General information

This scene contains less detail than my traditional Crofton Lake, so it is less demanding in judging lens resolution. Using artificial light, it also provides more controlled conditions. The shorter subject distance also may affect the results (under a close scrutiny), as it may affect the correction of some optical aberrations.

The camera used to shoot all samples presented here was the E-M1 Mk.II with firmware Version 1.1.

Camera settings:

• Images saved as full-size, 20 MP JPEGs in SHQ (1:2.7) compression;
• Natural Picture Mode with Sharpness, Contrast, and Saturation at default (zero) settings;
• Noise Filtering at Low (one notch below factory setting);
• White Balance set to Auto;
• ISO 200;
• Autoexposure (matrix metering), no exposure compensation;
• Medium tripod, remote release via OI Share on Android, IS off.
• Single AF mode; single AF spot chosen manually (center);

For each focal length I was adjusting the subject distance so that the scene was shot in (approximately) the same scale. There was one exception: for 300 mm I had no room to back off far enough, so the scale is double that at 150 mm. Sorry, I need a bigger house.

Samples for individual lenses

Full sets of samples for each lens, various focal lengths and apertures, are grouped in separate pages, accessible via the links below.

Here are just some remarks on particular lenses. They are independent of (and may slightly differ from) those derived from the Crofton Lake (or other) samples.

• Leica DG 8-18/2.8-4.0

  An outstanding super-wide (sorry for using superlatives, but there is no other way). Extremely high resolution, low distortion, no visible chromatic effects. Quite compact and not too heavy, takes filters(!). Joins the two Zuiko Pro zooms at the top of my list.

• MZD 12-40/2.8 Pro

  Looks like the lens, at all focal lengths, can handle more detail than the scene contains. The long end (40 mm) looks a bit better than that of 12-100/4.0. Optically best between F/4 and F/8, but very close already at F/2.8.

• MZD 12-50/3.5-6.3 EZ

  Very good in the mid-range, quite good at the wide end, good enough at tele. Optically best between F/5.6 and F/8, except for the long end, where I would recommend F/8. Performance at F/11 is very close to that at F/8, so this aperture also can be used if more DoF is needed.

• MZD 12-100/4.0 IS Pro

  Another champ. Best at F/5.6 and F/8, but the differences at F/4 are hard to spot with naked eye. Diffraction still at bay at F/11. Based on the Bookshelf Samples alone, I would rank this lens just behind the 12-40 (a role reversal compared to my Crofton Lake samples, interesting).

• MZD 14-42/3.5-5.6 EZ

  The pancake lens produces images good enough for most uses, but no match for other lenses shown here. Best resolution at F/8 to F/11, except for the wide end, where F/4 (or even F/3.5) is as good or close. At the long end, F/11 is somewhat better than F/8.

• MZD 40-150/4.0-5.6

  The lens does the best job at medium focal lengths (75 mm or so), but over the full range it performs much better than its price or make would suggest. Best at F/5.6 to F/8.

• MZD 75/1.8

  With this (low-detail) scene, the fully-open (F/1.8) results already look good, and there is hardly any improvement beyond F/2.8. This is clearly a low-light animal.

• MZD 75-300/4.8-6.7

  Perhaps because of the closer subject distance, the performance for the Bookshelf scene looks a bit worse than for the Crofton Lake. Still, quite good for a long tele zoom. F/8 is my preferred aperture (even if F/11 looks a bit better at 300 mm).

All lenses compared at full aperture

In the sample array shown below, each column shows samples for three different focal lengths of a zoom; for a prime only one sample is shown. The lens is identified at the top; the focal length (and corresponding full aperture) — beneath each sample.

Note that the 300 mm (bottom-right) sample shows the scene in a different scale, as explained in the top section above.

The 42 mm sample (14-42/3.5-5.6) looks like it suffered from camera shake. This is, however, not the case: the effect occurred in all four samples shot at F/5.6 and F/8, therefore it must be real (internal reflection?).

There is no doubt that, at full aperture, the three premium zooms look the best in this comparison. With the scene, however, not rich in detail, other lenses often come quite close for some focal lengths, especially the 40-150 medium tele.

All lenses compared at optimum aperture

In this comparison each lens is stopped down to the aperture at which it generated the best (in my naked-eye judgement) image of the Bookshelf scene.

Note that the pancake 14-42 performs best at F/11; this means that the loss of detail due to diffraction is more than offset by reduction in other optical errors at smaller apertures (greater F-numbers). I was expecting the "best" aperture to be F/8 here.

Surprisingly, the budget 40-150 zoom does its best job when fully open at 75 and 150 mm, something I would have never predicted.