8
\$\begingroup\$

I have a pretty high res sensor on my camera, but sometimes I have the feeling my lens doesn't even get to this resolution. How can I tell if my camera sensor has not enough resolution or if the lens is the limit?

For example I have a 5DMarkII with 22 Megapixel Sensor. And a Canon 35mm 1:2 Lens. (For example aperture at 11 shutter at 1/2000)

How can I visually see whether the bottleneck is the lens or the cameras resolution? Is there any place were I can look up the optical resolution of a lens?

Improve this question
\$\endgroup\$
0

6 Answers 6

Reset to default
10
\$\begingroup\$

There is no hard limit to resolution - details don't just vanish they fade out gradually as the contrast between light and dark is reduced. This reduction in contrast is expressed by means of the modulation transfer function (MTF). It's just a fancy way of saying details of size X will experience a loss of contrast Y.

The MTF of a camera system is the mathematical product of the lens MTF and sensor MTF. This means that any improvement in either lens resolution or sensor resolution will result in an improvement of the overall MTF.

So the technical answer to your question is, it is both.

However you do experience diminishing returns after a point, that is improvements in sensor resolution yield smaller and smaller improvements to system resolution (you can't take sharp images with a coke-bottle by using a sufficiently high resolution sensor.

Is there any place were I can look up the optical resolution of a lens?

Canon publish the MTF charts for all their current lenses on their website, your 35mm f/2.0 has been superseded by the IS version, however a google search turned up the MTF chart:

For information on how to read the chart, see this answer:

How do I interpret an MTF Chart?

What you can see clearly from the thin black lines is that wide open at f/2.0 that lens delivers pretty low contrast for high frequency details, less than 60% over most of the frame. This probably accounts for the poor results you're seeing.

As your sensor is currently pretty much the highest resolution available from Canon I don't think it's holding you back particularly. Switching to the 36MP Sony A7R would yield a measurable increase in resolution, but much less than upgrading the lens (to something like the Canon or Sigma 35mm f/1.4, or even the newer Canon 35mm f/2.0 IS) would.

Improve this answer
\$\endgroup\$
4
  • \$\begingroup\$ "There is no hard limit to resolution" - false. Wavelength is hard limit. We are far, far from it yet, but it is there. About 700 nanometres if we assume perfect detectors can exist. So it is 50 000 x 34 000 (1700 megapixels) for full frame sensor. \$\endgroup\$
    – Mołot
    Commented Jan 15, 2015 at 14:34
  • 1
    \$\begingroup\$ @Mołot How about "There is no practical hard limit to resolution in the foreseeable and reasonably imaginable future"? \$\endgroup\$
    – mattdm
    Commented Jan 15, 2015 at 15:53
  • \$\begingroup\$ @mattdm Would be closer to truth, but hey, in last 20 years max density for full frame grew from 0.5 Mpx to 50 Mpx. 100 times more. If this would to repeat, we should have 5000 Mpx by 2035, and that's significantly more than hard limit. \$\endgroup\$
    – Mołot
    Commented Jan 15, 2015 at 16:02
  • 1
    \$\begingroup\$ It may be worthwhile to note that if an image contains spacial frequency content above the Nyquist frequency of the sensor, such content may appear as content at much lower--and much more visible--frequencies. Even if increasing the resolution of a sensor beyond the useful acuity of a lens may not improve its ability to gather useful information, it may improve its ability to reject useless information. \$\endgroup\$
    – supercat
    Commented Jan 15, 2015 at 18:41
10
\$\begingroup\$

Lens resolution is much more complex than sensor resolution. It's not constant across the lens, and changes with aperture.

Here's a forum discussion that goes into the details of lens/sensor comparisons, and the important part is: even a perfect lens would have a lower resolution than your sensor at f/11 aperture due to diffraction!

Extremely low apertures are not good either in a real lens, due to aberration. So you'll get the best resolution probably around f/4 or f/5.6.

As for whether your specific lens will then outresolve the sesor: there are websites that have detailed data on lens performance, but you need a lot of technical knowledge to understand it (to be honest, I don't have it either). For example, here's Photozone's data on your lens:

enter image description here

Improve this answer
\$\endgroup\$
1
  • \$\begingroup\$ There are a few lenses that are designed to be as sharp wide open as they are at any narrower aperture. The Canon EF 300mm f/2.8 L IS II comes to mind. It's as sharp or sharper across the field at f/2.8 as it is at f/4, f/5.6, and f/8. But photozone (now "Optical Limits") doesn't seem to test any of the wide aperture super telephoto series from Canon. \$\endgroup\$
    – Michael C
    Commented Jul 14, 2020 at 17:09
1
\$\begingroup\$

Of course there are limits, just as there are limits to the amount of digital information that can be put into a CPU chip for computers. The limits usually increase in discreet steps, as the technology evolves. Nevertheless there are hard limits based on the molecular structure of matter.

There are also limits to the resolving power of lenses depending on the technology to design and manufacture them. In general, although the performance of zoom lenses has improved dramatically over the last 40 years, due to computerized design, the basic resolution available from optical glass has improved only marginally.

As I write this, there are only a few lenses that can match the maximum resolution available in sensors, but that will change.

Improve this answer
\$\endgroup\$
0
\$\begingroup\$

The DXOMark results for this lens and camera combination give quite good sharpness results, but they show that it already drops of at f/11. enter image description here

Best at f=35mm & f/2.8

Their "Perceptual MPix" score is supposed to be much more intuitive than those MTF50 charts.

Improve this answer
\$\endgroup\$
0
\$\begingroup\$

You can try using Imatest or sfrmat3, it helps answering this type of questions and better understand the subject.

Improve this answer
\$\endgroup\$
0
\$\begingroup\$

Transfer this lens to a body of the same grade or exactly the same model that you are sure of the health of its sensor, and then compare the focus resolution in the recorded output. If the result was better and the resolution was palpable and noticeable, the problem is with the current sensor of your camera, but If the results were similar to each other and the resolution was low, there is a high possibility of a problem with the accuracy of the lens. This issue should be examined a little more closely.

Improve this answer
\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.