[tuco]

I don't know Norm. I don't get into technical theory with my digital gear very much. If your lenses look good to you, perhaps just don't worry about it and enjoy using them. It certainly not worth getting all worked up over and is easier on your blood pressure too.

[normhead]

The Perceptual MPix measure confirms certain rules of thumb such as “a 12 MPix full-format camera is sharper than an 18 MPix APS,” and can verify if the kits offered by manufacturers will properly support a hardware update, etc. Many exciting issues are now quantifiable, and DxOMark will soon offer detailed analyzes based on these scores.

This is an easily refutable statement from the DxO site surrounded by lot's of text, legally worded to make it sound like other science agrees with it while not actually saying that, and stating that more great information is coming.

They never answer the question when is an APS-c camera shaper than an FF camera, hence, a bunch of biased hacks, presenting only one side of the equation. The reason I know these guys are absolutely full of it is th research i did before I opted for a K-3 instead of a D610 or 6D

Here are the test charts shot at base ISO, in "not in your mom's basement " conditions, over at Imaging Resources.









We are comparing a 24 Mp APS-c to a 24 Mp FF and a 20 MP FF. IN both cases the images are practically indistinguishable if you look at the sharpest parts of the image. I put this on the same level as Ford Motor Comapny's false mileage claims. With the exception that these guys are some kind of "test" ellectual freaks who's only excuse for being so wrong on such a large scale is they are so wrapped up in their own heads, they have no problem presenting such outrageous claims as facts. They claim they have done tests to support their data. I say they are faking it. Lots of real scientists have faked data for fame and popularity, (thats why peer review and replication are so important), do you really think it couldn't happen in a commercial business?

Or as my dad (the PhD) used to say, don't look at the conclusions look a the data. The corollary to that is, if the data isn't available, you have no way of checking whether you agree with their conclusions.

Sorry, but all the evidence (not measured in DxOs basement and conveniently unavailable for public viewing) points to these people being frauds.

I don't know about the laws in the States, but in Canada, there's enough evidence of them being frauds to defend the statement in a court of law.

[normhead]

I don't know Norm. I don't get into technical theory with my digital gear very much. If your lenses look good to you, perhaps just don't worry about it and enjoy using them. It certainly not worth getting all worked up over and is easier on your blood pressure too.

It is one of the curses of my life, is I was brought up by a very smart BS detector, who was endlessly outraged by the license taken in advertising. The DxO lens tests are an extension of their site advertising. If they can get you to their site, they have a chance of selling you their stuff. So why not just build a whole site based on bogus stats that reflect the status quo and conventional thought. IN other words, instead of doing any real research, why not just stroke the backs of those with the money for expensive gear and big wallets, tell them their stuff is better than anyone else's, then if it's not "buy our software, we'll restore you to your rightful place in the universe." It pains me to see people quoting this hogwash, as if it weren't selective research designed to sell software.

But in a way it brings me joy. The old adage, never trust a commercial enterprise to be truthful about anything.. holds true once again. Tested many times and never let's me down. Or another, if a commercial enterprise claims something is true, it might be true, but it's that small portion of the truth that interested their market, and is a tiny subset of the whole truth.

This case is a perfect example.

[Clavius]

How is this fraud if they disclose their method of testing openly on their website?

And how is it fraud if they consequently always use the exact same method and setting for testing?

So their score is based upon the best results of a certain lens under very controlled conditions that are always the same for each lens that they test. Sounds perfectly valid to me.

---

Last edited by Clavius; 01-31-2014 at 08:18 AM.

[normhead]

Because the represent their little corner of the truth, as if it were the whole truth. And that's the best I can say about them. I'm also suspicious that they actually selected which corner of the truth they were going to examine to create a distinct (read incredibly biased sample". There are many ways of being fraudulent.

Once you realize, this is not science it's advertising... then you'll understand the situation. Like the guys in the lab coats jumping onto airbags, and saying more airbags is better. This is advertising where the actors where lab coats to look like scientists. If you doubt it's fraud, read back through the site and see how many people assume that there is validity to the DxO tests for daylight shooting. See how many actually know, they are talking only about low light images, not about daylight , cloudy, your front room with light coming through the windows. Not one of them has ever mentioned that. That is because the DxO site is laid out in such a way as to obscure the truth. Sure they've done what it takes to legally protect themselves, but the fact that users don't know what they are doing, and assume their results mean something out in the world, just shows how effective they are at perpetrating their little fraud.

