[Socrateeze]

Belief is something that doesn't require evidence. We believe or we don't. Marketing plays a big role in that, guess why they get good salaries

That said, DXO test lenses with the best camera they have at the time they decide to test lenses, then they interpolate for lower resolution cameras. At the time DXO tested Pentax lenses, the Pentax K1 wasn't out yet, so they've tested most Pentax lenses with a K5 and some lenses with a K3. While Sony glass was test on A7r, Canon glass tested on 5DSr and Nikon glass tested on D800E or D810. No wonder why DXO scores are about half for Pentax glass vs Sony, Canon and Nikon.

For a belief to be *rational* it requires evidence (or, more generally, good reasons). If you don't accept that, then we really have nothing further to discuss. *Psychologically*, belief requires nothing more than a *cause* and many causes of beliefs are not good reasons: hypnotism, inculturation, cognitive biases, a bump on the head, marketing - you name it.

I'm interested in having *rational* beliefs, which means that I'm interested in the evidence. So should everyone be, because rational evidence-based beliefs have a greater propensity to be true than irrational ones. (I take it for granted here that truth is of some value.) So when people make claims I have no independent reason to believe are true (or, wrt which I have some reasons to think may be false) I like to ask: why do you think that? What is your evidence? That way I have a greater chance of learning something or being corrected (if I'm mistaken) - I always start out by assuming that the person has some rational basis for their claim, of course. Alas, I'm often disappointed.

Back to DxO. Unless you have grounds to believe that the lens tested is unrepresentative of that lens generally, then you have no basis for rejecting the findings based on that lens simply because it is a single instance of a lens. If I test a piece of chalk for its hardness, then I have some information about the hardness of chalk in general, even though there is some variation. And I can form reasonable opinions about whether, for example, chalk is harder or softer than some other sample of something else.

Merely saying that there is variability (which is true of all lenses, some more than others) merely introduces some degree of uncertainty into the results, insofar as those results purport to be about the lens generally. How much uncertainty it introduces depends on the variability in those lenses. The more known variability there is, the less information is carried in the results from a single instance. Testing more instances of a lens reduces the uncertainty and increases the likelihood of the results being representative, all else being equal (which helps to make lensrentals a valuable source of information). And that is a good thing because it allows the result to convey more reliable information about the lens in general, always assuming the the lenses are selected randomly.

Pointing to the better scores of the Sony lens and saying, in effect: "but that's only one lens and it could be by chance one of the better of a generally inferior lens" is true - it *could* be. Equally, however, it *could* be one of the inferior lenses of a generally better lens, in which case the lens is even more superior to the Pentax than the DxO mark would suggest. How do you pick between these two possibilities? You can't. Independently of any further evidence, there is no reason to suppose the former if more likely that the latter. It *might* be unrepresentative, but one ought not assume that it is unrepresentative simply because there is variation in the population. And it might just as well be unrepresentative to the detriment of the lens than in its favour. The same is true of the tests of the Pentax lens, of course. It could be that the lens tested was a particularly good example of the lens generally, and the Pentax lens is generally worse than the DxO score would suggest.

If your point is simply that there is some uncertainty surrounding the reliability of tests performed on a single instance of a lens, I agree. Uncertainty does not result in an absence of information.

You suggest that the scores listed as tested for lower resolution cameras (such as the NEX-7) were not actually tested on those lower resolution cameras - despite the fact that they say that they were - but were interpolated from scores on higher resolution cameras. What is your evidence for that? (I.e., what is your evidence that they are lying?) And even if they were interpolating, why think that the interpolation is not accurate?

---------- Post added 08-26-19 at 11:51 PM ----------

IDK about the Sony, but the Pentax K-S1 does not have an anti-aliasing filter. That's why it has Pentax's "AA Filter Simulator".

Thank you. I stand corrected. I'm not sure where I got that from.

[beholder3]

found the DXO ratings to be bunk. I don't know if DXO ever issued a mea culpa.

