[normhead]

8%? AN 8% increase in MTF numbers? How does removing a filter give you 8% more pixels?

[Digitalis]

8%? AN 8% increase in MTF numbers? How does removing a filter give you 8% more pixels?

Removing a filter doesn't magically give you 8% more pixels, you should know better than that. Perhaps what would have been better to say is that those 1,304,000 pixels now convey information that would have otherwise have been obscured by Bayer AA filter blur.

[normhead]

OK, so now we have the pixel thing cleared up...where did the 8% number come from?

By the way... the kit lens at 28 mm produces an MTF of 2604 @ F8 tested on photo zone. The 31 ltd tested best score anywhere in an MTF of 2563, less than the kit lens. The DA 18-135 tops them both at 2683. @ 5.6 22mm.

SO at least for Pentax, if you have the kit lens, the only possible way to increase your MTF is to go filterless. But I've yet to see any numbers indicating that going filterless increases MTF on a K-5.

The 35 2.4 tested on photozone tops out at 2723 MTF. Considering that the maximum possible is 3200 (based on alternating black and white lines of pixels over the whole image.) that's a pretty good number.

Just a brief glance at the photozone numbers would suggest that there's little gain to be had buying better lenses, in terms of MTF When you look at images from different lenses, you realize that sharpness and the preservation of detail are not the characteristics that make a great lens a great lens.

So I'd have to say, the evaluation of the K-5II and K-5IIs might have to be made on other values besides sharpness. Sharpness doesn't explain why the 31 is a great lens, and the 18-135 isn't. The 18-135 is sharper.

[Digitalis]

Read the report norm, there is a link to it in the OP. Enjoy.

[normhead]

Read the report norm, there is a link to it in the OP. Enjoy.

Ok I read the report, the magical 8% number came out of their lab, I have no idea what it means and neither do you, because no data are provided. They look like some kind of fantasy to me. I was hoping somebody had something better. Thanks for your bogus interjection.

[halfspin]

yeesh.

[deadwolfbones]

Ok I read the report, the magical 8% number came out of their lab, I have no idea what it means and neither do you, because no data are provided. They look like some kind of fantasy to me. I was hoping somebody had something better. Thanks for your bogus interjection.

As we describe on our "how we test" page, we shoot all our tested cameras on a standard resolution chart and run the shots through Imatest, like most other review sites that do lab testing.

For this review, since we had both cameras on-hand, we shot both the K-5 II and IIs at the same f-stops and focal lengths (RAW and JPEG (highest quality)) using several lenses. We compared the resolution figures at various points on the chart, at three focal lengths and aperture settings, and found similar differences at each ROI (region of interest). These differences were corroborated by the "Overall Res Average" figure that Imatest spits out. With the 18-55 kit lens, for instance, the overall res average was 1389.306 lw/ph @ MTF50 from the K-5 II, compared to 1507.488 lw/ph from the K-5 IIs. This is roughly an 8% increase in resolution—or, to be more accurate, an 8% decrease in resolution for the filtered sensor. The results were similar with the other tested lenses (DA 35/2.4 and FA 77/1.8), keeping the average across the three lenses right around 8%.

Hope this helps.

[Docrwm]

Removing a filter doesn't magically give you 8% more pixels, you should know better than that. Perhaps what would have been better to say is that those 1,304,000 pixels now convey information that would have otherwise have been obscured by Bayer AA filter blur.

And indirectly impact the surrounding pixesl frankly. The sharpness difference between the K-5 and K-5iis is simply stunning. The conclusion that it's not worth the Princely sum of $100 (or $330 in Europe) is simply silly. Sorry.

[normhead]

Terrific, exactly what I was looking for. One of the things I've noticed deadwolfbones is that every site seems to come up with different MTF numbers.... yours seem to be about half what the others are. Is there any light you can shed on that? I've always assumed that every lab is constant, but that you can't compare across different labs. Is there any light you can shed on that? Thanks for your information.

[brntoki]

Yes, the difference is small -- but so are the differences between lenses whose cost varies by hundreds or thousands of dollars.

Ding, ding, ding! We have a winner! Who on earth would conclude that a $100 difference isn't worth a not insignificant increase in detail and sharpness??? A non-photographer is the only reasonable answer that comes to my mind. Anyone who knows about photography knows that photographers spend waaaay more than a $100 for even less improvement. Heck! We'd pay twice that amount even if we KNEW it was a placebo effect we were getting.

Not throwing stones, but really, this goes back to the stone-age of photography. Photographers are obsessive about detail and sharpness almost without exception . . . and they'll mortgage their souls to get it.

[normhead]

And indirectly impact the surrounding pixesl frankly. The sharpness difference between the K-5 and K-5iis is simply stunning. The conclusion that it's not worth the Princely sum of $100 (or $330 in Europe) is simply silly. Sorry.

I considered the difference between a D800 and K-5 to be stunning, maybe you could quantify stunning on that scale. Like a If a K-5 II to a D800 is 100% stunning, how stunning is a K-5II to a K-5 IIs or even say to a D600? I need to define stunning.

