It's quite natural that folks should be excited with their new toy and be pleased with all the ergonomic improvements they see.
I was also impressed with the fast frame rate of the D3 when I tested it in a store in Bangkok. However, such features in themselves are not necessarily going to have any bearing on the ultimate quality of the shot, although they might in certain circumstances, in action shots for example, or hand-held shots for the purpose of image stacking or merging to HDR where the auto-alignment feature of CS3E might work more successfully.
I also find the following statement from Panorama somehow resonant with my own experience with the D3.
There are so many examples on the internet that can easily be compared. At low ISO ratings the 5d wins with more detail, and color/clarity/DR are either equal or better IMO. Obviously once ISO goes up the 5d has more noise, but at lower ISOs I'm giving the nod to Canon or believe it's equal.
I got the impression that the 5D shots in my testing, shooting the same dark corner inside the store, were marginally sharper than the D3 shots. At high and very high ISO, the D3 shots were noticeably cleaner. However, curiously, I found that after applying chroma-only noise reduction in Noise Ninja to the 5D file, the noise levels were much improved whilst resolution was still maintained.
Applying a similar amount of chroma noise reduction to the D3 file also improved the level of apparent noise,
but not without noticeable softening of the D3 image. These were not jpegs but Tiff conversions from RAW in ACR.
After playing around with various noise reduction settings in Noise Ninja for both 5D and D3 images, it was clear that the 5D level of noise could be improved to a greater extent than the D3 level of noise before image resolution was significantly affected in both images.
After such adjustments, the noise level differences between the two cameras at ridiculously high ISOs, between 6,400 and 25,600, sometimes seemed as little as 1/4 of a stop, although on average I would place it closer to 1/2 a stop.
One has to be careful about personal assessments of the performance of new camera models from people who have not carried out thorough, direct comparisons themselves. It's so easy just to believe and repeat the positive and promotional statements one finds on the internet and on the manufacturers website.
I remember well the glowing reports of the performance of Canon's 40D; improved autofoussing; lower noise due to 14 bit A/D conversion, and general image quality on a par with that of the 5D, if not better.
The reality is something else, as I see it.
(1) Despite 14 bit processing, no significant reduction in noise or increase in DR compared with the 20D.
(2) No noticeable increase in resolution compared with the 20D, except at 200% magnification on the monitor.
(3) Definitely slightly more noise and less DR than the 5D, comparing equal size image files or prints.
(4) No noticeable improvement in autofocussing whatsoever, that I've noticed.
The last point is a real disappointment for me. I never intended to buy a 40D when I went out shopping one day for the Canon EF-S 17-55/2.8, because I knew already that I could not expect any fundamental image quality improvement over the 20D, except at a pixel-peeping level, which is not enough for me.
However, I was seduced by an exceptionally good price. I also rationalised to myself that the superior autofocussing attributes of the 40D at F2.8 could be worthwhile. The faster frame rate and Live View would be icing on the cake.
So far, I haven't discovered any better autofocussing qualities of the 40D, compared with either my 20D or 5D, and Live View has mainly served the purpose of confirming just how inadequate autofocussing can be with the EF-S 17-55 at F2.8.
