そのような試みの一つとして、 2019 年に講習会を開催した。しかし、その講習会で紹介した内容の幾つかは、その後の技術の発展と検討により陳腐化しただけではなく、誤りであることが明らかになっている。ここでは、それらの発展と問題について説明し、新しい講習会をどのような内容・形式にすべきか議論したい。

DOI： 10.60274/asjsc.sc-2023-25

使える音声資料の作り方を再考する準備・録音・分析など

河原英紀, 榊原健一, 水町光徳, 矢田部浩平, 北村達也, 森勢将雅

音声コミュニケーション研究会資料 3 ( 4 ) SC-2023-25 2023年09月

遠隔異常検知システムのための機器稼働音の周期性に基づいた特徴量抽出の検討

執行律和, 野林大起, 塚本和也, 水町光徳, 池永全志

信学技報 IA2023-28 2023年09月

音場伝達関数の簡易な測定と臨界距離の推定のための試験信号について

河原英紀, 矢田部浩平, 榊原健一, 水町光徳, 北村達也

聴覚研究会資料 53 ( 6 ) 261 - 264 2023年08月

Subband Optimization of Nonlinear Beamformer with Long Short-Term Memory in Low-Frequency Region 査読有り国際誌

M. Mizumachi, K. Nashiro

Proc. 29th International Congress on Sound and Vibration (ICSV29) Paper ID: 483 2023年07月

音メディアを用いた実験を支援するための対話的測定ツールについて

河原英紀, 矢田部浩平, 榊原健一, 水町光徳, 程島奈緒, 松井淑恵

音楽音響研究会資料 42 ( 3 ) 41 - 44 2023年07月

Singing Voice Range Profiling Toolbox with real-time interaction and its application to make recording data reusable 査読有り国際誌

H. Kawahara, KI. Sakakibara, K. Yatabe, M. Mizumachi, T. Kitamura

Proc. Stockholm Music Acoustics Conference 2023 (SMAC 2023) 160 - 166 2023年06月

Relationship between perceptual width of stereo stage and inter-channel cross-correlation 査読有り国際誌

M. Mizumachi, T. Horiuchi

Proc. 154th AES Convention Express Paper 68 2023年05月

Analysis of Temporal Changes in Prosodic Features of Positive and Negative Emotions in Spontaneous Speech 査読有り

Y. Aoki, M. Mizumachi

Proc. ICIAE2023 VS-2 2023年03月

Relationship between the Temporal Variation of Fundamental Frequency and Subjective Hoarseness 査読有り

S. Tomimatsu, M. Mizumachi

Proc. ICIAE2023 PS-4 2023年03月

SUBBAND OPTIMIZATION OF NONLINEAR BEAMFORMER WITH LONG SHORT-TERM MEMORY IN LOW-FREQUENCY REGION 査読有り国際誌

Mizumachi M., Nashiro K.

Proceedings of the International Congress on Sound and Vibration 2023年01月

担当区分：筆頭著者,　責任著者記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

Acoustic beamforming is essential to achieve signal enhancement, noise reduction, sound source localization, and acoustic imaging in various applications. Traditional beamformers, such as delay-and-sum beamformers, have been designed by linear signal processing and diversified in adaptive filtering, statistical signal processing, and machine learning. Nonlinear beamformers optimized by neural networks have great potential to optimize their beampatterns. The nonlinear beamformer trained with a single sinusoidal signal succeeds in sharpening its beam-pattern at the specific frequency. The authors have proposed a wideband nonlinear beamformer to sharpen the mainlobe in the speech band. It was, however, challenging to achieve a sharp mainlobe in a lower frequency range below 500 Hz with an 8-ch small-scale microphone array. This paper investigates sub-band optimization of the nonlinear beamformer in the lower frequencies. It is confirmed that the nonlinear beamformer with long short-term memory, suitable cell size, and unit number could sharpen the mainlobe without sidelobes in the lower frequency range.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85170649321&origin=inward

Relationship between perceptual width of stereo stage and inter-channel cross-correlation 査読有り

Mizumachi M., Horiuchi T.

AES Europe 2023: 154th Audio Engineering Society Convention 2023年01月

担当区分：筆頭著者,　責任著者記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

It is desirable to properly generate inter-channel crosstalk while reproducing legacy stereo sound sources through earphones and headphones. The authors proposed amplitude-based and phase-based crosstalk generators, of which parameters were perceptually optimized by performing a listening test. The optimal parameters were determined based on the perceptual width of the stereo stage by comparing the proposed method with binaural rendering. This paper investigates the relationship between the perceptual stereo stage width and inter-channel cross-correlation. It is confirmed that the inter-channel cross-correlation has non-linear relation to the perceptual stereo stage width. This result contributes to efficiently optimizing the crosstalk generators without subject evaluation.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85174508475&origin=inward

Higher accuracy self-supervised visual odometry with reliable projection 査読有り

Zhou S., Yang Z., Zhu M., Li H., Serikawa S., Mizumachi M., Zhang L.

Artificial Life and Robotics 27 ( 3 ) 568 - 575 2022年08月

記述言語：英語掲載種別：研究論文（学術雑誌）

A complex outdoor environment includes variable illuminations, dynamic objects, and occlusion, making visual odometry in outdoor environments remains a challenging problem. This work trains a self-supervised pose estimation network based on the projection between adjacent frames. The pose between two adjacent frames is a 6D vector. It can be better solved with reliable pixels, instead of all the pixels. Thus, only reliable pixels are under consideration when training. This research proposed a boundary mask and inferior projection mask to eliminate the influence of unreliable pixels. Furthermore, we evaluated the proposed method on the KITTI datasets and compared it with other state-of-the-art deep learning-based visual odometry methods. The result shows that the proposed method can detect inferior projection and boundary pixels well. Moreover, it achieves higher accuracy in pose estimation.

DOI： 10.1007/s10015-022-00766-7

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85131664759&origin=inward

Optimization of Nonlinear Beamformer with Long Short-term Memory 査読有り

M. Mizumachi, S. Arase

Proc. 28th International Congress on Sound and Vibration (ICSV28) Paper ID: 707 2022年07月

Perceptual optimization of hybrid stereo width control method compared with loudspeakers reproduction 査読有り国際誌

M. Mizumachi, Y. Ueno, T. Horiuchi

Proc. 152nd AES Convention Paper Number: 10571 2022年05月

Perceptual optimization of stereo width shrinkage method in headphone reproduction 査読有り

Ueno Y., Mizumachi M., Horiuchi T.

Acoustical Science and Technology ( 一般社団法人日本音響学会 ) 43 ( 2 ) 170 - 172 2022年03月

特集―高臨場感オーディオ―誌上座談会

安藤彰男, 小澤賢司, 亀川徹, 中原雅考, 西口正之, 青木秀一, 杉本岳大, 水町光徳

日本音響学会誌 ( 一般社団法人日本音響学会 ) 78 ( 3 ) 108 - 113 2022年03月

Inter-channel Cross-correlations of Stereo Width Shrinkage Methods Based on Amplitude and Phase Differences 査読有り

Y. Ueno, M. Mizumachi, T. Horiuchi

Proc. ICIAE2022 VS2-2 2022年03月

Investigation of Suitable Training Conditions for Nonlinear Beamformer 査読有り

S. Arase, M. Mizumachi

Proc. ICIAE2022 PS-2 2022年03月

音声収録におけるスマートデバイスなどを用いた音圧較正の検討

河原英紀, 榊原健一, 水町光徳, 北村達也

音声コミュニケーション研究会資料 ( 音声コミュニケーション研究委員会 ) 2 ( 1 ) n/a 2022年01月

記述言語：日本語掲載種別：研究論文（研究会，シンポジウム資料等）

音声研究のための利用価値の高い音声素材の収録のためには、収録の物理的条件を記録する必要がある。録音系の音圧較正は、そのための出発点となる重要なステップである。ここでは、広く普及しているスマートデバイスを用いた較正の可能性についての検討状況について報告する。日常の環境において適切な較正を行うためには、機器そのものとあわせて、実施可能な手順と、それらに伴う誤差を把握しておく必要がある。ここでは、デジタルデバイスと離散時間系の信号処理の特徴を活かした、幾つかの方法について紹介する。

DOI： 10.60274/asjsc.sc-2022-1

OPTIMIZATION OF NONLINEAR BEAMFORMER WITH LONG SHORT-TERM MEMORY 査読有り

Mizumachi M., Arase S.

Proceedings of the International Congress on Sound and Vibration 2022年01月

担当区分：筆頭著者,　責任著者記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

Nowadays, increasing demand for teleworking needs audio capturing systems that can suppress noises and emphasize the target speech. Beamforming, which achieves spatial filtering using a microphone array, is one of the widely-used techniques for speech enhancement and noise suppression. The state-of-the-art nonlinear beamformers are designed with neural networks. In this study, the neural network with the Long Short-Term Memory as a recursive neural network is employed for dealing with acoustic signals as time series. In this study, the proposed beamformer using the neural network is optimized based on the frequency characteristics of the training data and the results of objective evaluation. The training data were prepared using an 8-ch small-scale microphone array with 10 mm microphone spacing in a soundproofed room. The experimental results show that the training data, which consist of the mixture of 1 kHz and 4 kHz sinusoidal signals, achieve suppressing noises coming from non-target directions and enhancing the target signal with low distortion.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85149853200&origin=inward

A Method to Automatic Create Dataset for Training Object Detection Neural Networks 査読有り

Zhou S., Yang Z., Zhu M., Li H., Serikawa S., Mizumachi M., Zhang L.