If it's not a fraud, point me to the part of their site where the results I posted above are acknowledged, that using a 70mm lens on both cameras and shooting a test chart under commercial lighting, 24 Mp APS-c is virtually identical to 24 Mp FF. The IR images establish that, irrefutably, yet there is nothing on the DxO site that suggests that is even possible, IN fact they suggest that the science says it's impossible. All by omission, innuendo etc which is perfectly legal as far as I know, but still fraudulent in the widest sense of the word. It's a deliberate effort to mislead.

I don't see what's so difficult about this... commercial company stretches the truth for their own commercial gain. Happens all the time. DO you folks really believe DxO has any interest in being completely objective in these numbers? They are selling you product, they have motive and opportunity to mould the truth to benefit their commercial model. Are you naive enough to think they are playing this straight up?

If someone asked you why you drive a ford truck, would you play them an F-150 commercial, would you consider that to be "science"? Would you quote the commercial. Well the DxO "science" is the commercial for their software.

[Clavius]

Admitted, I don't understand the problem. And I may well be the weak factor in that.

They use the results of their "crappy" measurements in their software. Then why is their software so good? And why does it sell so well?

[normhead]

Admitted, I don't understand the problem. And I may well be the weak factor in that.

They use the results of their "crappy" measurements in their software. Then why is their software so good? And why does it sell so well?

Their software from what I understand is excellent, I personally don't use it, I believe correcting every lens to some artificial standard is lunacy, and I don't care to buy what they are selling. Lot's of software is excellent. My choice is Apple Aperture. If the folks that make Aperture start rating cameras based on the custom profiles they do for each camera am I going to consider them to be some kind of experts on the subject?

I'm guessing from the structure of their software, that they are really good at producing images, I believe they said "8x12" from 10 Mpix, when I can get 20x30 from 10 real Mpix, so, I'm not really understanding the benefit to me. Listening to their BS would have cost me, close to $700 dollars. I never would have printed that point and shoot 10 Mp at 20x30, well beyond their standard, and sold three copies at $300 each. Because, according to them 10 Mpix is at best a good 8x12, and 10 Mp on the camera is probably only 6 Mpix on their rating system. They say so right on their site.

I gotta say, producing software that can produce an 8x12 from a 10 Mpix image is nothing special.

What stands out on their site, is how many times they imply things that are simply flat out wrong, and then cover their butts with legal disclaimers on another page somewhere else on the site. Nothing on any front page on their site is flat out what it appears to be, they are all heavily influenced by disclaimers written on other parts of the site. The only pages that are straight up, are the ones with the disclaimers, and even they go to great efforts to legally sugar coat the truth, and justify their bizarre methodology.

But your question does demonstrate the effectiveness of this particular marketing strategy. But a question for you, in the images comparative images I posted from IR, why is their prediction so far off? Why were they unable to predict that at base ISO and commercial lighting, 24 MP APS-c would be almost identical to 24 MP FF?

Why did they say

“a 12 MPix full-format camera is sharper than an 18 MPix APS,”

The images prove them wrong. Which begs the questions, do they really have any images that prove what they say? and why can't we see them? My guess is that images taken in your mom's basement without commercial lighting are so bad, no one would care if one was better than the other.

---

Last edited by normhead; 01-31-2014 at 09:29 AM.

[tibbitts]

I'm not understanding the eight images in the example. The four resolution chart images seem to be pretty comparable, while the one green clothing image is way off. Maybe somebody has a link to the original article, and there's an explanation there?

[drugal]

I think they use 10 megapixels because an 8x12, 300 dpi print is just under 10 Mpix without resampling in software.

[normhead]

I'm not understanding the eight images in the example. The four resolution chart images seem to be pretty comparable, while the one green clothing image is way off. Maybe somebody has a link to the original article, and there's an explanation there?

Yes, the FF images suffer from narrow DoF.

I think they use 10 megapixels because an 8x12, 300 dpi print is just under 10 Mpix without resampling in software.

