I am not mixing anything. Changing the ocular of the telescope does not change the f-stop of the lens. The brightness falls by cropping of the ocular with more magnification. Yes, we could get the image back to the same brightness, but we would have to do something about it.

The camera engineers already made the cameras appear similar in brightness, but they did something to achieve that.

@aki_hartikainen

Until yo go back to reading that other people posted and check all definitions you are prohibited to post in this topic.

Part of the problem is that we have not defined brightness accurately. Intuitively, the brightness of an image is the luminous flux per unit area, which is either illuminance or luminous emittance, depending on whether we are talking about light falling on a surface or being emitted by it.

When you take the same lens with the same lens field of view and a smaller sensor, the illuminance of the image falling on the sensor does not change. The luminous flux of the light falling on the sensor may be lower, depending on the distribution of luminous flux in the area of the image plane. If the parts of the image that fall outside of the crop border were all black, then there is no difference in the luminous flux of the light falling on the sensor.

If you hold up a photographic print in front of your face, cut it in half and let half fall to the floor, and then hold up the remaining half, did it get any darker? No, not in the intuitive sense of the word brightness. Less total light might reflected from the print onto your face. Someone looking at your face might say that your face is not as bright. But the part of the image that remains is just as bright as before. Cropping changes the area of an image, not its brightness.

Now if we compare an m43 and a full-frame camera where the full-frame camera has a lens with double the focal length so as to give the same field of view, if the two lenses have the same f-number and ignoring differences in light transmittance between the two lenses, the larger lens collects four times as much light from each object in the field of view because it has an aperture four times as large. This gives a two-stop advantage for noise performance. The advantage in noise performance comes from the lens having a larger aperture, not from the sensor being bigger. These cameras are not equivalent, and their fields of view are technically not the same because one has an aperture covering a larger area. The depths of field will be different, assuming the same output resolution. Here's a good read on camera equivalence: http://www.falklumo.com/lumolabs/articles/equivalence/index.html

.

Alfi, depth of field depends on the output resolution, magnification, and focus distance, not just the lens's aperture. The circle of confusion size of out-of-focus objects relative to the size of in-focus objects in the image depends only on the aperture diameter, focus distance, and object distance. But depth of field depends also on lens magnification and on your circle of confusion criterion. Generally we take the output pixel pitch scaled to the image plane to be the circle of confusion criterion. If you take the same lens with the same field of view and put it on a smaller format camera with the same output resolution, the depth of field becomes shallower, because the resolution in the image plane goes up. If the image plane resolution is the same, then the depth of field is the same. This sums it up pretty well: http://en.wikipedia.org/wiki/Depth_of_field#DOF_vs._format_size

"but you can't say that it's because the lens is "brighter" on the larger sensor. It's the same god-damn lens, or else, it's the same god-damn f-stop. Stop confusing people, please."

The lens and exposure will be the same, no disagreement there.

But if the larger sensor has better dynamic range, then yes, one could say the lens will be brighter on the larger sensor.

This is because there is magnification taking place inside the camera and magnification has the penalty of reducing brightness. The smaller the original, the more magnification required for the final size and thus bigger penalty for brightness. Both cameras need to produce the same image size of 1920 x 1080, but the full frame camera saw four times the light for that final size compared to cropped 1/4 area sensor. It has already been mentioned that it does not seem to have impact on the final image brightness or quality, but the basic principle of magnification is still there.

It is also true this does not seem to require any user intervention between different sized cameras, but that is because the camera engineers adjusted the output to be the same very carefully. The larger sensor that indeed captured more light could still have advantage for dynamic range, for example.

This is fantastic! Arguing that all the light-meters must be wrong. Creating a flame instead of spending 10 minutes with a camera to test.

Not only proof that the F-stop we've all known, loved and trusted can be found somehow faulty using words alone. But even more fantastic, we now have proof of the inexorable laziness of humans to postulate from their armchairs, even against all odds.

We could argue the hind-leg off a donkey!

+1. BMCC has larger pixel pitch than 5DmkIII.

See http://www.personal-view.com/talks/discussion/4403/pixel-pitch-size-sensor-refresh-rate/p1

@aki_hartikainen To clarify, does the wider FOV picture on the right use a focal reducer? The brightening effects of focal reducers are well known. They're also completely irrelevant to this debate. Focal reducers are rarely if ever used on video cameras, due to inherent issues with coma.

