Forgot Password
Pentax Camera Forums Home
 

View Poll Results: Do you think your K10D has autofocus issues?
No, and I haven't bothered doing any kind of testing. 2731.03%
No, and I have tested for it in one form or another. 2933.33%
Yes, but I think that just from eyeballing my images. 1213.79%
Yes, and I have tried to do some sort of formal testing. 89.20%
Who cares? My camera takes great pictures as is. 1112.64%
Voters: 87. You may not vote on this poll

Reply
Show Printable Version Search this Thread
05-08-2007, 02:03 PM   #1
Senior Member




Join Date: Apr 2007
Location: SE BC and NE Montana
Photos: Gallery
Posts: 198
Unscientific test of firmware versus autofocus

Sorry about pushing the margins way out there, but I thought the included images should show as much detail as possible for the subject at hand...

Still playing with my new K10D while messing with the manual. With some photos today, I got a strong impression that I may have a back focus situation with - at the least - this lens. Still being new to the camera, and needing to update the firmware anyways, I did a very unscientific and rough test of the firmware version compared to the camera autofocus.

Basically, I set up my camera on a tripod, next to the computer, and wired into the computer via the USB terminal. The focus test sheet was placed down with the camera as close to being on a 45 degree angle as I could get it with the Mk1 ver.2 eyeball (I told you this was unscientific). Selected the center focus point, flash on, lowest f stop I could get at that distance, anti shake off, and 12 second shutter delay. What you see here is the results with the original firmware, and then with the same setup after upgrading to 1.20. I did move the camera a little bit while going through the upgrade button sequences, but very little.

I will leave it to the viewer to decide if there hasn't been some improvement in the autofocus programming at some point between the original firmware and the latest version. With this lens only, of course. I suppose I could have also done it with my other zoom at the same time, but frankly, this was about all I had the time and patience for.

Presuming you could BACKDATE your firmware to the original - or find a willing person with a new in box K10D to play around, you could do this in a more scientific fashion with a variety of Pentax lenses while doing each progressive update to make a more accurate assessment.

As for me, I think I still have a bit of a back focus problem with this lens at least. Whether it will make a difference with real world images remains to be seen. Anyways, the results:

Please click the images to view the full versions.

Original firmware:


After 1.20 firmware upgrade:


05-08-2007, 03:49 PM   #2
Senior Member




Join Date: Jan 2007
Location: Dahlonega, GA
Photos: Gallery
Posts: 141
Interesting test, IF we knew which lens you were talking about! Give us a hint, please! 8o)
Rob W
05-08-2007, 05:23 PM   #3
Veteran Member
*isteve's Avatar

Join Date: Nov 2006
Location: London, England
Posts: 1,187
Whats wrong with it?

If the line is in focus, you cant expect a lot more of the AF system. AF or BF will have the focus line looking blurry.

AF will stop as soon as it gets reasonable sharpness - somewhere within the depth of focus. It cant do much better than that unless you reduce the tolerance, in chich case you add to the time...
05-08-2007, 08:25 PM   #4
Veteran Member




Join Date: Feb 2007
Posts: 1,299
QuoteOriginally posted by Rick Quote
As for me, I think I still have a bit of a back focus problem with this lens at least.
No, you don't. The line, which you focused on, is in sharp focus; so what else do you want?

05-08-2007, 09:18 PM   #5
Senior Member




Join Date: Apr 2007
Location: SE BC and NE Montana
Photos: Gallery
Posts: 198
Original Poster
QuoteOriginally posted by nosnoop Quote
No, you don't. The line, which you focused on, is in sharp focus; so what else do you want?
I don't think I'm going to lose any sleep over it as it is now. I'm guessing at this point any remaining focal issues won't be noticeable in images. The image with the original firmware, however, was something else again.

BTW, I did five shots with the original firmware, five shots after updating. Each shot had the lens refocused so it wasn't using the original focusing again and again. Each series of five shots was indistinguishable from the others in its' set when viewed.

Having said that, the depth of field is helping a lot here - f4.5 is as far as this lens will step down. The entire focusing trial would probably be meaningless at f32 for example, and f1.7 would probably magnify what truly was in focus much more concisely. I suspect that the true point of focus is actually somewhat beyond the "focus here" line.

So what else do I want, you ask? Nothing: I'll probably be more than happy with things as they are now. But...

If I was really picky, I don't think it would be terribly unreasonable to expect the useable depth of field to extend equally as far in front of the point of focus as behind. If image sharpness extends further in one direction than another from the chosen focal point, doesn't that tell us the focus is off? That the sharpest point of focus will actually be somewhere towards the direction where useable depth of field extends the greatest?

It seems to me it does - in my ignorance I would presume that the useable depth of field always extends equally in front of and behind the point of sharpest focus.

Anyways, the trials were more about whether or not the firmware updates made some adjustments to the autofocusing programming. Given that the above two results show significant differences in focusing accuracy, this one off, one example trial suggests there has been some adjustments going on.
05-08-2007, 09:30 PM   #6
Senior Member




Join Date: Apr 2007
Location: SE BC and NE Montana
Photos: Gallery
Posts: 198
Original Poster
QuoteOriginally posted by *isteve Quote
If the line is in focus, you cant expect a lot more of the AF system. AF or BF will have the focus line looking blurry.
As a strictly academic discussion, perhaps the focus line would have been blurry with a lens capable of opening up greater than f4.5... In contrast, with something like f32 for example, you probably couldn't begin to tell whether you had anything other than absolutely accurate autofocus - the depth of field would be so great the focus line would always seem to be in sharp focus.

QuoteQuote:
AF will stop as soon as it gets reasonable sharpness - somewhere within the depth of focus. It cant do much better than that unless you reduce the tolerance, in chich case you add to the time...
I can understand that explanation. So here's another academic thought... would it matter which direction the lens moved towards focus from? If it always moves from infinity inwards, for example, would it mostly achieve "reasonable sharpness" on the further side of the focal point, prior to actually getting sharpest possible focus at the exact chosen spot?

I suppose you could test this by setting this up again, manually focusing beyond the line, and then reverting to autofocus. Then do the same thing, manually focusing before the line, and then once again reverting to autofocus. See if there was a trend to focus a bit on the side the lens approached focus from.

These questions probably aren't particularly germaine to use and enjoyment of the camera, but I do have a curious turn of mind about technical issues in general.
05-08-2007, 11:39 PM   #7
Pentaxian




Join Date: Apr 2007
Location: Sweden
Posts: 1,940
QuoteOriginally posted by Rick Quote
As a strictly academic discussion, perhaps the focus line would have been blurry with a lens capable of opening up greater than f4.5...
Perhaps, or more logically, with the shallower DOF the auto focus would have turned a little bit more until the line was in focus, and not stopped in the same spot so to speak.
05-09-2007, 05:27 AM   #8
Veteran Member
*isteve's Avatar

Join Date: Nov 2006
Location: London, England
Posts: 1,187
You have to remember how AF works

QuoteOriginally posted by Rick Quote
As a strictly academic discussion, perhaps the focus line would have been blurry with a lens capable of opening up greater than f4.5... In contrast, with something like f32 for example, you probably couldn't begin to tell whether you had anything other than absolutely accurate autofocus - the depth of field would be so great the focus line would always seem to be in sharp focus.

I can understand that explanation. So here's another academic thought... would it matter which direction the lens moved towards focus from? If it always moves from infinity inwards, for example, would it mostly achieve "reasonable sharpness" on the further side of the focal point, prior to actually getting sharpest possible focus at the exact chosen spot?

I suppose you could test this by setting this up again, manually focusing beyond the line, and then reverting to autofocus. Then do the same thing, manually focusing before the line, and then once again reverting to autofocus. See if there was a trend to focus a bit on the side the lens approached focus from.

These questions probably aren't particularly germaine to use and enjoyment of the camera, but I do have a curious turn of mind about technical issues in general.
It compares two versions of the same (part of) an image taken through a split mirror. The images align if the subject is in focus, and dont if it is not. The amount of the deviation and the direction of the deviation indicates how far to move the lens and in whcih direction.

Now the AF sensor is not that brainy, nor is it that small. If the lens is wide and the aperture is small, the DOF will be quite big so the deviation - and the lens movement - will be small and may end up anywhere withing the plane of acceptable focus - most likely nearest to the point where it started because that will be in focus sooner. So if you focus on infinity, it will probably focus further away, and if closest it will focus nearer. Use a wider aperture or a longer lens and the deviation will be greater and the DOF smaller, so the AF will be more accurate.

AF tolerance is largely a matter of manufacturing precision and engineering tolerance. Pro cameras are more precise than amatuer ones for instance.

