dummy-link

Conv

Multi-dimensional time-domain convolutions and cross-correlations using FFTW.jl

Readme

Operations

This package aims for a fast implementation of the following operations: $$d_{i}=QF^{-1}[(F P g_i)(F P s_i)], \text{where}$$

  • s: source signature in time
  • g: Green's functions in time
  • d: data in time
  • F: FFT operator
  • P: zero-padding matrix
  • Q: truncation matrix

Usage

nt=10; nr=10
g=randn(nt,nr); s=ones(nt, nr)

Create parameter variable and allocate memory.

pa=Conv.Param(dsize=[nt,nr], ssize=[nt,nr], gsize=[nt,nr], g=g, s=s) # memory allocation

And finally, perform a convolution.

Conv.mod!(pa, :d) #updates the data matrix in pa.d

Build Status codecov

First Commit

04/03/2018

Last Touched

4 months ago

Commits

59 commits

Used By: