The standard 1.5D internal multiple prediction algorithm generates output in the ( k g ,ω ) domain. Through some manipulation versions in the ( k g ,t ) and ( x g ,t ) domains are derived, along with the 1D time domain formula. Algorithms for calculating the predictions in these domains centre on partial convolutions in vertical time or pseudodepth. This is implemented by forming masking matrices to overlie standard dispersion or convolution matrices. The result are some intuitive codes which produce very clean predictions, even in the presence of gathers prior to deconvolution.
View full article as PDF (1.38 Mb)