In this review of the Nikon D3300 w/ 18-55mm f/3.5-5.6G, Tony Northrup says that because the kit lens of Nikon D3300 is 9 mega pixel, you won't use all 24 mega pixel provided by the sensor.

What is the lens resolution? I know that the sensor is made of small dots or pixels which define the sensor's resolution. How do we define lens resolution?

  • 1
    I changed the link to start the video at 19:00 – scottbb Jul 13 '16 at 3:41

In the video, Tony is probably referring to DxOMark's Perceptual Megapixel rating that they developed, because as they say,

Most of our readers were not looking at MTF graphs

While DxOMark hasn't reviewed the DX 18-55mm ƒ/3.5-5.6G VR II lens yet, here are its 2 predecessor reviews, both mounted on the same D3300 that Tony is reviewing:

The "perceptual megapixel rating" hasn't really caught on very much. It mostly appeals to people who like to publish or consume marketing feature numbers or metrics. But really, it's an oversimplification of a more complex description of lens sharpness (MTF charts).

  • DxO Mark's lens test can sometimes be a joke. They apparently have no mechanism in place to insure that a lens is properly aligned before or after testing it. See photo.stackexchange.com/questions/71023/… – Michael C Jul 13 '16 at 9:05
  • Look at their scores for pretty much the entire Canon 70-200mm "L" series compared to every other review site on the planet. Especially the EF 70-200mm f/2.8 L II. It's a joke. Or look at the tests and other comparisons of the EF 16-35mm f/2.8 L II cited in this answer. Compare the DxO Mark data to the results linked in the question and at TDP: photo.stackexchange.com/a/80001/15871 – Michael C Jul 13 '16 at 9:10
  • Look at the question and comments to the question here: photo.stackexchange.com/questions/42882/… – Michael C Jul 13 '16 at 10:09
  • @MichaelClark Agreed, it's pretty meaningless. I'm glad the P-MPix rating hasn't caught on. – scottbb Jul 13 '16 at 12:14
  • I wish the same could be said for accepting DxO Mark test results as the inspired gospel. They're useful most of the time, but sometimes the overall score compared to the actual test data has one scratching their head. Sometimes the actual test data is so far off compared to other tester/reviewer's results that it's obvious something besides the nominal performance of a particular lens design affected the results. – Michael C Jul 13 '16 at 22:41

The resolution of lenses has been deliberated over the years. John William Stutt, 3rd Barron Rayleigh (English scientist Astronomer Royal 1812 – 1842 Nobel Prize 1904). He published the Rayleigh Criterion, the theoretical resolving power of lenses. This study, and method is still valid today. The lens is caused to image ruled lines. The width of the lines and the spaces between are the same. The lens is rated based how the lines are resolved. The resolving power (R. P.) of a lens in lines per millimeter (lpm) = 1392 ÷ f/number This is the value for green light chosen for photographic lens evaluation.
Table of R.P for wavelength 589 millimicrons as to resolved lines per millimeter.

f/1 = 1392 lpm

f/2 = 696 lpm

f/2.8 = 497 lpm

f/4 = 348 lpm

f/5.6 = 249 lpm

f/8 = 174 lpm

f/16 = 87 lpm

f/22 = 63 lpm

The resolving power of any lens decreases as the lens is stopped down due to diffraction. This is due to the light rays passing in close proximity to the blades of the Iris diaphragm (aperture). These close passing rays bleed into the shadow regions and into the path of light that clears the aperture. The result is the boundaries become less clearly defined.

Note: A lens operating at apertures f/8 and larger exceeds the resolving power of pictorial film.

  • This is the theoretical limit of lens resolving power. Poor lens quality will cause a variety of blurring effects, as will a lens system with uncorrected chromatic aberration. Further, the Rayleigh criterion needs to be adjusted for off-axis points, as the image "plane" of most lenses is in fact a concave sphere. (unless some quality flat-fielding was designed into the lens as well) – Carl Witthoft Jul 13 '16 at 18:57
  • @ Carl Withorft -- Nevertheless, we are hard pressed to match much less exceed this theoretical limit. – Alan Marcus Jul 13 '16 at 19:23
  • 1
    @AlanMarcus Which is exactly why the answer might give the false impression that all lenses set to f/4 can resolve 348 lpm. – Michael C Jul 13 '16 at 22:44

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