To further negate the legitimacy of that example, compare the brightness of a given lens at a given f-stop mounted to a GH2 with a focal reducer, with that of the same lens at the same f-stop mounted to a larger sensor that gives the same FOV as the GH2 + focal reducer. The GH2 will be brighter, because the focal reducer focuses the projected image into a smaller projection. In other words, using a focal reducer is NOT the "exact same" as what "would happen if the sensor by some miracle grew in size."

As another example, take a projector. The brightness of the projected image doesn't magically change if you're projecting the full image onto a 6' x 10' wall, or a crop of it on a 3' x 5' screen mounted to that wall. This is analogous to the [non-existant] difference in brightness between sensor sizes.

The brightness does change if you either move the projector closer to create a projected image that is 3' x 5', or if you use a lens / adjust the focal length to focus it onto a smaller area. This is analogous to using a focal reducer.

@aki_hartikainen I've lost patience with your inability to comprehend optical physics. I'm out.

Actually I have studied optical physics, and you're dead wrong about this.

Cameras with smaller sensors do not have built-in focal reducers. There is no process inside the camera that is analogous to using a focal reducer. Focal reducers have specific optical properties that are not replicated in any way, to "match" the "final image sizes viewed on a monitor". That is achieved by electronic, rather than optical, means.

You're treating image sensors as single, discrete entities. To understand the physics at work, you need to stop thinking of them that way and start thinking of them as matrices of pixels. In this view, and notwithstanding more sophisticated algorithms like pixel-binning, if anything it is the pixel size that determines the relative brightness (which is normalized by ISO, and thus actually determines only the ISO performance). Again, pixel-binning algorithms make pixel size less important, as they effectively merge pixels with their otherwise-inactive neighbors during video recording.

@ignatiusreilly oh I mean @aki_hartikainen lost a whole factory's worth of marbles today...

Now that you've scared away everyone who knows what they're talking about you've only got this obstinate log to deal with. I don't think anyone should be dismissed for not understanding optical physics, as it's not particularly elementary. I do, however, think common sense denial is a deadly sin, and so I'll just quote something...

A 50mm f/1.4 will project the same image, at the same brightness, irrespective of whether it's mounted to a Canon 5D or a Panasonic GH2. The only difference is that the 5D's sensor will capture most of the projected image, whereas the GH2 will capture a smaller "cropped" section.

Nothing about this makes me think that the exposure would be different, or the amount of light absorbed by the respective sensors would change outside of the actual sensor real-estate. The only thing that changes is your field of view. And if you take a focal length reducer, attach it to that same 50mm f/1.4 and attach that newly minted 40mm f/1.4 to the GH2, you will just be allowing more light to enter, because your field of view is wider so more light is projected onto the same size sensor. The same thing would happen if you put that lens on a 5D. Lets say the differences your talking about when bringing up the f/stop, and crop factoring in were at all accurate. Then, when comparing a GH2 and 5D with the same lens, the exposure difference you'd see between the two would be similar to the difference in DOF: drastic. You're basically suggesting that arbitrary iso ratings account for the fact that it is even possible to expose an image on a non FF camera.

No, the main reason why the Black Magic Camera can produce more dynamic range, is that it has larger pixels. The sensor size REALLY DOESN'T MATTER, at all. For example, while the 5D mark II has a much larger sensor than the Black Magic Camera, the pixels are smaller, and it therefore has a smaller light-sensitive area during video recording than the Black Magic Camera. Pixel-binning would have negated that, but unfortunately Canon instead opted for the vastly inferior line-skipping method of downscaling an image, which involves ignoring most of the pixels.

The increased image brightness of using a focal reducer comes from focusing a large image protection into a smaller one, which means more photons per pixel. You could not use a focal reducer on a 5D mark II with standard lenses, because it would focus it onto just the center of the sensor, and you'd have serious vignetting. Using a focal reducer on a smaller sensor increases the brightness, only insofar as it increases the field of view. More that would otherwise have missed the sensor are focused onto it, but these photons exclusively come from the part of the scene that would have been left out of the narrowed FOV. The exposure of any given element in the scene will not increase, it's just that your scene widens. That's where the extra light goes. The real reason that the picture you posted shows a brighter histogram, is that it includes more of the white wall behind the man, due to the widened field of view.

@sangye now lets argue about whether that's a ceiling or a wall. Judging by the posture of the man, it is obviously the ceiling. It's not impossible, though, that the man is horizontal, in which case it could maybe be the wall... but it's probably the ceiling.