[monochrome]

I think DxO pisses people off because of this feature of their analysis -- they try very hard to avoid using "resolution" numbers precisely because of what you state. Instead, they add biases that they believe more closely aligns with how people see. Thus, fast lenses generally get higher DxOmark scores because noise is one of those things that detracts from perceived sharpness. (not that noise is bad per se, but when it comes to perceived sharpness is certainly is)

Michael

I try very hard never to shoot wider than f/5.6 with any lens on any camera so a lot of that wide aperture perceived resolution is wasted on me.

The biggest argument I have with DxOmark is their Standard* that resolution itself (or sharpness, if one prefers) is the central metric that defines a good lens. If that must be the case then sharpness must, by definition, be the central metric that defines a good photograph, and the sharper the better. I disagree with that Standard altogether.

But in fairness to DxOmark's methodology, they can measure their metrics, can claim to measure them objectively (theycan't due to camera and sensor differences, but they can claim to) and so they can make a case for a scientific evaluation, ranking and sorting of lenses one to another - a kind of ordered meritocracy, if you will.

None of which explains anything about the special character of the DA15/4 Limited - the 'worst' lens in this little contrasting chart.



* Hierarchy:

Standard = Good - we all have to agree what is Good or what comes after can never be agreeable
Comparative = Better - rank accordingly and scientifically
Superlative = Best - in an Ordered Meritocracy, all other variables being equal, there should be no other rational selection.

That is why there is Price (and Marketing).

[wildman]

they add biases that they believe more closely aligns with how people see.

A laudable goal but If so they must have a very large Psychology department - trying to define human visual perceptions like "pixie dust" for instance. In my over 50 years of messing about with glass, to the best of my knowledge, I have never had a lens that exhibited that elusive quality of "pixie dust".

But I digress.
Interesting thread whatever it's trying to express.

[aurele]

It's more interesting to see the "field map" for sharpness.
You'll se the DAL 35 and the DA40 perform equally on the whole image, whereas the DA*55 perform very well in the middle only, wide open.

Pentax lenses don't score very well on DxO due to this character : sharp in the middle, softer in the corner.

[rawr]

While I do like the presentation of the data, I don't totally buy the results for the DA40. In my experience, it performs better than the chart suggests.

[MJSfoto1956]

Ok. What I *REALLY* was interested in was to see how different Pentax's MTF Program Line diverged from the DxO data. While I don't have all the above lenses, I *DO* have a good handful. What I now have compiled (see below) is a DELTA between how Pentax computes the optimal aperture and what the DxO data is hinting at. More importantly, for some time now I have had my doubts about the Pentax MTF suggestions, often scratching my head wondering "what are they thinking"? (for example: in good light Pentax was selecting f/4 for the 16mm setting of the 16-45mm lens whereas DxO is suggesting it should have been f/5.6) The good news is that the two data sets often align quite well. Anyway, if any of you have the missing lenses, simply put your K-3 (it has to be a K-3 btw) into MTF Program Line mode, put the camera into P mode, then walk around observing what the aperture chosen is for a variety of dark and well-lit subjects. I will add your observations to the diagram and repost.

Michael

---------- Post added 02-21-14 at 06:04 PM ----------

While I do like the presentation of the data, I don't totally buy the results for the DA40. In my experience, it performs better than the chart suggests.

It is entirely possible their web-posted data is completely wrong, When using others data, we should be looking at trends rather than any specific data point. For example, DxO's data for the DFA 100mm macro to my eye looks entirely bogus and I didn't post it because my own experience of macro lenses over the past 35 years from a variety of manufacturers (Leica, Canon, Mamiya, Nikon, Schneider) have demonstrated to me that they are all great performers without exception. I might very well remove the 40mm because it simply doesn't look "right" to me (i.e. I can't imagine that Pentax would ship a "Limited" lens that stood out like that).

Michael

---------- Post added 02-21-14 at 06:19 PM ----------

The biggest argument I have with DxOmark is their Standard* that resolution itself (or sharpness, if one prefers) is the central metric that defines a good lens.

The published DxOmark scores are not about sharpness or resolution. If you are curious their published methodology to calculate their lens+body scores is available for your reading pleasure.
That being said, the chart I am publishing in this thread *IS NOT* about DxOmark Scores (which is very very different) but rather it *IS* about sharpness (or resolution if you will).

