Focusing issues on Pentax cameras

I use 3-5 metres away for lenses up to say 100mm , a bit further away for longer lenses. The important thing is to make sure that the AF centre point (which is all you should use) covers the target on your test chart with no ambiguity.

With zooms I test at the longest and the shortest. In my experience if you get the adjustment right for one position the lens/camera combination is correct at all FL. There is the risk that you have a "wobbly" element in your zoom but you should identify that during testing.

I use the test chart at the end of this article http://www.kscameraclub.org/docs/pdfs/focus_test_chart_edited.pdf

One thing to remember is that a single test is not enough. You should do about 10 tests for each lens, and redo them for each adjustment. Each time you test you should be making the lens autofocus from either infinity or closest focus distance....so put your hand in front of the lens so it misfocusses or manually turn the focus ring. You have to make the AF system work hard to get best results.

If you have not done this testing before you will be surprised at how innacurate the AF system can appear to be. You would expect perhaps that each time you get the camera to AF that it would always pick the same focus. Well have a look on your lens distance scale after the camera has focussed each time. Out of 10 attempts the majority should be showing the same position, but there will be a few that don't. This is why it is important not to rely on a single test but to base yoour adjustments on a "best out of 10"

What you focus on should be parallel to the back of your camera. This cheap chart from Amazon looks good at £4.20 for two.... LINK

--

IMHO:
- the test distance should reflect, as far as possible, those at which you generally use the lenses - and, of course, the min test distance will likely vary according to the actual min focus distances of the various lenses.
- re the zoom lenses, not so sure about that, but probably you should do tests at the min, mid-point, and max zoom settings, and see what you find, and then decide where to optimise the adjustment for each lens according to how you are likely to use it.

If you have several bodies, then it may help to repeat the tests with each one to see if a pattern develops, as that could help to identify if any issues are lens or body specific.

BTW: there is a lot of info in various threads on the US forum - which is how I worked out how best to do it myself.

You've revealed a central issue with the whole process of AF. If you were to routinely use say a 55-300mm lens at 200mm and distances of about 30 feet, then it would make great sense to carry out the calibration at those settings, but of course we use these lenses at all focal lengths and many different distances. We only have one AF setting, so I would test at various points in the range and then apply an average correction. It could take some time.....

If AF works the same as the BEEP does in MF then it only works for the distance you test it at That's because it works similar to DOF and BEEPS as soon as it enters the nearest or farthest point. Not of much use to a photographer

--

PS: did anyone mention that you need to do the tests with the camera on a sturdy tripod with a head that can be firmly locked in place - otherwise you will never keep the single AF point on the target!

And, Yes, John R is right - it takes quite a long time, especially if you are testing several lenses/bodies = keep a morning or afternoon free and don't hurry!

I suppose technically you're right, but isn't how I do it! I normally hand hold the camera so if I detect I'm regularly getting a back/front focus issue due to my technique, I simply adjust the lens setting accordingly until the lens works right for me to eliminate the focusing issue. One gets a feeling for it taking real life photos which is what causes one to suspect it isn't right and needs adjustment.

Its always possible to revert it to zero and start over again if need be! It is possible to over complicate things.

I've tried that and quickly gave up because whilst it might work with short focal length lenses with a reasonable depth of field wide-open, it's a no-no for long lenses being tested wide-open (which is how it should be done to minimise the depth of field) because you will never keep the desired calibration "ruler" marking in focus before you release the shutter, let alone when you do; and so you can never be certain whether an apparent back/front focus issue is due to the lens, the body or YOU

The tripod makes things much easier, quicker and less frustrating - albeit that it takes a few extra minutes to set the whole test process up.

You can't be sure that the RED X is exactly over the detection point, or that it doesn't lock onto something nearby, that's why you need to focus on something parallel to the camera back. It's also easier with a tripod with a focusing rack.

Most of my tests are MF based on the BEEP. There's usually a beep moving in towards the target and another beep moving backwards into the target with about 25mm between the two. You have to decide which one to use. I normally set it so that I catch the beep moving into the target. +/- 7mm is usually possible.

--

That is a criticism of autofocus systems in general, not of the AF fine adjustment procedure.

The most accurate way to focus is to use live view. Either using the Contrast-Detect AF system, or by manually focussing and using the screen zoomed at 100%. This is because you are then detecting the focus point directly on the cameras sensor.

When using the viewfinder you are relying on the camera's Phase Detection AF system which has its own sensor. If the AF sensor and digital sensor are not in perfect alignment you can get back/front focus issues. Hence the AF Fine adjustment menu.

If your camera/lens combo is front/back focussing then using the AF Fine adjustment will correct this. It wont be as accurate as Live View, but it will be a heck of a lot better than not making the adjustment.

In my experience I have found that setting the Fine adjustment while focussing from about 3-5 metres ( on lenses below 100mm) gives me accurate AF at any focus distance. It is lens and camera dependent. So on my K1 most lenses required +7 to +9 to correct backfocus. One only needed +3 and one required +10

Important point to note here is that John's method has a logic because he is basing it on taking many many photos "in the wild". I would not recommend hand holding for actual AF testing which needs a more scientific approach.

No it isn't It can be quite good if you focus on something that causes moire on the screen, but it still isn't accurate. See....
http://www.slrgear.com/articles/focus/focus.htm

Summary at end...
"While the overall results were much better than with the other focusing methods though, the worst-case example was still 22% high, well within the "unacceptable" range. Given that, manual focusing via even a highly-magnified Live View display simply isn't reliable enough to use for lens testing."

And he used a powerful magnifying glass as well

--

Trying to set a focus on the tiny screen is tricky, especially if one's sight isn't what it used to be. If you zoom in to 100% you can no longer see the whole image you're trying to frame so that's tricky too.

Just as an aside, don't try and set the focus adjust using Contrast Detect in liveview as it's using a different focussing method to what is causing the TTL focus problem, so won't solve that. It has to be set using the TTL Phase Difference method.

Another massive problem with LV focussing is that the lens isn't wide open it gets stopped down so the image has a big DOF

Anyone who ever did printing with an enlarger knows that you have to open the lens fully to get exact focus even when using a grain magnifier.

--

Good point, and so I suggest the following:
Put the body in Aperture Priority Mode and minimise the max ISO setting that it can choose, open the lens aperture as wide as possible (either on the lens where possible, or from the body where not) and increase the shutter speed until you get under-exposed shots, then slightly decrease the shutter speed to get around about the correct exposure. That should force the lens aperture to stay wide-open.