I did look at the pictures posted as well as the ones at Imaging Resource... I'm still sitting at "better but false colour is holding me back". Which again returns to the unanswered question, how much of a role does sharpness play in how much we enjoy a good photograph? What is the value of an 8% increase in sharpness?

This line of thought is being encouraged for my preference for some of my 21 ltd images ahead of my 18-135 images taken in the same location, even though, the 18-135 is technically sharper.

[deadwolfbones]

Terrific, exactly what I was looking for. One of the things I've noticed deadwolfbones is that every site seems to come up with different MTF numbers.... yours seem to be about half what the others are. Is there any light you can shed on that? I've always assumed that every lab is constant, but that you can't compare across different labs. Is there any light you can shed on that? Thanks for your information.

I'm not sure whether other sites are using peak or average numbers. And there are also weighted and unweighted figures. (In this case, we used weighted figures—the highest unweighted figures were up to 600 lw/ph higher on a single ROI.) Then you have sample variation among lenses and bodies, as well as variance in test procedures, lighting, chart distance, etc. Finally, Imatest spits out a veritable mountain of data that can be used and picked and chosen from in all kinds of ways—unless a site gives a step-by-step description of how they get their results, it's pretty hard to tell which data they're drawing from. (And most sites won't make that info public.)

In general, I think your instinct to compare only among results from a given lab is correct.

[deadwolfbones]

By the way, I'd like to thank (almost) everyone for the civil and constructive discussion about the article's conclusion so far. You've definitely brought up some arguments we hadn't considered, and that's helpful for us going forward and writing similar pieces in the future. (Personally, based on the test shots, my experience shooting both cameras, and my usual use cases—almost entirely online display, with very little printing—I still think I'd buy the K-5 II if I were in the market. But as others have suggested, any choice involving money is a very personal decision.)

[normhead]

Ding, ding, ding! We have a winner! Who on earth would conclude that a $100 difference isn't worth a not insignificant increase in detail and sharpness??? A non-photographer is the only reasonable answer that comes to my mind. Anyone who knows about photography knows that photographers spend waaaay more than a $100 for even less improvement. Heck! We'd pay twice that amount even if we KNEW it was a placebo effect we were getting.

Not throwing stones, but really, this goes back to the stone-age of photography. Photographers are obsessive about detail and sharpness almost without exception . . . and they'll mortgage their souls to get it.

Andy Warhol was not at all obsessive about sharpness. My studio teacher at Ryerson PI recommended keeping a throw away camera in your car glove box... he said " Those plastic lenses are better than you think (because you can mold plastic to parbolic shapes you can't produce with glass) and produce some interesting shots." Maybe more photographers need counselling when it comes to their obsession on sharpness. Many of the world's greatest images were not taken with the world's sharpest lenses. Sorry to be argumentative, but lets say there is an 8% increase in sharpness. Let's use the increased pixel theory as espoused by Digitalis, let's say I'm printing at 100 DPI on my K-5 II, so a 49 inch print. (4900 x 1.08 =5229) I'm assuming I can now print to 52 inches on my K-5 IIs with the same clarity. So we're talking another 3 inches, bigger, 1.5 inches at 200 dpi. That would be my perspective, am I wrong here?

[brntoki]

I considered the difference between a D800 and K-5 to be stunning, maybe you could quantify stunning on that scale. Like a If a K-5 II to a D800 is 100% stunning, how stunning is a K-5II to a K-5 IIs or even say to a D600? I need to define stunning.

I did look at the pictures posted as well as the ones at Imaging Resource... I'm still sitting at "better but false colour is holding me back". Which again returns to the unanswered question, how much of a role does sharpness play in how much we enjoy a good photograph? What is the value of an 8% increase in sharpness?

This line of thought is being encouraged for my preference for some of my 21 ltd images ahead of my 18-135 images taken in the same location, even though, the 18-135 is technically sharper.

Strangely perhaps, I somewhat agree with this. What I mean is, for me it isn't about overall resolution, but about the sharpness of what is resolved. I will probably order a D800 (not the E) in the next couple of days, but I still long for a K-5IIs. The D800(e) has loads more resolution, but it doesn't come out as sharp (well defined, if that helps some) as the IIs. After years of scanning 35mm and MF film, it's just so difficult to get the files as sharp as they should be.

And what is worse: How do you get the sharpness problem corrected??? I mean really corrected??? This is a HUGE deal for me. When something isn't as sharp as it seems it is supposed to be, it's like black magic to try to get it right. Did I go far enough? Did I go too far? I just hate it. With the files that the 5IIs seems to deliver (never used one), and the ones I've downloaded and tested, it is a breeze to get it properly sharp. With false color or miore, it's not a concern at all of how far or too far? It's easy to see when it's there, and it's easy to see when it's gone, i.e., fixed. Really fixed. Knowing I can "really" fix false color and moire poses a much, much, much more manageable issue than dealing with soft files that need voodoo magic sharpening.