Originally posted by northcoastgreg However one chooses to define and test sharpness, the question remains: is there a 100% correlation between a given test (following a given definition of sharpness) and what we actually perceive? The DA 18-55 is a "sharper" lens (i.e., it would score higher on a resolution test) than the DA 10-17; yet the DA 10-17 produces images that look sharper to human perception (because the lens is much contrastier, with considerably better color rendition).
Another example of this is the lenstip review of the DA 17-70, where the reviewer gripes that the DA 17-70 is not much sharper than the DA 18-55. That may be true, but images from the DA 17-70 certainly look "better," and even "sharper," than images from the DA 18-55. More curiously, they look better and sharper even at web resolution, where sharpness differences should not even be perceivable.
But is there really any other metric that can be used? Is it possible that most metrics are "useless" (or, more plausibly, "problematic" or "potentially misleading") and that the only way to judge a lens is to evaluate the images it produces with one's own eyes?
How the typical review is done:
- One page to describe the lens physically
- A big sharpness chart test at all appertures on center, border, sometime even extreme borders... At several focal lens for zooms.
- Test on typical optical aberations like chromatics aberations, vigneting, geometrical aberations. All theses are automatically corrected by tools like DxO and except if pretty severe you'll not even notice it in practice.
- Some reviewers include 1-3 shoots aiming to test flare and extrapolate the flare performance from that.
- Some reviewers do say something on autofocus performance but that not all.
- There typically a few samples shoot that more often than not look bad. The subject is uninterresting, there no corrections done in post processing so you can see how bad the lens is, but not what to expect from it in practice. If you are in bad luck the day of the outdoor shoots was simply cloudy and the image look bad, because the lighting is bad.
Honestly theses tests are still interresting but you have to cross check with forums, with review like what we have on Pentax Forum and read a bunch of comments, by looking at real examples shoots (if possible from good photographers).
Honestly, I have this FA77 I decided to buy. It is because I wanted something really good and I decided to invest. Oh my god! This lens is sooo great. The contrast is fantastic, the bokeh is really realy smooth and nice. The 3D pop is easy to get in good ligtht (and that not something you get nearly as often from other lenses I tested (my other lenses are DA17-70, DA50-135, DA15, DA21, DA35, FA50). But if it was only from review like photozone and lenstip, this just look like an overpriced lens that uninterresting.
When you read article on how the FA ltd where designed, Pentax lens designer explain that perfectly corrected lenses perform well on tests charts but lack spice in practice. They made many tries with different levels of optical correction and used the advice of many professionnal photographers on what looked best in practice. In the end the chart result are good, but not perfect... But oh man... The photos are just georgous. There this pixie dust ! That very difficult to trust it and check without trying yourself. I did see many of very fine shoot on FA ltd lens club through at level other lenses fail to provide... But even that not easy, there more chance the photographer know his stuff when he has such kind exotic lens anyway.
I remember at the begining 2 year agos, I would not see at all what this kind of lens would provide. Now If I had to choose again, for sure I would not even try the already good DA70... And I would buy FA43 instead of FA50!