[Erictator]

I'll be the first to admit, I haven't really followed DXOMark or how they rate things over the years...but I'm really really confused now.

DXO just released their lens performance ratings on the K50, and every lens tested, tests way better on the K50 body than on the K5, like if I went out and swapped my K5 for a K50, all of a sudden my lenses would work 15% better or something.

BUT... this lead me to get curious about the sensor, so scrolling down their list of camera sensor ratings, the K50 only scores a 79, and the K5 scores an 82. WTH?

OK, I figured their testing was subjective, but this still is kind of whacky and leaves me lost.


Then I checked my Tamron SP AF 28-75mm F/2.8 XR Di LD Aspherical [IF] Pentax on DXO on a K5 body, and they rate it a 4P 12 on a K5 (Below a snakes belly) and 5P and 13 (still looking up at the snake) on a K50 ! Evidently the only Pentax lens that makes a middling score that I checked was the DA 50 f1.8 on a K50 body getting a 10P and 21 rating.

Good thing my ego isn't too fragile or I'm not too gullible or I'd have to believe that all my lenses are junk!

OK, what gives? Can anyone explaine this to me?

[kiwi_jono]

I noticed the lens ratings don't seem to stack up but did not see the K50 / K5 thing!

I find DXO results a bit dubious (particularly with Pentax it seems). I thinks it comes down to their testing practices which are different to other review site practices.
Certainly when you compare some of the lenses on other review sites they generally compare better to the competition than DXO would indicate.

At the end of the day there is plenty of evidence to show that most Pentax glass can capture stunning shots and thats good enough for me.

[calsan]

Lens reviews are mostly wrong or incomplete or miss the main point of the lens. They always come out with the result that a 100mm macro is sharper than a wide angle zoom. Great, I'll take all my architectural photos with a 100mm macro.

[Kunzite]

Being a synthetic benchmark, even if their testing is correct a single characteristic can skew the results significantly. In this case, I'd guess the K-50 has a lighter AA filter and/or a more precisely adjusted AF system.

[aglet]

The published detailed data DxO provides is (mostly) useful, their overall score number is not so useful.
In the case of lens scores, they attribute some of the lens' score to the image noise levels produced by the body they're testing it on.. Really.
So, other subtle changes in sensor and signal processing between otherwise similarly spec'd bodies can affect the outcome of their overall score.

I was also surprised at the very poor score of that old Tamron lens, I've seen some lovely images shot with it and I've been thinking of getting one for my collection.
Use DxOMark's overall lens scores with a big pinch of salt. I also suspect they only test one variant of each lens in many cases, so a better or worse than average one can easily show some skewed results.

FWIW, for the past number of years, one of the primary reasons I avoided coming over to the Pentax camp was the poor metrics of many of their lenses.
After the arrival of the K5 series and discovering how much I liked the ergonomics and UI I decided to try the brand anyway. My K52s is now my favorite and most often used camera, a D800e is a close second and an old bargain price used k10d is getting more use than I expected.

What I've found, with the limited shooting I've done so far with Pentax systems, is that the lenses may not always measure up to the best of the competition for sharpness, distortion or CA, but they still produce terrific images. In many cases I find I like the bokeh and sometimes the microcontrast of Pentax glass better than the competition. Every system has its own particular quirks and personality and I find Pentax is a very enjoyable system to use that's capable of providing a competitive level of performance for typical shooting and has only a few areas where competitors may perform better.
But then again, I wouldn't put the K5 in the same league as some of the competitors' more specialized or high-end cameras; in it's class, it's a top contender and so are the lenses.

[manntax]

One thing to remember is also the fact that some lenses will render better on some cameras and worse on the others. The best example I can think of is my Sigma 28-105, which gives super sharp images from wide open (2.8 ) on K10D but slightly more softish on K-01. Similarly Sigma 70-210 - K10D produces better and sharper images from that lens. I am not sure if this is down to a sensor differences (CCD vs. CMOS) or the fact that K10D has lower resolution with the same sensor size - therefore less demanding on lens resolution with less pixels packed in the same area, but I do know that the same lens would be rated by me differently depending on a camera body that was tested on.

[azimuthal]

I noticed the lens ratings don't seem to stack up but did not see the K50 / K5 thing!

I find DXO results a bit dubious (particularly with Pentax it seems). I thinks it comes down to their testing practices which are different to other review site practices.
Certainly when you compare some of the lenses on other review sites they generally compare better to the competition than DXO would indicate.

At the end of the day there is plenty of evidence to show that most Pentax glass can capture stunning shots and thats good enough for me.

Dxomark is fairly transparent regarding their testing methods, they are also very thorough, testing each lens on a number of bodies, and they provide a great resource for people who are interested in technical data related to cameras and lenses.

The megapixel of the sensor used has a big impact on the lenses final score. Try comparing the nikon 85 1.8g or sigma 35 1.4 on the d800 VS d7000 (sigma goes from 39[d800] to 26[d7000], 85 1.8g goes from 40[d800] to 25[d7000]). The highest sharpness a lens mounted on d7000 can achieve is 16 M-PIX whereas a lens mounted on d800 could potentially get to 36 M-PIX. So do not compare a lens tested on a high MP body to one tested on lower MP body, realistically even testing between bodies of similar MP isn't perfect since the strength of AA filters could differ.

Regarding k-5 vs k-50:
I briefly looked at the MTF profile of two lenses (sigma 35 and pentax 50 1.8) when mounted on k-5 VS k-50. the solid lines seem to be the same on both cameras but the dotted lines are higher on the k-50. i think dotted is either vertical or sagittal? I'm not too knowledgeable about MTF charts and I'm not sure what would cause that kind of difference.

[Barry Pearson]

[B]What I've found, with the limited shooting I've done so far with Pentax systems, is that the lenses may not always measure up to the best of the competition for sharpness, distortion or CA, but they still produce terrific images. In many cases I find I like the bokeh and sometimes the microcontrast of Pentax glass better than the competition. Every system has its own particular quirks and personality and I find Pentax is a very enjoyable system to use that's capable of providing a competitive level of performance for typical shooting and has only a few areas where competitors may perform better.

Lightroom and ACR can reduce distortion & CA very well. So for me they are among the least important aspects of lens performance.

I feel that my Pentax lenses plus Lightroom give results like more expensive lenses without Lightroom. It is hard for reviewers to take this into account.

[Erictator]

Thanks everyone, that does help a little. I guess I'll put DXO on my semi-ignore list. Perhaps, like most things, it's best to just go with what works for me, regardless of what the talking heads say. Seems to have worked for me so far.