Originally posted by wkraus I tend to think that most of what we are discussing as sensor characteristics really results from raw processing. Comparisons do not make much sense if we don't know what we are comparing: sensors, or jpeg engines, or raw converter settings... Today I did my own test (K10D vs. K-5II, both with FA31 Limited, f/8), and while there are subtle differences, I can get results than are very similar with appropriate settings in ACR:
Nice comparison shots - thanks for doing this test. I'm one of folks flapping their gums about CCD sensors giving punchier, more film-like results. I've owned a bunch of Pentax DSLR bodies, but never a CCD and CMOS body at the same time (I currently own an *ist DS2 and a K10D, but have also owned a K100D, K200D, K-x, K20D, and K-5). I had progressed through to the K-5, and then decided to back down to the K10D to get the CCD results I loved so much out of the older models I'd owned. So I'm going to now play devils advocate about the comparison you present.
First thing is from your description of doing the test, it sounds like you set out with the intent of making the images look the same ("I can get results that are very similar with appropriate settings in ACR"). I'm not saying that's a bad approach - it's a worthy comparison strategy, but if you set out to see if you could get the images to look the same, then you might be subduing the punchier color we claim the CCD sensor is capable of.
The second thing is that you scaled both images down to a smaller equal size, so differences other than color may be equalized. For example, I'm convinced that by packing 16 million pixels versus 10 onto the same size sensor and also working hard to keep noise levels low at high ISO settings makes for softer images at low ISO settings with the K-5 or K-5II. But if you scale both images down to 1000 pixels or less, however, the possibly softer image from the K-5II may tighten up and look very similar to the K10D image. I'm willing to bet if you did a 100% crop out of the middle of both test images, while the K-5II crop would show details at a larger size, it would also appear softer. From personally owning both cameras, I'm still convinced that what you gain in mega pixels you lose in sharpness, and that when pixel peeping, it's only when you scale down a K-5 image to the size of a K10D image do you see similar sharpness at 1:1 pixel ratio.
Lastly there's my own claim that CCD images from the K10D are more "film-like", which I love because I still love and still shoot film. My thinking there is that as noise does creep in with the K10D sensor, as long as you remove the color noise in ACR or equivalent, the luminance noise that's left is what's giving the images a more film-like look. I definitely know this is subjective, and I know some people would rather see less noise - period - regardless of the nature of that noise. But personally, I sometimes shoot at ISO 800 on the K10D in broad daylight because I like the film-like noise it will give my images.
Again, this is all subjective and just my opinions. The K-5 or K-5II are fantastic cameras capable of amazing images - better in some way, different in other ways.