Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
01-28-2009, 04:12 AM   #1
Forum Member




Join Date: Jan 2009
Location: Finland
Posts: 60
AF adjustment - am I doing it right?

Hi all, again

I was a bit worried about my DA21 performance, so I decided to try the AF adjustment. I put the camera on a tripod,and shot test shots - target was a newspaper ad some 1,5 meters away, lens fully open. There was a huge difference between 0 ja +1, but growing the adjustment did not add the sharpnesss. The difference was as big from really close range (appr. 40 cm) @ f5.6. Just to be sure, I made another set of test shot, this time again fully open but target being (a house) appr. 400 meter away. However, I was not able to detect any difference in sharpness in the range 0 - +5, not even in a 300 % crop. Could have thought that difference in the close range sharpness would be visible in the long range, too. Ok, I left the adjustment as +1, but begin to wonder if I'm doing this right...

Regards,

Jari

01-28-2009, 11:51 AM   #2
Pentaxian
Marc Sabatella's Avatar

Join Date: Sep 2006
Location: Denver, CO
Photos: Gallery
Posts: 10,686
Is there a reason you didn't use an actual focus test chart like the ones you usually see posted in threads about AF adjustment? See focustestchart.com for one I like (the site was down recently but appears to be back up now).
01-28-2009, 12:02 PM   #3
Forum Member




Join Date: Jan 2009
Location: Finland
Posts: 60
Original Poster
No reason. I did the test for DA40 with MTF test chart, but I sincerely don't know does it really matter - as long as you can see the difference. But, I'm still wondering shouldn't it be visible in long range shots, too.
01-28-2009, 12:07 PM   #4
Veteran Member
PentaxPoke's Avatar

Join Date: Jan 2008
Location: Oklahoma
Posts: 1,411
You are doing it the right way. The primary uncertainty in the angled chart methods is that the center AF point is not an infinitessimally small dot. This means that when you focus on the line, AF might lock when the line is at the bottom of the sensor region, or at the top. I have seen this myself on my k10d and k20d. When the chart is angled, this causes the focus points to be at two different distances which can give false readings, especially for lens conditions with very small DOF. This is why manufacturers use methods where the focus object is at 90deg.

The best way to fix BF or FF on the k20d is to not use the angled charts, but simply set up your camera, and use text on a wall. Adjust the lens correction + or - until it is perfect. Exactly what you are doing. Do the test with the aperture wide open, and preferably not in tungsten light. Make sure the camera is at 90deg to the wall. Get reasonably close, but not so close that your are anywhere near the min focusing distance. The shorter the distance, the more sensitive your adjustments will appear to be.

01-28-2009, 02:17 PM   #5
Pentaxian
Marc Sabatella's Avatar

Join Date: Sep 2006
Location: Denver, CO
Photos: Gallery
Posts: 10,686
QuoteOriginally posted by jesalonen Quote
No reason. I did the test for DA40 with MTF test chart, but I sincerely don't know does it really matter - as long as you can see the difference. But, I'm still wondering shouldn't it be visible in long range shots, too.
The difference, unless I'm missing something about how you are doing the tests, is that a focus test chart shows you exactly where the focus is, with no guesswork or multiple trials needed. You can see if it's in front or in back, and by how much, and you can monitor how that changes with each change to the AF adjustment.


Tough to reproduce that particular test at infinity, but I'd still be looking to simulate it somehow - a basketball placed on a well-trimmed lawn, for instance and shot from a bit of a height to also allow you to see exactly where the the focus zone is.

Bottom line: it's not enough to just check to see if your target is in focus or not. If it isn't, you need to be able to see where the focus zone actually is. Otherwise, you're relying too much on guesswork.
01-28-2009, 02:35 PM   #6
Forum Member




Join Date: Jan 2009
Location: Finland
Posts: 60
Original Poster
Thanks Marc, now I see what you mean. Excellent link, by the way. When I did the test, I took first a serie of pics at -10, -5, 0, +5 and +10. Based on that, I was able to tell that 0 and +5 had a significant difference, but +5 and +10 not. So, I took another serie at +4, +3, +2 and +1 and started to compare 0 and +1, and there was a clear difference. Between +1 and +2 I didn't notice any, hence I decided to adjust by +1. I printed the chart from the site you suggested and will try it tomorrow (it's nearly midnight here, and I prefer natural light...)
01-28-2009, 02:42 PM   #7
Site Supporter
smigol's Avatar