IEEE Access 10 80505 - 80517 2022年01月

記述言語：英語掲載種別：研究論文（学術雑誌）

Numerous high-accuracy neural networks for object detection have been proposed in recent years. Creating a specific training dataset seems to be the main obstacle in putting them into practice. This paper aims to overcome this obstacle and proposes an automatic dataset creation method. This method first extracts objects from the source images and then combines them as synthetic images. These synthetic images are annotated automatically using the data flow in the process and can be used directly for training. To the best of our knowledge, this is the first automatic dataset creation method for object detection tasks. In addition, the adaptive object extraction method and created natural synthetic images make the proposed method maintain strong adaptation and generalization ability. To validate the feasibility, a dataset that includes 44 categories of objects is created for the object detection task in a vending supermarket. Under the strictest metric AP75, both the trained EfficientDet and YOLOv4 achieve higher than 95% in accuracy on the common difficulty testing set and higher than 90% in accuracy on the high difficulty testing set.

DOI： 10.1109/ACCESS.2022.3195490

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85135766558&origin=inward

Perceptual optimization of hybrid stereo width control method compared with loudspeakers reproduction 査読有り

Mizumachi M., Ueno Y., Horiuchi T.

AES Europe Spring 2022 - 152nd Audio Engineering Society Convention 2022 236 - 240 2022年01月

担当区分：筆頭著者,　責任著者記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

Legacy stereo music sources cause unnatural spatial impressions through the earphones and headphones reproduction due to the lack of crosstalk, which is naturally produced in loudspeakers reproduction. The amplitude-based and phase-based hybrid stereo width control method is proposed to generate the crosstalk components between channels in headphones reproduction. The control parameters for both the previously-proposed amplitude-based and hybrid methods are perceptually optimized compared with the loudspeaker reproduction. Suitable parameter values are successfully obtained for both methods with classical music sources. It is confirmed that the suitable parameter values have a different trend in those methods depending on the music sources.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85136323424&origin=inward

公共空間の音声明瞭度を改善するEGG解析に基づく話者指標の検討

程島奈緒, 榊原健一, 水町光徳

日本音響学会春季研究発表会講演論文集 - 1123 - 1124 2022年01月

Study of Ultrasonic Data Collection System for Anomaly Detection for Mechanical Equipment 査読有り国際誌

Shigyo R., Nobayashi D., Tsukamoto K., Mizumachi M., Ikenaga T.

Proceedings - 2022 International Conference on Computational Science and Computational Intelligence, CSCI 2022 1237 - 1238 2022年01月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

loT-based acoustic data collection systems have been proposed for detecting anomalies in equipment. However, such systems do not work well in outdoor environments due to interference caused by various sound sources (e.g., wildlife sounds and traffic noise). Outdoor noise has unique frequency characteristics, with energy localized in the low-frequency domain. In contrast, mechanical equipment such as motors mainly generates harmonics, including ultrasonic components, during operation. This paper proposes a noise-robust ultrasonic data collection system for outdoor use. To reduce the size of collected data, which increases in proportion to the sampling frequency, the proposed system uses bandpass filtering and data compression.

DOI： 10.1109/CSCI58124.2022.00222

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85172008371&origin=inward

公共空間の音声明瞭度を改善するEGG解析に基づく話者指標の検討

程島奈緒, 榊原健一, 水町光徳

日本音響学会春季研究発表会講演論文集 - 1123 - 1124 2022年01月

雑音下と残響下での発話におけるEGG解析による声帯振動様式の比較

程島奈緒, 榊原健一, 水町光徳

日本音響学会秋季研究発表会講演論文集 - 1615 - 1616 2022年01月

EGG解析から見た緊迫音声の声帯振動様式

程島奈緒, 榊原健一, 水町光徳

日本音響学会聴覚研究会資料 52(8) 625 - 630 2022年01月

Hardware Development of Color Image Contrast Enhancement by High-level Synthesis 査読有り

Yamawaki A., Mizumachi M., Serikawa S.

TENSYMP 2021 - 2021 IEEE Region 10 Symposium 2021年08月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

To realize image processing as hardware module is an effective way to achieve high performance and low power processing. However, hardware development is a heavy burden for software engineer handling image processing usually. High-level synthesis, HLS, converting software program to hardware automatically is a desirable technology for software engineer with less skill of hardware design. On the other hand, the HLS has a significant problem such as it tends to generate a large and low performance hardware when the software program input has never consideration of hardware organization. This paper, focusing on color image contrast enhancement based on histogram equalization, investigates software description methods considering hardware to make HLS generate a good hardware module with high-performance and low power.

DOI： 10.1109/TENSYMP52854.2021.9550810

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85117467339&origin=inward

Non-linear beamformer with long short-term memory network 査読有り国際誌

M. Mizumachi, R. Oka

Proc. Inter-noise 2021 2021年08月

Power-saving Reproduction Algorithm for Speech Signals 査読有り

M. Nakagawara, M. Mizumachi

Journal of the Institute of Industrial Applications Engineers 9 ( 3 ) 97 2021年07月

Performance evaluation of the nonlinear beamformer with the small-scall mems microphone array 査読有り国際誌

M. Mizumachi, R. Oka

Proc. 27th International Congress on Sound and Vibration (ICSV27) 2021年07月

Stereo Matching Based on Features of Image Patch 査読有り

S. Zhou, H. Li, M. Zhu, Z. Li, M. Mizumachi, L. Zhang

Journal of the Institute of Industrial Applications Engineers 9 ( 3 ) 91 2021年07月

Self-supervised monocular depth estimation with occlusion mask and edge awareness 査読有り国際誌

S. Zhou, M. Zhu, Z. Li, H. Li, M. Mizumachi, L. Zhang

Artificial Life and Robotics 26 ( 3 ) 354 - 359 2021年05月

記述言語：英語掲載種別：研究論文（学術雑誌）

Depth estimation is one of the basic and important tasks in 3D vision. Recently, many works have been done in self-supervised depth estimation based on geometric consistency between frames. However, these research works still have difficulties in ill-posed areas, such as occlusion areas and texture-less areas. This work proposed a novel self-supervised monocular depth estimation method based on occlusion mask and edge awareness to overcome these difficulties. The occlusion mask divides the image into two classes, making the training of the network more reasonable. The edge awareness loss function is designed based on the edge obtained by the traditional method, so that the method has strong robustness to various lighting conditions. Furthermore, we evaluated the proposed method on the KITTI datasets. The occlusion mask and edge awareness are both beneficial to find corresponding points in ill-posed areas.

DOI： 10.1007/s10015-021-00685-z

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85106505382&origin=inward

Perceptual optimization of stereo width control methods in headphones 査読有り国際誌

Y. Ueno, M. Mizumachi, T. Horiuchi

Proc. 150th AES Convention ( Audio Engineering Society ) Paper Number: 10480 2021年05月

Proposal of Speed-Up Method for Pulse Type of Digital Switch and Expansion of Application Scope to Them 招待有り査読有り

R. Iwaizono, M. Mizumachi, S. Serikawa

Proc. ICIAE 2021 VS-6 2021年03月

Proposal of a Digital Switch for Easy Changes of Number and Types 招待有り査読有り

R. Iwaizono, M. Mizumachi, S. Serikawa

Proc. ICIAE 2021 VS-3 2021年03月

[Keynote Lecture] Acoustic Signal Extraction Relying on Spatial Cues 招待有り

M. Mizumachi

Proc. ICIAE 2021 KS-1, Keynote Lecture 2021年03月

Development of a Wireless Communication System for Reliable Acoustic Data Collection Toward Anomaly Detection in Mechanical Equipment 査読有り国際誌

Takano C., Fujino S., Nobayashi D., Tsukamoto K., Mizumachi M., Ikenaga T.

Proceedings - 2021 International Conference on Computational Science and Computational Intelligence, CSCI 2021 1544 - 1547 2021年01月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

With the recent proliferation of Internet of Things (IoT) devices that can send and receive data via wireless communication, we are able to monitor and operate these devices remotely. An example of an IoT system using wireless communication is a system for anomaly detection in mechanical equipment using acoustic data. In order to detect anomalies using acoustic data, continuous recording is essential, thereby increasing the data size. Although Wi-Fi networks provide high-capacity data transfer, performance degradation cannot be avoided due to reasons such as packet losses caused by collisions with data from other devices using the same frequency and the increase in distance between two communicating devices. In the present study, we developed a wireless communication system for reliable acoustic data collection for anomaly detection in mechanical equipment. First, as preliminary experiments, we investigated the communication characteristics for the transmission of large-size data by Wi-Fi in indoor and outdoor environments. The results indicated the communication performance was insufficient for transferring all recorded data handled by this system. Therefore, we developed a simple heuristic transmission timing control method and a method that can reduce the amount of transmission data in order to realize a stable acoustic data collection system. Finally, through demonstration experiments using mechanical equipment in the field, we verified the feasibility of the acoustic data collection system.

