Originally posted by Not a Number I would look n the K-S1/K-S2 and K-30/k-50 section of the forum at all the images that have been posted from these cameras as well as the comments from the posters and judge for yourself rather than rely on a bunch of numbers that have little to do with human perception.
Looking at images is a good thing, but much of what you will be seeing is how good the photographer was. A bad photographer can make even the best camera look bad. That's why lab numbers can be helpful.
I've owned numerous DSLRs over the years, starting with old 4/3 DSLRs, then Canon APS-C, then Pentax, along with various generations of m43 sprinkled in along the way. I've also owned some large-sensor compacts, such as the Olympus XZ-1, Sony RX100 Mark I and II, and the FF Sony RX1, and the Panasonic LX100. And I've gotten to use many other great cameras along the way.
What I've learned is that the results from the DXoMark sensor score track very closely with the relative sensor performance I have seen with my own eyes over years of use and thousands and thousands of images as I upgraded from one camera to the next. So for me, DXoMark is a very handy means to compare sensor performance, and I don't get why some people are so quick to dismiss it like it's all smoke and mirrors.
Now when we're talking about lenses, then it's true that things get much more subjective, and hard numbers simply cannot tell the story of a particular lens. But sensors, being the electronic devices they are, are much easier to benchmark in the lab.