Originally posted by MadMathMind Each camera is different.
Originally posted by jatrax Every camera is different.
What they said.
With your K-3 the task is easier since you can confirm with focus peaking in magnified live view. The trick is to separate random error from focus bias. I generally recommend that users do a simple evaluation before concluding that a focus adjustment is needed.* Using a
flat** target at 20x the focal length distance and with the aperture set to maximum*** :
- AF turned off
- Manually de-focus the lens to the minimum focus distance
- AF turned on
- Acquire focus using the PDAF (non-live view) system
- Turn AF off
- Switch to Live View (focus peaking turned on)
- Press the OK button to magnify to 100%
- Evaluate whether the image is in focus and whether you adjust closer or farther to correct
- Record "0" for no change, "-1" for closer, or "+1" for farther
- Do the above steps five times
Repeat the above except with the lens de-focused to infinity. What the above flow does is provide a simple three-state statistical sampling of the AV results. A preponderance (4 or more) for any of the three possible results for the total of 1o focus attempts indicates whether there is a need to attempt calibration and whether you need to add or subtract. If you need stronger proof, add 10 more attempts for a total of 20. Note that for lenses having moving elements (zooms and/or internal focus), the results are specific to the focal length and/or target distance.
Should you determine that front/back focus exists, I would suggest making corrections in single increments. Checking the results using the above flow between increments. I know, it is tedious, but at least there will be a minimum of rework.
Steve
* My conclusion after reading literally hundreds of AF adjustment threads is that many (most?) users are chasing shadows and making things worse rather than better. As Jatrax so nicely put it,
Originally posted by jatrax Often the margin of error in these tests, if not done precisely, is larger than the original error. This just leaves with a randomly adjusted lens.
As for myself, I have a single lens correction stored for my K-3 for my Sigma 17-70 (C), one which I intend to get rid of once I calibrate that lens using the Sigma dock.
** While the various 45-degree angle charts are quite popular and also very useful, it is difficult to use them to assess precision at a level smaller than the smallest increment on the chart scale. The fine adjustment on the camera uses smaller increments.
*** There is anecdotal evidence that the K-3 uses a less rigorous AF algorithm when the camera is set to narrower taking apertures.
Last edited by stevebrot; 06-21-2015 at 11:12 AM.