DOI： 10.1109/CSCI54926.2021.00300

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85133931264&origin=inward

Sharpening Directivity of Beamforming based on Neural Networks using Virtual Observations 査読有り

Kanekiyo Y., Mizumachi M., Miyazaki R.

2021 IEEE 10th Global Conference on Consumer Electronics, GCCE 2021 512 - 513 2021年01月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

Beamforming based on neural networks (NN) has been proposed as nonlinear beamforming. This method improves the performance with more microphones, but it may not be possible to easily increase the number of microphones due to hardware restrictions. Therefore, we focus on virtual observation, which is an estimation of the signal observed at a location where the microphone is not actually placed, using only actual observations. In this paper, we propose to apply virtual observation to NN-based beamforming. From the evaluation experiment, we confirmed that the proposed method can obtain sharper directivity than the case of using only actual microphones.

DOI： 10.1109/GCCE53005.2021.9621789

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85123489127&origin=inward

Performance evaluation of the nonlinear beam-former with small-scale mems microphone array 査読有り国際誌

Mizumachi M., Oka R.

"Advances in Acoustics, Noise and Vibration - 2021" Proceedings of the 27th International Congress on Sound and Vibration, ICSV 2021 2021年01月

担当区分：筆頭著者,　責任著者記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

Beamforming has been one of the most important techniques in noise reduction under adverse acoustic environments. Traditional beamformers were designed in analytical and adaptive schemes. On the other hand, recent neural network-based beamformers achieve great success in signal enhancement. The authors have also proposed a broadband neural network-based beamformer, which can deal with wideband speech signals. The proposed beamformer employs the long short-term memory to model the temporal relationship among multichannel observations. The performance of the proposed beamformer has been confirmed with an 8-ch small-scale linear microphone array, of which the spacing between neighbouring microphones is 10 mm. It is confirmed that the proposed nonlinear beamformer could sharpen the mainlobe and decrease the sidelobes, although an analytical linear beamformer, that is, delay-and-sum beamformer, could not achieve signal enhancement with the small-scale microphone array.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85117473124&origin=inward

Perceptual optimization of stereo width control methods in headphones 査読有り国際誌

Ueno Y., Mizumachi M., Horiuchi T.

150th Audio Engineering Society Convention, AES 2021 2021年01月

担当区分：責任著者記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

Legacy 2-ch stereo music sources cause unnatural spatial impressions through earphones and headphones reproduction due to the lack of the crosstalk produced in loudspeakers reproduction. Perceptual widths of the stereo stage in amplitude-based and phase-based stereo width control methods for headphones reproduction were matched with that of the binaural rendering which simulates loudspeakers reproduction. Paired comparison between the stereo width control methods and binaural rendering was carried out to achieve the perceptual optimization for classic, jazz, and pop music sources. Concerning the amplitude-based stereo width control method, the suitable parameter value for each sound source could be determined from the results of the listening test. It is also found that the phase-based stereo width control method should be optimized in different parameter spaces.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85113887802&origin=inward

Non-linear beamformer with long short-term memory network 査読有り国際誌

Mizumachi M., Oka R.

Proceedings of INTER-NOISE 2021 - 2021 International Congress and Exposition of Noise Control Engineering 2021年01月

担当区分：筆頭著者,　責任著者記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

Acoustic beamforming with a microphone array enables spatial filtering in a wide frequency range. It is a challenging issue to sharpen the mainlobe in the lower frequency region with a small-scale microphone array, of which the number and spacing of microphones are small. A neural network-based non-linear beamformer achieves a breakthrough in sharpening the main-lobe. The non-linear beamforming works well for the narrowband signals but is weak in wideband beamforming. The nonlinear beamforming with the long short-term memory is proposed to deal with wideband speech signals. The long short-term memory network is trained in the recurrent neural network architecture with the sequence of audio data such as speech signals. The performance of the proposed beamformer is confirmed using a small-scale 8-ch MEMS microphone array, where eight microphones are linearly arranged with the neighboring spacing of 10 mm, under a real environment. The beam-pattern of the proposed non-linear beamformer succeeds in sharpening the main-lobe although the linear delay-and-sum beamformer could not achieve frequency selectivity. The feasibility of the proposed beamformer is also confirmed in speech enhancement.

DOI： 10.3397/IN-2021-2673

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85117405196&origin=inward

Interactive and real-time acoustic measurement tools for speech data acquisition and presentation: Application of an extended member of time stretched pulses 査読有り

Kawahara H., Yatabe K., Sakakibara K.I., Mizumachi M., Morise M., Banno H., Irino T.

Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH 3 2197 - 2198 2021年01月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

Objective measurements of speech data acquisition and presentation processes are crucial for assuring reproducibility and reusability of experimental results and acquired materials. We introduce setting and measurement examples of those conditions using an interactive and real-time acoustic measurement tool based on an extended time-stretched pulse. We also introduce supporting tools.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85119206055&origin=inward

Simultaneous measurement of time-invariant linear and nonlinear, and random and extra responses using frequency domain variant of velvet noise 査読有り

Kawahara H., Sakakibara K.I., Mizumachi M., Morise M., Banno H.

2020 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2020 - Proceedings 174 - 183 2020年12月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

We introduce a new acoustic measurement method that can measure the linear time-invariant response, the nonlinear time-invariant response, and random and time-varying responses simultaneously. The method uses a set of orthogonal sequences made from a set of unit FVNs (Frequency domain variant of Velvet Noise), a new member of the TSP (Time Stretched Pulse). FVN has a unique feature that other TSP members do not. It is a high degree of design freedom that makes the proposed method possible without introducing extra equipment. We introduce two useful cases using two and four orthogonal sequences and illustrates their use using simulations and acoustic measurement examples. We developed an interactive and realtime acoustic analysis tool based on the proposed method. We made it available in an open-source repository. The proposed response analysis method is general and applies to other fields, such as auditory-feedback research and assessment of sound recording and coding.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85100949390&origin=inward

Performance evaluation of power-saving audio reproduction 査読有り国際誌

M. Mizumachi, T. Nakashima, M. Nakagawara

Proc. E-Forum Acousticum 2020 Paper ID: 321 2020年12月

[Keynote Lecture] Acoustic Beamforming for Selective Listening Under Noisy Environments 招待有り査読有り

M. Mizumachi

Proc. 4th International Conference on Technology and Social Science (ICTSS2020) IPS-12 Keynote Lecture 2020年12月

Perceptual Evaluation of Power-saving Audio Playback Algorithm Based on Auditory Characteristics 招待有り査読有り

M. Nakagawara, M. Mizumachi

Proc. 4th International Conference on Technology and Social Science (ICTSS2020) IPS-12-04 2020年12月

Comparison of Subjective Characteristics Between Binaural Rendering and Stereo Width Control Method 招待有り査読有り

Y. Ueno, M. Mizumachi, T. Horiuchi

Proc. 4th International Conference on Technology and Social Science (ICTSS2020) IPS-12-03 2020年12月

Proposal of Digital Touch Panel Switch for Easy Addition and Deletion of Panels 招待有り査読有り

R. Iwaizono, M. Mizumachi, S. Serikawa

Proc. 4th International Conference on Technology and Social Science (ICTSS2020) IPS-12-01 2020年12月

利用価値の高い音声データの録音手順招待有り査読有り

榊原健一, 河原英紀, 水町光徳

日本音響学会誌 ( 一般社団法人日本音響学会 ) 76 ( 6 ) 343 - 350 2020年06月

Relationship between objective and subjective evaluation of heavily-distorted speech signals 査読有り国際誌

M. Mizumachi

Proc. 12th Speech in Noise Workshop P-19 2020年01月

Perceptual evaluation of binaural rendering and stereo width control in headphone reproduction 査読有り

Ueno Y., Mizumachi M., Horiuchi T.

148th Audio Engineering Society International Convention 2020年01月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

Two-channel stereophonic sound reproduction is one of the most important issues in audio engineering. Most of stereo music sources, however, cannot naturally reproduce spatial impression through earphones and headphones. In this study, stereo sources are converted by either binaural rendering or stereo width control. The former convolves head-related impulse responses into the original stereo sources, and the latter shrinks the perceptual spatial width by adding the monaural signal. Perceptual differences in between the above methods are investigated on several spatial attributes. It is confirmed that the binaural rendering is suited to reproduce the spatial impression related to the whole sound field and the stereo width control is suitable for naturally characterizing the spatial attributes for individual sound sources.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85091558861&origin=inward

Frequency domain variant of velvet noise and its application to acoustic measurements 査読有り国際誌

H. Kawahara, K. Sakakibara, M. Mizumachi, H. Banno, M. Morise, and T. Irino

Proc. APSIPA ASC 2019 2019年11月

Frequency domain variant of Velvet noise and its application to acoustic measurements 査読有り国際誌

Kawahara H., Sakakibara K.I., Mizumachi M., Banno H., Morise M., Irino T.

