sucddecon.su.main

su 的源代码库

 SUCDDECON - DECONvolution with user-supplied filter by straightforward
 	      Complex Division in the frequency domain			

 sucddecon <stdin >stdout [optional parameters]			

 Required parameters:							
 filter= 		ascii filter values separated by commas		
 		...or...						
 sufile=		file containing SU traces to use as filter	
                       (must have same number of traces as input data	
 			 for panel=1)					
 Optional parameters:							
 panel=0		use only the first trace of sufile as filter	
 			=1 decon trace by trace an entire gather	
 pnoise=0.001		white noise factor for stabilizing results	
	 				(see below)		 	
 centered=0		not centered, =1 center the output on each trace
 verbose=0		silent, =1 chatty				

 Notes:								
 For given time-domain input data I(t) (stdin) and deconvolution	
 filter F(t), the frequency-domain deconvolved trace can be written as:

	 I(f)		I(f) * complex_conjugate[F(f)]			
 D(f) = ----- ===> D(f) = ------------------------ 			
	 F(f)		|F(f)|^2 + delta				

 The real scalar delta is introduced to prevent the resulting deconvolved
 trace to be dominated by frequencies at which the filter power is close
 to zero. As described above, delta is set to some fraction (pnoise) of 
 the mean of the filter power spectra. Time sampling rate must be the 	
 same in the input data and filter traces. If panel=1 the two input files
 must have the same number of traces. Data and filter traces don't need to
 necessarily have the same number of samples, but the filter trace length
 length be always equal or shorter than the data traces. 		

 Trace header fields accessed: ns, dt					
 Trace header fields modified: none					


 Credits:
	CWP: Ivan Vasconcelos
              some changes by John Stockwell
  CAVEATS: 
	In the option, panel=1 the number of traces in the sufile must be 
	the same as the number of traces on the input.

 Trace header fields accessed: ns,dt
 Trace header fields modified: none