K5 tungsten vs daylight AF calibration?? GordonBGood, awaldram, RPulley?

Started Nov 11, 2012 | Questions thread
GordonBGood Veteran Member • Posts: 6,306
Re: K5 tungsten vs daylight AF calibration?? GordonBGood, awaldram, RPulley?

Brian Thomas wrote:

I'm trying to calibrate a new k5 before a big trip to India and I'm no engineer. I photograph a lot under low light. Basically, I first ordered when they were announced a K5iis and got an incorrectly assembled model (only me!). I saw that the K5 had plunged in price and some people here had gotten it to focus correctly so I thought I'd take a chance with it.

(By the way, k5 users, did you notice that the k5 has one cool feature that was removed from the k5II series? -- Image comparison, a great display mode that lets you compare any 2 images on your card side by side magnify sections of each. This is great for those long train or bus rides.)

I have upgraded the k5 firmware. I was using the chart at 45 degree angle as recommended by GordonBGood and downloadable from here: http://photo.net/learn/focustest/

but had a hard time discerning the result with that. I may return to it. I'm now using moire live view method versus phase detect method described here:


Gordon states that he has gotten his low light tungsten AF within a few millimeters of his low light daylight focus.

.."firmware at versions 1.13 have a very slight shift to front focus between daylight and tungsten illumination of about three mm. at f/2.8 and minimum focus distances of about 40 cm. This is little enough that one can bias the micro focus adjustment just enough to the rear that both daylight focus and tungsten focus are within the Depth Of Field (DOF) for f/2.8, with the daylight DOF slightly biased toward the back and the tungsten DOF slightly biased toward the front."

I am doing this test at the photo.net article recommended 25 times the focal length or, for the DA 40 about 39 inches.

I have some maddeningly specific questions, especially for GordonBGood & awaldram & RPulley,

Brian, as others have stated, one needs to first determine whether your camera actually needs calibrating or not. For this, awaldram's shooting of the Pentax manual straight on at "low light" is fine: test if you get adequate sharpness shooting the cover of the manual straight on from a tripod or rest as compared to locking focus and shifting the camera slightly forward or backward before the shot, using flash with two second delay to avoid any effects of Shake Reduction (SR), mirror slap (the mirror retracts before the shot), or hand induced errors. When I did this for both my wife's and my K-5 using our range of lenses, I found that it seems a few microadjustment units in error, so did do the adjustment for the different lenses.

1. Can you quantify the low levels of light, the ev level, that I should be able to get to work with a sample iso, f-stop & shutter speed?

The EV value that qualifies for low light is dependent on the wide open aperture for the particular lens used. The Pentax specification for the K-5 is -1 EV using a lens of f/1.4 so with my P-DA 40 Limited f/2.8 the low light limit would be at EV 1 and I did my testing for this lens at about EV 2 to 3, which means f/2.8 at ISO 100 and an exposure of a half to one second, ISO 1600 at 1/32 to 1/16 second, etcetera. For my DA 55-300 zoom at the long end of f/5.8, the shutter speeds would be just over four times slower. However, these would have been the exposure settings not using flash, I disabled the focus assist lamp and used flash to freeze the shots better for this testing.

2. Can you quantify the micro bias adjustment between tungsten & daylight in Pentax micro adjustment numbers, rather than millimeters? i.e. The Pentax k5 adjustment goes between -10 to +10. My first tests show the tungsten test at quite bright levels of light calibrated at "0" and the low light daylight test at "-7" How far a distance can be safely split as Gordon recommends?

For our cameras, the difference between using tungsten bulb illumination and screened outside daylight illumination as only about two micro adjustment units so I just split the difference. However, other cameras such as yours may have a greater difference and you may need to make a decision about where you need the sharpest focus, or make a note of how the focus is in error and make a compensation before shooting. A difference of seven adjustment units when spit would mean that focus wouldn't be absolutely sharp at either end of the range of illumination when shooting at close distances or large zooms where the Depth Of Field (DOF) is the most shallow, as you can confirm for yourself.

3. If a lens that you have been successfully using with kx requires a bias of -10 or +10 to be close to right, would you send the camera back and get the k5II series?

In my experience, all lenses will have a slight error and as long as I was able to micro adjust them satisfactorily, I would accept them. Although the K-5 II is more sensitive, there is no guarantee that it is any more accurate and is likely to still need a micro adjustment. Our lenses, which consist of a DA 15-55 II kit lens, DA 55-55 WR kit lens, DA 18-135 WR, two DA 55-300's, a P-DA 40 Limited, a DA 70 Limited, and a Tamron 17-50 f/2.8, required a range of between about plus and minus 5 micro adjustment units to be calibrated, and talking to TheCameraStore in Calgary where we buy much of our Pentax gear, the sales/technical expert said that this is a normal range and within manufacturing specifications.

***Note, the reason we have some duplicate ranges of lenses is that we aren't often in the same country at the same time.

I have been doing the moire test AF on my laptop and then putting the laptop to sleep and then focusing on my dark tungsten lit dirty laptop screen. Amazingly it locks focus. Nonetheless, it is extremely front focused. The final question is..

The above test isn't really a good Auto Focus (AF) test as the camera is likely focussing on the dust on the surface of the screen rather than embedded lit LCD shutters as per the northlight images test.

4. If I slip a piece of paper on top of the screen to make the tungsten light focus easier, is the width of that sheet of paper so insignificant that the test is still valid? (f2.8 at 25x the focal length)

The width of the paper isn't really the problem but rather that this will still be testing at a different focus surfaces between the top of the screen and the embedded LCD shutters.

To confirm a shift between the different lighting, I would do the basic micro focus adjustments using the northlight procedure, since you are comfortable with that, then confirm those results under daylight illumination using the paper target and see if there is a shift when testing using the tungsten illumination with everything else remaining the same.

Now, it is likely that the K-5 II models have less shift of focus due to changes in illumination colour temperature, so if you confirm that your current K-5 does have this problem, it may be worth upgrading to one of the newer models.

EDIT_ADD:  Note that when using an angled test with the paper target as linked, one must be careful to print that target on smooth surface photographic paper or risk the AF system locking focus on the texture of the paper, which can make the results invalid depending on the AF algorithm used.  You should confirm the camera's inability to lock AF on a plain white area of the paper just above or below the black focus line for the very lowest and brightest illuminations used.

Hope this helps.

Regards, GordonBGood

selected answer This post was selected as the answer by the original poster.
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow