| Author |
 Replies: 47 / Views: 6,568 |
|
|
|
Pillar of the Community
 2087 Posts |
Quote:
There is also a trade-off between more megapixels and low light performance. The benefits of that extra resolution are marginal to non-existent; you get more noise and a larger file while the real resolution doesn't reach that of the sensor in many situations (diffraction is just one of many possible limiting factors.) If you're a pro photographer shooting sports or a rock concert or a wedding reception in natural light, which would you choose? A lot of pros who do rock concert photos are going for the Sony A7s(to illustrate your point). Full Frame 12mp with an ISO range of 50 to 409600, I dimly recall the "natural" ISO range for the A7s is actually around 100-102400. The pixels are large to ensure low noise. Edited by austrokiwi
02/03/2015 05:51 am
|
|
Pillar of the Community
 United States
4038 Posts |
Quote:
I disagree strongly you are both confusing the ability to produce a high pixel density in a small area with the ability to apply the same density on a large sensor. Its not just the pixels but the design and placement of the micro lenses. Using your arguments We could say the technology for going to the moon had been available for years before the apollo program. In actual fact the science had been available for years the application of that science ( technology took along time to follow) No confusion here at all. The problem isn't lack of technology, but market. If Sony suddenly produced a 200MP FF camera, they would not be able to sell it except to a very small crowd, so it would end up being extremely expensive, further reducing market size. They would be accused of being "ahead of their time". We have long been in the mode of manufacturers needing to make incremental improvements, and their competitors doing the same, in order to maintain a stable market. This slow motion development has the advantage of giving time to work out manufacturing issues such as yields and supply chains on a stable new product release schedule. Certainly yields are more of a problem for larger sensors, and for higher pixel density, and this must be dealt with in order to produce the camera for a price that the market will bear. Bottom line, I guarantee you a Sony engineer has a prototype of a 200MP (or perhaps 400MP) FF camera in their hands right now, and so does a Nikon engineer and a Canon engineer, working out the bugs. Contact me for photographic equipment or visit my home page at: http://macrocoins.com |
|
Moderator
United States
23522 Posts |
Austrokiwi, none of your arguments change the size of a Circle of Confusion, or the amount of processing which can be crammed into a camera. The only difference between Sony's RX100 sensor and a FF sensor is the size they chose to build it. The reason a 140MP (or a 50MP) FF sensor won't work for coin imaging is the exact same reason why a 16MP, 1/2.7" sensor in a P&S won't work: The pixels on the sensor are smaller than the Airy Disk. That's a strict mathematical certainty. We're already limited only by diffraction at current pixel densities, and more pixels in the same space will only make it worse.
|
|
Pillar of the Community
2087 Posts |
Ssuper dave Rmpsrmps has already told us all there is no major problem with diffraction on a 50mp FF sensor: heres what he said...go argue with him! (I understand the pixel size is only slightly smaller than an 18mp APSc camera....My APSc C camera is 24MP and it doesn't have any of the problems you predict!} Quote:
It's easy to calculate what the pitch will be.
FF sensors are 1:5:1 aspect ratio (36 x 24 mm). So assuming square pixels, TotalMP = Xpixels * Ypixels = 1.5*Ypixels * Ypixels = 1.5Ypixels^2
So for a 50MP sensor... 50e6 = 1.5Ypixels^2, or Ypixels = 5774 and Xpixels = 8660. The sensor will actually be some nicely-divisible number, perhaps 8832 x 5888. This actually calculates to 52MP and divides nicely by 2's down to 69x46. It also has LOTS of 8's so will be a hit with folks who love them!
So now we know what the sensor size will be, 8832 x 5888, for 52MP. What is the pixel size?
