Abstract
The reference in my paper to “Surface consistency: A solution to the problem of deconvolving noisy seismic data” by D. Hutchinson and B. Link was to illustrate that averaging trace spectra does not eliminate noise contamination, but merely averages it. This is true whether trace autocorrelations or trace log spectra are used, and whether the data reduction is performed with a mean, median, alpha trimmed mean (as proposed in my paper), or a modal analysis. The sensitivity of these methods to excursions in signal and noise level and character varies. The mean is the most susceptible to error from highly contaminated data; the median is the most stable, provided adequate data points are available (more than 100).
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
