SPEECH ENHANCEMENT USING CONSTRAINED SPECTRAL AMPLITUDE SUBTRACTION BASED ON NONCAUSAL A PRIORI SNR

Two gain forms of spectral amplitude subtraction are derived theoretically without neglecting the correlation of speech and noise spectrum during the period of a fralne. In the implementation, the constrained gain is expressed as a function of noncausal a priori SNR （Signal-to-Noise Ratio）. Noise and noncausal a priori SNR are estimated from the multitaper spectrum of the noisy signal with algorithms modified to be suitable for the multitaper spectruln. Objective evaluations show that in case of white Gaussian noise the proposed method outperforms some methods based on LSA （Log Spectral Amplitude） in terms of MBSD （Modified Bark Spectral Distortion）, segmental SNR and overall SNR, and informal listening tests show that speech reconstructed in this way has little speech distortion and musical noise is nearly inaudible even at low SNR.

Noise Removal in Speech Processing Using Spectral Subtraction

Spectral subtraction is used in this research as a method to remove noise from noisy speech signals in the frequency domain. This method consists of computing the spectrum of the noisy speech using the Fast Fourier Transform (FFT) and subtracting the average magnitude of the noise spectrum from the noisy speech spectrum. We applied spectral subtraction to the speech signal “Real graph”. A digital audio recorder system embedded in a personal computer was used to sample the speech signal “Real graph” to which we digitally added vacuum cleaner noise. The noise removal algorithm was implemented using Matlab software by storing the noisy speech data into Hanning time-widowed half-overlapped data buffers, computing the corresponding spectrums using the FFT, removing the noise from the noisy speech, and reconstructing the speech back into the time domain using the inverse Fast Fourier Transform (IFFT). The performance of the algorithm was evaluated by calculating the Speech to Noise Ratio (SNR). Frame averaging was introduced as an optional technique that could improve the SNR. Seventeen different configurations with various lengths of the Hanning time windows, various degrees of data buffers overlapping, and various numbers of frames to be averaged were investigated in view of improving the SNR. Results showed that using one-fourth overlapped data buffers with 128 points Hanning windows and no frames averaging leads to the best performance in removing noise from the noisy speech.

Single-Channel Speech Enhancement Based on Improved Frame-Iterative Spectral Subtraction in the Modulation Domain

Aiming at the problem of music noise introduced by classical spectral subtraction,a shorttime modulation domain(STM)spectral subtraction method has been successfully applied for singlechannel speech enhancement.However,due to the inaccurate voice activity detection(VAD),the residual music noise and enhanced performance still need to be further improved,especially in the low signal to noise ratio(SNR)scenarios.To address this issue,an improved frame iterative spectral subtraction in the STM domain(IMModSSub)is proposed.More specifically,with the inter-frame correlation,the noise subtraction is directly applied to handle the noisy signal for each frame in the STM domain.Then,the noisy signal is classified into speech or silence frames based on a predefined threshold of segmented SNR.With these classification results,a corresponding mask function is developed for noisy speech after noise subtraction.Finally,exploiting the increased sparsity of speech signal in the modulation domain,the orthogonal matching pursuit(OMP)technique is employed to the speech frames for improving the speech quality and intelligibility.The effectiveness of the proposed method is evaluated with three types of noise,including white noise,pink noise,and hfchannel noise.The obtained results show that the proposed method outperforms some established baselines at lower SNRs(-5 to +5 dB).

Single-Channel Speech Enhancement Using Critical-Band Rate Scale Based Improved Multi-Band Spectral Subtraction

This paper addresses the problem of single-channel speech enhancement in the adverse environment. The critical-band rate scale based on improved multi-band spectral subtraction is investigated in this study for enhancement of single-channel speech. In this work, the whole speech spectrum is divided into different non-uniformly spaced frequency bands in accordance with the critical-band rate scale of the psycho-acoustic model and the spectral over-subtraction is carried-out separately in each band. In addition, for the estimation of the noise from each band, the adaptive noise estimation approach is used and does not require explicit speech silence detection. The noise is estimated and updated by adaptively smoothing the noisy signal power in each band. The smoothing parameter is controlled by a-posteriori signal-to-noise ratio (SNR). For the performance analysis of the proposed algorithm, the objective measures, such as, SNR, segmental SNR, and perceptual evaluations of the speech quality are conducted for the variety of noises at different levels of SNRs. The speech spectrogram and objective evaluations of the proposed algorithm are compared with other standard speech enhancement algorithms and proved that the musical structure of the remnant noise and background noise is better suppressed by the proposed algorithm.

