Actually that's the problem. To do a good test, you need the following:
First, you need to ask the uncle to give you a tripod to put the camera on.
Second, you need a high-res body (like a 1Ds Mark II)-- I hope you happen to bring it along.
Third, you need a notebook or PC to view the image at full res or even 200%-- LCD display not good enough.
Fourth-- hopefully you were able to shoot a good test target (eg USAF 1951 chart or newspaper), otherwise evaluating sharpness is more difficult.
If you just anyhow use a 300D, shoot your friend's face handheld, if the picture is not sharp, how would you know if it's the lens, the camera or your lousy handholding?
To answer your questions specifically:
1. Do macro test?
* Only if your lens is a macro lens
2. Check at which aperture?
* That's the problem. All lenses improve when you stop down. If you test at f8, the lens will be sharper than at f2. But you may not be able to shoot at f8 in the shop without a lot of noise introduced in the image.
3. What subject to take for the test? Newpapers?
* I guess any target with straight lines will be good.
4. Must be tripod mounted else test is void?
* Of course. Else who know if your qi gong good enough or not. And the target must also be still. That's the tough one.
5. Test at minimum focal or max focal length? Or Macro?
* That's the other problem. All zoom lenses have an optimal focal length, and it's usually not the longest or shortest, so if you test at those focal lengths, you may find it's not as sharp as somewhere in the middle.
6. If you can only do 1 picture test, how would you take the picture?
* I would not take just one picture. It would be a waste of time.
* I would try the following:
* Choose a target at the appropriate distance for lens (eg if it's a portrait lens, maybe choose 2 m)
* Measure distance from sensor to target.
* Turn off AF.
* Set distance on lens.
* Shoot at various focal lengths.
* Shoot at various apertures
* Repeat above, adjust the distance setting slightly over and slightly under the measured distance.
* Check sharpness of pictures.
Note: The above procedure is basically to check if your lens is calibrated correctly. That is, if the real distance (measured with a ruler, etc) is 2 meters and the lens is set to 2 m, is the picture sharp? Or must the lens be set to 2.1 or 1.9 for the picture to be sharp?
This is because the capability of a lens is already determined when it was designed. In other words, an L lens is sharper than a non-L lens simply because it is designed to be sharper (and more expensive). So there is no point measuring sharpness per se.
What is important is whether the lens is correctly calibrated. If someone dropped the lens, for example, the calibration could be off, so that the focused distance may differ from the true distance to the target. This will result in unsharp pictures.
That's what I'm trying to measure here.
Wai Leong
===