2019 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2019 1523 - 1532 2019年11月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

We propose a new family of test signals for acoustic measurements such as impulse response, nonlinearity, and the effects of background noise. The proposed family complements difficulties in existing families, the Swept-Sine (SS), pseudo-random noise such as the maximum length sequence (MLS). The proposed family uses the frequency domain variant of the Velvet noise (FVN) as its building block. An FVN is an impulse response of an all-pass filter and yields the unit impulse when convolved with the time-reversed version of itself. In this respect, FVN is a member of the time-stretched pulse (TSP) in the broadest sense. The high degree of freedom in designing an FVN opens a vast range of applications in acoustic measurement. We introduce the following applications and their specific procedures, among other possibilities. They are as follows. a) Spectrum shaping adaptive to background noise. b) Simultaneous measurement of impulse responses of multiple acoustic paths. d) Simultaneous measurement of linear and nonlinear components of an acoustic path. e) Automatic procedure for time axis alignment of the source and the receiver when they are using independent clocks in acoustic impulse response measurement. We implemented a reference measurement tool equipped with all these procedures. The MATLAB source code and related materials are open-sourced and placed in a GitHub repository.

DOI： 10.1109/APSIPAASC47483.2019.9023247

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85082395969&origin=inward

Neural Network-based Broadband Beamformer with Less Distortion 査読有り

M. Mizumachi

Proc. International Congress on Acoustics （ICA 2019） 2019年09月

Neural network-based direction-of-arrival estimation with cross correlation function 査読有り

Mizumachi M., Okita K.

Proceedings of the 26th International Congress on Sound and Vibration, ICSV 2019 2019年07月

担当区分：筆頭著者記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

© Proceedings of the 26th International Congress on Sound and Vibration, ICSV 2019. All rights reserved. Direction-of-arrival (DOA) estimation has been one of the important issues in acoustic signal processing. It is well known that the DOA can be estimated based on the cross correlation with stereo observations. It is, however, difficult to accurately estimate the DOA because the cross correlation-based function could not yield the sharp spatial feature. On the other hand, the multiple signal classification (MUSIC) algorithm provides the pointed spatial feature but it requires the exact number of sound sources in advance. In this paper, the cross correlation-based spatial feature is sharpened using a neural network. The neural network is trained using the cross correlation-based spatial feature calculated from the stereo observations and the binary teacher signals. It is confirmed that the non-linear cross correlation-based spatial feature has an extremely sharp peak at the direction of the target sound source without the information on the number of sound sources

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85073037682&origin=inward

Performance evaluation of a neural network-based beamformer with a small-scale microphone array 査読有り

M. Mizumachi,K. Eguchi

Inter-noise 2019年06月

Proposal of power-saving audio playback algorithm based on auditory masking 査読有り

Nakashima T., Nakagawara M., Mizumachi M.

AES 146th International Convention 2019年03月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

© 2019 The Audio Engineering Society (AES). All rights reserved. Power consumption is an important issue while listening to music using portable audio devices. The authors have previously proposed a power-saving audio playback algorithm, which has adjusted filter-bank outputs according to our auditory characteristics. It succeeds in reducing power consumption but causes perceptual distortion. In this paper, the power-saving audio playback algorithm is improved based on auditory masking, which attenuates audio components below the masking threshold. As a result of a listening test, it is confirmed that the proposed method is subjectively superior to the previous method with the same power consumption.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85067129142&origin=inward

Neural network-based broadband beamformer with less distortion 査読有り

Mizumachi M.

Proceedings of the International Congress on Acoustics 2019-September 2760 - 2764 2019年01月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

Beamforming has been one of the important issues in the field of multi-channel signal processing including acoustic signal processing. A wide variety of beamformers have been proposed for each application. In general, acoustic beamforming deals with broadband signals such as speech signals compared to narrowband beamforming for antenna array and radar applications. Recently, neural network-based non-linear beamformers become popular but have a problem that causes an annoying non-linear distortion on the output signal. In the case of speech enhancement, it is a serious problem because our auditory system is highly sensitive to artificial non-linear distortion on speech signals. This paper proposes to solve the problem with the relaxed dual cost functions in the neural network-based beamformer for speech enhancement. The primary cost function aims at sharpening the beam-pattern, and the second cost function is introduced to achieve decreasing speech distortion. Those cost functions are alternatively used for optimizing the beam-pattern in the frequency range of speech signals. The feasibility of the proposed method is confirmed by carrying out a listening test.

DOI： 10.18154/RWTH-CONV-239926

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85095154960&origin=inward

Neural network-based direction-of-arrival estimation with cross correlation function 査読有り

Mizumachi M., Okita K.

Proceedings of the 26th International Congress on Sound and Vibration, ICSV 2019 2019年01月

担当区分：筆頭著者,　責任著者記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

Direction-of-arrival (DOA) estimation has been one of the important issues in acoustic signal processing. It is well known that the DOA can be estimated based on the cross correlation with stereo observations. It is, however, difficult to accurately estimate the DOA because the cross correlation-based function could not yield the sharp spatial feature. On the other hand, the multiple signal classification (MUSIC) algorithm provides the pointed spatial feature but it requires the exact number of sound sources in advance. In this paper, the cross correlation-based spatial feature is sharpened using a neural network. The neural network is trained using the cross correlation-based spatial feature calculated from the stereo observations and the binary teacher signals. It is confirmed that the non-linear cross correlation-based spatial feature has an extremely sharp peak at the direction of the target sound source without the information on the number of sound sources

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85084016380&origin=inward

Neural network-based beamformer with dual cost functions 査読有り

M. Mizumachi,K. Eguchi

Proc. WESPAC 2018年11月

Noise Reduction using Neural Network Trained with Amplitude Spectra 査読有り

M. Mizumachi,R. Otani

Inter-noise 2018年08月

What is speech intelligibility? 査読有り

M. Mizumachi

Proc. 25th International Congress on Sound and Vibration (ICSV25) 2018年07月

Power Saving Audio Playback Algorithm Based on Auditory Characteristics 査読有り

M. Mizumachi, W. Kubota, M. Nakagawara

Proc. AES 144th Convention Paper No. 9933 2018年05月

Noise reduction using neural network trained with amplitude spectra 査読有り

Mizumachi M., Otani R.

INTER-NOISE 2018 - 47th International Congress and Exposition on Noise Control Engineering: Impact of Noise Control Engineering 2018年01月

担当区分：筆頭著者記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

© INTER-NOISE 2018 - 47th International Congress and Exposition on Noise Control Engineering: Impact of Noise Control Engineering. All rights reserved. Noise reduction has been one of the most important issues in acoustic signal processing. There are a wide variety of techniques to enhance the target signal and reduce interference. In recent days, neural network-based approaches achieve a great noise reduction performance. The authors have also proposed a neural network-based beamformer, that is, an autoencoder in the temporal domain. Channel-dependent waveforms obtained by an 8-ch linear microphone array were prepared as training data. It succeeds in forming a sharp main-lobe and suppressing side-lobes when the interference comes from the direction away from the target direction. Its performance drops when the direction of the interference is near that of the target signal. In this paper, the neural network-based beamformer is trained with amplitude spectra of the harmonic signals such as musical instruments. In other words, phase information is not used for noise reduction. The performance of the proposed method is evaluated based on signal-to-noise ratio and spectral distortion. The proposed method has an advantage in noise reduction for interferences, of which arrival directions are close to the target direction.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85059360819&origin=inward

What is speech intelligibility? 査読有り

Mizumachi M.

25th International Congress on Sound and Vibration 2018, ICSV 2018: Hiroshima Calling 6 3633 - 3636 2018年01月

担当区分：筆頭著者記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

Copyright © (2018) by International Institute of Acoustics & Vibration. All rights reserved. Speech intelligibility is an important factor but is partly considered for each purpose. The intelligibility of clean speech is difficult to be quantified, although the speech intelligibility indexes in noisy and reverberant environments are widely standardized. In this study, the relationship between the vowel space and subjective impressions is investigated for Japanese five vowels. The vowel space, that is the pentagon defined by the first and second formant frequencies, is considered in the Equivalent Rectangular Bandwidth scale, which is an auditory-oriented, non-linear frequency scale. Concerning the subjective evaluation, a listening test is carried out with five grade scale on subjective fluency and so on. It is found that the skewness of the vowel space is highly correlated with the subjective fluency of vowels.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85058781649&origin=inward

Power saving audio playback algorithm based on auditory characteristics 査読有り

Mizumachi M., Kubota W., Nakagawara M.

