As the title says, looking at the relative performances of all the cameras and lenses is only part of the story. Shooting raw files opens up all sorts of possibilities, and we are all free to make as good (or bad!) a job of processing the files as we do, using various software programmes.
In terms of getting the best possible quality out of my images I use Adobe Camera Raw in Photoshop CS6. This allows me to produce what I see as the best files possible for the markets I serve.
The Sony NEX-7 came 4th = on my sharpness list. However I am able to get better results than from my test shoot by using various presets I have set up in ACR and Photoshop itself.
The image on the left was processed under the conditions of the test. The one on the right using my own preferences for NEX-7 files. Below are some 100% blowups from each.
As can be seen, the bottom shot (my presets) is both sharper and has a cleaner smoother sky.
So, basic testing using default parameters is just a starting point, and one of the reasons why its difficult to make judgements on cameras with only jpg. files to look at. Its always astonished me why review sites, in their eagerness to get their reviews out, make judgements on cameras from out of camera jpg. usually shot on pre-production firmware. It also always astonishes me that manufacturers seem quite happy to let them do that.
Fortunately not all sites do this. Dpreview always wait for a full production camera and ACR support before doing their full review.
So are tests like the ones I carried out useful? Well yes, of course they are. Its always useful to have a like for like comparison between alternatives to see what the differences might be, as it gives an indication as to what might be required to produce optimum results. For example, I never apply any sharpening to Sigma DP2 Merrill files. There is no need to anyway and doing so doesn't improve the image, in fact it degrades it. However for all the other cameras and lenses I used sharpening is a requirement to compensate for their AA filtered bayer sensors.
I've always found different files respond differently to sharpening. Files from my Nikons - D800E and D600 respond well. A fair amount of sharpening can be applied without degrading the image in a noticeable way. However with my NEX-7 and m4/3 cameras this has to be applied carefully and the files usually require some kind of "smoothing" for blocks of colour like those found in blue skies to remove the luminance noise ("graininess") that this sharpening creates.
However over time I've learned how to get optimum results from smaller sensor cameras in post-processing and get them to "punch above their weight". So much so that I now upload my 16MP m4/3 images upsized to 24MP. And no they aren't as good as my D600 files for example, but they are accepted by the libraries and people buy them.
So ultimate image quality is worth pursuing and worth testing for initially within the gear, but for me it comes as well from raw file development and Photoshop post-processing. If you are stuck with just jpgs. then you are just that, stuck. My finished processed files from raw when outputted to jpgs. are always an improvement as far as I'm concerned and indeed its very rare for me to get an out of camera jpg. accepted by a library. They are just too soft mostly. I've always thought they are a bit like the results from having your film processed at some mini-lab as opposed to an image being carefully developed and printed.
So what I do tends to equalise my various camera / lens combinations somewhat. My DP2 Merrill and D800E will always produce sharper files but with care I can get my other cameras to come close to that standard. It does require some work, but then quality usually does.