Dxomark is an inferior tabloid page for kids.

Anyone with a hint of skill just needs to add 1 and 1 to see their blunder.

Here is an MTF comparison of the Pentax 100mm macro versus the Canon L glass on similar res, similar AA filter sensors:
Pentax smc D FA 100 mm f/2.8 Macro WR review - Introduction - LensTip.com

As a result the K-5 gives exactly the same MTFs as the Canon 50D. The graph below shows it very clearly – we compared the results of the Canon EOS 50D plus EF 100 mm f/2.8L IS USM tandem with those of the Pentax K-5 plus the smc D FA 100 mm f/2.8 Macro WR lens. As macro 100 mm lenses are often the sharpest devices in a given system, they are perfect to show peak achievements of any detector.

From our point of view it would be difficult to get anything better. The values, the K-5 and the 50D show here, are identical to such a degree that for two out of four presented apertures the results were the same within an accuracy of 0.1 lpmm. Taking into account the fact that measurement errors on this level amount to 0.5 lpmm the accuracy is perfect indeed.

This is in sync with:
Pentax-D FA 100mm f/2.8 Macro WR Lens Review | ePHOTOzine

Pentax-D FA Macro 100mm f/2.8 WR Performance
Sharpness is already excellent in the centre of the frame and good towards the edges of the image at f/2.8 and stopping down as far as f/5.6 improves performance further. Here sharpness is outstanding in the centre of the frame and excellent towards the edges.

And now you can compare the Canon L glass which performs exactly like the Pentax macro with the Sony:
Canon EF 100mm f/2.8L IS USM Macro Lens Average MTF

Dang. The newer 1,000 EUR Sony underperforms the Canon L glass (and so the 500 EUR Pentax 100mm macro) by quite some margin.

So we find the Sony is inferior and ridiculously overpriced (> 100%).

And we again find that Dxomark hasn't got a clue and their "scores" can not be trusted at all.

Oh, did we already mention that the Pentax lens is 340g and the shorter Sony is 602g / 62mm filter? Sony FE is twice the weight, twice the price and worse performance. Didnt someone claim mirrorless is lighter?

[biz-engineer]

Merely saying that there is variability

Your reply is to my comment. I wasn't referring to test of one unit of a lens. I wrote that DXO tested Pentax lenses with a K5 or K3 (apsc), and DXO tested the Sony macro lens you mentioned in your post with Sony full frame high resolution camera. You say that testing with anything is better than no testing at all. Well, that's not true, at least no testing doesn't influence someone opinion as wrongly as a wrong test result based on a flawed test setup.

---------- Post added 27-08-19 at 09:48 ----------

You were saying: Sony lenses get consistently better scores than Pentax lenses by DXO mark. I replied: DXO tested Pentax lenses with apsc cameras , and tested Sony lenses with full frame higher resolution cameras. There is nothing to argue about, that's pretty clear, end of story.

---------- Post added 27-08-19 at 09:49 ----------

What is your evidence for that?

How can DXO have tested the lenses on a Pentax full frame camera while they tested the lens the Pentax K1 did not exist! By the way, I don't know what you are talking about, because it is not possible to get DXO test results of the DFA100 macro with the Pentax K1 since the K1 is not available in the list: https://www.dxomark.com/Lenses/Pentax/smc-D-FA-MACRO-100mm-F2.8-WR

[Socrateeze]

You make a claim based on some ad driven tabloid and then earnestly believe the world is flat and will spend effort to prove the tabloid wrong?
It is so convenient for you to try to turn the burden of proof around to people you have to convince, not the other way around.

So where is your evidence?

I'm not sure what you mean by "ad driven tabloid". Do you mean the DxO website? That's where I got my information about their scores.