基于机器学习的语音增强技术

主要研究基于机器学习的语音增强技术,以提升语音信号的质量。首先,介绍基于机器学习的语音增强系统框架。其次,详细探讨谱减法与深度神经网络(Deep Neural Network,DNN)相结合的语音增强方法的数学原理。最后,采用NOISEX-92数据集测试与评估提出的方法。实验结果表明,基于谱减法与DNN的语音增强方法在提升信噪比和语音清晰度方面取得显著的效果,能够有效提升语音通信质量。

改进粒子滤波跟踪的视听双模态语音识别仿真

噪声环境下视听语音不易被识别,为提升语音识别效果,提出改进粒子滤波跟踪的视听双模态语音识别方法。采用谱减法去除噪声数据,完成视听双模态语音的消噪处理;根据人语和唇动信息之间的相关性,采用改进粒子滤波跟踪方法提取视听双模态语音特征信息,构建transformer语音识别模型,将提取的特征信息输入到模型内实施并行训练,实现视听双模态语音的有效识别。实验结果表明,通过对上述方法开展信噪比测试、识别性能测试,验证了上述方法的可行性高、可靠性强。

肺音信号去噪技术的比较与分析

肺音听诊是医生诊断各种呼吸系统疾病的一种简单、无创的方法,可为检测和鉴别呼吸病理提供及时而有用的信息。然而,肺音中的噪声污染是采集肺音时不可避免的问题,在医生诊断患者时,噪声的存在会对肺音分析产生严重阻碍,因此有效去除多余噪声更有助于准确、客观地评价肺音信息。论文对不同去噪技术进行了讨论,包括小波去噪、最小均方自适应滤波、谱减法,评估了三种去噪性能指标,得出了相关比较性结论。

基于角度压制比谱减的环境自适应双麦语音增强

[目的]针对智能终端小型化、使用场景多样化的发展趋势,研制一种既能满足严苛的尺度、算力、存储空间限制,又能实现环境自适应的双麦语音增强算法.[方法]考虑到麦克风阵列波束形成算法可以增强期望方向信号,同时抑制非期望方向的噪声,但小尺寸阵列波束主瓣波束宽度较宽、影响增强效果.在小尺寸双麦对目标方向进行波束对准增强的基础上,参考干扰方向噪声,进一步对目标方向语音进行谱减处理,并引入角度压制比实时检测干扰方向噪声的能量估计,实现对不同混响、噪声类型的自适应处理,从而提升语音增强效果.[结果]角度压制比随混响时间增加而增大,与信噪比不相关.相对于原始带噪信号、滤波-累加波束形成(filter-and-sum beamforming, FSB)信号、FSB结合固定对向谱减的语音增强信号,通过FSB结合角度压制比自适应对向谱减得到的语音增强信号,在不同噪声类型、不同信噪比和不同混响时间下,均能得到最高的分段信噪比得分和大多数的最高客观语音质量评估得分.[结论]角度压制比能一定程度地反映不同的混响情况,利用角度压制比得到的谱减阈值具有一定的环境适应性.

一种自适应的地面震动信号降噪方法

针对传统谱减法应对地面震动信号的干扰和噪声时降噪效果差、信号失真的问题,研究和分析不同类型干扰和噪声的功率谱特性,提出一种基于干扰和噪声的功率谱特性自适应选择降噪处理操作的方法,既满足了不同干扰和噪声的降噪要求,也避免了降噪处理造成的信号失真。

基于MobileNetV3卷积神经网络的供水管道漏损音频分类

为了对城市供水管网漏损音进行准确识别,提出一种基于MobileNetV3的供水管道漏损音频分类识别方法。首先将ROPP数据集中的音频文件进行离线数据增强,将漏损信号转变为对数梅尔谱图并采用谱减法实现数据降噪;然后使用注意力机制模块与MobileNetV3网络训练识别并提取图像特征;最后使用Softmax函数对漏损音频进行分类。实验结果表明,该方法可以使漏水类别的分类精确度达到99.40%,召回率达到99.20%。

强噪环境下分频段数字音频信号精细化采集研究

在强噪环境下,为了获取高质量、低失真度的数字音频信号,提出一种强噪环境下分频段数字音频信号精细化采集方法。通过麦克风设备获取音频信号,利用LM4550芯片对其作采样、编码等处理后生成数字音频信号,基于AC-97单元接收数字音频信号,利用频段分割器对其作频带分解,获得互不重叠子频段数字音频信号。采用多窗谱谱减法对其去噪,利用数据通信接口将其传输给LM4550芯片,在完成模拟信号转换后,通过耳机输出,实现数字音频信号精细化采集。实验结果表明,该方法处理后的各子频段数字音频信号有用信息得以完整保留,并提高了信号波形的规整度和规律性,强噪声环境下数字音频信号的PESQ指标达到4.08以上,最大失真度为3.74%。

基于梅尔频率倒谱系数的语音清晰度DRT识别

语音清晰度在通信终端、设备系统语音识别方面具有重要意义。本文对110dB噪声干扰下采集到的语音信号进行谱减法降噪,双门限端点检测提取发音字段,然后提取梅尔频率倒谱系数(MFCC),再将其进行差分计算,得到一阶和二阶分量,结合短时能量作为语音信号的特征参数,最后通过动态时间归整(DTW)进行相似度识别。实验表明,本文算法对汉语清晰度诊断押韵测试(DRT)字表的测试结果高达92.90%,有良好的识别率。

基于声学检测技术的刀片储能锂电池热失控行为研究

随着电化学储能的规模化应用,开展储能锂电池热失控检测与及时预警研究对于保障储能电站的安全运行具有重要意义。搭建储能锂电池单体热失控性能检测试验平台,对刀片储能锂电池进行热失控条件下的声音信号检测,分析电池鼓包、泄压阀打开、泄气、爆炸起火等热失控不同发展阶段的声音特性与变化规律,表明声音特征对于热失控检测的有效性。针对热失控声音特征易受储能电池舱内通风散热风机、储能变流器等其他设备噪声干扰的问题,提出基于小波包分解与谱减法语音增强的热失控声音信号抗干扰分析方法。分析结果表明,所提出的方法能够有效剔除热失控声音信号中的风噪干扰,可准确还原出电池热失控声音特征,还原声信号与实际声信号相似系数达到0.96,为储能锂电池热失控声音检测与早期预警提供技术参考。

Application of gemstone spectral imaging for efficacy evaluation in hepatocellular carcinoma after transarterial chemoembolization

AIM: To assess the value of gemstone spectral imaging (GSI) in efficacy evaluation in hepatocellular cancer (HCC) after transcatheter arterial chemoembolization (TACE) treatment.METHODS: Thirty patients with HCC underwent GSI, including nonenhanced, arterial, portalvenous and delayed phase scans, after TACE treatment. Arterial phase images were acquired with GSI for reconstruction of virtual nonenhanced images and color overlay images. Digital subtraction angiography (DSA) was performed in all these patients. Two blinded and independent readers evaluated the data in two reading sessions; standard nonenhanced, arterial, portalvenous, and delayed phase images were read in session A, and the optimal monochromatic images, iodine/water based images and spectrum features were read in session B. Sensitivity and specificity were calculated with the DSA data as the reference standard. The sensitivity and specificity were compared using the χ2 test.RESULTS: DSA revealed 154 lesions in 30 patients, and 100 of them had blood supply. Overall sensitivity and specificity were 72% (72/100) and 77.8% (42/54) for session A, and 97% (97/100) and 94.4% (51/54) for session B, respectively. The sensitivity and specificity of the two reading sessions were significantly different (χ2 = 23.04, χ2 = 7.11, P < 0.05).CONCLUSION: Compared with conventional CT, GSI could significantly improve the detection of small and multiple lesions without increasing the radiation dose. Based on spectrum features, GSI could assess tumor homogeneity and more accurately identify residual tumors and recurrent or metastatic lesions during efficacy evaluation and follow-up in HCC after TACE treatment.

