About z-step size:
 
The Niquist Sampling Criterion applies to sampling in the Z direction just
as it does in the X-Y plane.  This theorem says that, to presenve in the
digitized data, all on the information originally present in in the analog
image, one should sample at Z-plane spacings that are at least 2x smaller
than the Z-resolution.  If we assume a Z-resolution of 0.6 =B5m (which seems
about correct for blue/green light and NA 1.3), this implies that you
should set your Z-spacing at 0.3 or even 0.2 =B5m.
 
Aside from the constraints imposed by proper sampling, the other
consideration on inter-plane spacing has to do with data homogeneity.  It
is often easier to deal with 3-D data (in terms of subsequent computer
analysis and display) if the voxels are cubic rather than parallelograms
and this implies that the interplane spacing should be the SAME as the
linear dimension of a pixel in the X-Y plane.  Although this logic is
unassailable, it is seldom put into practice because it means recording and
storing about 3x as much data with the consequent costs in terms of
collection time, bleaching and storage space.  The exception is X-Z imaging
where the need to get a side-view of the specimen that is as undistorted as
possible takes precedence.
 
More common is rather severe undersampling (inter-plane spacing too large)
because of lack of time or specimen degredation.  These practical
constraints are understandable reasons that can be supported in terms of
being able to look at more objects and therefore increase the sampling rate
of the "Biological Variability" dimension but they are somewhat sub-optimal
in terms of theoretical "Best Practice".  It is also true that sampling at,
say, twice the recommended interplane spacing (recording half as much data
as one "should".) will still record  MORE than 50% of the data (about 67%)
so, moderate undersampling can be defended in terms of more
"Information/dose-of-irradiating-illumination".
 
********NEW ADDRESS**************
Prof. James B Pawley,
Room 1235, Engineering Research Building,
1500 Johnson Dr.
Madison, Wisconsin, 53706.
Ph.  608-263-3147,  NEW FAX 608-262-6707
[log in to unmask]