With the default minization in dem_error.py switched from phase to phase velocity in #875, I noticed the time-series residual after dem_error got much bigger for one dataset I have, as shown below. Although this change of behavior is reasonable given their different objective functions, they are not suitable anymore for the following timeseries_rms step in smallbaselineApp.

The right way would be: keep phase velocity as default in correct_topography, move the residual_RMS step after the time function fitting in the velocity step, which has more powerful / advanced time func capability; then we won't need the TS residual from dem_error anymore. Before this change, I would like to revert #875 to minimize its impact on timeseries_rms.

@mirzaees @hfattahi let me know what do you think please.

• Residual time-series with phase minimization

• Residual time-series with phase velocity minimization