144th Audio Engineering Society Convention 2018 2018年01月

担当区分：筆頭著者記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

© 2018 Audio Engineering Society. All Rights Reserved. Music appreciation can be achieved with a variety of manners such as smartphones, portable music players, car audio, and high-end audio systems. Power consumption is one of the important issues for portable electronic devices and electric vehicles. In this paper, a power saving audio playback algorithm is proposed in restraint of perceptual distortion. An original music source is passed through filterbanks and is reconstructed after increasing or decreasing the narrow-band component in each channel. The channel-dependent manipulation is carefully done not to cause perceptual distortion. The feasibility of the proposed method is evaluated by both measuring consumption current while music playback and carrying out a listening test.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85050035197&origin=inward

Relationship among speech intelligibility and non-parametric distortion measures under various noise conditions 査読有り

M. Mizumachi, Y. Moriyama

Proc. Inter-noise Paper ID: 1324 2017年08月

Non-linear broadband beamformer with deep neural network 査読有り

M. Mizumachi, M. Origuchi, and Y. Nishijima

Proc. 24th International Congress on Sound and Vibration (ICSV24) Paper ID: 873 2017年07月

音響・超音波研究がひらいた今日の生活と今後の展望

水町光徳

電子情報通信学会誌 ( 電子情報通信学会 ) 100 ( 6 ) 419 - 423 2017年06月

加齢による聴力低下と発話変化の相互作用招待有り査読有り

水町光徳

日本音響学会誌 73 ( 5 ) 297 - 302 2017年05月

Discussion on Subjective Characteristics of High Resolution Audio 査読有り

M. Mizumachi, R. Yamamoto, K. Niyada

Proc. AES 142nd Convention EB02-07 2017年05月

Relationship among speech intelligibility and non - Parametric distortion measures under various noise conditions 査読有り

Mizumachi M., Moriyama Y.

INTER-NOISE 2017 - 46th International Congress and Exposition on Noise Control Engineering: Taming Noise and Moving Quiet 2017-January 2017年01月

担当区分：筆頭著者記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

© 2017 Institute of Noise Control Engineering. All rights reserved. Speech intelligibility under noisy environments is one of important issues in noise control engineering. It is necessary to subjectively evaluate speech intelligibility, although listening tests require enormous time and costs. Then, objective speech distortion measures have been investigated with parametric models under noisy environments. It is, however, difficult to design an all-purpose parametric distortion measure for severely distorted speech. This study aims at investigating the relationship among subjective speech intelligibility and non-parametric speech distortion measures. The non-parametric distortion measures include signal-to-noise ratio, spectral distortion, and cepstral distortion, which are calculated in each short-term frame. The spectral distortion is prepared in each octave frequency band. The short-term distortion measures yield the means and standard deviations of speech distortion. Multiple linear regression analyses are carried out, where the speech intelligibility and the objective distortion parameters are dependent and independent variables, respectively. The signal-to-noise ratio dominated speech intelligibility as expected. It is also confirmed that speech intelligibility was strongly influenced by the temporal variation of the cepstral distortion under non-stationary noise conditions.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85042060858&origin=inward

Non-linear broadband beamformer with deep neural network 査読有り

Mizumachi M., Origuchi M., Nishijima Y.

24th International Congress on Sound and Vibration, ICSV 2017 2017年01月

担当区分：筆頭著者記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

Speech enhancement is indispensable for human-to-machine speech interfaces, because speech signals are extremely fragile against acoustical interferences. Spatial filtering, that is, beamforming, is one of the popular solutions using multiple microphones. In the case of linear beamformers, a number of microphones are required to make the main lobe sharp at the desired direction. On the other hand, non-linear beamformers efficiently form the sharp main lobe compared with the linear beamformers. Kobatake et al. proposed a pioneering non-linear beamformer with a neural network. This method considered the narrowband optimization of the network structure, and succeeded in both achieving a sharper main lobe and decreasing grating lobes. Narrowband nonlinear beamformers have been steadily developed in the field of antenna and propagation. In this paper, a non-linear broadband beamformer with a deep neural network is proposed for speech enhancement. The proposed beamformer substitute a deep neural network for a conventional neural network, and non-equally-spaced microphone arrangement is employed for making grating lobes disappear. It is confirmed that the proposed beamformer can achieve ideal beam-pattern under a real environment.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85029461029&origin=inward

加齢による聴力低下と発話変化の相互作用招待有り査読有り

水町光徳

日本音響学会誌 ( 一般社団法人日本音響学会 ) 73 ( 5 ) 297 - 302 2017年01月

Advanced delay-and-sum beamformer with deep neural network 査読有り

M. Mizumachi and M. Origuchi

Proc. 22nd International Congress on Acoustics (ICA2016) Paper ID: 696 2016年09月

Speech intelligibility under in-car distant-talking environments 査読有り

Mizumachi M., Takuma S., Ohsugi I., Hamada Y., Nishi K.

Proceedings of the INTER-NOISE 2016 - 45th International Congress and Exposition on Noise Control Engineering: Towards a Quieter Future 389 - 393 2016年08月

担当区分：筆頭著者記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

© 2016, German Acoustical Society (DEGA). All rights reserved. Speech intelligibility has been widely investigated for various purposes. Hands-free speech communication without any close-talking microphone becomes popular inside a vehicle for the safety reason. It is necessary to carefully install microphones so that speech signals are clearly captured under noisy in-vehicle environments. In this study, the relationship between the intelligibility of distant-talking speech and acoustical interferences is investigated inside car cabins. Target speech signals are prepared as the utterance of Japanese 4-digits random sequences. The acoustical interferences consist of exterior and interior noises. Difficulty in speech perception is defined by the speech reception threshold, which corresponds to the sound pressure level of speech at the 50 % correct score in an articulation test. It is suggested that the interior blower noise caused by an air-conditioner is the most dominant interferences in speech intelligibility. For further investigation, multiple regression analyses have been carried out using subjective indexes related to speech intelligibility. It is found that speech intelligibility can be modeled with thickness of speech, loudness of speech, and loudness of noise.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84994560619&origin=inward

Speech intelligibility under in-car distant-talking environments 査読有り

M. Mizumachi, S. Takuma, I. Ohsugi, Y. Hamada, K. Nishi

Proc. Inter-noise Paper ID: 408 2016年08月

Perceptual restoration of intermittent speech using human speech-like noise 査読有り

M. Mizumachi, S. Imanaga, and T. Horiuchi

Proc. 23rd International Congress on Sound and Vibration (ICSV23) Paper ID: 1050 2016年07月

Subjective evaluation of high resolution audio through headphones 査読有り

M. Mizumachi, R. Yamamoto, and K. Niyada

Proc. AES 140th Convention Paper No. 9529 2016年06月

Singing support system using LPC residual wave for laryngectomee 査読有り

Himeno T., Tanaka Y., Mizumachi M., Matsui K.

2016 IEEE International Conference on Consumer Electronics, ICCE 2016 395 - 396 2016年03月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

© 2016 IEEE. In this paper, we propose a singing system to transmit the pseudo-vocal sound source to the vocal tract. The purpose is the quality improvement of singing voice when using an Electric-Larynx (EL), which is an alternative vocal apparatus of laryngectomee. Thus, we focus on the LPC residual wave, but it has a problem that the signal-power of the vowels /i/ and /u/ are largely reduced. Therefore we tried to improve the quality of sound by adding a signal-power compensation. We carried out listening experiments for the four sound sources «LPC residual wave», «LPC residual wave (signal-power Compensation)», «Rosenberg wave» and «EL». As a result, power-compensated LPC residual wave had the highest naturalness. However, the spectrum of the vowels /i/ differed greatly from the singing voice of the normal utterance.

DOI： 10.1109/ICCE.2016.7430662

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84965132816&origin=inward

Perceptual restoration of intermittent speech using human speech-like noise 査読有り

Mizumachi M., Imanaga S., Horiuchi T.

ICSV 2016 - 23rd International Congress on Sound and Vibration: From Ancient to Modern Acoustics 2016年01月

担当区分：筆頭著者記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

Mobile phones have caused an explosive increase in packet distribution. It causes a serious problem, that is, packet loss. Packet loss concealment is indispensable for achieving smooth speech communication. As a packet loss concealment method on a client side, a waveform substitution is popular and standardized by ITU. The ITU-T G. 711 conceals packet loss by inserting a phase-adjusted amplitude-attenuating previous packet into each break, but cannot deal with long-term breaks over 60 ms, that is, burst loss. The authors have previously proposed an alternative packet loss concealment relying on a human auditory capability. This method does not aim at restoring a waveform of the intermittent speech signal, but achieves perceptual restoration relying on phonetic restoration. When a gap of a intermittent speech signal is filled up with a loud arbitrary signal, we can smoothly listen to restored speech even if some segments of the original speech signal are completely lost. There is a trade-off between smoothness of the restored speech and noisiness of the gap-filling signal. Previously, the gap-filling signal was composed of a harmonic complex and ambient noises. In this paper, a human speech-like noise is substituted for the gap-filling signal. The speech-like noise can be prepared by repeatedly overlapping short-term human speech signals. It is confirmed that the proposed gap-filling signal succeeds in reducing its noisiness.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84987899770&origin=inward

Subjective evaluation of high resolution audio through headphones 査読有り

Mizumachi M., Yamamoto R., Niyada K.

