8-16mm HSM focal range is weird, calibrate?

spiffariffic wrote:

xpatUSA wrote:

Imatest states that the closest focus distance is 24cm / 9.4 in, see:

Indeed, it didn't actually focus, I wasn't actually trying to shoot it, I was trying to see what would finally make the lens go to <0.8m. It definitely wasn't 80% of a meter, was my point.

What focal length? What aperture?

Various. I tried several combinations and little to nothing was ever in focus at all when the distance indicator was set to < 0.8m.

I have taken photos at 8mm @ f/4.5 where I had great DoF and immediately after took another shot 8mm @ f/7.1 with shallow DoF. (See attachments.)

(view full size to see that the second one is blurrier in the distance)

At 8mm f/7.1, hyperfocal distance should be 1.9 feet. That's not how it worked out!

Thanks for the examples.

The f/7.1 shot is blurred from top to bottom, not just in the distance:

7.1 vs 4.5, near:

7.1 vs 4.5, far:

I calculate** the hyperfocal distance at f/4.5 to be 786mm and at f/7.1 to be 487mm - both of which are more the published minimum focusing distance. Therefore, shooting focused at those distances should theoretically get you from "infinity" to half those distances.

Does this seem like something that calibrating the camera can fix?

Failure to focus any of the scene at f/7.1 indicates a faulty lens to me and I'm not certain that playing with the SD1 AF adjustment would somehow fix it.

Exactly how I shot quite a few shots before I realized anything was wrong.

But I was originally trusting the autofocus here.

I rarely use autofocus and never did on the 8-16, so can't help there, sorry.

Here's a street shot at 8mm and f/8 on an SD15:

Focused on the end of the street and sharpened in RawTherapee. I have another taken around the same time with the lens set to the infinity mark. It is softer, which tells you nothing because modern infinity marks are no longer accurate unlike older lenses for film cameras ....

If you got the lens cheap, you could stick to f/4.5 and focus manually ignoring the markings; otherwise send it back if possible.

** http://kronometric.org/phot/iq/DepthOfField-Lyon.pdf

This might be of interest:

http://www.trenholm.org/hmmerk/DOFR.html