Join Date: Sep 2008
Location: Menlo Park, CA
Photos: Albums
Posts: 394
QuoteOriginally posted by jesalonen Quote
(it's nearly midnight here, and I prefer natural light...)
Important to note is that the AF sensor will perform differently under light that has a strong amount of Infrared. This means that tungsten lighting will give a different set of results than sunlight.
01-28-2009, 08:35 PM   #8
Pentaxian
Marc Sabatella's Avatar

Join Date: Sep 2006
Location: Denver, CO
Photos: Gallery
Posts: 10,686
QuoteOriginally posted by PentaxPoke Quote
The primary uncertainty in the angled chart methods is that the center AF point is not an infinitessimally small dot. This means that when you focus on the line, AF might lock when the line is at the bottom of the sensor region, or at the top.
I am not sure what you mean. If you center the line horizontally across the sensor, it's going to focus on the line, period. The only issue is the *thickness* of the line, but that's what, a millimeter? Very few lenses have a DOF so shallow that this is is going to be an issue (although to be sure it can happen at macro distances). Usually we're looking at a zone of acceptable focus that is probably like 100 times larger than the thickness of that line.

FWIW, angled charts in which the target is the border between solid black and white blocks rather than a black line of non-zero width against a white background are also available - I just haven't liked other aspects of those charts that I've seen, so I keep coming back to Jeff's.

QuoteQuote:
The best way to fix BF or FF on the k20d is to not use the angled charts, but simply set up your camera, and use text on a wall. Adjust the lens correction + or - until it is perfect.
That works too, but it has issues too. For one thing, requiring more trial and error to figure out out what's going on. But also, if there is any variation - say a lens that needs different focus adjustments at different distances, or a zoom that needs different adjustments at different focal lengths - you'll drive yourself crazy trying to figure out what is going on. The angle method lays it out on the line for you. But once you've identified the nature of the problem that way, I'd agree the type of method you're describing could work well for actually doing the calibration, because it's not like you can actually *solve* either of the problems I mentioned.

01-28-2009, 10:10 PM   #9
Veteran Member
PentaxPoke's Avatar

Join Date: Jan 2008
Location: Oklahoma
Posts: 1,411
QuoteOriginally posted by Marc Sabatella Quote
If you center the line horizontally across the sensor, it's going to focus on the line, period. The only issue is the *thickness* of the line, but that's what, a millimeter? Very few lenses have a DOF so shallow that this is is going to be an issue (although to be sure it can happen at macro distances).
I agree this issue can be small for many lenses that have a large enough DOF to not notice. My 50 f/1.4 however, is not such a lens, and you can see how this uncertainty affects the result. To get the most accurate result, one wants to test wide open (that is the aperture the AF uses, no matter what the exposure setting) and close. The total DOF for the 50 f/1.4 under these conditions is only 4mm, which translates to 2.8mm on the chart because of the 45deg angle. There is also a relatively wide window over which this lens will focus on the line. The thickness of the line in that chart is actually 2mm which by itself is 70% of the total "chart" DOF! A camera angle error of less than 0.25 degree more than makes up the total DOF for this lens!

Bottom line is that the uncertainty for this lens, is on the order of the total DOF. It is easy to be fooled. Running a test where the subject is 90deg to the camera axis really improved the results and I could finally get a consistent reading because you don't have the extreme sensitivity to angle. It is a lot easier to control distance from the lens variations, than it is to control angle variations.

Last edited by PentaxPoke; 01-28-2009 at 10:36 PM.
01-29-2009, 12:38 AM   #10
Veteran Member
blende8's Avatar

Join Date: Feb 2007
Location: Bremen, Germany
Photos: Albums
Posts: 1,484
I agree 100% with PentaxPoke.
I did a lot of testing with the AF adjustment and finally settled on the target-on-wall method.
Those 45 degrees targets are good for checking roughly for strong back- and front focus. But for the accurate calibration it does not work.
That's my experience.
  • Use a target (newspaper, USAF chart) at about 50 * focal length.
  • Do four shots at 0, four at -4 and four at +4 (the AF has slight variances and not all shots are alike.)
  • Compare the sharpness of the shots: If the Zeros are better than both the +4 and -4 leave it at that. If not, check further ...
  • If you have found the best value, test it again at infinity (steeple or antennas) if it is ok there, too.

