These are fascinating subjects.
Graham graciously shared his stack with me, when I asked if I could get a copy for testing purposes. Indeed, HF is fairly easy to use on this stack -- just increase the Radius to its maximum value of 20. Doing the same thing in ZS gives about the same result but takes one more step, and the output is more affected by dust.
On the other hand, ZS has some tricks of its own.
Here is a PMax synthetic stereo (crossed-eye), with the background cleaned up and a few dust streaks removed. This structure is probably old hat to folks who study diatoms, but since I've never seen this diatom before, I was surprised to see the exquisitely thin lip around the end of the beast, as well as those four little projections in the center of the end.
Helicon's 3D rendering is not so helpful in this case. Its fitted-surface approach smears out the exquisitely thin lip into a robust bulwark, and the four central projections turn into a pair of hills.
Color differences are entirely due to post-processing, by the way. The originals were yellow due to illumination color. For the ZS pair, I just converted the whole image to grayscale. For the HF output I also tried Photoshop's auto-color adjustment. I liked the visual appearance of the residual color so I'm showing that version.
--Rik