Okay, so you raise an interesting question about burden of proof. Who has the burden of proof? As a *general* principle, it is with the person making the claim. If someone claims that DxO marks are meaningless, that is a claim and they shoulder a burden of proof for that claim. So there is nothing wrong with me asking people for evidence for such claims - i'm placing a burden of proof on the shoulders of those who make that claim. As I'll argue below, this general principle, however, cannot be extended universally, on pain of radical skepticism.

What burden of proof do *I* have? *At most*, I have a burden to provide reasons for only those claims that I have made. Me asking those who have claimed, for example, that DxO marks carry no information, for grounds for their claim, does not result in any burden of proof for me to establish anything.

But I have made some claims. My first claim was that DxO gave the scores that I said they do. I could provide you with links, but I'm going to assume that that is not your concern. I think what you are asking for evidence for is for my assumption, later made explicit, that DxO marks carry some information (and are therefore not meaningless). Here is my response:

All reasoned argument has to start somewhere with premises (not always explicitly stated, of course) that are not ultimately subjected to further questioning. (In order to get anywhere in an argument, you have to start somewhere.) If you express a belief I *could* ask you for a reason for that belief, then a reason for your reason, and then a reason for your reason for your reason, and so on ad infinitum. The result of demanding that reasons be provided for *any and all* beliefs is to demand something that *cannot* be given and leads to radical philosophical skepticism. (Which is bad!)

Our premises almost always include someone saying something (something on the internet, in a newspaper, on tv, on the news, in Pentaxforums, in a peer-reviewed scientific paper, etc.) The vast majority of the things that we humans believe are based ultimately on the testimony of others, and that requires a large degree of *trust*. Without trust there is almost nothing we could learn about the world. For example, I believe that gold atoms have 79 protons in its nucleus. I believe that only because lots of sources that I trust have said that. I haven't got any personal experience that corroborates that, and neither have even the vast majority of scientists. I certainly have not experimentally tested it. This does not make our beliefs irrational however, for reasons that I'll try to make clear.

The challenge lies in deciding where to place our trust.

In the case of lenses, people here have pointed to, for example, Lensrentals as a source of information about lenses or to MTF charts sourced on the internet. To use those as sources of information requires some trust. One could equally ask the same question there: where is your evidence that Lensrentals actually carried out the tests they say that they did? Where is your evidence that, even if they did carry out those tests, that their instruments work or are well-calibrated? Etc. If one were to adopt such a radical skeptical approach then no one has any reason for anything if it is based on the testimony of someone else - which almost all of what we believe about the world is. I take that kind of radical skeptical approach to be coherent, but also very corrosive for the general enterprise of critical enquiry. In other words, I take it that at some level we have to trust others if we are to have informed opinions - as we desperately need to have - that extend beyond our own immediate and very narrow personal experience.

So I believe that the *default* should be that a speaker sincerely expressing a declarative statement raises the probability of the statement being true. In the absence of grounds for thinking that the statement is either a) false, or b) without epistemic warrant (i.e., without rational justification), it is then reasonable to believe that the person's statement carries information. This is *not* to say that the statement is true - it is simply to say that it raises the probability of the statement being true. (This is related to what is sometimes called the principle of credulity, or principle of epistemic charity.) Very roughly speaking, without getting into to much detail, we should give epistemic credit to testimony *unless* we have grounds to think that the testifier is lying, stupid, or ill-informed. Without that principle we could not learn anything from anyone giving testimony. Since I think we *can* and *do* learn things from testimony, I endorse that principle. I think pretty much everyone does, implicitly.

So far, I have no reason to think DxO are lying, no reason to think that they are stupid, and no reason to think that they are ill-informed, when they make statements about, say, the sharpness of the lenses they test. The same goes for Lensrentals, or ePhotozine, or Optical Limits or Lenstips, etc. I therefore give them epistemic credit and treat their statements as carrying some information. There are, of course, many factors that modulate the reasonableness of crediting their statements with carrying information, and many factors that modulate the amount of information their testimony is likely to carry (such as the number of lenses they say they test), and I strive to take those factors into account.

