Canon 1Ds Mark III diffraction limit

[bjanes]

Hi,

I made a small experiment. I ran a series of tests with Imatest on my two DSLR-s an KM Dimage 7D and a Sony Alpha 100. The 7D is a 6 MP camera and the Alpha 100 is 10 MP. Pixel pitch is about 7.8 microns on the 7D and 6.1 microns on the Alpha 100.

I then plotted uncorrected LW/PH for both cameras. I think it is quite obvious that this lens achieves it's peak performance about aperture 8. The Alpha looses sharpness faster then the 7D, even if it is essentially always sharper than the 7D.

Sharpness at full aperture may be negatively affected by focusing errors, I did no focus bracketing to find optimal focus.

Airy disk size for f/8 would be 10.7 micron, for f/11 14.8 micron and for f/16 21.5 micron.

[{POST_SNAPBACK}][/a]

I missed that observation before. This is yet another report from actual experience that
diffraction starts to reduce resolution at aperture ratio about twice the pixel pitch in microns
which is 8 microns for the 6MP Sony sensor in the Konica Minolta 7D.
[a href=\"index.php?act=findpost&pid=141501\"][{POST_SNAPBACK}][/a]

Erik & BJL,

Very nice series of tests, Eric. The Alpha 100 and the Nikon D200 both have the same pixel pitch, 6.1 microns. My own tests and those of Klaus at Photozone.de both demonstrate that good lenses have peak resolution at f/4 or f/5.6 on the D200. The diffraction spot for green light is 5.2 microns for f/4 and 7.2 microns for f/5.6 and 10.3 microns for f/11.

Klaus's tests have focus follow through and are done meticulously. For the [a href=\"http://www.photozone.de/8Reviews/lenses/zeiss_zf_50_2/index.htm]Zeiss Makro-Planar 50 mm f/2 [/url] peak resolution the center is at f/4 and there is a definite fall off at f/8.

This would indicate to me that BJL's criterion of twice the pixel pitch is somewhat lax. Resolution begins to fall off when the diffraction spot exceeds the pixel size. The two times criterion is reminiscent of the Rayleigh limit where the first minimum of one image point is at the maximum of another as shown:



I'm sure the analogy is simplistic, but it makes intuitive sense that when the diffraction spot begins to spill over into the next pixel, contrast would be lowered and MTF would suffer. The Rayleigh criterion MTF if about 9%, which is low for photographic work. Perhaps someone with a better physics background (BJL is proficient here) can comment.

Bill

[BJL]

This would indicate to me that BJL's criterion of twice the pixel pitch is somewhat lax. Resolution begins to fall off when the diffraction spot exceeds the pixel size.
[a href=\"index.php?act=findpost&pid=141660\"][{POST_SNAPBACK}][/a]

Sorry, I did not make myself clear: I was thinking of a rule of thumb for when the decrease in resolution becomes sufficient to be of practical concern, not the first onset of a decline that is measurable in the lab, but might be of only very small magnitude. Erik's graphs above for the 6 micron pitch A100 sensor show a rather modest decline at f/11, the measurement closest to twice pixel pitch, so I doubt that f-stops lower than about f/12 would show a particularly noticeable fall-off in print sharpness.

In fact Erik's graphs might really show that improving sensor resolution does not move the aperture ratio giving maximum resolution significantly with this particular lens (80-200/2.8 APO?), but simply raises the resolution at all f-stops. Maybe this lens has too strong aberrations at f/8 and below to give a clear answer and a high quality prime is needed for a better experiment. (Resolution decreasing once aperture ratio gets below f/8 is not state of the art these days, even for good zooms designed for 'APS-C' formats.)

With the Zeiss 50/2 and D200, f/4 is the peak, but at what point does the fall-off produce visible effects? By f/22 definitely and by f/16 probably, but at f/8 and f/11 it is not clear to me: what percentage decline is needed to be visibly less sharp? That is why I prefer judgments based on print viewing.

For a perfect aberration free lens, measured resolution will be highest at minimum aperture ratio, no matter how low, falling off at least slightly as soon as one stops down from wide open, even at aperture ratios far less than the pixel spacing. But the point at which the decline is noticeable might come only at distinctly higher aperture ratios. So clearly the effect of lens aberrations is also a factor in where the resolution drop becomes significant: proposed theoretical answers in terms of only pixel spacing and aperture ratios can only be loose approximations, valid for lenses of sufficiently low aberrations.

[ErikKaffehr]

Hi!

I think that the lens I used is a pretty decent one (Minolta 80-200/2.8 APO, original version) . At full aperture I think that focusing might have been an issue. Some focus bracketing would probably be needed. I checked both the Canon 80-200/2.8 and the Nikon 80-200/2.8 at Photozone and both reach maximum performance at f/5.6 while my lens seems to max out at f/8. I cannot compare my LW/PH figures with Photozone, because I don't know exactly which settings Klaus uses. I might repeat the test with a better lens, but I don't have any high quality fixed focals laying around. I sort of find that I prefer the convenience of zooms over the quality and speed advantages of fixed focals.

