Cross-power Spectrum Phase Research Articles

The Combination of Spectrum Subtraction and Cross-power Spectrum Phase Method for Time Delay Estimation

In order to solve the problem of large error of delay estimation in low SNR environment, a new delay estimation method based on cross power spectral frequency domain weighting and spectrum subtraction is proposed. Through theoretical analysis and MATLAB simulation, among the four common weighting functions, it is proved that the cross-power spectral phase weighting method has a good sharpening effect on the peak value of the cross-correlation function, and it is verified that the improved spectral subtraction method generally has a good noise reduction effect under different SNR environments. Finally, the joint simulation results of the whole algorithm show that the combination of spectrum subtraction and crosspower spectrum phase method can effectively sharpen the peak value of cross-correlation function and improve the accuracy of time delay estimation in the low SNR environment. The results of this paper can provide useful help for sound source localization in complex environments.

Two High-Precision Ultrashort Baseline Location Methods Based on Phase Difference

In this paper, two ultra-short baseline location methods for underwater broadband signal are proposed. In order to improve the positioning accuracy of acoustic positioning, the coarse time delay is measured by parabolic interpolation of cross-correlation results. According to the coarse time delay, the ambiguity number of the phase at the cross-correlation peak is solved and the ambiguity is solved twice, so as to obtain the high-precision time delay difference and improve the positioning accuracy. This method is suitable for the case without near body multipath. Aiming at the near body multipath problem, its influence on ambiguity resolution and positioning accuracy is analyzed. Based on the coarse time delay, adjusting time delay and the Fourier transform of cross-power spectrum phase joint method is proposed to further improve the accuracy of coarse time delay, so as to improve the accuracy of ambiguity resolution. On this basis, the location information feedback and real-time attitude information are used to estimate the coarse time delay difference, and the ambiguity number is further constrained to further improve the accuracy of ambiguity resolution. In order to reduce the influence of multipath on the phase and improve the positioning accuracy, a Kaiser window is added to the source and the phase difference is changed to the phase difference at the first point of threshold. Numerical simulation and field test demonstrate that the first algorithm proposed in this paper improves the positioning accuracy and the accuracy of ambiguity resolution. The second algorithm improves the accuracy of ambiguity resolution and positioning accuracy in the presence of near body multipath.

듀얼 마이크를 이용한 교차 파워 스펙트럼 위상 기반의 음성 구간 검출 알고리즘

듀얼 마이크를 이용한 교차 파워 스펙트럼 위상 기반의 음성 구간 검출 알고리즘

Frequency-Sliding Generalized Cross-Correlation: A Sub-Band Time Delay Estimation Approach

The generalized cross-correlation(GCC) is regarded as the most popular approach for estimating the time difference of arrival (TDOA) between the signals received at two sensors. Time delay estimates are obtained by maximizing the GCC output, where the direct-path delay is usually observed as a prominent peak. Moreover, GCCs play also an important role in steered response power (SRP) localization algorithms, where the SRP functional can be written as an accumulation of the GCCs computed from multiple sensor pairs. Unfortunately, the accuracy of TDOA estimates is affected by multiple factors, including noise, reverberation and signal bandwidth. In this paper, a sub-band approach for time delay estimation aimed at improving the performance of the conventional GCC is presented. The proposed method is based on the extraction of multiple GCCs corresponding to different frequency bands of the cross-power spectrum phase in a sliding-window fashion. The major contributions of this paper include: 1) a sub-band GCC representation of the cross-power spectrum phase that, despite having a reduced temporal resolution, provides a more suitable representation for estimating the true TDOA; 2) such matrix representation is shown to be rank one in the ideal noiseless case, a property that is exploited in more adverse scenarios to obtain a more robust and accurate GCC; 3) we propose a set of low-rank approximation alternatives for processing the sub-band GCC matrix, leading to better TDOA estimates and source localization performance. An extensive set of experiments is presented to demonstrate the validity of the proposed approach.