Speech detection method based on a multi-window analysis

Aiming at the poor performance of speech signal detection at low signal-to-noise ratio(SNR),a method is proposed to detect active speech frames based on multi-window time-frequency(T-F)diagrams.First,the T-F diagram of the signal is calculated based on a multi-window T-F analysis,and a speech test statistic is constructed based on the characteristic difference between the signal and background noise.Second,the dynamic double-threshold processing is used for preliminary detection,and then the global double-threshold value is obtained using K-means clustering.Finally,the detection results are obtained by sequential decision.The experimental results show that the overall performance of the method is better than that of traditional methods under various SNR conditions and background noises.This method also has the advantages of low complexity,strong robustness,and adaptability to multi-national languages.

强噪环境基于谱减法的录音数字音频信号降噪

由于强噪环境中存在大量的冗余信号和噪声信号,且噪声信号种类繁杂,影响信号降噪效果,为此提出基于谱减法的录音数字音频信号降噪方法。建立联合稀疏录音数字音频信号重构模型,根据目标函数得出含噪信号与正常信号的峰值变化规律。为了降低算法误差,引入交叉项估计法对长时信噪比和瞬时信噪比实施交叉替代,使算法在低、高噪声环境下都能实现高效适应。以谱减法降噪理念,分别计算含有和不含有噪声的频谱密度值,求解得到二者间的相位差值,通过差值剔除噪声信号,实现高效降噪。仿真结果表明,所提方法所需运算量较少,去噪效率更高,在不同噪声环境中均能完成精准降噪,且不会破坏原始信号细节,实用价值较高。

基于SOM神经网络的摩擦状态识别研究

为提取摩擦振动的特征和实现摩擦副摩擦状态的识别,在往复摩擦磨损试验机进行摩擦副混合摩擦和干摩擦状态的摩擦磨损试验。应用谱减法对试验采集的摩擦振动信号进行降噪,计算降噪后的摩擦振动15个特征参数。应用自组织映射(Self-organizing map, SOM)神经网络对摩擦副不同摩擦状态的摩擦振动特征参数进行分析,得到摩擦振动的SOM神经网络神经元分类。研究结果表明,谱减法能消除摩擦磨损试验机的背景噪声,SOM神经网络算法能够有效分析摩擦振动信号的特征,实现摩擦副摩擦状态的识别。

光纤监测系统入侵信号的检测和识别

为改善光纤监测系统监测信号特征复杂及入侵信号检测效率和识别精度低等问题,提出了一种改进的谱减法进行去噪处理,从而有效减少噪声干扰并获得具有高信噪比的监测信号;其次,由于光纤监测信号大部分为静默信号,因此对其进行端点检测从而有效提取入侵信号;最后,基于提取的入侵信号的时频域特征,提出了一种多特征融合深度学习识别模型,从而准确识别入侵类型.

基于谱减法的隧道地震波超前探测去噪方法

为快速有效地去除隧道地震超前探测数据中具有随机、频带广和能量强特征的噪声,提升偏移成像质量,进而有效探明隧道掌子面前方灾害源,充分利用谱减法在压制随机噪声上的优势,将其引入隧道地震波探测数据处理中,形成基于谱减法的隧道地震波超前探测去噪方法。该去噪方法可以有效去除随机噪声且计算效率较高,可以满足隧道的快速施工要求。通过对典型不良地质模型进行数值模拟,表明利用谱减法可有效压制隧道环境中的随机噪声干扰,地震记录中反射波信号信噪比更高,在此基础上的偏移成像结果伪像干扰更少,可有效提升探测准确性。将该方法应用至实际隧道地震探测中,成功探明了隧道掌子面前方的岩体破碎带分布位置和范围,证明应用该方法进行超前探测具有有效性和实用性。

一种舰载低信噪比环境下的音频端点检测算法

针对舰载环境下音频端点检测准确率及鲁棒性较低的问题,提出了一种谱减法和朴素贝叶斯分类器相结合的音频端点检测算法。首先提取纯净音频信号MFCC0与GFCC0构建融合特征,与能熵比特征一同作为朴素贝叶斯分类器的输入进行训练及建模,再利用多窗谱谱减法提升待测含噪信号信噪比,提取信号相关特征,朴素贝叶斯分类器根据待测信号特征判断该信号的类别。仿真实验结果表明,该算法针对舰载低信噪比含噪音频信号与传统方法相比有效降低了虚检和漏检,具有更好的准确性及鲁棒性。