05-09-2007, 07:45 AM   #9
Veteran Member
HogRider's Avatar

Join Date: Sep 2006
Location: Arizona, USA
Photos: Gallery
Posts: 608
Hm - I wish there was a "I don't use Auto Focus" selection in that poll......
05-09-2007, 08:59 AM   #10
Senior Member




Join Date: Apr 2007
Location: SE BC and NE Montana
Photos: Gallery
Posts: 198
Original Poster
QuoteOriginally posted by HogRider Quote
Hm - I wish there was a "I don't use Auto Focus" selection in that poll......
Mea culpa... I don't use manual focus much anymore because, between aging eyes and the impossibility of predicting where running dogs and flying pheasants will move, autofocus generally meets my personal needs much better. I do use manual focus when autofocus can't seem to get 'er done, but I can't see me personally ever going back to manual focus only by choice.

Maybe that's why that option never occurred to me.
05-11-2007, 09:43 PM   #11
Inactive Account




Join Date: Apr 2007
Posts: 133
Mine improved in almost exactly the same way as yours.

It was spot on with the 43mm limited , 40mm limited, but backfocused like yours with the 31 limited, and the 70mm macro(sigma).

It got better with both of the worse ones, at no cost to the spot on ones. The 31 is still not quite perfect though



One thing I need to mention is that it is unfair to use that chart if it is not lined up correctly, as one side will front focus, the other will back focus, etc etc etc.
05-11-2007, 09:47 PM   #12
Inactive Account




Join Date: Apr 2007
Posts: 133
QuoteOriginally posted by *isteve Quote
AF tolerance is largely a matter of manufacturing precision and engineering tolerance. Pro cameras are more precise than amatuer ones for instance.


I would argue this slightly, as there is also the speed/precision compromise in the gearing of any stepper-type motor.

I believe pentax went too far on the speed side of this equation for the k10d and would trust the focusing mechanism of a previous pentax over the k10d.
05-12-2007, 06:32 AM   #13
Veteran Member
*isteve's Avatar

Join Date: Nov 2006
Location: London, England
Posts: 1,187
QuoteOriginally posted by SloPhoto Quote
I would argue this slightly, as there is also the speed/precision compromise in the gearing of any stepper-type motor.

I believe pentax went too far on the speed side of this equation for the k10d and would trust the focusing mechanism of a previous pentax over the k10d.
I cant agree - my K10D is fast AND accurate compared to my *istD and DS even in low light where my *istD would not even bother...

Its true that some lenses are not built well enough to handle the torque of the new motor - specifically the 24-90 sometimes performs a 2 degree image shift when focusing - very disturbing....
05-12-2007, 07:05 PM   #14
Senior Member




Join Date: Apr 2007
Location: SE BC and NE Montana
Photos: Gallery
Posts: 198
Original Poster
QuoteOriginally posted by SloPhoto Quote
Mine improved in almost exactly the same way as yours.

One thing I need to mention is that it is unfair to use that chart if it is not lined up correctly, as one side will front focus, the other will back focus, etc etc etc.
Yup. I just made miniscule adjustments until the resulting image was square - or as close to appearing as square as possible in the resulting image. If you look at them, they're pretty close.

The bottom line is that there was a signifcant improvement in autofocus with that lens after the firmware update.
05-12-2007, 11:24 PM   #15
Inactive Account




Join Date: Apr 2007
Posts: 133
QuoteOriginally posted by *isteve Quote
I cant agree - my K10D is fast AND accurate compared to my *istD and DS even in low light where my *istD would not even bother...

Its true that some lenses are not built well enough to handle the torque of the new motor - specifically the 24-90 sometimes performs a 2 degree image shift when focusing - very disturbing....
yes, that is the difference between focus precision, and the AF system being able to find something to focus on to.

completely different topics.

Torque is a non-issue. Step size and or gearing is the issue.
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!
autofocus, bit, camera, dslr, firmware, focus, images, lens, photography, test, time, upgrade
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
K-x versus K20D ISO test rparmar Pentax DSLR Discussion 68 08-26-2010 02:19 PM
Pentax DA* 55mm 1.4 SDM (unscientific test) asdf Pentax SLR Lens Discussion 7 01-29-2010 02:20 PM
Interesting test - k7 versus k10. ISO800 Dubious Drewski Pentax DSLR Discussion 22 09-02-2009 09:10 PM
my unscientific lazy man's ltd lens test nostatic Pentax SLR Lens Discussion 4 12-30-2008 07:35 PM
Don't update to the 1.01 firmware! It breaks autofocus jms698 Pentax DSLR Discussion 45 10-01-2008 01:22 PM



All times are GMT -7. The time now is 10:15 AM. | 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