EOS 5D Mark III vs. EOS 1D X - Differences?

[neuroanatomist]

I'm assuming that for now, we don't know anything the relative performance of the two sensors when it comes to noise, other than a few jpg samples. That means we really don't know if one camera will have better read noise than the other, though I assume that if there is a difference, the 1DX will have the edge here. Am I wrong? (If I knew for sure that the 1DX had far lower read nose, I might seriously consider the 1DX.

True. There are now some 5DIII RAW available for download (Imaging Resource), but they're from a pre-production model, and there are none from the 1D X. Chuck Westfall has stated that the 1D X would be superior to the 5DIII in that regard, but gave no indication of the magnitude of the difference.

For me, the frame rate and build are important factors, as is what I presume will be better tracking with input from the metering sensor and the associated face tracking feature.

[Daniel Browning]

True. There are now some 5DIII RAW available for download (Imaging Resource), but they're from a pre-production model,

Nice. Now to see if it has the same low-ISO noise issues as all the other Canon cameras...

[HDNitehawk]

True. There are now some 5DIII RAW available for download (Imaging Resource), but they're from a pre-production model, and there are none from the 1D X. Chuck Westfall has stated that the 1D X would be superior to the 5DIII in that regard, but gave no indication of the magnitude of the difference.

Chuck Westfall's statement indicated that the 5D III would have unsurpassed IQ, with the exception of the 1D X which would handle noise and the higher ISO's better.

Taking in to account all the discussions about pixel density and noise I have seen on this site, and Canon's marketing propaganda about noise and how the camera handles noise, might this be correct:

If the 5D III and 1D X have the same sensor technology and the same type of firmware, that the 5D III will excel in most areas with the 1D X having only the edge with noise because of the larger pixels?

I deal with salesman all the time in my profession, over the years I have developed the habit of reading between the lines and hearing what they don't say. Maybe I am doing that here with Westfall's salesman speak and Canon's papers.

[andnowimbroke]

Would some of that be due to the number of processors? Not trying to throw this thread off track, but could you take three rapid shots, auto-align, and "merge" those with software that goes through comparing pixels of the three, basically throwing out the odd one. Assuming noise is random, that should already give you a cleaner shot. Any of that make sense?

[Daniel Browning]

[...] with the 1D X having only the edge with noise because of the larger pixels?

Well, the difference in pixel diameter is only about 10%, so if that does give it an edge, it's not a very big one.

[Jon Ruyle]

Nice. Now to see if it has the same low-ISO noise issues as all the other Canon cameras...

Any conclusions, Daniel?

(My viewer wouldn't display them properly... hardly surprising, I suppose)

[Daniel Browning]

Any conclusions, Daniel?

(My viewer wouldn't display them properly... hardly surprising, I suppose)

Unfortunately, the scene and lighting is setup so that even the very darkest tones possible are only about 5 or 6 stops below clipping, so we can't tell anything yet. The "outdoors" sample files that Imaging Resource does are the only ones that have enough dynamic range to be useful.

[Jon Ruyle]

Unfortunately, the scene and lighting is setup so that even the very darkest tones possible are only about 5 or 6 stops below clipping, so we can't tell anything yet. The "outdoors" sample files that Imaging Resource does are the only ones that have enough dynamic range to be useful.

Bummer.

[neuroanatomist]

Indeed. When I've looked at DR, I used a backlit Stouffer step wedge exposed so the clear block is just short of clipping. But (sigh) I don't have a preproduction 5DIII or 1D X.

[neuroanatomist]

It's ok, though - there's already a cadre of people who've done a 'careful analysis' of the IR RAW files, using a blown highlight to assign the max signal and the inactive pixels at the edges of the frame to measure the baseline noise, and 'determined' the DR that way. Ok, right.