I agree with John. (But wait... which John? Tee hee.)


Quote Originally Posted by neuroanatomist


But those two factors are not causally related.


No, but both noise and DOF are causally related to iris diameter. But only under certain circumstances/assumptions, and I think you two are assuming slightly different circumstances (e.g. whether exposure duration is the same or not), which I think may explain the disagreement. I'll try laying things out in the way I think of it:


For a given DOF, focus distance, and bellows factor, the only way to get thinner DOF is to increase iris diameter. That increase in diameter is going to reduce image noise by one of two ways: increasing the light intensity over the same amount of sensor area, or increasing sensor area exposed to the same amount of light intensity.


If both sensors have the same sensor technology (e.g. QE, or Quantum Efficiency), are exposed for the same duration, to the same scene (e.g. same flash power), and are intended to achieve the same image brightness, then noise will scale with DOF. Photon shot noise for certain, and often times read noise as well.


So why talk about DOF when the real cause is iris diameter? Because many photographers don't tend to care that much about iris diameter, but they do care about DOF, so explaining in terms of something that is an important factor in composition is more immediately useful. The reason it's possible is because DOF scales linearly with iris diameter when all formats are assumed to have the same AOV, focus distance, and bellows factor.


Quote Originally Posted by neuroanatomist


Take the statement, "Asuming you have the same sensor technology, the thinner the DOF the better the noise regardless of sensor size," let's assume you have exactly the same sensor - are you saying that a shot at f/2 would have less noise than a shot at f/5.6?


I don't know about John, but I would say yes. The reason you may disagree is because you may be assuming a context where it's possible to adjust exposure duration, flash, or ND filters to achieve the same brightness. In cases where that is true, DOF does not scale with noise.


But when duration and other factors are kept the same, then the f/5.6 shot will require either +2 EC in post or two stops higher ISO to have the same brightness as the f/2 shot.


Quote Originally Posted by neuroanatomist


In your example, you're comparing different sized sensors - and you seem to be saying that if one shoots with an aperture 1.33 stops faster on the 7D, that the noise will be equal to that of a 5DII. That's certainly not the case.


For the circumstances explained above, I think it is the case. At least for photon shot noise and the random part of the read noise. (The 7D and 5D2 do differ a bit in pattern noise, with my preference leaning toward the 7D, but a lot of photographers wouldn't ever go that high anyway.) For example, I think these two shots would have the same noise:


1. 7D, 125mm, f/2.8, 1/500, ISO 640


2. 5D2, 200mm, f/4.5, 1/500, ISO 1600