For analogy, suppose I want to know what the time is and I ask someone and they say: "it's nearly ten". I take it that I now have some reason to believe that it is nearly ten. It may well be that there are better and more accurate sources of information about the time out there, but it would be unreasonable for someone to simply *assume* that the statement carried *no* information about the time without some reason to do with, say, the person's poor sight, their high alcohol levels, their propensity to lie, the recent operational failures of their watch, etc. In other words, they would need some reason if their dismissal of the statement is to be reasonable.

None of this means that DxO marks *are* worth anything. BUT, it does mean that if someone claims that they are not worth anything, then it is reasonable *for me* to ask for the reasons for thinking that.

---------- Post added 08-27-19 at 02:18 AM ----------

All the 'measurements' are done with a single copy of the lens on a camera at printed out targets and they look at the resulting file. Remember, they're a RAW development software company.

What if the picture is out of focus?

Then the test would be worthless. I could equally ask: What if the apparatus LensRentals used was badly calibrated? Then their tests would be useless. How do such possibilities show that the DxO mark should be ignored? Have you grounds to believe that it was not in focus? Have you reason to think that DxO folks don't know how to focus a lens properly? I have no such reason to think such things.

A real test like the camera companies or Roger Cicala does is in a rig just for the lens:

Lens Rentals | Blog

Otherwise you're not testing the lens, you're testing that individual copy of the lens, the camera body, the software algorithm, the subject setup and lighting, and the skill of the individual photographer. It's called controlling variables.

You are testing the lens, but you are not testing *just* the lens. That is, differences in the lens makes a difference to the marks, but it is not the only thing that makes a difference. Claiming, with justification in my view, that Lensrentals is a *better* source of lens information than DxO marks is something I would certainly agree with. But it does absolutely nothing to show that DxO marks should be ignored. All you are saying is that there are other factors that we need to control for before interpreting the marks from any source as a measure of the lens' properties.

DPR had to apologise for their K-1 review when it was pointed out to them they were using an obviously decentred FA77. Yes, they do that for a living!

Right. And again, none of that shows that the DxO marks carry no information or should be ignored. Have you any reason to think that the DxO testing also used a decentred lens? Could they have? Sure!

MTF charts, as imperfect as they are, are best. They predict the sharpness, microcontrast and aberrations by position across the frame. We knew the DFA*50 was a stunning lens by looking at the MTF curve without seeing a single photo taken with it.

Yep, and again....

Sample photos make a useful adjunct as long as it's understood they're both scene and operator dependent.

I can tell you the DA35 and DA40 are not sharper than the DFA 100 as DxOMark claim - I own all these lenses!

What you possibly can tell me concerns your experiences of your particular copies of those lenses. Your testimony carries information, but it does nothing to show that DxO marks should be ignored, rather than, say, showing that Pentax lenses vary in their sharpness from one copy of a lens to another.

That 100 is a fantastic lens in its own right, WR, metal body and what it can achieve with just a 49mm filter size - surely the smallest of the brands for an f2.8 macro in the 90-105mm range! No excuse not to be carrying in the field if you love nature.

The Sony is a wonderful lens for what it is, but it's double the size, double the weight, and double the price!

I love that Pentax lens as well. And those are some good reasons to prefer the Pentax lens over the Sony, imo, all else being equal. What they are not are reasons to dismiss or ignore DxO marks. That there is a lot more to deciding which lens will be best for oneself than DxO marks (or MTF charts, for that matter) is not something I have remotely disagreed with.

[Rondec]

You say they smothered it. They say that they delayed it because of 'production bandwidth,' right? If they 'smothered' it, one should probably ask why? I mean, that smacks of malfeasance. Without knowing their internal workings, I don't know why their explanation shouldn't be believable.

They delayed it for over a year and only released it when another brand had matched it. It was a single entry in a whole list of cameras and they continued to add other cameras, while not putting the data for the 645z out there.

