Hi Bradley or Ben,
Thank you to release such a nice simulation tool.
I have a question about k-space and non k-space (set 'Use_kspace',false or true). Same inputs for two. I compare two results I found there is a small shift in phase. The longer simulation, the more obvious shift you can observe. Do you have any idea to explain that?
Thank you,
Ty
Hi Ty,
Well spotted. What you are seeing is numerical dispersion due to the finite difference timestep, ie. due to the fact that the time derivative in the continuous equations is replaced by a finite difference timestep in the discrete equations. Without the k-space correction term this will result in numerical dispersion, which will accumulate - as you observe. The k-space term corrects for this dispersion; in the case of a homogeneous medium the correction is exact, and so there is no dispersion at all. That is why we include the k-space term in the model.
Ben
Thanks, Ben. I like your explanation. It is really awesome.
Best,
Ty
You must log in to post.