简介:Inordertoresearchsurfacecharacteristicsofliquidmetalwhichbasedonhydrodynamicstheoryandinteractionprinciplebetweenlaserwaveandsurfacetensionstationarywaveofliquidmetal,itisfinishedtoprovideanentirelytheoreticalequation,apparatusandtechniquesofthemeasuringthesurfacetensionforliquidmetalathightemperature.Thisapparatusisanautomaticmeasuringsystemforinformationcollectingandprocessing,inwhichacomputerisattachedtochargecoupleddevice,drivingcircuitsandA/Dconverter.ApplyingthenewapparatusthesurfacetensionvaluesofSn-Pbbasedbinaryandternaryalloysoldershavebeenobtainedonthebasisoftheabovework.TheexperimentalresultsindicatethatthesurfacetensionvaluesofliquidmetalSn-Pbbasedsoldersaredecreasedby37.3%and31.9%respectively.When1%ofBiorInisaddedinthesolders,theelementsBiandInhaveevidentsurfaceactivityinSn-Pbbasedsoldersandthewettabilityofthesoldersinincreasedwithincreasingsurfaceactivity.
简介:TheperformanceofthetraditionalVoiceActivityDetection(VAD)algorithmsdeclinessharplyinlowerSignal-to-NoiseRatio(SNR)environments.Inthispaper,afeatureweightinglikeli-hoodmethodisproposedfornoise-robustVAD.Thecontributionofdynamicfeaturestolikelihoodscorecanbeincreasedviathemethod,whichimprovesconsequentlythenoiserobustnessofVAD.Divergencebaseddimensionreductionmethodisproposedforsavingcomputation,whichreducesthesefeaturedimensionswithsmallerdivergencevalueatthecostofdegradingtheperformancealittle.ExperimentalresultsonAuroraIIdatabaseshowthatthedetectionperformanceinnoiseenvironmentscanremarkablybeimprovedbytheproposedmethodwhenthemodeltrainedincleandataisusedtodetectspeechendpoints.Usingweightinglikelihoodonthedimension-reducedfeaturesobtainscom-parable,evenbetter,performancecomparedtooriginalfull-dimensionalfeature.