I was dismayed to see the explanation of the effect of a smaller light cone
on DPR!
Oh come on, Mr. Richard Butler, just because a smaller sensor captures a smaller portion of the cone of light does not mean that the light captured is less, from the point of view of the sensor!
Yikes! I lost my reply to page refresh! Rewriting.
So if I put one cm of butter on top of one sandwich and then one cm of butter on two different sandwiches, the amount of butter used will be equal regardless of if I look at the 1 or the 2 sandwich system?
A crop is a crop regardless of how you crop - and cropping throws away light.
Bad analogy, but alright. When it comes to crop
sensors, you're spreading the butter on a smaller bread. Compare it to your eating an equivalent area from the middle of a bigger slice (some people do that) and throwing away the sides. The real wastage in this case would be the unused FF sensor area, which does not exist in case of the smaller sensor. Using an FF lens on a crop sensor then is throwing away the extra butter. I said I didn't like this analogy.
It is the same amount of light on a per unit of sensor area basis!
Sure, but the amount of light collected is different.
Does not matter. The image is different.
It is the same amount of light for the image being captured.
This contradicts directly what you said one line above this
No. Same amount of light for the same view. Look at it from the POV of the smaller sensor - what lies outside its frame does not matter to it.
Different sized sensors have different amount of sensor area - if a fixed amount of light "per unit area" is used, how can the light captured be the same?
The debate is about the light intensity as measured by the f ratio. We are not discussing photovoltaics and the amount of power generated.
The only thing that's smaller is the field of view!
If you use the same lens the field of views will be different, as well as the amount of light captured. Or are you saying that anything outside of "APS-C" crop on a full frame gets no light at all?
The fov will be different between both the sensors unless the larger one uses a longer focal length.
Want to read a more reasonable interpretation?
Here's one!
Not really, but you might want too check
this out . It should provide a rather comprehensive explanation on the relevant subjects. The relevant mathematical formulas are in the most details article of them.
Your blog was the first one I read, and I didn't find it thoroughly agreeable.
There is of course one way in which a smaller sensor will have lesser light capture - on a per-pixel basis, if the resolution (and technology etc.) is the same as a larger sensor. The larger sensor's larger pixels will get more light, and there may be other physical or electromagnetic disadvantages for the smaller sensor.
If it is raining and I have a small cup and a large barrell, both get the same exposure to the rain, yet one of them will collect more raindrops.
Again, a flawed analogy when it comes to optics. It might look self evident when you describe it as barrels in the rain. What you really have in optics is akin to using a funnel to fill your barrel - the larger the barrel, the wider your funnel needs to be for collecting the same
proportion of water. Why? Because the larger your sensor, the wider the image circle, and therefore lesser amount of light per unit area for a given
aperture. Therefore, you have to increase the front element size and the focal length, akin to your catchment area. Yes, SNR improves, because you're measuring more photons collected for a given period of time, but the same catchment area would give a smaller sensor a lower f/ratio, upsetting the calculation.
Noise if almost entirely a function of light - capturing more light increases both signal and noise, but the signal increases much faster, thus the signal to noise ratio (SNR) improves.
Yes, the square root of photons collected, but what relationship does it have with the f ratio?
This is why big sensors in general are able to outperform the smaller ones when non-identical images are acceptable.
If we compare, say a 16MP APS-C with a 24MP FF, [edit: wrong comparison - area varies in ratio of the square of the diagonal crop factor. Substitute with appropriate resolution difference] the per-pixel light advantage would disappear. However, for an apple to apple comparison, we need to have the same resolution, right?
With equally efficient sensors a FF will capture about 2,25 times more light than a APS-C sensor and the SNR will be 1,5 times better.
"Amount" of light: irrelevant unless fov matches
SNR: Again, irrelevant in debates about "Aperture Equivalence". Let's keep it restricted to apertures and focal lengths for the purpose of this debate
Yes, if we view both the images at the same size.
There is usually little point in comparing output images of anything but the same size.
Not if you were stuck with a lens of a certain focal length and
had to crop the image. Can happen often, you know?
Not if we match the FoV, though there's no reason why you'd want to do that.
Huh, of course you want to match the angle of view. If I want to take a picture of cerain landscape I'll use for example 20mm lens of APS-C and 30mm on full frame. It would be silly to not do so, would it not?
FoV here meant the resultant image, not the capture. But then again, the previous point.
You'd like to view the smaller sensor's higher resolution at comparable viewing size, wouldn't you?
No idea what you mean here. The FF in your example has higher resolution. The APS-C has finer pixel pitch.
That was a bad example as immediately noted alongside. The correct comparison would be 36MP FF vs 16MP APS-C, if you find a pair with matching parameters otherwise.
The remaining big difference is in terms of the DoF, but if you want a deeper DoF, which one would you prefer, ceteris paribus? With the same lens being used on both sensors, the smaller one will have a greater advantage with sharp centres too, yadda yadda.
Nope - the smaller image of the APS-C (24mm by 16mm or a bit less) will need to be
enlarged 1,5 times more than the FF image (of 36mm by 24mm) for the same output image size. Thus the relevant aberrations of the lens are also enlarged 1,5 times more.
But would still avoid the aberrations at the extremes. The point then is that an FF lens on a crop sensor is a waste of money if the fov is acceptable.
Well anyway all of this is not that big a deal. We could avoid much of the confusion if we just started to talk about lenses in terms of Field of View rather than focal length. And stopped applying crop factors to apertures.
The same f-number creates a different effect of different systems - thinking that a cell phone f/2,8 and medium format f/2,8 are somehow the same just might be a bit silly.
Comparing the two as if a real life application existed isn't very smart either. Let's not measure outside boundary conditions.
The f-number is nothing but focal length divided by the diameter of the entrance pupil. If you change one, the other changes as well.
Yes, your typical, non-exotic, everyday ratio. The first thing I learnt about photography when I wanted to purchase a system camera.
