Old simulated lensing data based on another source galaxy. For that mock the source image background was insufficiently subtracted, which left a faint imprints in the data after being lensed. While unnoticeable to the eye in the data (left panel), the square boundaries of the source are clearly noticeable in the input source (middle panel) and even on the reconstructed source (right panel, using a correlated field model as an example). These artificial boundaries led to biases in lens potential parameters and source size. Note that a logarithmic scale is used in all panels.