[ Posted 23 May 2013 - 21:03 ]From today's work on the flowers of
Glechoma hederacea (Ground-Ivy), I was struck by what a modest lens could achieve given the proper technique.
The JML 50 mm f/3.5 is a
dirt cheap and unpretentious enlarger lens, comprising a simple triplet optical design and the very minimum if any at all in terms of coating. I got some samples from a sell-out sale where they sold for $10 incl. shipping each, so the financial strain was tolerable. The seller even combined shipping costs when I purchased a couple of them together.
I knew this lens did pass a lot of UV, but always have been in doubt about the optical quality. Earlier attempts trying to stacking the JML with a UV-Nikkor to achieve higher magnifications were not successful, though. However, today I needed a 50-ish lens to go on my PB-4 bellows so I could shoot the Ground Ivy flower at around 2X magnification. Nearest candidate on the overflowing studio table was the JML so that was the lens I used. To acquire some additional depth of field the lens was set to f/16. Yes, I know, this is way into diffraction domain, but with UV you may get away with it nonetheless (sometimes).
First, a reference shot in visible light with the Nikon D600. Apart from demonstrating my D600 has a dirty sensor and the lens is a terrible performer operated in the manner chosen, no surprise here. Next, I swapped the D600 for my Nikon D3200 (modified with internal Baader U Venus Generation 2) and started a stacking sequence shooting as quickly as the Broncolor heads would recycle in order not to roast the poor flower. As only the camera was exchanged image perspective and magnification obviously were unchanged going from D600 to D3200, but frame coverage became different (more on that later). The total stack comprised 52 images but PhotoNinja crashed after about 40 so I had to make do with them.Unfortunately, PhotoNinja is not a very reliable program for batching. Zerene Stacker was used to perform the final focus stacking.
In visible light, this 100% crop shows the awful performance of the JML. This is a crop from the lower lip of the flower corolla. Due to the inherent low contrast of the lens and its negligible coating, you get terrible noise as well.
Horrible, isn't it? It's been a good many years if not decades since I last encountered this kind of 'quality' (sic). Maybe even $10 was to much to pay for such a lens?
Next, the D3200 output. I cropped so the covered area was similar to what is shown for D600. The cameras have different pixel pitch and in order to match the perceived scale, I downsampled the D3200 image by a factor of 0.7. Thus you can compare the crops directly. Note this entails putting the D3200 at a
disadvantage as you effectively lose some of the resolved details.
Now, this is like opening a door into a new world.
Every detail is there and crisply defined. You can identify each and every conical cell on the corolla surface, and you can study their shape as well. Some are flat-topped, some are papillate, some are chained in rows, and so on.
Again, keep in mind both shots were of the same subject from the identical position, the identical magnification, and with the lens set to the same aperture, f/16. The UV capture truly unleashes the potential of this $10 lens. Food for thought, I would suggest.
