This Page

The parameters for computing delays.

Module doubletParamIO


#parametre de florent DictOfParamComputationDoublet = {

‘freqSamplingRate’:20.0, ‘lengthCorrelation’: 4801,#length of the correlation ‘maxPointsInCoda’:800,#Number of points in the coda from the middle of the correlation ‘indexBegin’:1600, #index of the beginning of the first temporal windows ‘lengthTemporalWindows’:210, #length of the temporal windows in points ‘shiftTemporalWindows’:16, #shift of the temporal windows ‘lengthTaperCos’:10, #length of the taper of the temporal windows for each border ‘lenghtFft’:1024, #length of the fft: should be great than lwin ‘minFactorMeanSpectrum’:0.0, #minimun factor of the two mean spectrums for accepting a points in the two spectral windows ‘minPointsForDelay’:0, #minimum number of points accepted for compute regression which compute dt ‘minAcceptedMeanCoherence’:0.0,#minimum of the mean spectral coherence accepted for keeping the dt ‘factorLenWindowsTemporal’:1, ‘minTemporalCorrelationCoef’:0.0,#minimum of the coefficient of correlation of the two temporal windows for computing the dt ‘minPrecisionDelay’:100000,#minimum of the error of the dt to keep it ‘lengthSmooth’:3,#length+1 of the smooth of each border: if lengthSmooth=3 the signal is smoothing with square cos in a windows of length 5 ‘freqMin’:0.1, ‘freqMax’:0.9, ‘isDoubleRegression’:0,#set 1 for double regression, otherwise only regression for Y=aX ‘typeOfErrorRegression’:0,#set 1 for the classic error regression otherwise it is the original doublet error regression ‘isSmoothPhase’:1,#set 1 for getting the phase after smoothing (otherwise get the phase before smoothing) ‘numberMaxIterations’: 20,#set 1 to have the option “sr_only” ‘maxDiffDoubleRegression’: 10.0,#maximum of the difference between the two regressions for keeping it ‘lenUnwrap’:3,#number of previous values in order to compare the mean and the current value to decide the unwrap ‘minCohrAcceptedForRegressionDt’:0.0,#minimum coherence accepted for take into account the point in the regression which compute dt ‘maxCohrTakeForRegressionDt’:0.99,#maximum of the coherence for the point in the regression which compute dt:

#all greater values are equal to the parameter ‘cohrValueForMaxCohrTakeForRegressionDt’

‘cohrValueForMaxCohrTakeForRegressionDt’:0.9990045045,#see the parameter ‘maxCohrTakeForRegressionDt’ ‘typeOfWeigth’:2,#set number in {1,2,3,4},

#3 first types of weight see paper Clarke and al, appendix C, p14, default is type 2 #type 4: the weigth square root of amplititude of the cross-spectrum

‘stopConvergence’:0,#set 1 for stopping the convergence if the windows comes back to the initial windows and come back to the last value compute }

Whisper project: FP7 ERC Advanced grant 227507 (WHISPER).
Codes are develop for the Whisper project (Seismic ambient noise).
If you use this code, specify it is provide by: Whisper project FP7 ERC Advanced grant 227507 (WHISPER).
Author: Xavier Briand. For questions, send email to with subject begin exactly with [whisper].