What I'd like to see is head-to-head RAW comparison between the
The only way your going to see anything like this that is relevent is to do it yourself and it will take an investment.
To test any camera in the style and genre you work is to pretty much do it in the worst conditions imagineable.
If you shoot 300 images a day, test it by shooting 1,000. If you post processes out small jpegs it batches in 30, do large tiffs in batches of 500, if you never change out of camera or sotware settings, then start moving the expsoure around +1 up to 2 stops and try to correct them.
Rename, edit, sort and move the images around in the software. Process images in the background why you shoot.
Do this on your primary AND backup computer.
If you work in bright sunlight, try shooting in a deep shaded area, with bounce light for fill, if you work with 400 watts of strobe pump it up to 2,000 watts.
Only under the most stressful, harsh, changing conditions will anyone know what works in a professional atmosphere.
All of these cameras and sensors can be scene specific. Some cross more territory than others, some are more affordable, some offer better service, some are more reliable but all of them take a learning curve that no review or microscopic examination of raw files is going to inform until you work in your specific genre under your specific conditions.
Pleasing a "client" is important, but remember the client usualy hires you for the highest level you can obtain, so pleasing yourself is really why you are in the room in the first place.
Basically, this conversation just rages on from year to year and quite honestly I've gotten caught up in it in the past, but it's just hyperbole and opinion until you try this yourself.
There are a lot of dealers that want to sell a lot of cameras and it won't take too long to find one that will let you demo any of this equipment to your satisfaction.
P.S. and don't just try the most expensive item on the shelf. I think you'd be surprised how good some of the entry level cameras and backs compare.