Ephotozine Review: Z 100-400/4.5-5.6 S VR

Eh....

It's another flat chart testing site, who is *not* reporting MTF information in the way a proper MTF optical bench test would, so I lump most of these into the same category of flat test chart sites that are minimally useful in the real world.

They all have their issues, and sure - with any of them you have to watch that the camera being used is the same, and then you also can't compare to another sites flat chart sites number-for-number since the testing protocols might be different, the definition of corner or edge might be different, and the distance might be different.

The bigger problem is, as I've typed endlessly about, that an MTF50 scoring from a test chart in NO way fully categorizes a lenses resolution performance nor it's image quality - it is but a small singular data point that sadly far too many forum readers use as some absolute proof of quality in an attempt to prove or defend their lens opinion. Useful, yes, but as an absolute indicator of resolution and/or image quality, not at all.

As for the 100-400, since I own one:

It's really sharp at the wide end, 105-135mm being the best, and it progressively loses things as one gets to the long end. I'm pretty picky, and I like it from 100-300; 400mm is just barely missing my own personal threshold for image quality, but I'm totally cool with it - it's a 4X zoom (harder to do than a 3X zoom) that is fairly priced, so I don't expect 400/2.8 FL exotic performance from it - it's a pretty cheap way to get to the longer lengths in the Z system without spending insane money.

Let's delve into it's 135mm performance for a bit. Compared to the Zeiss 135/2 Apo Sonnar at F/7.1, it's interesting. At about studio distances - call it 8 feet, the zoom gives the Zeiss a serious run for it's money. At moderate distances - think across a small pond, the Zeiss clearly runs away from the zoom, although the zoom is far from bad. At far distance, atmospheric haze tends to even the playing field, although if the air is clear enough. the Zeiss maintains a bit of a lead. So now let's think about that from a test chart point of view. We don't know the precise distance the lens was tested at, but it's at ONE distance. So if that distance happens to be in the realm of where the lens is at it's best, it's going to test well. But that could be misleading, because if the lens is used outside that distance range where another lens might be better, the chart now has obvious limitations.

We also have to remember that the lens designers are often doing a lot of horse trading and aberration balancing, often in an attempt to hit their design goals which may not always mean "best test chart sharpness" (the lenstip test of the 58/1.4G John noted is a classic case in point here). So a lens may have slightly different performance with regards to low/medium/fine/very fine MTF frequency structures dependent on distance for *intentional* reasons (the 85/1.8S, a favorite of mine, is just slightly suppressed in the very fine/high frequency detail in the closer ranges, whereas the competing Zeiss 85/1.4 Milvus is not - but just slightly so - it gives the 85S a very nice rendering for people, but then at infinity, the two lenses almost match throughout most of the range - indicating the designers have an idea of real life use, which is different than optimizing for an MTF 50 score. Another dagger in the heart of the usefulness of MTF50 test chart performance done at close (and often undefined) range.

So I look at the ephotozine scores with about the same weight as I do the photography life scores, photozone scores, etc, but I don't place too much weight on them. I tend to ignore lenstip (IMO a clear anti-Nikon bias exists there, along with a fundamental lack of understanding about lens performance) and DXO as their single scores are utterly worthless.

Moral of story: look at consensus of opinions, including the chart tests, look at subjective reviews, think about differences in distance performance vs close performance and obsess less about MTF50 scores.