Several months ago, Rik Littlefield asked me if I knew whether Olympus had published the pupil ratios for their bellows macro lenses. As far as I know, they have never published them. I own several of the lenses, and I have finally got round to measuring the ratios:
20 f/2 = 0.75
38 f/2.8 = 0.85
38 f/3.5 = 1.03
80 f/4 auto = 0.99 (0.83 with dedicated close-up lens)
135 f/4.5 = 0.93
I used the method that Rik recommended, photographing the iris from one side using a macro lens, then turning the bellows lens over and photographing its iris from the other side, re-focusing by moving the camera (not touching the focusing ring). Then counting the pixels in the 2 images of the iris and calculating the ratio.
I have added the pupil ratios to my pages about the Olympus bellows macro lenses:
http://www.alanwood.net/photography/oly ... enses.html
I hope someone finds these pupil ratios useful.
Alan Wood
Pupil ratios of Olympus bellows macro lenses
Moderators: rjlittlefield, ChrisR, Chris S., Pau
- rjlittlefield
- Site Admin
- Posts: 23562
- Joined: Tue Aug 01, 2006 8:34 am
- Location: Richland, Washington State, USA
- Contact:
Excellent -- thanks, Alan!
I'm pleased to see that you got the same numbers that I did for the 20 f/2 and 38 f/2.8 . It's always nice to have independent confirmation.
To briefly put one of these numbers into context, suppose that you set up the 20 f/2 to give 10X magnification and calculate the corresponding NA for comparison with a microscope objective. Ignoring pupil factor, the calculation comes out to NA 0.227, not much below a typical 10X objective at NA 0.25. But taking pupil factor into account, the correct value is NA 0.174, not nearly so good.
Following industry practice, the "f/2" rating of this lens is calculated on the basis of its focal length and entrance pupil. It represents the speed of the lens when focused at infinity. But this lens simply will not focus at infinity. Its minimum magnification is about 5X, at which point it acts more like you'd expect for an f/2.6 lens (with pupil factor = 1). For the Olympus 20 f/2, the rated f-number is one that you're guaranteed to never see in practice.
--Rik
I'm pleased to see that you got the same numbers that I did for the 20 f/2 and 38 f/2.8 . It's always nice to have independent confirmation.
To briefly put one of these numbers into context, suppose that you set up the 20 f/2 to give 10X magnification and calculate the corresponding NA for comparison with a microscope objective. Ignoring pupil factor, the calculation comes out to NA 0.227, not much below a typical 10X objective at NA 0.25. But taking pupil factor into account, the correct value is NA 0.174, not nearly so good.
Following industry practice, the "f/2" rating of this lens is calculated on the basis of its focal length and entrance pupil. It represents the speed of the lens when focused at infinity. But this lens simply will not focus at infinity. Its minimum magnification is about 5X, at which point it acts more like you'd expect for an f/2.6 lens (with pupil factor = 1). For the Olympus 20 f/2, the rated f-number is one that you're guaranteed to never see in practice.
--Rik
- Craig Gerard
- Posts: 2877
- Joined: Sat May 01, 2010 1:51 am
- Location: Australia
I've done some tests, and come up with the following (Alan's figures in brackets)
Method was to lay the lens on a lightbox and use the same exposure for both ways up.
In Photoshop, the Magic Wand selects the light aperture, and the Measure tool calculates its area directly. Four exposures, separately focused, gave 4 slightly different figures, so I ignored the highest and lowest and averaged the middle two. They all agreed to within about +/- 0.3%.
Square root of ratio of areas gave the figure I think I wanted.
The lenses were all stopped down one stop - it's much easier to see the edge of the aperture.
Admin edit 2024 Mar 6, rjl, to fix code block corrupted by phpBB3 conversion
Code: Select all
20 f/2 = 0.737 [0.75]
20 f/3.5 = 0.886
38 f/2.8 = 0.843 [0.85]
38 f/3.5 = 1.012 [1.03]
80 f/4 = 1.009 [0.99 (0.83 with dedicated close-up lens)]
135f/4.5 = [0.93]
In Photoshop, the Magic Wand selects the light aperture, and the Measure tool calculates its area directly. Four exposures, separately focused, gave 4 slightly different figures, so I ignored the highest and lowest and averaged the middle two. They all agreed to within about +/- 0.3%.
Square root of ratio of areas gave the figure I think I wanted.
The lenses were all stopped down one stop - it's much easier to see the edge of the aperture.
Admin edit 2024 Mar 6, rjl, to fix code block corrupted by phpBB3 conversion