full frame vs cropped sensor

Also, correct me if I am wrong, but to get the same photo with the same lens in a full frame camera I would have to be closer to the subject, and would have to close the lens more to get the same DOF, thus increasing diffraction, right?

 

Luiz

 

<{POST_SNAPBACK}>

You are correct that the diffraction ring size increases in this case, but not as much as the increase in the width of the pixels in going to a FF sensor with the same number of pixels. So, the net blurring in the final image is less.

Now here's an interesting point. If you take the comments and observations in this discussion to their logical conclusion, it should become obvious that at some point, the cropped sensor will run out of MPixels before the FF as optical considerations will mean that although you might be able to fit more pixels in, they won't be doing anything as the image will be limited by purely optical (physical) constraints. Now I've discussed this with a lens designer who has come up with an abritrary figure but based on various factors such as those discussed here, anf he things that around 15MPixels will be the economic and optical limit of the 1.5x sensor - interestingly this has already been exceeded by the Canon EOS1DSMkII BUT and it is a big but, I already find that the 1DS outperforms underwater optics and residual chromatic problems are difficult to entirely eradicate at 11MPixels. There anre many complaints too about Canon wide-angle lenses and the MkII. No doubt Canon will eventually upgrade their wide lenses but there will come a time when the economics of doing so will price them out of the market (look at Leica lens prices) as quality control issues tighten if MPixels increase much further. Nikon do not have the same problems with wide lenses - but then there is far less choice especially of fixed focal wide lenses for the Nikons. So what I am saying is that at 10MPixel+ we have pretty mature systems. Whether you prefer Nikon or Canon should now be down to personal preferences. I am very loath to say that one is better than the other. On balance I would say that Canon has the more potential for future development simply because of its bigger sensor, whether Nikon will finally go for a FF camera will depend on their designers and the constraints due to the use of the existing lens mount.

Enough said, back to reviewing the 4000 shots from my last trip!

Hi Luiz,

I think that's a nice shot, but I think it shows the same F32 softness as in the Bob Atkins example.

Cheers

James

James -

Regarding Luiz's photo...do you mean the original or the cropped image is soft?

I'm not sure I'm seeing what everyone else is seeing, especially Herb's example.. I can see a lot of noise in the last cropped picture but is Herb trying to show us a good example or poor example? If it is poor what is in the cropped photo that makes it a poor image? All the noise? What does diffraction look like in pictures?

Thanks!

M.

I'm really enjoying this thread but it's kind of like arguing religion: Nikonians vs Canonites. Most of the arguments fall into the realm of theoretical / hypothetical as opposed to practical. Fun to argue but of little practical difference. Even in the film era it was understood that sharpness falls of at f22 and f32. This is nothing new to digital. That's why I shoot most of my macros at f16.

I think if you are buying a one of the top models from Nikon or Canon you'd be hard pressed to show any REAL difference in image quality.

Larry, I agree with you to a point, it's a great thread but un-like religion, the stand on FF that's it's noticeably sharper with macro shots and that, at f22 on a cropped sensor, quote: "at f22 sharpness is so bad, you might as well be using a 5mp camera". has me dying to see a FF macro shot at f22 so I can see the difference.

Here (again) is my little Bartender: f22, 105mm + 2xTC @ 1/125 full frame:

And eye detail at 100%:

No PS work, but the TIff is from the original Raw file and Rawshooter is quite powerful in the amount of adjustments that can be made before conversion.

Rand

For what its worth (and I'm not posting examples!), I find that f/11~16 is optimal on macro shots with the 1DS (although I tend to use it nearer f/11). Going smaller than f/16 results in softer images. Even so they are usually usable.

But being honest, you have to really blow the images up to see the differences and I reckon that mostly, whilst we as the photographers look at images at 100% on screen and see the problems, there are many other factors that go into making up a good photograph and diffraction limitation may not be the biggest!

I rather like the quality of the out of focus background created by using wide-angle flash together with a 100 macro lens and the 1DS, this is a characteristic of the Canon FF which attracts me to it, the Nikon is slightly different. But as I'm not into black background images I am more concerned about the way the background is recorded than people who like the black background.

Ok, so, if I understood it so far, taking a 6mp camera and blowing it up to a D2x resolution creates similar results. But what happens if I take a D2x file and blow it up to 1ds resolution? Here is a 100% crop of one of my raw files (the same coral I posted before, but this time I processed the RAW file in Nikon Capture) with resolution increased to 16 megapixels:

100% crop (16mp):

I think I can go even further than the above, and that's f32. Here is the original uncropped, unprocessed JPG from the camera again:

Luiz