I have no idea why they didn't release it, but "bandwidth" seems duplicitous. If there was an issue for them, it was that they had talked about supporting Pentax medium format and then back tracked on that. It still would have been worthwhile putting the data out there, if they are truly interested in photographers having information about camera sensors.

I don't use DXO Mark much any more, even for sensor data. I find the Photons to Photos site to be more complete and accurate compared to DXO Mark's data.

Anyway, this is probably not a fruitful discussion. You, as you say, have belief in DXO Mark's data and I have more confidence in other sites. The reality is that if you are willing to spend money, you can get some high quality lenses from Sony. They'll have a little different look and feel from Pentax lenses, but they will be fine. Your photos will actually look pretty similar, but maybe you will be happier and that's the most important thing.

[Socrateeze]

Your reply is to my comment. I wasn't referring to test of one unit of a lens. I wrote that DXO tested Pentax lenses with a K5 or K3 (apsc), and DXO tested the Sony macro lens you mentioned in your post with Sony full frame high resolution camera. You say that testing with anything is better than no testing at all. Well, that's not true, at least no testing doesn't influence someone opinion as wrongly as a wrong test result based on a flawed test setup.

I don't think I did say that testing with anything is better than no testing at all. Whether it is depends, of course, on whether the testing provides any information. On the assumption that it does, then some information is better than no information. My issue lies with your contention that DxO marks provide no information.

Obviously, a "wrong test" leads us to believe false things, and that is (usually) worse than a state of complete ignorance.

---------- Post added 27-08-19 at 09:48 ----------

[/COLOR]You were saying: Sony lenses get consistently better scores than Pentax lenses by DXO mark. I replied: DXO tested Pentax lenses with apsc cameras , and tested Sony lenses with full frame higher resolution cameras. There is nothing to argue about, that's pretty clear, end of story.

What in the above do you think is something I disagree with?

DxO also tested the Sony 90mm macro on aps-c cameras. When tested on similar resolution aps-c sensors the Pentax scored markedly less for sharpness than the Pentax 100mm macro.

---------- Post added 27-08-19 at 09:49 ----------

[/COLOR]

How can DXO have tested the lenses on a Pentax full frame camera while they tested the lens the Pentax K1 did not exist! By the way, I don't know what you are talking about, because it is not possible to get DXO test results of the DFA100 macro with the Pentax K1 since the K1 is not available in the list: Pentax smc D FA MACRO 100mm F2.8 WR - DxOMark

Again, I'm not sure what your point is or why you think I disagree with the above. I never suggested that DxO tested the 100mm macro on a FF camera. Neither did I suggest that they tested it with the K-1.

[ffking]

Just to be a bit tangential here, is it just me, or does anybody else have the impression that nobody any longer talks about leaving Pentax (or any brand, really) for anybody except Sony these days? While that is probably rational in many cases, and I certainly don't ant to suggest anything to the contrary for any members here, it is a worrying trend if you value plurality and diversity among camera offerings, and it seems driven by marketing spend and influencing influencers. Sony are almost certainly the best bet for some photographers and some types of photography, but the same could be said of pretty much any brand when you get doen to the reality of images making rather than spec sheets and hype.

[biz-engineer]

it is a worrying trend if you value plurality and diversity among camera offerings, and it seems driven by marketing spend and influencing influencers

I see some lens makers offering new lenses only for Canon and Sony mounts, I guess it is related to current and future market share. Sony have succeeded long ago to federate third parties to make lenses for them (using the name of Zeiss to associate with dedicated lens designs for Sony), it's a marketing/business masterpiece from Sony. At some point, business size does matter, it is snowballing because business partners (including test lab and reviewers) tend to associate with who will potentially bring them the most revenue. So in the future, we may have two choices, Canon and Sony, for business reasons essentially. The kind of equipment we get is the kind of equipment we are told we should get, for a large part, not related to what we want, but it is business driven.

