So a f/1.8 on a 1" sensor would still gather more light than a f/3.5 on an APS-C?

Bob A L wrote:

Great Bustard wrote:

Aside from DOF and noise, why do we care about the f-number at all? It's not exposure time, because whatever exposure time the 1" sensor can use at f/1.8, the APS-C camera can use at f/3.5.

I don't understand. We care about f-number because, the size of the hole to let light through (f-number) and the shutter speed are the only way we can control the amount of light getting to the sensor. And we use f-numbers because physical size of the hole must vary according to focal length to allow the same amount of light through and f-numbers are derived from calculations that create the same amount of light reaching the sensor at f1.8 regardless of focal length, where the physical size of the aperture will be much different on a 5mm lens than a 500mm lens. And aperture physical size is based on focal length, not sensor size.

Despite the utility of f-ratios for predicting exposure and AF ability, they do not exist in the empirical world; they are an abstraction.  What does exist is pupils; entrance pupils for what is seen in the front of the lens, and exit pupils, for what is seen from the film/sensor perspective, and they are both the same when there is only one element, like a typical magnifying glass.

The real "thing" of capturing available light from any object or rectangular area of a scene to be cropped by the sensor or in post is the pupil size, its distance from the subject, and exposure time.  It is certainly possible for a landscape photographer to ignore this and still make good decisions, but ignoring this, unfortunately, degrades the quality of capture for people doing macro with needed DOF, or people who are mostly focal-length-limited, so I think that people really need to understand the principle of equal etendue (another facet of f-ratio and focal length "equivalence") if they really want to make the right choices.

I would say that almost every small wildlife photographer that I meet is unfamiliar with etendue, and thinks that cropping from a 100/2.8 will give less noise than a 600/4, and their incorrect judgment is unfortunately supported by the illusion created by the implicit DESTRUCTIVE noise reduction in under-sampling a subject with too few pixels.  Most people would not even dream of taking their 100/2.8 subject capture and upsample it to 600% to see how it compares to a 600/4 at the same displayed subject size, saying that 600% is ridiculous, but if you do it for them, the upsampling (if it isn't nearest neighbor) will create a smooth texture without the "dottiness" in the noise that they are looking for, as an indication of noise.  They will mistake the smoothness for a lack of noise, when in fact, it is a lack of ANYTHING at all narrower than 6 pixels.  Then if you show them how much NR and/or median filtering you can throw at the 600/4 capture and still have more detail than the 100/28, they will say, "you shouldn't have to do that; it should come out of the camera clean", totally incognizant of the fact that their "clean" option is only clean because it is SHRUNK on the display, in their normal viewing habits.

Someone might bring up the fact that they can use a slower shutter speed with the 100/2.8, but that really makes no good argument in favor of the 100/2.8, because you can use the slower shutter speed with the 600/4, too, with no blur any larger than with the 100/2.8, relative to the subject.  It is not a matter of the 600mm needing more shutter speed; it is a matter of a slow shutter speed LOSING MORE resolution relative to the focal length's potential with 600mm than with 100mm, as the 100mm has far less potential for subject resolution to begin with, at a given distance.  IOW, you've already lost most of everything that a faster shutter speed can give you with the shorter focal length, when you choose the shorter focal length.