K-5 Focus Questions

Well they made the K-5 so that it can remember the settings
for 20 Pentax Lenses
So they must have thought it was needed

'

I think you can adjust any lens, although there's no guarantee that the third party lenses will report their identity correctly.

No doubt somebody has done it and can tell us definitively.

Any lenses which don't report an ID, which includes M42 and M series can be adjusted under "Apply to All" but the setting has to be changed for the next lens used if that doesn't report an ID and needs a different setting.

'

Manufacturing tolerances = need for AF fine adjustment.

JohnX is right - unless we are all prepared to pay massive sums of money for kit that is manufactured to the highest possible standards, we will require some adjustment to allow for both manufacturing tolerances and wear/tear.

I can confirm it works with Pentax and third party lenses, subject to the lens identity being recognised by the camera (Sigma 105 DG EX and Sigma 150-500 are both recognised).

I can also confirm that even using F8, the differences in focus accuracy can clearly be discerned when using a focus accuracy chart and blowing the resulting images up to 200% or 300%. When comparing them side by side (in tiled view) within Photoshop it is a relatively easy task to see which settings produce the most accurate focus. I used F8 because the lenses tend to be reasonably sharp at this setting, which makes the comparison task easier.

There is no need for the chart to be at 45 degrees; this achieves no useful benefit when comparing the images in the way that I did; in fact it makes the tests more tricky to perform. The results when using the lenses for 'real world' shots bear out the fact that the adjustments made have been successful.

Regards
David

Has anyone sent their k-5 to johnsons for calibration???

If so how long?

How much?

What I've never quite been able to wrap my head around is WHY adjustment is needed per lens.

Consider: AF fine adjustment puts in an electronic offset to compensate for a difference between sensor and AF device path lengths. Sensor is "fixed", but both AF sensor and matte screen are at an adjustable length. Shims bring the screen into focus wrt sensor, a mechanical adjustment of the AF unit should zero it in. Any residual error here is dealt with in debug and the user adjustment fine-tunes that farther.

So if lens mount is fixed, sensor is fixed, AF block is fixed, and there's no real positioning feedback from lens to body, what is it that is a variable per lens? Phase shift of some kind?

The main differences are DOF and image quality wide
open for each lens...... basic stuff really.

A sharp lens wide open such as the 50mm A f/2.8 macro
(obviously used in manual focus) will trigger the
green hexagon differently than a poor lens wide open
such as some 50mm f/1.7-f/1.9's.

'

No particular bias against pentax or 3rd party lenses here on either of my K-5s. Lenses from both makes need some adjustment amd lenses from both makes don't need adjustment

Quite possible. The same lenses needs different adjustments on my two K-5s, and it's not a consistent difference between lenses either.

I think Gary is right to a large extent and Smeggy too. The new K-5II bodies are so much more accurate and consistent in their focussing, but I have to say that since calibrating my lenses on the original K-5 I now turn to it with much more confidence than I used to have, especially with lenses that have a shallow DOF.

It also seems more consistent with manual focussing since I adjusted the global AF in the Debug menu, perhaps the green hexagon is more reliably indicating 'in focus' that it used to?

Regards
David

I think I've posted this before but worth another post IMHO:

http://www.dpreview.com/articles/7333489584/variation-facts-and-fallacies

I have no trouble calibrating all the 3rd party AF lenses I've tried, these include Sigma 10-20, Sigma 24/2.8 AF, Tamron 90/2.8, the Tamron being the only one requiring +/- 0 on my K20D

John

That certainly was worth posting again

Following recent posts and focussing issues I've been having with my K-5, I wonder whether anyone knows the answer to the following question:

Would it be technically / theoretically possible for the Pentax engineers to release a firmware update for the K-5 mk1 that would improve the AF performance to something approaching that obtained with a K-5II or K-5IIs?

I'm referring here to focussing accuracy and consistency, not necessarily to improving low light sensitivity. If the answer to the above question is 'yes', does anyone think that there's any possibility of such a firmware update ever being released?

Regards
David

I've been trying to get to grips with calibrating the focus on a DA*50-135mm I picked up off eBay a couple of weeks ago. I've since tried to calibrate all my lenses. What a can of worms it's turned out to be.

I'm really glad to have found this thread and discovered the focus correction feature in the Debug mode. That's been very helpful. Thank you.

What I've uncovered is some fairly dire focus calibration issues with my lenses on the K-5, in particular with the 18-250mm and the 50-135mm zooms, although these are the two I've concentrated on. I wonder if what I'm seeing is normal or if I've got duff lenses/duff camera. I bought the camera new a year ago from SRS so if it's a duff camera I'd like to get it sorted. Here are a few of my findings:

1) The auto focus point varies according to whether focus is approached from the long end or the short end. If focus is approached from the long end the focus point is located considerably behind that of the focus point from the short end. The difference is in the order of 4 steps of AF adjustment. I reckon this roughly translates to a range of about 2/3rds of the standard accepted depth of field when wide open. I get similar results on all my lenses at all distances and focal lengths. If this is normal it begs the question as to why the auto focus algorithm wasn't adapted to add 2 steps when focusing from the long end, subtract two steps when focussing from the short end.

2) The amount of AF correction required varies according to focus distance. For example, my 18-250mm and 50-135mm at a focal length of 80mm behave similarly, requiring about +10 correction at a distance of 1.3m and infinity but +18 correction at a distance of about 15m.

3) On zoom lenses the amount of AF correction required varies according to focal length. I haven't tested this yet but I get the impression the variation may be quite significant on the 18-250mm. Perhaps this is because it isn't anything like parfocal?

Given the above variations there's no way I can see of optimising the AF to cover all situations, especially with the zoom lenses. I hadn't appreciated that poor AF was a price to pay for zoom convenience. If I were to send the camera and lenses for calibration would they be able to make adjustments at different focal lengths and distances, or would they basically just tweak the same AF adjustments as I'm doing?

I've always felt a bit underwhelmed by the sharpness I've been obtaining from the K-5 system, particularly at faster apertures, often thinking it wasn't better than my Canon point and shoot. However, being my first DSLR I didn't have anything to compare it with. What I've discovered in doing this is how good my lenses can be when they are properly focussed. It's therefore very frustrating to find that my camera body doesn't seem to be capable of taking advantage of them. Should I consider moving to the K-5II/IIs. Is focus accuracy really a lot better?

Unfortunately I wear glasses permanently (varifocals with high astigmatic correction) and I am unable to manual focus with any degree of accuracy. The hexagonal focus assist light is of course derived from the phase difference focussing, so it isn't much help if it isn't correct!