Science and data do not tell the whole story. Sometimes the data can tell us what to expect or give us a concept, but in absolute terms, you can "never" say how anything is going to perform until you perform tests yourself. But no, you do not rely on review tests because your data is superior and/or your personal skill set for doing scientific tests of photographic tools is superior to any review site.
Let me know how the Sony A57 does compared to the NEX-5N//Pentax cams/and the Nikon D5100, all that share the same sensor. I'll gladly enjoy hearing about your results of the Sony at ISO 1600 vs. the others...While you are at it, tell me about all the past cams sharing same sensors and how they could possibly be vastly different iq wise in some cases (i.e. Fuji/Nikon, Oly/Panny/Leica, Sony/Nikon, etc.)...even what was thought to be identical cams in the point and shoot world had very different results...why or how if they are the same cam?
I'm surprised a Photography Guru would even look at a silly site like DXO
)))). Maybe time to study ULF for a while and come back when you have sufficient technical data to back your claims.
What do you mean science doesn't tell the whole story? What do you think the scientists do, build a basic design and hope for the best? You sound like one of those guys that claims that his Monster HDMI and speaker cables improve his sound system, and there is no measurable difference in anything "but I can just tell." Photography overall is an art, but the act of converting light rays into a digital signal is pure
science and completely understood. These cameras are designed by people with a litany of physics, electrical engineering, and mechanical engineering degrees, and every aspect of the camera's function is deliberately chosen, not just left up to the hocus pocus faeries to work out.
The Sony a57 has a translucent mirror that cuts out about 1/3 of the light to the sensor, to divert it for use in the phase detect AF system! And when you look at the high ISO tests at ISO 1600, it performs, funny enough, about 1/3 of a stop worse! Now imagine that, what a funny coincidence that is probably completely unrelated! When you discuss "vastly different IQ results," in the grand scheme of things they are not that different. Simple choices on how to run the CMOS or CCD that is capturing light will affect image quality, just as two race drivers tested in the same car will come up with slightly different lap times. However, considering that both teams of engineers are about
the same competency level, except for a few small personal choices such as where the analog-to-digital converters are in the signal path, and which ones to use, whether the camera captures and works in 16 bit the whole time or whether it truncates to 12 bit at the ADC, which internal filters (such as anti-aliasing) to use, etc. Then, the rest of the differences relate to the JPEG processor, such as Canon's DIGIC whatever-number-they're-up-to, Nikon Expeed, etc. If you have Lightroom and shoot RAW, bypassing the individual company's color shift and white balance choices, these settings are bypassed, and to a large extent reversible even if you shoot in JPEG.
Overall, these differences cause relatively minor
differences in image quality. Realistically, you'd get wildly more different results using two different lenses or two different settings, than using two different camera bodies. For example, go look at that link I posted again. Check out both cameras at ISO 800. You will notice that when both cameras are set to it, they both underexpose and digitally push the image up (they do this since they have such good dynamic range). Pentax has taken a relatively conservative approach, actually giving the sensor an ISO of 700+ at that setting. Sony is a little more aggressive, having a measured ISO of only 529 when you set the camera to 800! This is cheating a little bit, as it will hurt your shutter speeds a little compared to what you'd calculate, but overall, less than a stop difference.
Here's the only difference you have to worry about between these two cameras: do you want ultimate small size with the NEX but are willing to pay more for lenses, or do you not mind having the larger size of this camera, but have the capability to use DSLR lenses (especially little gems like the Sigma/Tamron 70-200mm f/2.8, the Tamron/Pentax/Sigma midrange 17-50 ish f/2.8, etc)? You can use the Sony alpha DSLR lenses on a NEX, but you need to buy either the $150 LA-EA1 adapter, or the $280 LA-EA2 adapter.