Lens sharpness and distance from subject

[Lawliet]

On the other hand, my 100mm macro is better suited at close work.

If one goes through the detailed lens descriptions the term "floating elements" crops up. Thats a way to deal with the different needs of extreme close focus and infinity - move some elements around as if it where a zoom lens, but instead of changing the focal length its the aberration control that gets adapted. (Well, IF/RF adds another layer of complexity and there may be partial overlap, but the idea stays). Most of the time you find that in macros.

[LetTheRightLensIn]

No actually the original question was about distance affecting sharpness of a lens...and the answer is quite clear.  Distance does not affect sharpness.  The amount of light affects sharpness.  Quality of optics and alignment affects resolution of a lens.  What the camera can resolve is up to the sensor.  The lens resolution should be relatively fixed..but that is not always the case because no one can make a perfect lens.

nah some lenses go to mush near MFD (say 70-200 f/4 IS at 200mm, it's RADICALLY insanely sharper at infinity than near MFD) and some do better at say 4' than near infinity and some, say some tamron standard zooms do worse at 3' than at infinity for sure.

[Aglet]

A couple posters here have hit the answer to your question.
The answer is YES, lenses may have different abilities to resolve detail (defined as lines on the sensor at a given contrast ratio), depending on the focus distance.

As of yet, I know of no site that actually measures this characteristic of Canon or any other lenses.

I've certainly noticed this with a few lenses, typically zooms, where they will do a better job at close distances than at or near infinity.  I had a 17-40/4L that comes to mind first.
With the minute inconsistencies of manufacturing tolerances, various copies of a given lens model may exhibit more or less of this tendency as well with further variation possible at different focal lengths.
It sometimes seems that more low cost zoom lenses perform this way than others, likely because they're compromised to work better at the closer "people-are-the-subject" distances they're used with when kitted with consumer bodies.
Pro level lenses aren't exempt from this effect either, many workhorse zooms are a bit less sharp near minimum focus distance.

[RAKAMRAK]

Thank you guys for all the replies and discussions.

[Plainsman]

I was wondering whether and how much the distance from the subject affects sharpness of the image taken by different lenses (on same sensor). Is there any website or source where such information may be available for Canon lenses?

The lens designer has to have a reference point when designing his lens eg 50xfocal length.

The big telephoto lenses are really birder lenses and as such have max sharpness at relatively close range ie if you want really sharp shots of the moon you need a reflector telescope.

Read Thom Hogan's Nikon long lens reviews where he points out the lenses which aren't "distance" lenses like the 200-400/4 or even the 400/2.8. The same is very likely true for the Canon equivalents.

[insanitybeard]

It does. Use a lens at 1m, then at infinity - it is a different lens really. I have noticed this with my 17-55 - much better at 2-3m than at infinity.

Many people test lenses at close distances and make far reaching conclusions; and this is wrong.

This is a very good point, and not one I've seen covered in any depth on most review sites I've come across- anybody know any review sites where this is considered with more than a passing comment?

[Pi]

This is a very good point, and not one I've seen covered in any depth on most review sites I've come across- anybody know any review sites where this is considered with more than a passing comment?

I have not seen any. Maybe the reason is that every such site has one (expensive) chart only and this pretty much determines the distance. I believe that you can still test at different distances even with the same chart, in the center at least; but they do not do that.

[East Wind Photography]

That's because the test chart is designed to reveal the limits of the lens at the test distance.  Doesnt matter if you move the chart further away.  Likely the test chart wont show anything.  In order for the distance test to work you would need a chart that is proportionately larger to show the same pattern but larger scale.

For example moving the same chart out to twice the distance will double the number of lines shown per mm...so you are effectively trying to test the lens at an even finer resolution which most sensors would likely not be able to resolve.

This is a very good point, and not one I've seen covered in any depth on most review sites I've come across- anybody know any review sites where this is considered with more than a passing comment?

I have not seen any. Maybe the reason is that every such site has one (expensive) chart only and this pretty much determines the distance. I believe that you can still test at different distances even with the same chart, in the center at least; but they do not do that.

[Pi]

That's because the test chart is designed to reveal the limits of the lens at the test distance.  Doesnt matter if you move the chart further away.  Likely the test chart wont show anything. 

It would but you need software which would still analyze it. The slanted edge test allows you to vary the distance to some degree. An edge is an edge at any distance but you do not want the whole chart to be viewed at an extremely narrow AOV, so there is some practical limit how far you can get. This requires a curious tester, not the typical businesses type.

[East Wind Photography]

And likely a test of the lens OFF of a camera body since the limiting factor of most decent lenses is the chip.

That's because the test chart is designed to reveal the limits of the lens at the test distance.  Doesnt matter if you move the chart further away.  Likely the test chart wont show anything. 

It would but you need software which would still analyze it. The slanted edge test allows you to vary the distance to some degree. An edge is an edge at any distance but you do not want the whole chart to be viewed at an extremely narrow AOV, so there is some practical limit how far you can get. This requires a curious tester, not the typical businesses type.