changed default padding value to 0

functions/fresnelPropagation/FresnelPropagator.m

@@ -34,7 +34,7 @@ classdef FresnelPropagator
 %     Determines the method (= underlying discretization approach) used to Fresnel-propagate images.
 %     Admissible choices: {'fourier', 'chirp', 'chirpLimited', 'pixel2pixel'} (details are given 
 %     in the documentation of fresnelPropagationKernel)
-% padVal : Default = 'replicate'
+% padVal : Default = 0
 %     Number or string that determines how the images are optionally padded. Same admissible
 %     choices as for the MATLAB-function padarray.
 % singlePrecision : Default = false
@@ -256,7 +256,7 @@ methods (Static)
         defaults.method = 'fourier';
         defaults.sizePad = [];
         defaults.sizeOut = [];
-        defaults.padVal = 'replicate';
+        defaults.padVal = 0;
         defaults.fresnely = [];
         defaults.fresnelz = [];
         defaults.singlePrecision = false;

[p.jhagema]

may i ask why you chose 0 padding? in general 'replicate' or '1' padding should be better choices to avoid boundary artifacts.

[paulphilipp.meyer]

Sure 😄 . The reasoning behind was the usual use case being the propagation of just the disturbances of an otherwise uniform field distribution. If you look at this commit's parent 02fad5bf you can actually see the default value has been 0 in the first place, before i updated it to agree with the docstring (the docstring being the part that should have been changed in the fist place).