[Rondec]

Just to be a bit tangential here, is it just me, or does anybody else have the impression that nobody any longer talks about leaving Pentax (or any brand, really) for anybody except Sony these days? While that is probably rational in many cases, and I certainly don't ant to suggest anything to the contrary for any members here, it is a worrying trend if you value plurality and diversity among camera offerings, and it seems driven by marketing spend and influencing influencers. Sony are almost certainly the best bet for some photographers and some types of photography, but the same could be said of pretty much any brand when you get doen to the reality of images making rather than spec sheets and hype.

It has to do with the fact that we are focused more on cameras as gadgets than on cameras as still image creation machines.

As such, things like video, frame rates, and features like that end up taking priority and Sony is about the best at creating gadgets that happen to take photos. If you are focused on glass, I would argue that Canon is actually creating the most interesting line up of glass for their new mirrorless model -- most of it is priced way out of my range, but so are the G Master lenses.

[repaap]

I see some lens makers offering new lenses only for Canon and Sony mounts, I guess it is related to current and future market share. Sony have succeeded long ago to federate third parties to make lenses for them (using the name of Zeiss to associate with dedicated lens designs for Sony), it's a marketing/business masterpiece from Sony. At some point, business size does matter, it is snowballing because business partners (including test lab and reviewers) tend to associate with who will potentially bring them the most revenue. So in the future, we may have two choices, Canon and Sony, for business reasons essentially. The kind of equipment we get is the kind of equipment we are told we should get, for a large part, not related to what we want, but it is business driven.

This might just be exactly what Sony has been going for. And it does seem to work, so far. Canon has to ansver, because their biggest opponent at the moment is Sony.


Sony want to be the new Canon.

But that does not mean that smaller players are not needed. Dunno how this has been forgottern. Oh it has not. it is just again that loudest one streers the look towards something else.

[Socrateeze]

Dxomark is an inferior tabloid page for kids.

Anyone with a hint of skill just needs to add 1 and 1 to see their blunder.

Here is an MTF comparison of the Pentax 100mm macro versus the Canon L glass on similar res, similar AA filter sensors:
Pentax smc D FA 100 mm f/2.8 Macro WR review - Introduction - LensTip.com

Read more at: New zoom 70-200 F4 - Page 25 - PentaxForums.com
This is in sync with:
Pentax-D FA 100mm f/2.8 Macro WR Lens Review | ePHOTOzine

And now you can compare the Canon L glass which performs exactly like the Pentax macro with the Sony:
Canon EF 100mm f/2.8L IS USM Macro Lens Average MTF

Dang. The newer 1,000 EUR Sony underperforms the Canon L glass (and so the 500 EUR Pentax 100mm macro) by quite some margin.

So we find the Sony is inferior and ridiculously overpriced (> 100%).

And we again find that Dxomark hasn't got a clue and their "scores" can not be trusted at all.

Oh, did we already mention that the Pentax lens is 340g and the shorter Sony is 602g / 62mm filter? Sony FE is twice the weight, twice the price and worse performance. Didnt someone claim mirrorless is lighter?

Thanks for trying to provide some evidence, but I think you're misreading the text and the graphs. To use your analogy, you are adding 1 and 1 and getting 3 (or at least 2.5). Let me explain...

Firstly, the quote from Lenstip: "From our point of view it would be difficult to get anything better....". This refers to their (then) new ability to compare Pentax lenses with other lenses with similar size and resolution sensors, not to the quality of the lens. You might seem to think that that phrase was evaluating the Pentax lens as "difficult to get any better", as is possibly suggested by you putting it in bold and then saying that "This is in sync with" the ePhotozine review, which is a very positive review of the lens.

