CRCchemist wrote in post #17052479
So to get those crazy results on that link, you're saying those were shot at low ISO, and then the guy writing the review cranked up the exposure to show that there is some noise evident at low ISO on a Canon?
Even a simple curves adjustment will bring it up, and it's very evident if you print large.
If you don't shoot anything that has a wide dynamic range, then you can get around it by exposing to the right.
But, if you shoot architecture with natural lighting, landscapes or anything else with a wide dynamic range, you don't have that option. If you expose so that the highlights aren't blown out, then you have areas of deep shadow that are badly affected by the pattern noise. This forces you to use techniques such as blending bracketed exposures with intricate masks, which doesn't always work and is a pain to do.
That doesn't concern me as much because it's kind of a false alarm - since I never have ever had a problem with noise levels at low ISO in the shadows since I'm not trying to increase my exposure in the blacks. I thought those sample shots were both at 25,600 ISO, which I thought was absolutely crazy to see no noise from the Nikon.
It's not a false alarm if you need the dynamic range.
It reduces the usable dynamic range of the 5D2 or 5D3 to around 8 stops, vs the usable 12-14 stops of non-Canon sensors.
Can someone clarify this? Is the A/D converter on the Nikon or Sony cleaning this up before the RAW file is being written?
The Nikon/Sony A/D converter isn't 'cleaning it up'. It just isn't introducing the noise in the first place. Canon's pattern noise doesn't come from the sensor - it's introduced by Canon's A/D conversion process.