Reader Daniel Y writes to say:
You’ve commented on the full size jpegs from the D800, yet you've not made any mention of the fully size jpegs from the Canon 1D X.
I'm sure many of your readers will be interested to read your comments regarding the 1dx images and how they compare with the D800.
Esp. with the issues you've mentioned in the past e.g. lens mount tolerances, sensor pixel size, noise vs mp, lens limitations... it'll be interesting to read your take on 18mp vs 36mp given the existing lens and mount technologies for 35mm.
E.g. has Nikon gone over the top this time? Seems to me Canon and Nikon has just reversed roles. i.e. high vs low mp. And given the issues with high mp sensors what is the subjective mp of the 36mp images really equivalent to?
DIGLLOYD:New possibilities accrue with 36 megapixels. While 36 megapixels are not needed for many things, there are benefits to be had beyond the resolution.
A sensor with 36 megapixels can be thought of as just a higher sampling density, suitable for down-sampling to 12 or 16 or 18 or 24MP to reduce digital artifacts, or for cropping.
So even if technique and lenses are not scaled up to match the higher pixel density requirements of a 36MP sensor, the lens still delivers the same image to the sensor— so the higher pixel count can be used to produce a higher quality 24MP (or 18MP or 12MP) finished image from a 36MP sensor (higher quality per pixel). This technique is covered in DAP. It works very well to eliminate digital artifacts on the finest of details. And when good technique is used, the extremely fine detail opens new creative avenues without the bulk of medium format.
Just in the news today is the delay of the Nikon D4 until mid March, and the 1D X is not expected until April.
Note well that for the vast majority of shooting situations, 16-18MP is just fine. But surely few people are averse to 36MP at about half the price of the pro bodies, so that is probably more of a factor than the megapixels.
Canon’s sample images
Briefly on the Canon 1D X samples— Canon had the smarts to use the very high performing 135mm f/2L for the model shots. OTOH, the sports images are not very useful for evaluation, and movement is not a variable one wants for assessing image sharpness!
The “002” model shot is stunning, and one mark of high quality is how good an image renders in black and white (try it)— and here the results look superlative. What I don’t understand is the blur in the iris of the eyes; it just looks weird, sort of milky in appearance [is it a slight focus error, or some kind of contact lens? whose edges are clearly delineated in 001].
At any rate, it seems clear to me that the Canon 1D X looks to set an image quality bar far above the 5D Mark II and 1Ds Mark III, and apparently even a notch higher than the 1D Mark IV.
In short, Nikon’s image quality lead with the D3x/D3s might be over with the D4 and the 1D X (which is not to say that the Canon 1D X is superior, at present there is no basis to say either way at this point).
Per pixel quality
Both the Nikon D4 and Canon 1D X should produce exceptional per-pixel image quality, but it is only a matter of degree (relative to the D800) by which they will suffer similar lens limitations or mechanical tolerance issues. Because the sensor photosites are considerably larger (more than double the area per photosite), we can expect lower noise per pixel, superior color rendition (at least when the ISO is raised even somewhat), and perhaps higher dynamic range. But such things depend on both sensor design and the quality of associated electronics— I’d certainly expect that photosite size plus very high professional expectations for such workhorse cameras would mean that the quality would be top notch.
In a nutshell if one wants a simple answer: the D800 / D800E will deliver significantly more resolved detail with a top-notch lens focused perfectly at an optimal aperture. Whether the images have the same visual impact is another matter; in this sense the 12MP Nikon D3s still has no peer, until and unless the Nikon D4 and/or Canon 1D X show otherwise.
An interesting question is whether a Nikon D800/D800E image downsampled to the 16-18MP range will outperform (in total image quality) a native Nikon D4 or Canon 1D X image. Based on experience with the Nikon D3x and D3s, I’d say the answer is likely to be “absolutely” in terms of resolution and “perhaps” in terms of color and dynamic range and visual impact, the higher ISO range being more in favor of the lower-resolution cameras.
As I’ll be reviewing the Nikon D4, Canon 1D X, D800 and D800E, I intend to explore many of these areas (but I might or might not be able to have all these cameras together at once, so some combinations could be more difficult to arrange, even the D800 and D800E— I certainly cannot afford to buy them all!). Fortunately, all of them are full frame, making at least that aspect of the comparison easy.