A good point. I made smaller prints (crops corresponding to 53x83 cm). These were A4s with a 33% crop of the full image. Colors were adjusted in Photoshop so color and tonality was similar.
When looking at the print at short distance, like 25 cm the difference is very significant. At arms length distance I cannot see a difference.
Deep shadow detail seems quite similar to me, but the Pentax has very different color. I also got the impression that the Pentax lens is a better performer than the Nikon. Astonishing! Leping has tested many Pentax lenses on a Canon with adapters and found them truly excellent so it is perhaps no absolute surprise.
I agree. Lenses are a very important factor when buying a new system. I switched from the Minolta SLR 35mm film camera to Canon 35mm SLR years ago, before I bought my fist DSLR, basically because Canon had invented the Image Stabilisation system and offered a decent 100-400 IS zoom at an affordable price.
I now have a couple of Nikon DSLRs as a result of the excellent Nikkor 14-24/2.8. Any additional performance of the Nikon sensors, compared with Canon, is a bonus.
I've reconverted these RAW files from Miles using zero settings in ACR, a linear tone curve, same Daylight WB, but adjusted exposure to push both histograms equally far to the right without clipping.
I can see no differences in shadow noise, but these images are not ideal for DR comparisons.
What initially surprised me was the difference in exposure adjustments required to push the histograms to the same position, ie. +1.05 for the 645D, and +0.65 for the D3X. Whlist both images could be considered underexposed with regard to ETTR, the 645D is a good 1/3rd of a stop less
exposed than the D3X image, according to ACR 6.3.
This could imply a different T/stop rating for the two lenses used, or a different ISO sensitivity for the two sensors.
Having just checked again the DXOMark results for these two cameras, I see that the true sensitivity for the 645D at ISO 100 is actually 104, and that for the D3X, ISO 78.
The difference between ISO 78 and ISO 104 would account very closely for the exposure differences between these two images. This comparison once again confirms the accuracy and relevance of DXOMark testing, and I'm also impressed that ACR 6.3 is able to reflect such accuracy.
What is most puzzling is the significant differences in color hue and saturation between these two images at the same Daylight WB setting in ACR.
One or both of these cameras need calibrating in ACR. I tend to get the impression the most natural result would be about half way betwee the two renditions.