Inspired by the noise/zoom discussion in the first part of this thread, I decided to conduct a an objective trial. Seems that people on both sides of the discussion were talking in hypotheticals without any pictures to back up their points. And I admit I was equally guilty in that regard.
So the attached pics are as close to an apples-to-apples comparison as I could do. You're looking at crops of an identically-framed shot with identical exposure, taken with both the K10D and K100D. What you are seeing is a crop, but
not a 100% crop from either camera. It's the same angle of view taken from each identical original picture. The reason I didn't use the whole image is because resizing the whole image to fit as a forum attachment would have resulted in far more downsampling for both cameras than I would normally do, and so would be a pointelss comparison. This section exported to about 800 by 560 pixels represents a comparable amount of resizing compared to what I might publish as a large web image (say, 2000 pixels on the long side for the whole image).
Also attached is the original shot as it came out of the K10D, just so you can see how much of a crop we're talking. The original from the K100D looks so close to the same it's not worth taking up space with.
Both shots were RAW captures at ISO 1600, 1/15 sec, f/2, using an FA 35 f/2 AL. Lightroom settings were identical between them, and no noise reduction was performed.
As you can see, the K10D has a lot more chroma noise, especially visible in the shadow on the side of the amp. Subjectively, I'd say the K100D is less noisy overall, and what noise does exist is a little more appealing.
I don't have any noise reduction tools outside what Lightroom and Photoshop offer. But attempts to apply noise reduction to both images never made the K10's picture look as good as the K100's. The extra resolution doesn't seem to help the Adobe programs' efforts. Maybe a third-party application would work more magic...I'm open to that possibility.
Sure this is one shot, and doesn't tell the whole story. But I think it somewhat refutes the idea that the noise differences between the two cameras even out when you output to the same size. That's exactly what I've done here, and the K10D is still worse.