The photodiode size is irrelevant. What matters is the capture area of each microlens. The assumption, a perfectly reasonable one and one supported by experiment, is that the fill-ratio is approximately the same from pixel size to pixel size. There is no reason it shouldn't be. In fact, given the same manufacturing technology, there's reason to expect that the fill ratio would go up slightly with larger pixels.
Maybe but you'll still have the DOF issue.
From this and your other posts, I see where you're going. You're going to say that a bigger sensor gets more total signal, but also gets more total noise, so it all cancels out. However, there's one critical element to realize here: noise is random, whereas signal is coherent. The more signal you have, the more redundancy that can be "spent" fighting (canceling) the noise. Think about, for example, noise reduction by image averaging. By your theory, when you add more images, you add both signal and noise, so there should be no improvement. But the signal adds up, whereas the noise cancels itself out. Total light matters.
Even if these assumptions are correct (until proven otherwise these are just assumptions), there is NO evidence that there is a 1-1 correlation between microlens area and sensor sensitivity/noise performance. In other words, how can you be sure that doubling the microlens area results in a two-fold increase in sensitivity? Any scientific document to prove that?
I'm never satisfied...when it comes to image degradation...the better they can make it look,...I'm all for it..
I can totally relate!
There is no certainty.. Except that in the limit that the sensor is 'good', or that in both cases the sensor performs in the same regime of operation, the microlens that receives twice as many photons gains one extra "bit" of information over the smaller microlens. This is a purely physical constraint. Quantum mechanics cannot be overcome with any amount of engineering, alas.
I think LJ's experiments were reasonably rigorously conducted, yes.
Well, how much could the difference really be (from 2.56x)? We're talking of a linear scaling of about 1.6, so presumably Canon's manufacturing processes aren't that drastically different. They're also both CMOS. So it's a number somewhere between 1x and maybe
I’ve never argued that the 5D is a better camera than the 20D and that its ISO 1000 performance is equivalent to ISO 400 on the 20D. This has nothing to do with sensor size, though. It just happens that the 5D has a better sensor that the 20D. What if one day Canon release a 1.6x camera that takes the same quality shots at ISO 256000 as the 5D at ISO 400? How would sensor size matter than?
Wide angles stitch just fine. Distortion of the lens should be
removed even for teles. I have stitched great things with 35mm,
24mm and 16mm focal lengths. The end result distortion depends
strictly on the final FOV and not the lens used.
Steven
--
---
2006 Southern Arizona Monsoon Wildflowers
http://www.pbase.com/snoyes/images_summer_2006_ii
Summer 2006:
http://www.pbase.com/snoyes/gallery/images_summer_2006
The thesis here, supported by experiment, is that the reason the 5D has about 2.5x 'better' sensitivity w.r.t. ISO is not at all mysterious: it is the ratio of the total light collected by the respective sensors. (for const FOV and f-stop, as dictated in this case)
You cannot increase s/n or ISO or whatever you want to call it indefinitely. Eventually you'll end up counting photons, like telescopes do...
Actually, you are right: it is the total light collected for a given FOV that matters. It just so happens that if you keep the FOV and the f-stop constant, then the larger sensor must, by definition, collect more light. QED.
No.
) Futhermore, the images should contain about the same amount of information.Just wait for the results. We'll then see the merits of this annoyng bigger-sensor-collects-more-light conspiracy theory.Do the same experiment with ISO set to 260 and f4.5 in the 5D.
Then the exposures will be equivalent. (I think. It's late.
Futhermore, the images should contain about the same amount of
information.
Hehehe...Just wait for the results. We'll then see the merits of thisDo the same experiment with ISO set to 260 and f4.5 in the 5D.
Then the exposures will be equivalent. (I think. It's late.
Futhermore, the images should contain about the same amount of
information.
annoyng bigger-sensor-collects-more-light conspiracy theory.
Huh? None of the 'conspirators' have ever mentioned exposure.
Yes. There is something called the diffraction limit. When you are diffraction limited, stopping down your lens loses you resolution. If your xxD had the same pixel density as a G6, you would only be able to get good sharpness out of your lenses at f/4 or better. Larger sensors also have the ability to collect more photons per unit time. This increases sensitivity.
This is news to me! What format do you want me to use when I shoot
your wedding? 1.6X crop, FF (1x) or 2 1/4 film?
