You're using reason and logic which never works out for anyone.

Dpreview images softer then my 24-70mm VR images. Why i don't know. Cameralabs has two sets of images, the usual test images, and one set of samples on sunny southern europe Coast (those look insanely sharp). I wonder what makes the difference. So far only Cameralabs (and only their southern EU cost pics) convince me of the 'Z-mount ultra sharpness' in this 24-70mm F2.8S. But i know a lot of things can make image softer (like rainy weather). So i'll keep eye open.

I agree that consistency is the main issue. The downside to testing all lenses with corrections on is that all new lenses will appear better than old ones. That's not good info for consumers.

Providing uncorrected images for a couple of the most challenging shots would be a good idea IMHO.

To my eyes the Nikon seems to be a good lens even without corrections. Its just hard work to figure that out.

@NarrBL

Not if you can reprocess those old RAWs and also apply software correction.

'As a system' is nonsense. The same system is available for lenses from the 70's. Clearly lots of people don't understand this. Which is fine, but reviews should educate and help people with their purchases.

Its meaningless to comparatively evaluate lenses based on software. The next firmware or software update will change the performance of the lens. The performance might only last for a week, following week a new 'system' to review. This is in addition to the fact you can't compare with DSLR mounts at all.

You'd have to re-review all the lenses every time there's a firmware or software update! I doubt our dpreview folks would be up for that :p

@Lessiter, I'd take your concept precisely in reverse.

- again, your _system_ gets the performance all its contributors deliver.
- it matters not at all what a previous lens/camera used to deliver its performance
- if the newer lens/camera system does better, you simply get better images, and/or other measures of overall camera system ability (lightness, price, etc.)

As far as 're-reviewing' lenses which have computed portions to their design, it's rather unlikely, isn't it, that a firmware refresh would reduce performance.

If it significantly raised performance, then it's like improvement in IBIS, continuous focus tracking, etc.

In either case, significance governs editorial choice whether there's just a 'mention' or some kind of update on that corner of testing.

Just as you see all the time for the camera body part/s of the system...

@panther fan Sorry, missed your comment earlier. I did think about this aspect.

My conclusion was that manufacturers putting out information to allow processing of legacy lenses is unlikely to be allowed by their marketing/sales vector.

It seems not too likely that raw developing software would do the testing for this either.

Plus, per the comment above, that better manufacturing consistency is part of the package for these lens-camera correction systems...

Not meaning an avalanche here :) -- just the thoughts :) take care...

Sorry NarrBL but you've not getting some rather fundamental things about reviews, tech and software. Your reply to my post is so far out there I'm struggling with finding my way back.

1. Software improvements affect old lenses and new lenses equally.
2. Reviews should help us identify when we're being played into paying more for less. Help us identify when buying old cheap gear gives better results than new expensive gear.
3. You can do your own lens profiles see https://pixls.us/articles/create-lens-calibration-data-for-lensfun/ no need for either software devs or manufacturers blessing. It's not complicated.

More than CA. Flare, smearing, etc.

Missed the cat. Must take another look.

It looks like the DxO score is based only on 24mm wide open (I just skimmed, correct me if I'm wrong). It would be nice to know how it does at 50mm and 70mm

@ Joed700

Yes, I like the Field Map and have used it to good advantage when shopping for lenses for my Sony. There is much useful information in the individual measurement graphs within DxO.

The summary scores, on the other hand, are really sketchy. They're based on a whole pile of assumptions and unexplained algorithms that won't necessarily correspond to any particular person's needs or preferences.

And, of course, their tests suffer from the usual collection of shortcomings related to possible measurement error, testing of only one sample, etc. etc. That's why I would never — never — take any single source as definitive. For lens tests, I like to have 3-5 quality testing sources, if possible (often isn't).

It's also easy to misinterpret their data. For example, system resolution is the product of the MTFs of all components (including even demosaicing software.) So the Z 24-70/2.8 will, in fact, capture more detail on a Z7 than a Z6; its a good match for both cameras.

Joed700,

First, "Perceptual Megapixels" is an abstraction that DxO created; it has no real relationship to physical pixels on a sensor. So be careful with that.

Second, I suggest some research on the topic of Modulation Transfer Function. What I told you about system MTF being the product of the individual components (i.e. you multiply their MTFs) is correct, and it's established math, part of the discipline known as information theory.

But you can disprove your own claim simply by looking at DxO's own numbers. Find some lenses that have been tested on mutilple different bodies with different sensor resolutions. There are many for Nikon and Canon. Look at the way the PMP varies with the sensor resolution. It's a little tricky because the presence or absence of AA filters is a wild card, but you'll see the pattern; same lens scores higher on higher-res bodies in example after example.

@ Peripheralfocus - I've looking at the sharpness results of various lenses from different brand for a number of years. Yes, I'm fully aware of the different results of the same lens on different sensors. Keep in mind that the image quality is a result of both lens and sensor...this explains the different results. Anyhow, to cut the story short. The only thing that matters to me is the sharpness score. As I had stated earlier, maybe at my other posts, I tested the various lenses I owned just to either approve or disapprove DXO's findings. I have yet to be able to disapprove DXO. In fact, even Dpreview's own findings matched DXO's.

Let me give you an example. When Sigma 1st introduced their 50mm art, I tested it against Canon 50 f/1.8 STM, Nikon 50mm f/1.4 and f/1.8g and Sony 55 f/1.8. My results agreed with DXO, which means both Sony 55 f/1.8 and Sigma 50 art are similar in terms of sharpness and they outperformed both Canon and Nikon...there are others I tried...cont.

Some people must want the lens to sprinkle tiny pixie dust in every shot, put a smile where there wasn't...

Its a zoom lens and a very good one at that? What more do you want?

Maybe you can screw it on to your D700 to get that 3D pop!

I have a D850 as well as 16 F mount lenses. I have the lone 850 because I sold my pair of D750's and bought two Z6 bodies and one Z7 body. I also have and use the 14-30 F4S, 24-70 F4 S, 24-70 F2.8S and 50 1.8S. In my entire 30 year career I have never seen glass this good from Nikon and hardly anyone else for that matter and that includes Leica and Zeiss. The sharpness is undeniable but it is the light transmission, color, contrast and flare resistance that are complete game changers for me and my clients are really taking notice. And ever since the firmware update that has introduced face and eye detection, I could care less what the heck Sony or Canon does because my keeper rate even with adapted F mount glass like the 105mm 1.4 and 70-200FL are just insane. Some of the lifestyle and editorial shoots I have done in the past three months are having me completely redo my portfolio, it is that big of a gain.

They are only just getting started and it is outstanding in my book.