Positioning Method for Various Indoor Sound Sources by Detecting Reception Time Difference using Cross-power Spectrum Phase Analysis and its Experimental Evaluation

Positioning Method for Various Indoor Sound Sources by Detecting Reception Time Difference using Cross-power Spectrum Phase Analysis and its Experimental Evaluation

Sound Source Localization Technology Based on Generalized Cross-Power Spectrum Phase

Sound Source Localization Technology Based on Generalized Cross-Power Spectrum Phase

Time Delay Estimation Technique Based on Cross Correlation Function

Time delay estimation is the key to acoustic source localization based on time delay, and the time difference caused by the difference of the same signal arriving at different microphone distances is estimated. The main algorithms now include generalized phase days method, adaptive least mean method, the cross power spectrum phase method and so on, all of them get the time delay essentially by seeking the maximum position of the cross-correlation function. This paper discusses several of the above algorithms, and compares their advantages and disadvantages.

Localizing speakers in multiple rooms by using Deep Neural Networks

In the field of human speech capturing systems, a fundamental role is played by the source localization algorithms. In this paper a Speaker Localization algorithm (SLOC) based on Deep Neural Networks (DNN) is evaluated and compared with state-of-the art approaches. The speaker position in the room under analysis is directly determined by the DNN, leading the proposed algorithm to be fully data-driven. Two different neural network architectures are investigated: the Multi Layer Perceptron (MLP) and Convolutional Neural Networks (CNN). GCC-PHAT (Generalized Cross Correlation-PHAse Transform) Patterns, computed from the audio signals captured by the microphone are used as input features for the DNN. In particular, a multi-room case study is dealt with, where the acoustic scene of each room is influenced by sounds emitted in the other rooms. The algorithm is tested by means of the home recorded DIRHA dataset, characterized by multiple wall and ceiling microphone signals for each room. In detail, the focus goes to speaker localization task in two distinct neighboring rooms.As term of comparison, two algorithms proposed in literature for the addressed applicative context are evaluated, the Crosspower Spectrum Phase Speaker Localization (CSP-SLOC) and the Steered Response Power using the Phase Transform speaker localization (SRP-SLOC). Besides providing an extensive analysis of the proposed method, the article shows how DNN-based algorithm significantly outperforms the state-of-the-art approaches evaluated on the DIRHA dataset, providing an average localization error, expressed in terms of Root Mean Square Error (RMSE), equal to 324 mm and 367 mm, respectively, for the Simulated and the Real subsets.

複数フレームのCSP係数を用いたTDOA推定による複数人物定位·追跡

In this paper, multi-speaker localization methods using voice signal are proposed. Sound source localization systems consisting of many microphones have been widely researched. In our laboratory, we have researched sound source localization of persons based on TDOA (Time Difference of Arrival) between distributed microphones for the purpose of efficient use of air conditioner and lighting in the office . Localization systems based on sound such as footsteps or voice are superior to those based on image data in terms of privacy-consciousness and utility in the dark. As sound source localization methods, beamforming, subspace methods and TDOA-based methods are widely used. CSP (Cross-power Spectrum Phase) analysis is one of the most famous method to estimate TDOA. We have researched sound source localization methods using CSP analysis because it spends less calculation cost than beamforming or subspace methods. However, CSP analysis doesn't work well by itself in the real office environment because of sound lasting for a long time, reverberations, low SNR, and so on. The main cause of difficulty in sound source localization in such a case is the “weak correlation” between different sound signals. In this paper, it is shown that the bad effect of the “weak correlation” can be improved by our proposed methods, Wiener-filter-based CSP analysis and H-CSP (Histogram-CSP) analysis. H-CSP analysis is a statistics data analysis method for CSP coefficients which uses the histogram of CSP coefficients. In this research, the effectiveness of proposed methods are shown by sound source localization experiments in the case sound sources are stable and the case sound sources are moving.

A Time Delay Estimation Algorithm Based on the Weighted Correntropy Spectral Density

Accuracy and resistance to noise are two important aspects of time delay estimation algorithms. The cross power-spectrum phase algorithm is a classic time delay estimation algorithm. It is accurate but not robust to noise. Its performance deteriorates or even fails under impulsive noise. The improved cross power-spectrum phase algorithms based on fractional lower-order statistics rely on a priori knowledge of impulsive noise. In this paper, we propose a novel time delay estimation algorithm based on the weighted correntropy spectral density. The proposed method works well under impulsive noise because it is based on correntropy theory and its resistance to noise is further strengthened by the weighting process in the frequency domain. Our theoretical analysis and simulations demonstrate that the proposed method is very accurate in the presence of Gaussian or impulsive noise with a low generalized signal-to-noise ratio.