Which is as ridiculous as anything else they do. Many commercial printers resample our images up to 300 dpi for printing, and tests show, most people can't tell the difference between a 300 dpi image, and a 200 dpi image resampled to 300 dpi, at any size, just more of their technical mumbo jumbo. They don't really know anything, but what the results are with low light images and dismal lighting conditions, which is great if that's what you shoot.

[sniper29a]

I use DxO Optics Pro Elite for over 4 years. Best SW I ever purchased and worth well money invested. Lightroom and other SW are far away from quality DxO OP produces. I would say that you misunderstood what DxO does. These values are for e-experts and tech onanist. They are useless for real life photography. Whole reason DxO does this controlled test is to determine lenses and camera flaws. These theoretical values are then used in DxO Optics Pro to correct images.

Bottom line. Output of DxO OP based on controlled tests of cameras + lenses combinations is what matters most. It is "magic" ;-) Sure, I use also Adobe PS CR. But it never gives me what I want unless I fiddle with all fancy sliders. I rather take pictures and let DxO do the job. Since I have got this state of art SW, I don't bother about lens quality since all is corrected by DxO. Simply clever.

Don't bother about theoretical results and shoot ppl.

[normhead]

I use DxO Optics Pro Elite for over 4 years. Best SW I ever purchased and worth well money invested. Lightroom and other SW are far away from quality DxO OP produces. I would say that you misunderstood what DxO does. These values are for e-experts and tech onanist. They are useless for real life photography. Whole reason DxO does this controlled test is to determine lenses and camera flaws. These theoretical values are then used in DxO Optics Pro to correct images.

Bottom line. Output of DxO OP based on controlled tests of cameras + lenses combinations is what matters most. It is "magic" ;-) Sure, I use also Adobe PS CR. But it never gives me what I want unless I fiddle with all fancy sliders. I rather take pictures and let DxO do the job. Since I have got this state of art SW, I don't bother about lens quality since all is corrected by DxO. Simply clever.

Don't bother about theoretical results and shoot ppl.

I think we can agree that if you like their software, it's good at what it does. There are lots of testimonials on the site with happy customers. That being said, they should take down there lens and sensor ratings, which suggest they have an expertise they don't have, and serve no practical purpose, beyond misleading the public.

My biggest issue with DxO as software, is that of the 10 or so lenses I use on a regular basis, they support 3.

[Rondec]

I think we can agree that if you like their software, it's good at what it does. There are lots of testimonials on the site with happy customers. That being said, they should take down there lens and sensor ratings, which suggest they have an expertise they don't have, and serve no practical purpose, beyond misleading the public.

My biggest issue with DxO as software, is that of the 10 or so lenses I use on a regular basis, they support 3.

It's pretty impressive software, Norm. For lenses that may be kind of weak in points (16-50, 18-135) it can make a huge difference and it has some of the best noise reduction that I have seen.

I don't like it enough to want to pay for it and move away from Lightroom (not as good, but not as time consuming either), but I think it is good software.

I do think their sensor ratings are beneficial and often tell us what we can see with our eyes. You didn't have to be a brain surgeon, when the K5 was released, to realize that it had impressive dynamic range at iso 80. But DXO Mark did quantify it and say it was the best up to that point.

[normhead]

Their ability to mix what seems to be decent research with absolute hogwash is infuriating. That coupled with the inconstancy.

Like the K-5 being rated 2 points ahead of the d7100, yet the only common lens tested on both has a higher rating on the D7100 than it does on the K-5. They have made no attempt to bring any consistency to their findings.

I hope people will check out their products and not waste as much time as I have trying to figure out what anything on the rest of the site means or what it's relevant to. My conclusion, is, it's so disjointed as to be worthless. Hence the title of the thread.

[Rondec]

Their ability to mix what seems to be decent research with absolute hogwash is infuriating. That coupled with the inconstancy.

Like the K-5 being rated 2 points ahead of the d7100, yet the only common lens tested on both has a higher rating on the D7100 than it does on the K-5. They have made no attempt to bring any consistency to their findings.

I hope people will check out their products and not waste as much time as I have trying to figure out what anything on the rest of the site means or what it's relevant to. My conclusion, is, it's so disjointed as to be worthless. Hence the title of the thread.

It is sample variation and test variation. If you test three copies of a lens you may get three different results. If you run the same test on the same lens three times, you will get three results. That doesn't negate the usefulness of the test.