Originally posted by D1N0 There is a difference:
At iso 800 the k-1 appears sharper, while at iso 100 the K-1 II appears sharper (due to the D fa 50 2.8 instead of the FA 77) I don't think it matters as much hat higher iso's because sharpness is already eaten op by noise more, filtered or not. By iso 3200 i have a hard time choosing. Let's hope Ricoh is able to adjust the firmware so we can select the point at which nr kicks in (if at all). Since it is only applied from a certain point, i think that point must be able to be influenced through firmware. Then it is a best of both worlds situation.
Interestingly, the ISO 1600 image on the K-1 II looks sharper than the ISO 800 image in that spot you point out above.
So I think it might be a slight focusing error or de-centered lens.
It still matches the K-1 with a tiny bit of sharpening applied in post.
But there are also other areas of the image that seem to favor the K-1 II in the same ISO 800 comparison images.
The point is that the differences are more due to the use of 2 different lenses, each with their own set of strengths and weaknesses, faults, etc, and the camera settings than anything else. For areas with fine line detail, I like the rendering on the K-1 II much better since there is less false color to deal with in post; the usable result after processing is better out of the K-1 II, but in an admittedly minute sense.
At truly high ISO settings like 12800, the K-1 II is better at detail retention, with smoother looking files that respond much better to typical post-processing (sharpening, balancing, color noise reduction).
It's a close call but to alleviate concerns (for the record, I'm not one of the concerned individuals), Pentax should add a firmware setting to either greatly reduce or completely turn off the effect of the signal processor at ISO 12800 and below.
The second point I want to make is the concept of noise reduction as being viewed as a bad thing automatically, if the DPR review and most of the comments are to be taken at face value. I normally agree with this, but in this situation, the noise reduction is as a result of a hardware, not a software algorithm to smear away details. What do others feel about this? My feeling is that it's giving the photographer smoother images that respond slightly better to sharpening and need a lot less color noise reduction, which yields crisper images if you know what you're doing. I can't comment on people that just want to use JPGs straight out of the camera (don't get this at all), or just want to use some "automatic" RAW converter that decides how your images look for you (don't get this approach either).