Michael

[causey]

I just can't believe the DA 40mm is less sharp than the DA 21mm. All other reviews I've seen say otherwise. Anyway, my copy of the DA 40mm is plenty sharp from f2.8.

[MJSfoto1956]

I just can't believe the DA 40mm is less sharp than the DA 21mm. All other reviews I've seen say otherwise. Anyway, my copy of the DA 40mm is plenty sharp from f2.8.

It is entirely possible their web-posted data is completely wrong, When using others data, we should be looking at trends rather than any specific data point. For example, DxO's data for the DFA 100mm macro to my eye looks entirely bogus and I didn't post it because my own experience of macro lenses over the past 35 years from a variety of manufacturers (Leica, Canon, Mamiya, Nikon, Schneider) have demonstrated to me that they are all great performers without exception. I might very well remove the 40mm because it simply doesn't look "right" to me (i.e. I can't imagine that Pentax would ship a "Limited" lens that stood out like that).

Michael

[CDW]

I've personally always taken DXO lens ratings with a lot of grains of salt. Photozone.de findings regarding the DA40mm correlate with user comments who find it very sharp.

Pentax SMC DA 40mm f/2.8 Limited - Review / Lab Test - Analysis

[DeadJohn]

Thank you for posting this. My takeaway from this is, when using a prime lens that I don't have much experience with, and sharpness is my main objective (instead of fast shutter or controlling dof), then shoot at f5.6.

The 16-45 graph is surprising. It's sharper wide open than stopped down at middle focal lengths.

Everyone, remember that sharpness isn't everything. It's vital for some subjects and for heavy cropping, but many other lens factors contribute to a photo. Aberrations, flare, color, contrast can be more important depending on the photo. The OP isn't saying that photos taken in the red zones of the graphs are bad.

[dansamy]

It's very difficult to quantify that artistic quality that makes an image "pretty". DxO's strength seems to lie in measuring the quantifiable aspects of a piece of hardware. That's about as useful to an artist as counting the bristles on a paintbrush.

[MJSfoto1956]

That's about as useful to an artist as counting the bristles on a paintbrush.

I think this is too broad a generalization. I own the DA 16-45mm and not surprisingly my experience with this lens aligns well DxO's data. I was wondering why that lens didn't shoot like other zooms I have owned. Now I know why. Nobody gives you insight into the behavior of lens + body systems as insightful as DxO. The DxOmark site is worth digging around -- if you keep your mind open, you might be surprised what you learn!

YMMV

Michael

[dansamy]

I think this is too broad a generalization. I own the DA 16-45mm and not surprisingly my experience with this lens aligns well DxO's data. I was wondering why that lens didn't shoot like other zooms I have owned. Now I know why. Nobody gives you insight into the behavior of lens + body systems as insightful as DxO. The DxOmark site is worth digging around -- if you keep your mind open, you might be surprised what you learn!

YMMV

Michael

What I mean is that the quantifiable aspects of a particular lens or camera, which can be graphed, charted, etc aren't necessarily the only measures whereby an artist finds value in a tool. The lens is just a tool. Sure, knowing where the weak spot is allows one to exploit it for artistic advantage or avoid it as desired.

[audiobomber]

MTF results are interesting, but do not correlate very well with user opinions. The only lens test I've seen that correlates with actual Pentax user opinions is PopPhoto's SQF (Subjective Quality Factor) score.
Imatest - SQF (Subjective Quality Factor) and Acutance

[mikemike]

Thank you for posting this. My takeaway from this is, when using a prime lens that I don't have much experience with, and sharpness is my main objective (instead of fast shutter or controlling dof), then shoot at f5.6.

The 16-45 graph is surprising. It's sharper wide open than stopped down at middle focal lengths.

Everyone, remember that sharpness isn't everything. It's vital for some subjects and for heavy cropping, but many other lens factors contribute to a photo. Aberrations, flare, color, contrast can be more important depending on the photo. The OP isn't saying that photos taken in the red zones of the graphs are bad.

I totally agree with this, the more I learn and hone my skills the more I appreciate the non-sharpness aspects of a lens' optical characteristics.

[MJSfoto1956]

well, the 21mm limited is a damn fine little lens. Here is one @ f5.6 I printed 18"x27" and it is tack sharp:



Michael