Diffraction? To those smarter than me 20MP APC

I'm very interested by the new 20MP Samsung NX200. However as I understand it...(I don't really understand it at all)

Isn't this sensor going to be severely limited by diffraction when stopping down? I mean I remember reading somewhere that it kicks in on the 7D around F8 or so?(vague memory) The NX sensor is slightly smaller(though still APC) and has more MP.

I'm interested in this but have no real knowledge...please feel free to enlighten me!

Trying to answer my own question found this....

http://www.fredmiranda​.com/forum/topic/74268​6

(I'll remove if I'm not allowed to link to that site someone let me know)

(EDIT 1) and this:

http://www.cambridgein​colour.com …al-camera-sensor-size.htm

EDIT 2:

Even better this:
http://www.cambridgein​colour.com …ffraction-photography.htm

Big deal, the new Sony cams are 24MP and are APS-C as well. Also, diffraction should kick in at about f/5.6 on the 7D, but this is assuming two pixels are required to resolve detail (in other words theoretical), in practice it may be higher.

Ya...after reading about diffraction for the last hour+ again I'm realizing this is not a big deal again...

Sometimes there is just so much to know...and I forget about diffraction until a new camera comes out and it makes you think about it again...that hey aren't they pushing against some limits!

But when you think about the size prints you make...you realize that even though the limits are there most of the time they are not going to effect the image in its final viewing.

And if you have to stop down to get the shot...You have to stop down...not much you can do about it.

Thanks

You do also have to think about what exactly you'll be using the camera for, a diffraction limit of f/5.6 sounds limiting, but if you bought the camera to "shoot fast", is it really a problem?

People who buy the 7D almost certainly get it for the speed and reach it has over the 5D, and that can mean sports, wildlife, portraits and none of those things really require stopping down too much. If you want to shoot landscapes or anything else of that nature, you probably want to go full-frame, especially when lenses suited for those roles also perform their best on FF cameras anyway.

The f-Number at which diffraction will BEGIN (notice that I'm using the word "BEGIN") to inhibit a desired print resolution (specified in lp/mm) at an anticipated enlargement factor can be calculated as follows:

f-Number = 1 / desired print resolution / anticipated enlargement factor / 0.00135383

1) Thus, IF (notice that I'm using the word "IF") you personally want to prevent diffraction from inhibiting a desired print resolution of 5 lp/mm AND IF (again, I'm using the word "IF") you anticipate an enlargement factor of 16.4x (to produce a 10 x 15-inch uncropped print from the APS-C sized sensor), you would have to avoid stopping down below f/9.0

f-Number = 1 / 5 / 16.4 / 0.00135383 = 9.0

2) Another example (smaller print, but the same desired print resolution):

IF you personally want to prevent diffraction from inhibiting a desired print resolution of 5 lp/mm AND IF you anticipate an enlargement factor of 8.2x (to produce a 5 x 7.5-inch uncropped print from the APS-C sized sensor), you would have to avoid stopping down below f/18.0

f-Number = 1 / 5 / 8.2 / 0.00135383 = 18.0

3) Another example (larger print than the first example, but a lower desired print resolution):

IF you personally want to prevent diffraction from inhibiting a desired print resolution of 2.5 lp/mm AND IF you anticipate an enlargement factor of 32.8 (to produce a 20 x 30-inch uncropped print from the APS-C sized sensor), you would have to avoid stopping down below f/9.0

f-Number = 1 / 2.5 / 32.8 / 0.00135383 = 9.0

There is no one f-Number at which diffraction begins to visibly degrade resolution for a given camera or sensor.

Mike Davis

Obviously this is variable based on print size and print resolution, be we aren't talking about prints or the science thereof, we're talking about the absolute measurable point at which diffraction begins to degrade an image. This, of course, means viewing at 100% zoom in Photoshop of a subject we assume to have extremely fine detail with enough contrast. This is why in my first reply I wrote "this is assuming two pixels are required to resolve detail (in other words theoretical)".

In practice you really can get away with say, an aperture of f/22 for a product photo in a magazine, because it's easy to calculate at which point the camera's artifacts and aberrations are going to be visible. On reflection, I should have written "in practice it will be much higher" instead of "in practice it may be higher"... Some people seem to forget this is a photography forum and thus when someone says "lens is soft" or "diffraction sux" you can bet your ass they're looking at the equivalent of a 2-meter wide print.

Kolor-Pikker wrote in post #13117424
Obviously this is variable based on print size and print resolution, be we aren't talking about prints or the science thereof, we're talking about the absolute measurable point at which diffraction begins to degrade an image. This, of course, means viewing at 100% zoom in Photoshop of a subject we assume to have extremely fine detail with enough contrast. This is why in my first reply I wrote "this is assuming two pixels are required to resolve detail (in other words theoretical)".

Except that when an image is viewed at 100% there is still an amount of enlargement from the original optical image on the sensor involved which is not absolute. It is sensor resolution (sensel density) / monitor resolution, so the equation is still valid. For instance the 5D2 has a sensor resolution of nearly 4000 ppi and the 7D has around 5900 ppi. Assuming a monitor resolution in the 96 - 100 ppi range, the 5D2 image is enlarged x40 and the 7D is enlarged x60. There are no absolutes.

But it's still technically visible or at least measurable at which point detail is blurred to the extent where it crosses over several pixels instead of just a few. There may be a different amount of enlargement depending on the viewing medium, but I'm taking the viewing medium out of the equation, the true point at which diffraction starts to take place on the sensor is not the same as when it becomes visible on screen or in print.

I know it's just measurbating on my part, but a line of detail 1 pixel in width on the 7D is going to start blurring across more lines starting at f/5.6 and that's fact. This has little bearing on real photography, as debayering, noise, anti-aliasing, optical deficiencies and the circle of confusion will cloud all of this in practice, however, it's still a point of reference.

Well, then, it's a lot like discussions about noise and sensor resolution, as well as how sensor size compares in regards to resolving an image. A lot of the differences are technical but as to how it affect your images in real life usage, well, it depends. Large prints or close crops can show up, well, all that jazz. But for many/most people it will likely not be an issue. Your occasional poster-size prints, well, hang 'em in a way that won't enable close up viewing!

I have occasionally shot at very narrow apertures, both for some good close-to-distant depth of field on some scenes and then when doing macro photography. Even thought I prefer to stick to more "normal" apertures (flike f/16 on a full frame or, if I can, f/11 on a crop body) I have on occasion pushed to say f/22 and been quite satisfied. Yes, when viewing fine details at 100% there would be a bit more softness than I'm used to but when printed at one of my "normal" print sizes (like 12x18") then, well, a good shot makess a good print!