140th Audio Engineering Society International Convention 2016, AES 2016 2016年01月

担当区分：筆頭著者記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

Recently, high resolution audio (HRA) can be played back through portable devices, and spreads across musical genres and generation. It means that most people listen to HRA through headphones and earphones. In this study, perceptual discrimination among audio formats including HRA has been investigated using headphones. 36 subjects, who have a variety of audio and musical experiences in the wide age range from twenties to seventies, participated in listening tests. Headphones presentation is superior in discriminating the details to the loudspeakers presentation. It is, however, found that the headphones presentation is weak in reproducing presence and reality. Audio enthusiasts and musicians could significantly discriminate audio formats than ordinary listeners in both headphones and loudspeakers listening conditions.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84983565516&origin=inward

LPC残差駆動歌唱システムにおける音韻性除去による音質改善の検討査読有り

姫野剛至, 田中裕人, 水町光徳, 松井謙二, 中藤良久

産業応用工学会論文誌 ( 一般社団法人産業応用工学会 ) 4 ( 1 ) 19 - 25 2016年01月

記述言語：英語掲載種別：研究論文（学術雑誌）

We propose a singing system to transmit the pseudo-vocal sound source into the vocal tract. In this paper, we examine the pseudo-vocal sound source that is suitable for singing system. So we carried out a comparison of LPC residual wave and Bark-LPC residual wave that the phonological component included in pseudo-vocal sound source was removed by Bark spectrum. As a result, we found that although the singing voice using Bark-LPC residual is in the low natural tendencies, the Bark-LPC residual wave can reproduce a higher speaker information contributed by the speaker characteristics.

DOI： 10.12792/jjiiae.004.01.04

その他リンク： https://ci.nii.ac.jp/naid/130005263509

Study on the improvement of intelligibility for elderly speech using formant frequency shift method 査読有り

Tanaka Y., Mizumachi M., Nakatoh Y.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 9811 LNCS 683 - 690 2016年01月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

© Springer International Publishing Switzerland 2016. In general, aging is progressing in developed countries. Elderly people have difficulty controlling their articulation accurately due to aging. We need to improve the quality of elderly speech for smooth communication. In this paper, we analyzed the 1st formant frequency (F1) and 2nd formant frequency (F2) between the more intelligible speech and less intelligible speech of Japanese elderly people. In addition, we improved the intelligibility of less intelligible elderly speech by using the formant frequency shift method. This method is the correcting by shift value of formant frequency based on LPC. The shift value is the magnification such as expanding the F1-F2 size of less intelligible speech.

DOI： 10.1007/978-3-319-43958-7_83

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84984829622&origin=inward

Acoustical Sensing of Approaching Vehicle by Non-Linear Filtering 査読有り

M. Mizumachi

Proc. 12th Western Pacific Acoustics Conference 2015 (WESPAC) 2015年12月

Design of elder-friendly auditory signals for microwave ovens 査読有り

Kadota N., Mizumachi M., Nakatoh Y.

2015 4th International Conference on Informatics, Electronics and Vision, ICIEV 2015 2015年11月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

© 2015 IEEE. Recent house electrical appliances generate a number of auditory signals, when a user operates an appliance or an appliance draws attention to a user. In this paper, we focus on multi-functional microwave ovens, which commonly have several auditory signals for user's operation, error occurrence, completion of cooking, and so on. A wide variety of auditory signals for current microwave ovens have two problems: it is difficult for users to image what the auditory signals inform, and the auditory signals including high frequency components are not friendly for elderly users. In this paper, popular auditory signals for microwave ovens were subjectively evaluated with/without an elderly-simulated band-pass filter, which simulated an averaged audiogram of elder people, on the five grade scales of easiness of perception, unpleasantness, and emergency. Experimental results suggest that young and elderly people might imagine different function for a single auditory signal. Finally, suitable auditory signals both for young and elderly users are proposed in the viewpoints of frequency and temporal patterns.

DOI： 10.1109/ICIEV.2015.7333977

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84961778212&origin=inward

Perceptual restoration of intermittent speech under noisy environments 査読有り

M. Mizumachi; S. Motomura; T. Takakura; T. Horiuchi

Proc. 22nd International Congress on Sound and Vibration (ICSV22) 2015年07月

Subjective evaluation of high resolution audio under in-car listening environments 査読有り

Mizumachi M., Yamamoto R., Niyada K.

138th Audio Engineering Society Convention 2015 1 60 - 64 2015年01月

担当区分：筆頭著者記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

High resolution audio (HRA) becomes increasingly popular both for music production and consumers. It enables to record a music performance in a wide-band and precise digital audio format. It is, however, unclear in its perceptual advantage under some listening environments. In this study, listening tests were carried out inside cars, where 34 participants listened to the same music in four different audio formats. The participants chose an audio format with better quality in paired comparison among 192 kHz/24 bits PCM, 48 kHz/16 bits PCM, and two kinds of lossy-compressed MPEG audio formats. The participants, who are familiar with HRA and live music performance, could significantly discriminate among the audio formats.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84943236951&origin=inward

文字入力を用いた発話支援システムの開発査読有り

濱井裕也, 水町光徳, 中藤良久, 松井謙二

産業応用工学会論文誌 ( 一般社団法人産業応用工学会 ) 3 ( 1 ) 10 - 17 2015年01月

記述言語：日本語掲載種別：研究論文（学術雑誌）

Utterance is very important for us to communicate with others. However, many people have speech handicaps and can't produce utterance smoothly. They mainly rely on writing and sign language for communication. But they are not easy skills for them to master. In order to help them, we developed a new utterance support system which can function through an information terminal. We made predictions which includes two additional functions of fixed phrase dictionary and the class 2-gram – one of the statistical approaches for processing natural language. Fixed phrase dictionary includes phrases often used in our daily life conversations and class 2-gram uses co-occurrence frequencies of parts of speech in Japanese. There are two purposes for us to use class 2-gram, one is to generate next candidates and the other is to change frequency. We concluded two evaluation experiments. The results show that our system needs less numbers of inputs in helping with written utterance production and using our prototype permits us to input more characters in one minute than using iPhone IME. Hereafter, we intend to create more effective dictionaries and develop new GUI to make our system more effective and faster.

DOI： 10.12792/jjiiae.003.01.02

その他リンク： https://ci.nii.ac.jp/naid/130005263536

単語発話音声の周波数特徴の加齢変化 (福祉情報工学)

緒方一仁, 水町光徳, 中藤良久

電子情報通信学会技術研究報告 = IEICE technical report : 信学技報 ( 一般社団法人電子情報通信学会 ) 114 ( 261 ) 49 - 54 2014年10月

記述言語：日本語掲載種別：研究論文（研究会，シンポジウム資料等）

日本を含む先進諸国では,高齢化が深刻な社会問題となっている.加齢は,音声コミュニケーションにも影響を及ぼす.聴覚の加齢変化については精力的に研究が行われてきたが,発話器官の加齢変化についてはまだ明らかになっていない.本稿では,発話器官の加齢変化に着目し,発話音声の静的特徴について検討する.分析に用いた音声は,64歳から87歳までの男性話者21名による日本語単語発話である.各単語の語頭母音に対して,パワースペクトルを算出し,20代話者と高齢者話者グループの平均と分散を比較した.本分析の結果,加齢により2.5kHzから4.5kHzの周波数帯域でのパワー減少と,4.5kHz以上でのパワー上昇を確認し,すべての音素において周波数特性が平坦になることが判明した.

その他リンク： https://ci.nii.ac.jp/naid/110009958972

Robust Sensing of Approaching Vehicles Relying on Acoustic Cue 査読有り

M. Mizumachi, A. Kaminuma, N. Ono, and S. Ando

Proc. IEEE IS3C2014 Paper No. 265 2014年06月

Robust Sensing of Approaching Vehicles Relying on Acoustic Cue 査読有り

M. Mizumachi, Atsunobu Kaminuma, Nobutaka Ono, and Shigeru Ando

Sensors 14 ( 6 ) 9546 - 9561 2014年05月

Applicability of perceptual evaluation of speech quality in evaluating heavily distorted speech 査読有り

Mizumachi M.

137th Audio Engineering Society Convention 2014 224 - 229 2014年01月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

Speech quality assessment is indispensable to properly design a speech enhancement algorithm. Perceptual evaluation of speech quality (PESQ) is frequently employed as an objective speech distortion measure. The PESQ is a methodology for estimating subjective assessment of speech distortion caused by speech codecs for telephony systems. In case of noise reduction, however, a degree of speech distortion is heavier than those caused by the speech codecs. In this paper, applicability of the PESQ is investigated for noisy and noisereduced speech signals under severe noisy conditions. A relationship between PESQ scores and subjective mean opinion scores reveals that the PESQ can be applicable for heavily distorted speech only under nonstationary noisy conditions.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84938854653&origin=inward

Perceptual restoration of intermittent speech relying on illusory continuity 査読有り

Mizumachi M., Motomura S., Takakura T., Horiuchi T.

