Supertelephoto shootout: Nikkor VR 70-300 mm f/4.5-5.6 vs Panasonic Lumix G 100-300 mm f/4-5.6 (1)

Taking advantage of beautiful weather, I took two big guns (and some smaller ones) outside to see how they compare. That has actually already been done in excellent piece by Fredrik Gløckner, and more about 70-300 CX lens in his other post.
Why repeat the test? One of the assumptions and conclusions from my previous comparison attempt was to try to reproduce conditions the gear is going to be used in. I am no birder, in my case supertelephoto is used for landscape (kind of) and micro photography. Another thing is that testing itself actually shows many secondary effects, not often described in reviews, that occasionally can disqualify a piece of equipment for specific purposes. There is also sample variation involved, which at very close results can bias the scale.

Nikkor VR 70-300mm f/4.5-5.6 was attached to Nikon V1 body, MFT lenses to Olympus E-M1. The consequences are:
- 70-300 combination is handicapped by relatively lower spatial resolution of the sensor (10MPix, AA filter) when other lenses project on effective 14Mpix (3:2 crop of 16MPix sensor). All the files were downsampled to 2592px horizontal size for fair and direct comparison
- difference in sensor formats also slightly affected magnification: even though I tried to set the same equivalent (in 35mm format terms) focal length, because corresponding angle of view is defined over diagonal, the horizontal angles differs between 4:3 and 3:2 formats by about 10%
- V1 base ISO is 100, Olympus is 200. In reality Olympus is also close to 100 (and pulled in processing), but at the moment of exposure twice shorter shutter opening time is used, and that could affect sharpness by reducing any motion blur. Using base ISO is sensible anyway, to avoid reducing detail by added noise*
* in landscapes, especially with grass and leaves, false detail of noise grain can actually produce impression of increased resolution

Test 1: 'paparazzi' mode - subject at about 50m. 
That wonderful piece of architecture seemed to be a good target, with brick wall as well as piece of fine text.
Camera mounted on a tripod with 2 sections extended and central column lowered, bag hanged for extra stability. Aperture priority, mechanical shutter, central AF point, self timer 2s to avoid shutter actuation shake. Raw files developed in LR5.7 with default settings (all sliders at 0, sharpening 25), exported to Jpeg at 90% quality, sharpening to screen - none, resized to 2592px on long edge (to match V1 - actually I made a mistake: it is number of horizontal lines, but the bottom line is they were normalized to the same magnification). Crops for side-by-side comparison edited in IrfanView and saved as Jpeg at 90% (different percents than LR) quality.
I added other lenses I found in the bag: Panasonic 45-150 mm f/4-5.6 and Panasonic 14-140 mm f/4-5.6 as well as Leica branded glass of Panasonic FZ1000 to see how they stack up agains big guns within their reach.

At 'short' end frankly there is little difference in the centre.
When looking at the whole images, the vignetting reduces - slightly for 70-300, quite considerably (especially between f/4 and f/5.6) for 100-300 when stopping down, but acuity seems to stay the same even on the edge (bush). Panasonic on the other hand shows better contrast than Nikkor, especially in grass and bush.

At 300mm equivalent (end of story for most of the telephoto lenses) Panasonic 100-300 mm looks for me better than Nikkor 70-300 mm, but again with more visible vigneting. All the other lenses noticably worse (but not terrible), FZ1000 needs stopping down to f/5.6 for the best result

400 mm equivalent - only 3 contenders left. Panasonic 100-300 still better regarding detail, still worse regarding illumination uniformity. Nikkor looks exactly the same at f/5.6 and f/8, only depth of field gives away change in aperture setting. FZ1000, even with 20Mpix sensor, way behind Nikon 1 combo with 10Mpix under the bonnet. Clear evidence what optics at its worst (FZ1000, end of range) and its best (CZ 70-300, middle of range) brings to the total output.

Supertele: 600 mm equivalent. This is what these lenses were made for: the same reach on APS-C or full frame formats costs much more weight and money. Still Panasonic leads in acuity department, with huge change in corner shadowing between f/5.6 and f/8. Here Nikkor starts showing vigneting when wide open too. Panasonic lead is actually a bit of a surprise for me, as it always looked the softest at the long end, but maybe Nikkor starts losing in that range as well. What surprised me even more, was the fact that Panasonic kept acquiring focus immediately, whereas Nikkor at about a third of attempts was scanning slowly through the whole range (lens limiter was off). It is exactly opposite to my field experience, where I artificially limit Panasonic to about 250mm of its focal, because above that, focusing on moving object is pretty much impossible (and on static is slow). Nikkor on the other hand is fast (not instant, but fast) in those conditions.

800mm equivalent - one actor left: there is clear benefit of extending Nikkor CX 70-300mm lens to the full reach (not true for all the lenses), so very good result.

Conclusion: high contrast target is not a good test subject (unless you love shooting newspapers and abstracts). In this conditions lenses with higher abberations will lose (large influence of 'white' light), when in normal conditions (especially with landscapes) greater bias towards green/blue spectrum can shift the outcome. No way to judge microcontrast. Also as much as great focusing target for contrast dependant method, phase AF seemed to struggle with such pattern (and focus position is critical for sharpness) - see results for 600mm range.
Metering gets dumbed as well, Nikon seems to place more importance on centre in its matrix metering and it needed exposure compensation.
The worrying thing is that Imatest charts, so much loved by reviewers, are exactly that: high contrast targets, moreover photographed at short distance. How much are those results reliable then?

Other things noticed:
- it was occasionally windy (side wind) and it can be seen on the photos, especially at long end, that there are minute offsets between frames. It seems like area of a lens is large enough to act like a sail and move the setup slightly
- even though not felt in hand, there is sag in Nikkor 70-300 3-tube construction: lens set to the central point at tele end, showed the same point lifting upwards at short end. Other contributing factor can be flex on tripod bracket joint to the lens (secured with coin screw, I'd appreciate normal screwdriver slot) and tripod plate joint to tripod bracket.
- with lack of remote trigger I relied on 2s self timer and Nikon V1 implementation is terrible: not only it resets to normal mode after each exposure, to set self timer the option needs to be highlighted AND confirmed with OK (unnecessary operation, what for going to self timer menu if not with intention to choose one of the options?). It is faster to highlight by turning the dial, but it is counterintuitive: turning it clockwise (thumb travels down) moves the selection up...