Secondly, the MTF graph showing the sharpness of the Pentax as being in line with that of the Canon on the Lenstip webpage is *for the centre of the image*. If you check the MTF graph later in the review (Pentax smc D FA 100 mm f/2.8 Macro WR review - Image resolution - LensTip.com) you can see this to be the case. That matters because if you then compare the MTF charts you cite comparing the Canon with the Sony, you can see that the Sony is in fact *better* than the Canon at the centre of the image (by a relative measure of 0.6 for the Canon compared to 0.65 for the Sony for 50lp/mm). It follows that if you were to transfer the Sony to the original comparative Canon/Pentax MTF chart you would see the Sony as scoring *better* than both the Canon *and the Pentax*, not worse.

And the comparison between the Canon and Pentax is also not quite fair, either. The Canon 50D has 15 effective megapixels, whilst the Pentax K-5 has 16 effective megapixels. That not a huge difference, but it does mean that since the lenses resolve with equal sharpness on those cameras, that the Canon lens is actually a bit sharper, at least at the centre, than the Pentax lens.

Now the MTF charts comparing the Sony with the Canon certainly suggest that the advantage of the Sony is more towards the centre of the image, but that advantage is lost to the Canon towards the edges. Kudos to Canon for its superior edge sharpness, and kudos to Sony for its superior centre sharpness! Which, then, is the better lens? Well, on sharpness, it depends on whether centre sharpness is more important to you than edge sharpness. And before you overinterpret the MTF charts even more, remember that the poorer performance of the Sony at the edges than the Canon is not as important on the crop sensors used in the Lenstip article because they crop out most of the edge superiority of the Canon over the Sony.

So what we've learned from Lenstip is that if the DxO marks carry similar information we should expect similar scores from the Pentax and the Canon for the same lenses and bodies as Lenstip uses for their comparison. So let's look at the the DxO marks and what do we see?:

DxO mark for sharpness for the Canon EF 100mm f/2.8L Macro IS USM on the Canon 50D: 8

DxO mark for sharpness for the Pentax smc D FA MACRO 100mm F2.8 WR on the Pentax K-5: 8

Well, quelle friggin' surprise!! Exactly the same! Precisely what you'd expect given the Lenstip measurements! Ironically, far from undermining the suggestion that DxO marks contain information about relative sharpness, the Lenstip MTF graph actually corroborates it!

And what about the superior Sony scores? Well, DxO places greater weighting to the centre sharpness over edge sharpness. That would not be my choice, but hey, maybe that matters more to other people than to me. In any case, it is no great surprise that the DxO scores the Sony higher than the Pentax or the Canon on similar resolution sensors, especially on crop sensors that crop out the edges where the Sony would seem to be relatively weaker when compared with than the Canon. Unfortunately, DxO do not test the Pentax lens on the K-1. If they did, I suspect you'd find the Pentax closer to the Sony than the aps-c comparisons might suggest. It's possible that the Pentax macro would even beat the Sony, but that's rather speculative.

And as for the claim that "mirrorless is lighter", I never made any such a blank claim. My claims regarding weight savings refer to specific comparisons of lens sets that I might want with the K-1 vs the a7r4. And the Sony set-up would be lighter for me from between 800g and about 3kg. I wasn't including macro lens in that because I don't tend to carry a macro lens together with my other lenses.

[repaap]

Not at offence, but why all of this noise?

I think that you have ’proven’ your way of thinking and wether it is right or not, there has been discussion about that.

Sony seems to fit the list of what you want? Is that worth the money? Will you actually benefit of any of that? If yes. Why not let this go. And take that step? There is also a new discussion about this DxO thing at other place on this very same forum. If you want to have that round on and against Pentax?

This seems not to have anything to do about this rumor anymore

Thanks for trying to provide some evidence, but I think you're misreading the text and the graphs. To use your analogy, you are adding 1 and 1 and getting 3 (or at least 2.5). Let me explain...