Proceedings of Forum Acusticum 2014-January 2014年01月

担当区分：筆頭著者記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

A digital communication network makes our daily lives comfortable. It, however, causes a serious problem in speech communication due to a packet loss, which yields intermittent speech after decoding the packets. Packet loss concealment (PLC) is indispensable to achieve smooth speech communication over a digital network. There are various kinds of PLC methods based on wave substitution and feature prediction. Those methods aim to achieve intact restoration of a speech waveform. The ITU-T Recommendation G.711 offers a standard PLC method, but cannot conceal the packet loss, of which lost duration exceeds 60 ms. Then, G.711 PLC method is not enough to cope with the actual packet loss problem. On the other hand, the authors have proposed a perceptual restoration scheme of intermittent speech. We can listen to intermittent speech smoothly, when its gap is filled out with a loud, wide-band noise such as a white noise. This is an auditory illusion called illusory continuity, and is explained as the environmental adaptability of the human brain. In order to employ the illusory continuity for PLC, it is necessary to decrease the harshness of the loud gap-filling noise. A speech-like gap-filling signal is prepared with a mixture of less-harsh environmental noises in advance. Both dynamic and static properties of speech are incorporated into the proposed gap-filling signal. The feasibility of the proposed PLC method was examined by a listening test. Participants gave five-grade mean opinion scores on the smoothness of the restored speech, noisiness of the gap-filling signal, and the easiness of speech perception. It is confirmed that the proposed method has a significant advantage over the G.711 PLC method under long-term packet loss conditions.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84953311058&origin=inward

加齢による聴覚と発話の相互作用 (福祉情報工学)

水町光徳

電子情報通信学会技術研究報告 = IEICE technical report : 信学技報 ( 一般社団法人電子情報通信学会 ) 113 ( 272 ) 61 - 66 2013年10月

担当区分：筆頭著者記述言語：日本語掲載種別：研究論文（研究会，シンポジウム資料等）

高齢化は大きな社会問題であり、加齢による音声コミュニケーションに生じる障害も切実な問題である。加齢による聴力低下は、60歳代以前にも見受けられるが、70歳を超えると急激に高周波数成分の聴力閾値が上昇する。70歳代では、聴覚障害の自覚がなくても、加齢による高音障害型感音性難聴と診断されることが多い。これにより、子音が聴き取りづらくなったり、雑音のなかでの音声の聞き取り能力が低下する。一方、音声知覚において、我々は、声からある程度の年齢を推定できる。つまり、我々は、聴覚的印象として、高齢者音声らしさを表す複数の特徴を知っている。その一つが嗄声であり、これは高齢者音声を特徴付ける聴感印象として数多く報告されている。著者は、高周波数帯域における発話エネルギー上昇が、嗄声の音響的特徴であることを確認した。従来の研究では、高齢者の聴力特性と音声の音響的特徴は、それぞれ独立して調査されてきた。本稿では、「ことばの鎖」を拠り所として、加齢による聴力低下と発話音声変形の相互作用について総合的に検討する。

その他リンク： https://ci.nii.ac.jp/naid/110009783062

Speech enhancement by non-linear beamforming tolerant to misalignment of target source direction 査読有り

H. Hayashi, M. Mizumachi, Y. Nakatoh

Journal of the Institute of Industrial Applications Engineers 1 ( 2 ) 97 - 104 2013年10月

Investigation on objective quality evaluation for heavily distorted speech 査読有り

Mizumachi M.

134th Audio Engineering Society Convention 2013 668 - 672 2013年09月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

Demand for evaluating speech quality is on the increase. It is advisable for evaluating the speech quality to employ the common objective measure for the wide variety of adverse speech signals. Unfortunately, current speech quality measures do not suit for heavily distorted speech signals. In this paper, both the applicability and the limit of the perceptual evaluation of speech quality (PESQ) are investigated compared with the subjective mean opinion score (MOS) for noise-added and noise-reduced speech signals. It is found that the PEAQs are compatible with the MOSs for the noise-reduced speech signals in the non-stationary noise conditions.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84883431685&origin=inward

How do we perceive noisy and noise-reduced speech? 査読有り

M. Mizumachi

Proc. Inter-noise 2013年09月

Restoration of intermittent speech based on human auditory capability and temporal characteristics of speech 査読有り

M. Mizumachi, S. Motomura, T. Takakura, and T. Horiuchi

Proc. 20th International Congress on Sound and Vibration (ICSV20) 2013年07月

Relationship Between Subjective and Objective Evaluation of Noise-reduced Speech With Various Widths of Temporal Windows 査読有り

M. Mizumachi

Proc. 21st International Congress on Acoustics Paper ID: 1aSCb13 2013年06月

Investigation on objective quality evaluation for heavily distorted speech 査読有り

M. Mizumachi

Proc. AES 134th Convention CD-ROM (Paper ID: 8830) 2013年05月

How do we perceive noisy and noise-reduced speech? 査読有り

Mizumachi M.

42nd International Congress and Exposition on Noise Control Engineering 2013, INTER-NOISE 2013: Noise Control for Quality of Life 6 4818 - 4823 2013年01月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

It is important for speech applications to quantify the intelligibility and the quality of speech. There are some means of evaluating the intelligibility and the quality objectively and subjectively. The objective evaluation is economical, but the general-use distortion measure is not available. On the other hand, the subjective evaluation requires a cost-consuming formal listening test. The right evaluation should be carried out in the right purpose. No almighty methodology has been established at the moment. In this paper, brain activities involved in evaluating the noisy and noise-reduced speech are measured using the functional magnetic resonance imaging (fMRI) technique. It is found that the brain activity in evaluating the noise-reduced speech varies depending on the noise reduction approach. The noise-reduced signals, which are obtained by the linear processing such as the delay- And-sum beamforming, calm down the brain activity compared with the noisy speech signals. On the other hands, non-linear signal processing such as the spectral subtraction activates the auditory cortex intensely. It is necessary for designing the advanced speech distortion measure to independently consider how the noise is removed and how the speech is distorted by noise reduction.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84904515025&origin=inward

Restoration of intermittent speech based on human auditory capability and temporal characteristics of speech 査読有り

Mizumachi M., Motomura S., Takakura T., Horiuchi T.

20th International Congress on Sound and Vibration 2013, ICSV 2013 3 1896 - 1902 2013年01月

担当区分：筆頭著者記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

Recently, a new obstruction to speech communication, that is, the short-term intermittence of the speech signal, is occurred due to the congestion in the digital communication network. Mobile phones and smart phones have spread rapidly, and then the communication lines are occupied by the heavy traffic of the received/transmitted packets. Therefore, the packet associated with speech can also be lost in the congested communication line. It is not advisable to require the re-transmission of the lost packet in speech communication, because the re-transmission of the lost packet disturbs smooth speech communication and causes the further traffic in the communication line. The restoration of the intermittent speech should be completed on the user's client equipment. The copy of the previous packet should be the easiest approach in restoring the intermittent speech, although the restored speech must be perceived slowly and lazily. The authors have proposed a novel approach to the restoration of the intermittent speech relying on the human auditory capability [Proc. ICSV19, 2012]. This is not an exact physical restoration of the intermittent speech signal, but a perceptual restoration. When the gap of the intermittent speech signal is filled up with an extremely noisy noise signal, we can listen to the intermittent speech smoothly even if some segments of the original speech signal are lost completely. The authors have proposed a practical scheme of the restoration based on our phonemic restoration capabilities. The previously proposed method has employed the less harsh noise and the speech-like harmonic component designed in the frequency domain as the gap-filling signal. Furthermore, in this paper, the temporal characteristics of speech are introduced into the gap-filling signal. It is confirmed that the amplitude modulation and the temporal decay in the gap-filling signal improve the auditory impression of the restored speech signals.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84897035542&origin=inward

ハイレゾリューションオーディオの音質評価査読有り

山本竜太, 金只直人, 水町光徳

産業応用工学会論文誌 ( 一般社団法人産業応用工学会 ) 1 ( 2 ) 52 - 57 2013年01月

記述言語：日本語掲載種別：研究論文（学術雑誌）

In this paper, sound quality of high resolution audio (HRA) is investigated in the view point of auditory perception. Perceptual characteristics of HRA have been examined by listening tests compared with the standard audio CD and the compressed MPEG audio layer-3 (MP3) qualities. The listening tests were carried out by the paired comparison, and the participants were asked to discriminate among HRA, CD, and MP3. Jazz music was selected out as the original HRA. Both CD and MP3 quality audio files were prepared from the original HRA file by the authors. It is found that the participants prefer the audio format with richer information, that is, HRA is most desirable and CD is better than MP3. The questionnaires filled in by participants indicate that subjective characteristics of HRA include much presence, high spatial clarity, and wide spaciousness. Additionally, perceptual discrimination in quantization distortion is also focused on in between 16 bits and 24 bits resolution. As the result of the listening test with 28 participants, the difference of the quantization resolution was discriminated with the accuracy of 60.3 %. Concerning the perceptual discrimination of HRA, quantization resolution is an important factor as well as its wide frequency range.