Also I think that the difference in pitch is pretty small on the KM 7D and the Alpha 100. The original question was about about diffraction limit on the Canon 1DsIII and that camera seems to have a pixel pitch of 6.4 microns so it would be in the same ballpark as the Alpha 100 or the Nikon 200D.

Erik

Sorry, I did not make myself clear: I was thinking of a rule of thumb for when the decrease in resolution becomes sufficient to be of practical concern, not the first onset of a decline that is measurable in the lab, but might be of only very small magnitude. Erik's graphs above for the 6 micron pitch A100 sensor show a rather modest decline at f/11, the measurement closest to twice pixel pitch, so I doubt that f-stops lower than about f/12 would show a particularly noticeable fall-off in print sharpness.

In fact Erik's graphs might really show that improving sensor resolution does not move the aperture ratio giving maximum resolution significantly with this particular lens (80-200/2.8 APO?), but simply raises the resolution at all f-stops. Maybe this lens has too strong aberrations at f/8 and below to give a clear answer and a high quality prime is needed for a better experiment. (Resolution decreasing once aperture ratio gets below f/8 is not state of the art these days, even for good zooms designed for 'APS-C' formats.)

With the Zeiss 50/2 and D200, f/4 is the peak, but at what point does the fall-off produce visible effects? By f/22 definitely and by f/16 probably, but at f/8 and f/11 it is not clear to me: what percentage decline is needed to be visibly less sharp? That is why I prefer judgments based on print viewing.

For a perfect aberration free lens, measured resolution will be highest at minimum aperture ratio, no matter how low, falling off at least slightly as soon as one stops down from wide open, even at aperture ratios far less than the pixel spacing. But the point at which the decline is noticeable might come only at distinctly higher aperture ratios. So clearly the effect of lens aberrations is also a factor in where the resolution drop becomes significant: proposed theoretical answers in terms of only pixel spacing and aperture ratios can only be loose approximations, valid for lenses of sufficiently low aberrations.
[a href=\"index.php?act=findpost&pid=141674\"][{POST_SNAPBACK}][/a]

[BJL]

I might repeat the test with a better lens, but I don't have any high quality fixed focals laying around. I sort of find that I prefer the convenience of zooms over the quality and speed advantages of fixed focals.
[{POST_SNAPBACK}][/a]

Me too. Especially with telephoto zooms, which can be very good: I love my Olympus 50-200/2.8-3.5, which has resolution improving as you open up until about f/4: [a href=\"http://www.photozone.de/8Reviews/lenses/olympus_50200_2835/index.htm]Olympus 50-200/2.8-3.5 photozone test[/url]

The original question was about about diffraction limit on the Canon 1DsIII and that camera seems to have a pixel pitch of 6.4 microns so it would be in the same ballpark as the Alpha 100 or the Nikon 200D.
[a href=\"index.php?act=findpost&pid=141771\"][{POST_SNAPBACK}][/a]

And in fact your graph answers the original concern very nicely. It suggests that increasing pixel density does no harm to resolution, instead improving it at any given f-stop, just with less improvement at very small apertures.

Also, there continues to be some improvement al the way to about f/22, so even beyond my rule of thumb limit of "twice the pixel spacing", increased pixel density still gives some rewards in increased resolution. Maybe once the pixel spacing in microns is down to about one quarter of the aperture ratio, further reduction of pixel spacing will give no further significant increase in resolution at that f-stop. I doubt that DSLRs will ever hit that limit at good old "f/8 and be there."

[Mark D Segal]

The 12.7 MP 5D has a pixel size of 8.2 microns and the 21 MP 1Ds MIII pixel size is 6.4 microns.
Bill
[a href=\"index.php?act=findpost&pid=140688\"][{POST_SNAPBACK}][/a]

Bill, do you think this difference in pixel size has any necessary implications for the anticipated image quality from the 1DsMKIII compared with the 5D (defining "image quality" in any manner you see fit - there are of course a number of dimensions!)

Mark

[ErikKaffehr]

Hi!

Resolution is significantly higher on on the 1DsIII than on the 5D. Pixel surface is something like 1.6 times larger on the 5D than the on the 1DsIII, which would give a small (< 1stop) advantage for the 5D regarding noise. The 1DsIII is a later generation of sensor and electronics and I would suggest it's probably improved quite a lot, so I would expect it to be at least as good as the 5D

From the viewpoint of diffraction limitations the 1DsIII should be behave like a 10 MPixel APS-C camera, albeit with a much bigger sensor surface. The 5D is more like a 6MPixel APS-C camera. That really means that resolution should be pretty similar to what I have in my figure for diffraction effects. There should not be a significant loss of sharpness up to around f/11.

To fully utilize the full frame you would need high quality optics. Most lenses should yield good performance on axis (in the center) but demands on corner sharpness is hard to meet in wide angles.

Another way to see it is that Michael Reichman and others say that the D40 is very close to the 5D with regard to image quality. If Canon wants to sell a camera for five times the price of the 40D it needs to have something going for it.

Best regards

Erik

Bill, do you think this difference in pixel size has any necessary implications for the anticipated image quality from the 1DsMKIII compared with the 5D (defining "image quality" in any manner you see fit - there are of course a number of dimensions!)

Mark
[a href=\"index.php?act=findpost&pid=142056\"][{POST_SNAPBACK}][/a]