I (think I) understand the principle that as a crop sensor is physically smaller than FF, the image has to be blown up, say, twice as much for typical viewing size.
However, my possibly wrong idea is twofold. One, that diffraction at a viewable print size, rather than at sensor level or 100% crop size, comes into play at a narrower than commonly thought aperture (f/11 for m4/3 and f/22 for ff). So the lens can be stopped down more than thought without incurring visible loss of resolution. See below:
Second is that resolution is not “thrown away” if you increase the megapixels in your sensor with a lens that supposedly cannot resolve the increase. I’m taking this from Roger Cicala at Lens Rentals, pretty much on authority but I guess that’s what he is.
During the nearly 20 years I’ve been photographing with FT/MFT cameras, there has been a nearly endless discussion on dpreview on this and related topics on the differences between the various formats. So far, though, I don’t think I have ever seen anything regarding image quality that has any real practical importance.
As image resolutions of cameras are much higher than for the prints and screens we eventually view images on, images are anyhow scaled down before viewing, so that much noise is cancelled out and max resolution is not important.
So if your hobby is to pixel peep or you have some special technical needs – or if you plan to skip buying longer lenses and rather want to “zoom in” by extreme cropping – you can spend a lot of time studying the technical discussions to which there seldom seems to exist conclusive answers.
Else I suggest one selects what fits a sweet spot between price and preferred handling of camera during transport and shooting.
(The real difference in my mind is between DSLRs and mirrorless cameras, as the latter provides a whole lot of new ways to make pictures.)
Agree that resolution is just one factor. My main takeaway so far may be that I don’t need any new camera given my output. But I enjoy trying to understand, somewhat, the principles, if only so that I can know what’s real and what’s marketing spin. Cheers
Nice rabbit hole — I symphathize, as I would have a tendency to get lost in this
But from a practical perspective, I would consider
the desired level or “sharpness”,
the typical focal lengths and apertures you will be shooting at,
the budget, and
tolerance for bulk/weight first.
Practically, some full frame lenses (more expensive Nikon primes, Sony GM series) are indeed amazingly sharp. But they are fairly expensive, and bulky too (especially the higher-end zooms). A standard zoom will set you back about $1–2k, more expensive if you want higher apertures. For the price of a pro telephoto lens for wildlife & birds, expect to sell a kidney and carry a lot of stuff.
On the other end of the spectrum, if you are happy with micro 4/3, you can expect to carry about 1/2–1/3 of the weight, with about a 50% saving in price. This of course comes at a cost… but is this relevant for you photography? Do you make super-large prints?
Personally, compared to my FF/DSLR days, I am so much happier with micro 4/3. After some initial enthusiasm, all my FF gear was gathering dust of the shelf because it is not something I would just casually carry in case I encounter something interesting, and I was always coveting lenses I could not justify buying.
Some micro 4/3 lenses are amazingly sharp too, eg the Olympus 75mm f/1.8 is a real gem. But in practice “sharpness”, especially outside the center, is not the first thing I care about. Lenses have a lot of character worth exploring beyond MFT.
I agree. I don’t have much tolerance for spending that kind of money on lenses for resolution that I don’t need. I learnt a bit in the process, I think, fwiw.
Unfortunately the weight reduction seems to only be really worth it in more compact cameras. For example, the OM-1 is almost as heavy/large as a sony a7IV. The OM-1 has more computational features, better stabilization, and M43 lenses are still smaller when it comes to longer focal lengths, but it’s not always as clear cut when it comes to the higher end cameras.
Either way, smaller formats have always been used and they have proved their worth. Super 35 was used for decades and is still used for cinema. Call it super 35 and nobody bats an eye, call it aps-c and they all lose their minds and start comparing it to full frame
Yes, that is currently a sore point for micro 4/3. Both manufacturers are focusing on large body flagships (OM-1, G9ii), neglecting the smaller bodies that are unique to the format.
I don’t know if that is an engineering problem (so much a stuff like ML coprocessors having to fit into the body), trying to influence brand perception (we are a serious camera brand, because we make large cameras that look very professional), or simply a matter of priorities (flagships first, smaller cameras later).
Fortunately, the used market has a lot of smaller bodies, but that is the case for APS-C formats too (eg gems like the Sony A5xxxx, still very capable cameras).
Indeed. With hindsight, Panasonic & Olympus should have called the format something like “super 4/3”
This may be way too simple an approach, but here goes:
There are two things involved here: the lens and the sensor.
The lens projects an image on the sensor, which samples the image. The smaller the sensor elements (“pixels”) the more resolution you can get, but the lens has to provide that resolution.
So for a given pixel size there is no difference between crop and full frame sensors.
So the question in itself is poorly formulated, as you need to take into account the pixel size.
Of course, the full frame sensor will have many more pixels for a given pixel size. So at the same final output size, a full frame sensor could give more detail. But that’s at large print size…
My APS-C sensor provides 6000×4000 pixels. An A4 print would require about 3000×2000 pixels. That’s half my original resolution…
I think the Cambridge in Colour excerpt and the later threads argue that the image on a smaller sensor with the same number of megapixels has to be enlarged to a greater extent for any particular print size so therefore resolution is inherently lower unless that is compensated for by a lens with greater resolving power (than the one on the larger sensor). I think it’s effectively what you said but with different assumptions.
I agree that you need to know your resolution needs, too. Perhaps the relative advantages of the smaller format (ignoring for a moment whatever manufacturers actually produce) are somewhat less than I had previously assumed as I had only really thought in terms of a lower dynamic range and signal to noise ratio. Thanks
Well yes, but if you have the same number of pixels on two sensors (with the same aspect ratio, to keep it simple), the smaller the sensor, the smaller the pixels. So yes, in that specific case, to get the same level of scene detail, you’ll need a sharper lens.
And that doesn’t require any “assumptions”. Assumptions come into play when you compare different sensors without knowing all the relevant facts about them…
Those two concepts are related: the dynamic range is determined by the highest signal a sensor can handle, and the lowest signal you can distinguish from noise, so a lower signal to noise ratio will lead to a lower dynamic range…
All current cameras, including 50MP smartphones are resolution-limited by their sensel size, not diffraction. Diffraction is generally still an order of magnitude smaller than even the smallest sensels (for open-ish apertures). Note that that f/1.8 smartphone lens is equivalent to an f/4 MFT lens. All else being equal, you’d expect these two to perform similarly.
But a $500 fist-size lens will always outperform a $5 pinkie-nail-size lens. I reckon that lens resolution, aberrations, veiling glare etc, ultimately play a much larger role than the marginal differences between our essentially-perfect sensors.
Yes, that’s my understanding after all this. I think the point made in the first post on the resolving power of the lens is roughly correct (ignoring the real world complications that the single MTF figure can’t capture), but the diffraction limiting argument isn’t really relevant as that doesn’t come into play in any meaningful sense at reasonable f stops. I think.
Probably. But smartphone lenses are not miniature versions of camera lenses (as we know them). They frequently made from plastic, which sounds bad but actually it allows very tight manufacturing tolerances in high volumes, with the minimum number of moving parts. A lot of effort went into designing them well, given the constraints (usually no aperture control, fit into increasingly narrow smartphones, that get bumped around all the time). Tons of weird-looking asphere elements are used commonly these days, much more than in camera lenses.
I would love to see some of these lens designs in MILC cameras, but probably the size makes the cost prohibitive. In this case smartphones lucked out because they are small.
AFAIK it is not uncommon to photosite size to comparable with the Airy disc diameter (eg see this calculation), but since most smartphones combine pixels into clusters this is not a big problem (eg that 48 Mp sensor is actually delivering 12 Mp images).
That’s what I, barely, understand about phone lenses. There’s also a distinction worth making between the diffraction limited aperture and the diffraction cutoff frequency, according to the text above, so the actual viewing medium is relevant, I guess. That former Kodak guy i posted in a separate thread points out that people who say their phones take pictures as good as an ILC are looking at the pictures on their phones. The pics I took recently of coastal sunsets on my phone looked pretty decent on the phone. On my monitor, less so.