Formica and Ajuga

[apt403]

Some recent test shots w/ the Olympus LMPLFL10X objective using my automated XYZ stacking rig. Both at about 11x with a DCR-150 tube lens, focused to infinity on a PB-5 bellows, D5600. .00025" step size (~6.35 micron). Standard wisdom says the sensor is seriously out resolving the objective, but w/ some post processing the results don't seem too terrible. Definitely have to switch to mirrorless before I move up to higher magnifications; 150k rated shutters seem like a lot until you need 5000-some-odd exposures for a single S+S.

Flashpoint budget strobe, 2 second exposure (a la Charles Krebs' method described in the Tabletop 'macro' setup thread). My hotshoe -> PC port adapter is lacking some necessary features to trigger the strobe, so I've got it set to slave mode, running off the on-camera flash at the moment. This seriously limits the speed, as the recharge rate on the on-camera flash is terrible. For 100% reliability, I'm stuck at around 8 seconds per exposure. When I'm not hampered by that recharge rate, I think 2.5-3 seconds is feasible. Unsurprisingly, the recharge rate isn't very reliable for timing critical operations. In my initial tests, I had to keep pushing the amount of dwell time between shots up, as the flash would recharge reliably for 100-150 shots, then suddenly need an extra few hundred ms between shots.

Looks like the compression hit the flower shot a little hard. I'll add some links to full size versions.

[rjlittlefield]

Nice images. I especially like the closeup of the antenna base. Such wonderful ball-and-socket joints these little critters have!

Some recent test shots w/ the Olympus LMPLFL10X objective using my automated XYZ stacking rig. Both at about 11x with a DCR-150 tube lens, focused to infinity on a PB-5 bellows, D5600. .00025" step size (~6.35 micron). Standard wisdom says the sensor is seriously out resolving the objective, but w/ some post processing the results don't seem too terrible.

By standard formulas, using 550nm green light and 11X NA 0.25 (effective f/22), the diffraction limit will occur at 12.1 microns per cycle. Your pixels are 3.9 microns, so about 3.1 pixels per cycle at cutoff, assuming no Bayer filter effects. In recent years I've come to think of this as comfortably capturing all the detail that's in the optical image, rather than something like "seriously out-resolving".

Straight off the sensor I expect the images look very soft from diffraction blurring. But with appropriate sharpening to restore the contrast of fine detail they can look quite good. I'm guessing that you've applied an amount of sharpening to these images that would turn say an f/8 landscape image into oversharpened junk. Can you quantify how much sharpening you've done?

--Rik

[apt403]

Thanks Rik! I agree, the ball-in-socket joint is one of my favorite features of these little guys.

Reflecting after reading your post, I agree, 'seriously out-resolving' is overzealous. I guess my quest for giant prints is making me a little too sensitive to resolution. Before I started testing this objective out in earnest, I was assuming I'd be looking at downsizing each 24mp APS-C sized frame down to about ~12mp (for the sharpest 300 DPI prints), based on this chart @ Nikon's microscopy info site:

https://www.microscopyu.com/tutorials/m ... resolution

(11.1 micron minimum feature size / 2, for NA .25 @ 10x magnification. 3.9 pixel size on sensor / 5.5 micron = ~.71x scale factor)

I was pleasantly surprised, to say the least, when I saw the first finished stacks come out of Zerene!

As far as amount of sharpening goes, the vast majority comes from Topaz AI Sharpen. I believe I had both the 'denoise' and 'sharpen' (w/ the 'sharpen' setting, not 'stabilize' or 'focus') set to either .7 or .8. A hint of the clarity and dehaze sliders in Lightroom (15 or 20 on each). LR also automatically applies some sharpening on import to the raw files before I export to TIFF for Zerene (40, 1.0, 25 on the Amount, Radius, and Detail settings, respectively).

I agree, unless one were printing on a media that bleeds quite a bit (eg. news print), the amount of sharpening I'm performing would result in some less-than-spectacular results w/ a more 'normal' type of image. Everything gets crunchy and jagged looking.

I made some much smaller 1:1 pixel crops of the antenna joint area, to hopefully avoid too much image compression. First shot is a crop of the raw DMap file (w/ PMax retouching, but I don't think there's any in this area) from Zerene, second shot is the finished version after all adjustments. I believe the TIFF->JPG conversion is still robbing some sharpness, but these are essentially correct, comparing them to the files as displayed in LR.