The three most serious limitations of the current Canon 5D Mark III are:
- Noise (including pattern noise) even at ISO 100.
- Dynamic range two stops or so short of the Nikon D810. For a lot of conditions in the field, this is a real headache. And relates directly to exposure latitude, and thus noise (e.g., reducing exposure to retain highlights).
The new Canon EOS 5DS / 5DSR address resolution, but Canon confirms that the dynamic range is no different from the Canon 5D Mark III.
Noise: since all past full-frame Canon DSLRs (including the EOS 1D X) also exhibit pattern noise (documented in my reviews), the ISO 100 noise question is also open. However, Canon does state that noise is improved, so on the whole (more pixels, “improved” noise), the total picture should be better on the noise front.
In terms of luminance and chroma noise, the only appropriate metric is per pixel noise. Because a 50 megapixel camera has to be enlarged a lot less (for the same size print or image). Testing will have to demonstrate whether per pixel noise is improved when the 5DS / 5DS R images are downsampled to match the resolution of the 5D Mark III. See Comparing Noise Between Cameras.
Peter M writes:
For those mixing flash with daylight it's imperative to bring the ISO down to a minimum of 25
I just about get away with using the current D810, which has two thirds of a stop advantage over this camera. With regard to working in the studio, studio strobe is not always that efficient on low pin pricks of light, and a low ISO would also be an advantage here too.
DIGLLOYD: the 36-megapixel Nikon D810 has a stunning dynamic range with a true base ISO of 64. It is odd that Canon’s new 5DS does not have an available base ISO of 50, commensurate with its smaller photosite size and likely need for higher signal/noise ratio to minimize noise.