Tradeoff between sampling and bleaching
While sampling close to the Nyquist Rate is a very good idea, it is in many practical situations hard to attain. If practical considerations (bleaching, data size) don't allow these ideal sampling densities, you'll just have to do the best you can. In our experience, unless you undersample dramatically (see the Nyquist Calculator for limitations) the Image Restoration will always improve your data.
One typical reason to reduce the Sample Size is to avoid Bleaching Effects. Still it is good to remember that the Image Restoration is capable to reduce noise considerably, and that this capability improves with higher → Sampling Density. You might find you need less signal than you thought in order to obtain good deconvolution results.
Having more samples means that you need less photons per sample: the overall photon dose can remain largely constant. As a result bleaching does not need to get worse. For example, in cases where one uses averaging it is possible to decrease averaging and increase sampling density.
Sampling densely, even OverSampling, may require that you reduce the excitation intensity. With a scanning device like a Confocal Microscope, the confocal zoom factor will localize the scanning in a smaller area the higher you set it. As the scanning time remains the same, the photon dose per volume will actually increase. Therefore to reduce bleaching you might need to reduce the laser intensity.
If such an exchange of sampling and averaging or excitation intensity is not possible, best spread the UnderSampling factor over all dimensions, not just along Z.
With denser sampling the data size grows, though for a noisy image 8 bits are enough, and disks are still getting cheaper per GB!!!