DOI： 10.12792/jjiiae.001.02.003

Kyutacar

その他リンク： https://ci.nii.ac.jp/naid/130005263583

Relationship between subjective and objective evaluation of noisereduced speech based on temporal variation of speech distortion 査読有り

Mizumachi M.

41st International Congress and Exposition on Noise Control Engineering 2012, INTER-NOISE 2012 1 109 - 117 2012年12月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

It is indispensable for designing a noise reduction algorithm to evaluate the quality of the noise-reduced speech signals accurately and efficiently. Subjective evaluation is an ideal approach, but requires a time-consuming listening test with a number of subjects. Then, objective distortion measures are often employed for efficient evaluation. However, almost all the distortion measures do not consider the temporal variation of the speech distortion. In this paper, the temporal aspect of the short-term speech distortion is investigated based on higher-order statistics such as variance, skewness, and kurtosis for short-term segmental SNR distribution. It is interesting that the skewness of the segmental SNR gives a good explanation for the discrepancy between subjective and objective evaluation.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84883590418&origin=inward

Relationship between subjective and objective evaluation of noise-reduced speech based on temporal fluctuation of speech distortion 査読有り

M. Mizumachi

Proc. Inter-noise CD-ROM (Paper ID: 808) 2012年08月

Relationship between subjective and objective evaluation of distorted speech 査読有り

Mizumachi M.

131st Audio Engineering Society Convention 2011 1 351 - 356 2011年12月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

It is important for designing a noise reduction algorithm to evaluate the quality of the noise-reduced speech signals accurately and efficiently. Subjective evaluation is ideal, but requires a time-consuming listening test with a lot of subjects. Then, objective distortion measures are often used for efficient evaluation. However, almost all the distortion measures do not consider the temporal variation of the speech distortion. In this paper, the temporal aspect of the short-term speech distortion is investigated based on higher-order statistics, that is, variance, skewness, and kurtosis. It is interesting that the skewness of the short-term SNR gives a good explanation for the discrepancy between subjective and objective evaluation.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84866309724&origin=inward

Versatile beamforming for speech enhancement in non-stationary acoustic environments 査読有り

Mizumachi M.

Proceedings of Forum Acusticum 311 - 316 2011年12月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

Speech enhancement is indispensable for designing high-quality speech interfaces. Various kinds of beamformers have been proposed for speech enhancement by spatial filtering. Delay-and-sum beamforming is usually expensive for speech enhancement with numbers of microphones, but provides less-distorted speech signals due to linear signal processing. Null-steering subtractive beamforming achieves efficient speech enhancement, but resultant speech signals include artificial distortion by non-linear signal processing. Each beamformer has both an advantage and a drawback. Directionof- arrival (DOA) is the most important spatial feature for beamforming. The proposed versatile beamforming aims at integrating delay-and-sum beamforming with null-steering beamforming adaptively under time-varying acoustic environments. Confidence for each DOA estimate can be estimated based on a spatial feature, e.g., a cross-correlation function for stereo observations, and is the key for achieving the adaptive integration. Null-steering beamforming can be employed with a reliable DOA estimate, and delay-and-sum beamforming should be used with an unreliable DOA estimate. The DOA confidence varies time by time under non-stationary acoustic environments. The propose versatile beamformer chooses between delay-and-sum and null-steering subtractive beamformers relying on the confidence of each DOA confidence in non-stationary acoustic environments.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84866005187&origin=inward

Speech enhancement with adaptive integration between spatial and spectral filtering 査読有り

Mizumachi M.

40th International Congress and Exposition on Noise Control Engineering 2011, INTER-NOISE 2011 2 1432 - 1437 2011年12月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

Speech enhancement is indispensable for robust speech processing in noisy environments. It is, however, difficult to achieve effective speech enhancement under practical, non-stationary, and unknown noisy environments. Challenge to design an almighty algorithm is not realistic for speech enhancement in real environments. In other words, the right scheme should be prepared for speech enhancement in the right noisy environment. In this paper, a novel approach is introduced to achieve robust speech enhancement. Provided that two or more microphones are available, we can estimate a direction of arrival (DOA) from a spatial feature, which can be also used for spatial filtering as a means of speech enhancement. The author has already proposed to give reliability, which is also estimated from the spatial feature, to each estimate of the DOA. The proposed method achieves speech enhancement by adaptively selecting the suitable approach relying on the reliability of each DOA estimate. Beamforming and spectral subtraction are prepared as candidates for spatial and spectral filtering, respectively. Each output is merged with the time-varying mixing ratio depending on the reliability of each DOA estimate. Then, the proposed method can deal with non-stationary noisy environments by introducing the adaptive integration between spatial and spectral filtering. Feasibility of the proposed idea is confirmed under a less-reverberant, non-stationary noisy condition. Copyright © (2011) by the Institute of Noise Control Engineering.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84867960160&origin=inward

Relationship between subjective and objective evaluation of distorted speech 査読有り

Mitsunori Mizumachi

Proc. AES 131st Convention 2011年10月

Speech enhancement with adaptive integration between spatial and spectral filtering 招待有り査読有り

Mitsunori Mizumachi

Proc. Inter-noise 2011年09月

Versatile beamforming for speech enhancement in non-stationary acoustic environments 査読有り

Mitsunori Mizumachi

Proc. Forum Acusticum 2011年06月

Statistical confidence measure for direction-of-arrival estimate 査読有り

Mizumachi M.

European Signal Processing Conference 666 - 670 2010年12月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

This paper proposes to estimate the reliability of each direction-of-arrival (DOA) estimate explicitly. Observed spatial features, which give DOA estimates, are usually distorted by acoustical noise and reverberation. Then, distorted spatial features are filtered out by particle filters with a system model, which describes sound source dynamics on spatial state space. After particle filtering is carried out, set of the weighted particles can be considered as the discrete approximation of the true spatial feature. It is reasonable that particles with large weights concentrate around the true DOA on spatial state space. In this paper, crest factor, skewness, and kurtosis are calculated based on the second, third, and forth order moments from the particle set for representing how particles concentrate simply. It is confirmed that the skewness confidence measure is the most suitable for estimating DOA reliability under both directional and distributed noise conditions. © EURASIP, 2010.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84863793327&origin=inward

Speech enhancement based on spatial and acoustical features in non-stationary noisy environments 査読有り

Mizumachi M.

17th International Congress on Sound and Vibration 2010, ICSV 2010 5 3603 - 3608 2010年12月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

Speech enhancement is essential for many audio applications with speech input. Those devices are supposed to be used under a wide variety of acoustic conditions. Then, speech enhancement should be effective even in unknown, time-varying, and non-stationary noisy environments. The author has proposed an efficient noise reduction method, which aims at making the most use of stochastic spatial information [Mizumachi, Proc. ICSV16, 2009]. The spatial and acoustical features have been prepared based on directions of sound sources and spectral characteristics of a target and acoustical interferences, respectively. In this paper, those stochastic features are extended including much richer spatial information. New spatial features are derived from higher order statistics of the spatial information extracted from the 2-ch observations. The proposed method can deal with non-stationary noisy environments, because it takes the suitable scheme in noise reduction relying on the spatial and acoustical features flexibly. In this paper, the feasibility of the proposed method is investigated under non-stationary noisy environments.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84871439200&origin=inward

Self-organized approach to noise reduction based on acoustic and spatial features 査読有り

Mizumachi M.

39th International Congress on Noise Control Engineering 2010, INTER-NOISE 2010 7 5645 - 5652 2010年12月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

Noise reduction has been one of important issues in acoustic signal processing. Until now, various kinds of algorithms have been proposed to achieve noise reduction, but we have not established an almighty algorithm, which always works well in the real world. Each algorithm has both advantage and disadvantage in noise reduction. In this paper, a practical approach to noise reduction is proposed by introducing an idea of self-organization for determining the suitable algorithm in time-varying acoustic scenes based on acoustic and spatial features. Time evolution of acoustic and spatial features and reliability of direction-of-arrival estimates are used for selecting the suitable approach out of spectral subtraction with acoustic features and beamforming with spatial features. Feasibility of the self-organized approach has been confirmed under non-stationary noise environments.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84868667154&origin=inward

Confidence measure based on higher-order statistics for spatial feature in particle filtering 査読有り

Mizumachi M.

2010 World Automation Congress, WAC 2010 2010年12月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=78651417406&origin=inward

Adaptive source separation based on reliability of spatial feature using multi-channel acoustic observations 査読有り

Mizumachi M.

128th Audio Engineering Society Convention 2010 2 1092 - 1098 2010年12月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

Separation of sound source can be achieved by spatial filtering with multi-channel acoustic observations. However, the right algorithm should be prepared in each condition of acoustic scene. It is difficult to provide the suitable algorithm under real acoustic environments. In this paper, an adaptive source separation scheme is proposed based on the reliability of a spatial feature, which gives an estimate of direction of arrival (DOA). As confidence measures for DOA estimates, the third and fourth moments for spatial features are employed to measure how sharp the main-lobes of spatial features are. This paper proposes to selectively use either spatial filters or frequency-selective filters without spatial filtering depending on the reliability of each DOA estimate.