DxO's tests have little to do with real world imaging, unfortunately. They measure DR in a totally artificial way that doesn't show up in real photographs, taken by real photographers, and is unduly influenced by base ISO.
The DXO results certainly show up in
my photographs. If they don't show up in
your photographs, maybe your technique is lacking, or your understanding of the DXO results is lacking.
Of course, I hope you realise, when I write that the DXO results show up in my photographs, I don't mean that I see a big DXO sign in the deepest shadows reading
13 EV Dynamice Range. That would spoil the appearance of the photo, don't you think?
Rather, I find if DXO claim that Camera A has 2 EV better DR than Camera B, then I can underexpose a shot from Camera A by two stops, and find when raising the deepest shadows in both images of the same scene, that the shadow noise, detail and general quality in those shadows, is about the same in both images.
You should try it some time. However, DR is only one parameter. SNR in the midtones, around 18% grey, is another parameter, and the effects can also be very noticeable.
Quite often two different models of camera can have significantly different DR but very similar SNR at 18% grey. A case in point would be the Nikon D800 and the Canon 5D3.
The D800 has about 2.5 stops greater DR than the 5D3, at base ISO. But what I'm certain is frequently overlooked when some people refer to the DXO scores, rather than the graphs, is the fact that SNR at 18% grey (the midtones) is
about the same for both cameras across all ISOs.
This is a significant measurement which DXO does not disguise. It's all there for your benefit. What this means is, if you were to underexpose the D800 shot by 2.5 stops and find that the deepest shadows were just as clean (or noisy) as in the fully exposed 5D3 shot, the 5D3 shots would nevertheless have significantly cleaner midtones, skin tones etc, which would be quite obvious.
The
comparative significance of the DXO results needs to be understood. When this is accomplished, I'm quite confident you will fine that their results
do correspond with, and show up on, real world photographs.