Originally posted by Class A From the
Wikipedia article on "Lens Speed":
For controlling depth of field, especially in portrait photography[1], lens speed is a key variable in combination with other variables such as focal length and camera format size.
Also from wikipedia:
A lens with a larger maximum aperture (that is, a smaller minimum f-number) is a fast lens because it delivers more light intensity (illuminance) to the focal plane, allowing a faster shutter speed.
(emphasis mine)
Quote: What?
So the mathematical model is wrong according to you?
If it is correct (Hint, it is. Hint, Falk is a physicist; he understands more than "basic physics"), how can one derive "counter factual claims" from it?
Because he was speaking of "equivalence" only. My point has been to illustrate that "equivalence" is an analogy. Besides, I'd really be interested in Falk's input, but he's chosen not to become involved. And second, this is called "appeal to authority".
I would LOVE to hear Falk explain how an 8.5mm f1.9 lens passes less light to the FF sensor than a 47mm f1.9 passes to the same FF sensor.
I mean, I get that there can be slight variance due to angle of incidence, and that lens designers have struggled to control vignetting because of that, but that's a lens design problem; an ideal lens, regardless of focal length, will pass the same amount of light to the same sensor at f1.9, and ideal lenses with the same image circle will pass the same TOTAL light (even light that doesn't end up on the sensor) at f1.9 regardless of focal length.
I suspect that Falk would agree with that - because it's true; otherwise you'd have to tell your hand-held light meter what the focal length of your lens was.
Quote: I never claimed that. I claimed that the 8.5/1.9 Q-lens passes less total light than a 47/1.9 FF-lens.
A 8.5/1.9 Q-lens will pass the same total amount of light to an FF-sensor as it passes to a Q-sensor. But try mounting the 8.5/1.9 Q-lens on an FF camera. It won't produce the image circle required for the FF-camera. OK, let's enlarge it with extra optics. This, however, will spread the same amount of photons over a larger area. This will reduce the exposure. In other words, the exposure will not be the same as that from a 47/1.9 FF-lens that doesn't need enlarging of its image circle.
There you go again. This is the ENTIRE root of the counter-factual claim. A FF sensor at ISO 100 requires a certain total amount of light to create an exposure, as you have insisted over and over - and I agree with. If I put that 47mm f1.9 lens on the camera and meter it at f1.9, I get a reading. If I remove it and replace it with an 8.5mm f1.9 lens with appropriate optics and image circle, it will pass the same total amount of light to the sensor, giving me the same exposure at ISO 100 and the same shutter speed. Can we agree on that? Or do you believe that you get DIFFERENT total exposures from two different focal lengths @ f1.9?
Quote: If the Q-lens passed through the same total amount of light that is produced by the 47/1.9 FF-lens then I wouldn't call it slow. Note that for this to have any relevance for the Q-system, the Q-lens would have to concentrate all these photons onto the size of a Q-sensor. In order to be able to collect all the photons required for that, however, it would have to be a lot faster than f/1.9. It would have to be a f/0.3.
Right - it would be the SENSOR size, not the f-stop of the lens. See, I *understand* that you're saying that 1) if you had a sensor with sufficient capacity and 2) sufficient pixels at that capacity, you could concentrate the total light from the larger sensor onto it and achieve the same total SNR; the problem with that is that no such sensor exists, period. You could build the fastest glass in the world, but you couldn't USE it this way, because the SENSOR couldn't sink that many photons. It would require a DIFFERENT KIND OF SENSOR.
Quote: Yes, but that doesn't matter if you have more of them.
Pixel-pitch does not determine dynamic range. Sensor size does.
LOL. Yep. You're right. So a larger sensor must have higher SNR, therefor, it *is* about the sensor size, not the lens, right?
Quote: Yes, you'd need very low ISO values (or neutral density filters). This is part of the limitation of the Q-system. It doesn't have fast lenses and it wouldn't be able to cope with them either.
Which would reduce your exposure, negating all of the relevance of "equivalency" except for DOF.
Quote: If you think it doesn't fit the real world then you don't understand it.
There's this concept that is foreign to many people. "I understand, but disagree."
Quote: Strawman.
Strawman.
The fact that equivalency doesn't work in the real world because no such sensor exists - because no such sensor *can* exist - is a straw man?
Look... Let's say you invent a sensor that can accept enough light at the size of the Q to use your f0.3 lens. You get the same SNR as a current FF sensor, same DOF, same noise floor. Great. But NOW we can use that same technology to build a FF sensor that has *exactly* the same improvement in SNR over your Q that a current FF has over the Q now.
That's not a straw man, it's a *fact*.