An Improved Image Stitching Algorithm

Proposed a new stitching algorithm to suit for coal mine. Cross-power spectrum phase information is used to coarse matching fast and overlapping regions can be found. Invariant features are extracted by using the dimensional reduced SIFT algorithm-sample principal component analysis. RANSAC combined with iterative L-M nonlinear optimization algorithm can improve projection matrix. New algorithm can improve the matching efficiency, and it’s suit for coal mine environment.

Direction of arrival estimation by cross-power spectrum phase analysis using prior distributions and voice activity detection information

Direction of arrival estimation by cross-power spectrum phase analysis using prior distributions and voice activity detection information

Artifact Reduction in Tomographic Images for Nondestructive Testing of Square Billets Using Ultrasonic Computerized Tomography

To reduce the intensity of artifacts in visualized images for detecting defects in square billets by ultrasonic computerized tomography, we propose a method, which utilizes the crosspower-spectrum phase (CSP) function between a measurement signal and a reference signal, to measure the time-of-flight (TOF) difference. The artifacts arise owing to measurement errors of the TOF difference, which are caused by the effects of propagation characteristics and side lobes in correlation waveforms. The effects of propagation characteristics may be reduced using the reference signal, which contains the same characteristics as those of the measurement signal. The side lobes in correlation waveforms may also be decreased using the CSP method. The results obtained by experiment suggest that the intensity of artifacts was reduced because the proposed method reduced the side lobe level by about -14 dB and the number of measurement errors. The visualized images showed that the defects whose diameter was more than 2.0 mm were detected using the input signal whose wavelength was in the range from 2 to 6 mm.

Speech Activity Detection for Multi-Party Conversation Analyses Based on Likelihood Ratio Test on Spatial Magnitude

This paper proposes a microphone array-based speech activity detection (SAD) method for analyzing multi-party conversations recorded in the presence of noise. In particular, the proposed method considers conversations where the number of speakers and speaker locations cannot be restricted, such as when standing and talking, and at poster sessions. When we observe such conversations, there are directional noise sources and diffuse noise that affect the direction of arrival estimations of the target speech signals. To detect speech activity without a priori knowledge about the speakers and noise environments, a likelihood ratio test (LRT)-based SAD method is applied to spatial magnitude, which are estimated by using the time-frequency masking of the observed spectra. The proposed method can exploit the enhanced signals obtained from time-frequency masking, and works even in the presence of environmental noise. Experiments with recorded simulated poster sessions confirmed that the proposed method could outperform conventional methods based on the LRT for a single channel, magnitude coherence, or crosspower spectrum phase.

DOA Estimation with Local-Peak-Weighted CSP

This paper proposes a novel weighting algorithm for Cross-power Spectrum Phase (CSP) analysis to improve the accuracy of direction of arrival (DOA) estimation for beamforming in a noisy environment. Our sound source is a human speaker and the noise is broadband noise in an automobile. The harmonic structures in the human speech spectrum can be used for weighting the CSP analysis, because harmonic bins must contain more speech power than the others and thus give us more reliable information. However, most conventional methods leveraging harmonic structures require pitch estimation with voiced-unvoiced classification, which is not sufficiently accurate in noisy environments. In our new approach, the observed power spectrum is directly converted into weights for the CSP analysis by retaining only the local peaks considered to be harmonic structures. Our experiment showed the proposed approach significantly reduced the errors in localization, and it showed further improvements when used with other weighting algorithms.

Talker Localization Based on Interference between Transmitted and Reflected Audible Sound

In many engineering fields, distance to targets is very important. General distance measurement method uses a time delay between transmitted and reflected waves, but it is difficult to estimate the short distance. On the other hand, the method using phase interference to measure the short distance has been known in the field of microwave radar. Therefore, we have proposed the distance estimation method based on interference between transmitted and reflected audible sound, which can measure the distance between microphone and target with one microphone and one loudspeaker. In this paper, we propose talker localization method based on distance estimation using phase interference. We expand the distance estimation method using phase interference into two microphones (microphone array) in order to estimate talker position. The proposed method can estimate talker position by measuring the distance and direction between target and microphone array. In addition, talker's speech is regarded as a noise in the proposed method. Therefore, we also propose combination of the proposed method and CSP (Cross-power Spectrum Phase analysis) method which is one of the DOA (Direction Of Arrival) estimation methods. We evaluated the performance of talker localization in real environments. The experimental result shows the effectiveness of the proposed method.