I experimented three days on this.
In the end I concluded that all of my 13 lenses are on zero, except one.
The FA50 needed a -6.
01-29-2009, 03:10 AM   #11
Forum Member




Join Date: Jan 2009
Location: Finland
Posts: 60
Original Poster
Yes. I have to say, that I found that angled chart much harder to use, so I'll keep testing with a local supermarket ad...
01-29-2009, 06:33 AM   #12
Inactive Account




Join Date: Jan 2007
Location: Long Island, NY
Photos: Gallery
Posts: 150
Wieland:

Not sure I understand this:

QuoteQuote:
Use a target (newspaper, USAF chart) at about 50 * focal length.
So for an FA50 that's 50 * 50 or 2,500. 2,500 what? Can't be inches as that's 208 feet! Is it milimeters?
01-29-2009, 07:52 AM   #13
Forum Member




Join Date: Jan 2009
Location: Finland
Posts: 60
Original Poster
I suppose that focal lenght is millimeters...
01-29-2009, 09:09 AM   #14
Pentaxian
Marc Sabatella's Avatar

Join Date: Sep 2006
Location: Denver, CO
Photos: Gallery
Posts: 10,686
Well, if it works for you, I guess that's what matters. Requires a whole heck of lot more patience than I have. I'm not interested in dealing with AF errors on the order of a millimeter or two. Not that I have any choice at the moment - I'm using a K200D which doens't have the AF adjustment feature. But I do test my lenses and bodies to be sure they don't have any serious issues. Most of the real world situations where I am dealing with DOF that shallow, I'm using MF, and in other real situations, an error of 1mm just isn't doing to be noticed, because as has been observed, the AF sensor itself is so big that you can't really guarantee focus with that kind of accuracy.
01-31-2009, 03:38 AM   #15
Site Supporter
GeneV's Avatar

Join Date: Oct 2008
Location: Albuquerque NM
Photos: Albums
Posts: 9,772
QuoteOriginally posted by PentaxPoke Quote
I agree this issue can be small for many lenses that have a large enough DOF to not notice. My 50 f/1.4 however, is not such a lens, and you can see how this uncertainty affects the result. To get the most accurate result, one wants to test wide open (that is the aperture the AF uses, no matter what the exposure setting) and close. The total DOF for the 50 f/1.4 under these conditions is only 4mm, which translates to 2.8mm on the chart because of the 45deg angle. There is also a relatively wide window over which this lens will focus on the line. The thickness of the line in that chart is actually 2mm which by itself is 70% of the total "chart" DOF! A camera angle error of less than 0.25 degree more than makes up the total DOF for this lens!

Bottom line is that the uncertainty for this lens, is on the order of the total DOF. It is easy to be fooled. Running a test where the subject is 90deg to the camera axis really improved the results and I could finally get a consistent reading because you don't have the extreme sensitivity to angle. It is a lot easier to control distance from the lens variations, than it is to control angle variations.

That makes perfect sense to me. I think the 45 degree method is good for finding out whether you have BF or FF issues, but I noticed the same thing about the focus on a particular line. Three or four exposures can give me three or four slightly different results with an FA77 at 1.8 as the camera locks on planes that are ever so slightly different each time
Reply

Bookmarks
  • Submit Thread to Facebook Facebook
  • Submit Thread to Twitter Twitter
  • Submit Thread to Digg Digg
Tags - Make this thread easier to find by adding keywords to it!
adjustment, af, af adjustment, camera, close, dslr, photography, range, sharpness, shot, target, test
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
Af adjustment Alex00 Pentax SLR Lens Discussion 8 04-20-2010 03:36 PM
Landscape A re-do / WB adjustment for Jimbo Post Your Photos! 14 03-03-2010 07:23 AM
AF Adjustment skid2964 Pentax DSLR Discussion 14 10-22-2009 03:51 PM
AF Adjustment cupic Pentax DSLR Discussion 8 08-12-2009 09:30 AM
No AF- adjustment in K7 ? lock2nl Pentax News and Rumors 6 05-20-2009 04:03 AM



All times are GMT -7. The time now is 05:29 PM. | See also: NikonForums.com, CanonForums.com part of our network of photo forums!
  • Red (Default)
  • Green
  • Gray
  • Dark
  • Dark Yellow
  • Dark Blue
  • Old Red
  • Old Green
  • Old Gray
  • Dial-Up Style
Hello! It's great to see you back on the forum! Have you considered joining the community?
register
Creating a FREE ACCOUNT takes under a minute, removes ads, and lets you post! [Dismiss]
Top