36mm / 8832 = 4.08um. This is a bit smaller pitch than my 18MP APS-C at 4.3um, but not enough to worry about diffraction problems being any worse. The DLA, which is just the onset of when diffraction can start to degrade the image, will be f/6.4, so not much of a problem for current lenses. Edited by austrokiwi
02/03/2015 2:17 pm
|
|
Pillar of the Community
United States
4038 Posts |
AK...I was referring to portrait/landscape photography, which is the main use for cameras and represents the biggest market. SD is correct that we're already at the point of having issues with diffraction at existing pixel resolutions for macro use, though 50MP won't be much if any worse than where we are today. I made another post where I discussed that for f/2.8 operation we are theoretically still good to 67MP even for macro, but you of course need a lens that is good at f/2.8 and 1:1. Got one of those?
|
|
Moderator
United States
23522 Posts |
It's a practical situation, not a hard limit. Either glass of sufficient quality, or in-camera software capable of using all those spare pixels to artificially create smaller Circles of Confusion, will break us loose. I expect the latter sooner than the former.
|
|
Pillar of the Community
United States
4132 Posts |
Quote:
but you of course need a lens that is good at f/2.8 and 1:1. Got one of those? I know you do, but I am not keen on doing a focus stack just to take a full-coin image of a dime. :) |
|
Moderator
United States
23522 Posts |
Yeah, that's another thing.  |
|
Pillar of the Community
2087 Posts |
Now let me get this right you are saying the pixel size on a 50mp0 ff sensor isn't going to work for Macro so why does a 24mp APSC sensor not have any problems from what you indicated,RMPSRMPS, on a 50mp FF the pixel size is only slightly smaller than your 18mp APSc camera so the 24mp APSCs pixels must be smaller than the FF50mp. BTW: SSuperdave who was the person who said this: Quote:
"Wow. A 46MP FF pro-level MILC. The camera geek in me is giggling uncontrollably - that's practically enough sensor to abandon >1:1 magnification. Mag? I don't need no stinkin' mag" BTW my standard coin size ( for photography) is 42mm in diameter Edited by austrokiwi
02/03/2015 9:52 pm
|
|
Pillar of the Community
United States
4132 Posts |
Quote:
so why does a 24mp APSC sensor not have any problems It does have a problem, namely, it's diffraction limited at macro range unless you have a lens that can perform at like f/2.8 and you're willing to stack. Those megapixels are being wasted. |
|
Pillar of the Community
2087 Posts |
What is macro range? For most of us we are photographing at sub macro ( magnifications 1-1 or less) when I take 5-1 magnification shots I am stacking!
|
|
Moderator
United States
23522 Posts |
That quote was about the effective magnification available from a sensor of that size. The implication of that huge pixel count means you can reduce the magnification - and the onset of diffraction - while not compromising size. Conversely, you could face the diffraction dragon by stacking a wider aperture and have images so large you don't need detail shots. Still gotta obey the laws of physics, and it should be noted that a 46MP FF with 2.6x as much surface area as APS-C should have less-dense pixels than the 18MP Canon Rebel sensor we like so much.
|
|
Pillar of the Community
United States
4038 Posts |
Quote:
BTW my standard coin size ( for photography) is 42mm in diameter So you're mostly shooting at around 0.5:1 on FF. Certainly you're not going to have the same issues with diffraction that I do. I mostly shoot Cents, which are 19mm diameter. So on APS-C I'm typically at around 0.7:1 (leaving a little room around the coin). On FF I'd be around 1.1:1. At 0.5:1, the best lens (other than the exotics) is probably the 55mm f/2.8 Micro Nikkor. The low magnification means you can get a reasonable working distance, so no major lighting problems. The 55Micro is also diffraction-limited from f/4 so at M=0.5 you are within the DLA of your 24MP APS-C sensor. Seems like all-around goodness. I can't use the 55Micro at 1:1 since it really only does well up to 0.5:1 and the working distance is unworkable for good lighting control. Your 42mm coin size preference is a distinct advantage in being able to use the available technology, and you won't have as many issues going forward as those of us shooting smaller coins. |
|
Pillar of the Community
2087 Posts |
Edited by austrokiwi
02/04/2015 02:34 am
|
|
Moderator
United States
23522 Posts |
Diffraction is a strict mathematical function. No subjectivity is involved. If the characteristics of your equipment and methodology add up to diffraction, you're experiencing diffraction whether your eyes think so or not. Create a series of images at steadily-smaller apertures so you can see how diffraction looks as it progresses; it'll show you what you're missing at the beginning of the process. It's only visible in 100% crops, of course. We're talking about artifacts a few microns in size - it's not exactly obvious.
I'm not seeing how the answers you're getting in that thread are any different.
|
| |
Replies: 47 / Views: 6,568 |
First Page
Posted 02/03/2015 09:42 am