Binaural Active Audition for Humanoid Robots to Localise Speech over Entire Azimuth Range

We applied motion theory to robot audition to improve the inadequate performance. Motions are critical for overcoming the ambiguity and sparseness of information obtained by two microphones. To realise this, we first designed a sound source localisation system integrated with cross-power spectrum phase (CSP) analysis and an EM algorithm. The CSP of sound signals obtained with only two microphones was used to localise the sound source without having to measure impulse response data. The expectation-maximisation (EM) algorithm helped the system to cope with several moving sound sources and reduce localisation errors. We then proposed a way of constructing a database for moving sounds to evaluate binaural sound source localisation. We evaluated our sound localisation method using artificial moving sounds and confirmed that it could effectively localise moving sounds slower than 1.125 rad/s. Consequently, we solved the problem of distinguishing whether sounds were coming from the front or rear by rotating and/or tipping the robot's head that was equipped with only two microphones. Our system was applied to a humanoid robot called SIG2, and we confirmed its ability to localise sounds over the entire azimuth range as the success rates for sound localisation in the front and rear areas were 97.6% and 75.6% respectively.

Human Tracking System Integrating Sound and Face Localization Using an Expectation-Maximization Algorithm in Real Environments

We have developed a human tracking system for use by robots that integrate sound and face localization. Conventional systems usually require many microphones and/or prior information to localize several sound sources. Moreover, they are incapable of coping with various types of background noise. Our system, the cross-power spectrum phase analysis of sound signals obtained with only two microphones, is used to localize the sound source without having to use prior information such as impulse response data. An expectation-maximization (EM) algorithm is used to help the system cope with several moving sound sources. The problem of distinguishing whether sounds are coming from the front or back is also solved with only two microphones by rotating the robot's head. A developed method that uses facial skin colors classified by another EM algorithm enables the system to detect faces in various poses. It can compensate for the error in the sound localization for a speaker and also identify noise signals entering from undesired directions by detecting a human face. A developed probability-based method is used to integrate the auditory and visual information in order to produce a reliable tracking path in real-time. Experiments using a robot showed that our system can localize two sounds at the same time and track a communication partner while dealing with various types of background noise.

An evaluation of hands‐free voice activity detection algorithm based on weighted cross‐power spectrum phase analysis and zero crossing detector

VAD (voice activity detection) is necessary for recognizing hands‐free speech. However, conventional VADs could not achieve enough detection performance in real noisy environments because they utilize only time‐sequence information or only spatial‐sequence information of captured signals. To cope with this problem, we have already proposed noise‐robust hands‐free VAD by composing the time‐sequence and spatial‐sequence information. The conventional ZCD (zero crossing detector) was employed to acquire the time‐sequence information, and weighted CSP (cross‐power spectrum phase) analysis that was a specialized DOA (direction of arrival) estimation algorithm for speech signal was also employed to acquire the spatial‐sequence information. However, this algorithm can insufficiently improve the VAD performance in higher noisy environments. To overcome this problem, in this paper, we try to utilize talker direction information and attempt the environmental adaptation of our former algorithm. As a result of evaluation experiments in real acoustic environment, the proposed hands‐free VAD could achieve a more effective performance than the conventional VAD and our former VAD. [Work supported by MEXT of Japan.]

A study of robust omnidirectional audio‐visual talker localization algorithm with microphone array and omnidirectional image

In video conferencing environments, it is very important to localize the talker. However, conventional audio signal‐based algorithms often suffer from audio interference, and conventional visual signal‐based algorithms fail in the presence of visual interference. To deal with these problems, this paper proposes a robust omnidirectional audio‐visual talker localization algorithm that not only exploits audio feature parameters, but also subordinately uses visual feature parameters. To achieve omnidirectional audio‐visual talker localization, paired‐omnidirectional microphones are employed as an audio sensor, and an omnidirectional camera is employed as a visual sensor. For robust talker localization, audio feature parameters are extracted using weighted cross‐power spectrum phase (CSP) analysis and CSP coefficient subtraction, and visual feature parameters are extracted using background subtraction and skin‐color detection. The talker is finally located by the fusing of weighted audio/visual feature parameters, and the weight of this feature parameter fusion is automatically controlled based on the reliable criterion of audio feature parameters. The results of localization experiments in an actual room revealed that the proposed audio‐visual talker localization algorithm is superior to that of conventional localizers using only audio or visual feature parameters, but not both. [Work supported by MEXT of Japan.]