Firstly, the quote from Lenstip: "From our point of view it would be difficult to get anything better....". This refers to their (then) new ability to compare Pentax lenses with other lenses with similar size and resolution sensors, not to the quality of the lens. You might seem to think that that phrase was evaluating the Pentax lens as "difficult to get any better", as is possibly suggested by you putting it in bold and then saying that "This is in sync with" the ePhotozine review, which is a very positive review of the lens.

Secondly, the MTF graph showing the sharpness of the Pentax as being in line with that of the Canon on the Lenstip webpage is *for the centre of the image*. If you check the MTF graph later in the review (Pentax smc D FA 100 mm f/2.8 Macro WR review - Image resolution - LensTip.com) you can see this to be the case. That matters because if you then compare the MTF charts you cite comparing the Canon with the Sony, you can see that the Sony is in fact *better* than the Canon at the centre of the image (by a relative measure of 0.6 for the Canon compared to 0.65 for the Sony for 50lp/mm). It follows that if you were to transfer the Sony to the original comparative Canon/Pentax MTF chart you would see the Sony as scoring *better* than both the Canon *and the Pentax*, not worse.

And the comparison between the Canon and Pentax is also not quite fair, either. The Canon 50D has 15 effective megapixels, whilst the Pentax K-5 has 16 effective megapixels. That not a huge difference, but it does mean that since the lenses resolve with equal sharpness on those cameras, that the Canon lens is actually a bit sharper, at least at the centre, than the Pentax lens.

Now the MTF charts comparing the Sony with the Canon certainly suggest that the advantage of the Sony is more towards the centre of the image, but that advantage is lost to the Canon towards the edges. Kudos to Canon for its superior edge sharpness, and kudos to Sony for its superior centre sharpness! Which, then, is the better lens? Well, on sharpness, it depends on whether centre sharpness is more important to you than edge sharpness. And before you overinterpret the MTF charts even more, remember that the poorer performance of the Sony at the edges than the Canon is not as important on the crop sensors used in the Lenstip article because they crop out most of the edge superiority of the Canon over the Sony.

So what we've learned from Lenstip is that if the DxO marks carry similar information we should expect similar scores from the Pentax and the Canon for the same lenses and bodies as Lenstip uses for their comparison. So let's look at the the DxO marks and what do we see?:

DxO mark for sharpness for the Canon EF 100mm f/2.8L Macro IS USM on the Canon 50D: 8

DxO mark for sharpness for the Pentax smc D FA MACRO 100mm F2.8 WR on the Pentax K-5: 8

Well, quelle friggin' surprise!! Exactly the same! Precisely what you'd expect given the Lenstip measurements! Ironically, far from undermining the suggestion that DxO marks contain information about relative sharpness, the Lenstip MTF graph actually corroborates it!

And what about the superior Sony scores? Well, DxO places greater weighting to the centre sharpness over edge sharpness. That would not be my choice, but hey, maybe that matters more to other people than to me. In any case, it is no great surprise that the DxO scores the Sony higher than the Pentax or the Canon on similar resolution sensors, especially on crop sensors that crop out the edges where the Sony would seem to be relatively weaker when compared with than the Canon. Unfortunately, DxO do not test the Pentax lens on the K-1. If they did, I suspect you'd find the Pentax closer to the Sony than the aps-c comparisons might suggest. It's possible that the Pentax macro would even beat the Sony, but that's rather speculative.

And as for the claim that "mirrorless is lighter", I never made any such a blank claim. My claims regarding weight savings refer to specific comparisons of lens sets that I might want with the K-1 vs the a7r4. And the Sony set-up would be lighter for me from between 800g and about 3kg. I wasn't including macro lens in that because I don't tend to carry a macro lens together with my other lenses.

[Serkevan]

God, such walls of text for a sharpness measurement that ends up being mostly irrelevant in any pictures that people will actually want to look at...

[Sandy Hancock]

I thought this thread was about the roadmapped DFA70-200/4.0

Can we get back on topic? Thread bans may ensue for recalcitrants.

[Tri_BK]

Hope someone can bring some news about this DFA70-200.