Radiolaria in reflected polarized light

[Pau]

Some time ago I bought an antique slide of Policystina radiolaria from Springfield, Bermuda. It's a deep dry mount whith a coverglass at more than 1 mm of the specimens. Radiolarians are glued whith a black and reflective gum, so I find it a challenge to photograph.

Illumination was done whith two fiber optic light guides, each one whith a polarizer and a focusing lens at the end. Another polarizer (analizer) inside the microscope. Cross polarized light was the key to avoid most reflections and flare

Zeiss Standard trinocular microscope, Canon EOS 20D coupled whith Zeiss 0.25X adapter. Leiz Pl Apo 4X and 6.3 Periplan eyepiece (total magnification 6.3X on sensor).
Stack of 62 images whith CombineZP. Some artifacts cleaned up in PS.
First picture, near whole frame, second one croped to beter show details.





Any comments and advice welcome.

[rjlittlefield]

Any comments and advice welcome.

The radiolaria are lovely!

I see quite a few areas that look like "stacking mush" -- places where I would expect detail but don't see any. In the crop, examples include the areas at Photoshop pixel coordinates (857,115), (373,517), (347,415), and (798,327), among many.

If you look at the original images, are these areas actually free of detail in all images, or did CombineZP fail to find it?

If the areas are free of detail in all images, then can you tell what is causing that? It seems like 62 images at 6.3X should be enough to have gotten everything in focus.

--Rik

[Pau]

I see quite a few areas that look like "stacking mush" -- places where I would expect detail but don't see any.
If you look at the original images, are these areas actually free of detail in all images, or did CombineZP fail to find it?


--Rik

Rik, thanks for your analysis.
I revised the original frames and I can identify both causes: few zones out of focus (marked in red), out of the reach of the stack, and all the others (I marked someones in blue) "stacking mush", mainly in zones whith highlights and low contrast where the stacking software was unable to retain detail. I need to try later other software and/or parameters, but perhaps i will need update my old and slow computer to do this work at a reasonable speed.
I was aware of the imperfect image, in fact I posted it mainly as an aplication of my cross pol illumination setup.

[rjlittlefield]

Pau, thanks for investigating the fuzzy areas.

In CombineZP, are you using the "Do Stack" macro? The parameter to play with there is the one in the Find Detail command that establishes a noise threshold. I am not sure which one that is in the current version, since the parameters have changed order and meaning since the last time I played much with them.

Also in CombineZP, you might try the "Pyramoid Maximum Contrast" macro. That does a better job of preserving detail wherever it occurs, though it may blow out highlights.

For this sort of thing I mostly use Zerene Stacker (which I wrote to handle these problems). The PMax method might give you a good result with no fiddling, or you could visually adjust the DMap contrast threshold slider so as to preserve detail in the radiolaria bodies while ignoring the dark background. ZS also allows resizing on-the-fly (Options > Preferences > Preprocessing). That can speed up processing if you don't need so many pixels to capture detail that you care about.

I have never tried cross-polarization for front-lighting through a microscope. It seems to have done a good job on these very shiny radiolaria. I will have to keep it in mind for future work. Thanks for the illustration!

--Rik