[zxaar]

In the viewfinder, it looked as though the people were in focus. I wasn't trying to photograph the flowers. I get shots like this fairly often.

I know what you are talking about. I had k100d and that time i had my old 50mm F1.2 lense. I had this issue very often when things looked focused in viewfinder but finally in photo focus was shifted.

Of course it would be dismissed as technique problem. But read on.

Anyway soon i bought another 50mm F1.2 lense. Nothing changed other than that everything become alright. The things that were in focus in veiwfinder were again in focus.

Anyway i now have k-x and i have never tried old f1.2 lense on it so far so i am not sure if it still exists.

[Sakura]

If diopter adjustments didn't do the trick, I'd be reluctant to pursue any shimming until I was completely sure it wasn't a technique issue. OP needs to do more practice with manually focussing in low-light, and shoot a lot more scenes, to make sure it is not just a technique issue related to poor illumination, dim viewfinders and perhaps tired eyes.

If I was a eye doctor, I'd be reluctant to immediately advocate eye surgery if someone came to me and said my vision is sometimes a bit fuzzy. That's what some of these shimming and other suggestions sound like to me.

Of course, but you would advise without any thought that he should first try to cleen his glasses.
And replacing a shim at a K100D is not more difficult as cleaning you're glasses.

TS is lucky that he has BF problems, that means his shim is to thin, so it is extremely easy te fix. T1 is longer than T2, so the focus screen needs to be closer to the lens. (see the picture and the link i posted before)

I should advice TS the same thing as i had done, make a setting with bright daylight, you're cam on a tripod and a focus chart and make some test shots. (for the K7 i did some tests with the viewfinder and some with lifeview.

Also, make some shots with you're right eye and some with you're left eye, just to extra rule out an eye problem.
Make some 10 or so shots with each eye. If all the shots got more or less the same amount of BF or FF, you can be sure it is the shim and not you're eye.

Start with two 1 mm small strips of postit, put you're cam on its back without a lens mounted, and carefully push the release tab of you're focus screen. Flip the focus screen in the direction of the mirror (if it didn't already is in that position), stick the two post-it strips on right and left edge of the focus screen and carefully push the screen and screen plate back to its position and start the test all over.

Repeat this procedure till you you're focus is right on the spot.

Here a link how to replace a focus screen of the K100D, it shows nicely how easy it is to replace the screen or the shim.
http://www.katzeyeoptics.com/files/install6.pdf

Here a link for the best test chart for you're job.
http://www.k10dbook.com/newchart.pdf

There are a lot of charts on the net, but most of them use a small line to focus, but the thickness of the line itself produces focus problems because you don't know on with side of the line you're focus is.
The above chart doesn't has this problem.

[Marc Sabatella]

The best way i know to get a handle on MF is to try focusing on a sheet of text on angle angle. make an effort to focus on a specific line, but also take note of which lines above and below also appear to be in focus in the viewfinder. Then take the shot (at wide open aperture) and compare. There's a very good chance the line you thought you were focusing on won't be in focus. But there is no camera or lens problems if the lines that *are* in focus are within the range of lines that *appeared* in focus. That is, it's normal for less to be in focus in the shot than appears in focus in the viewfinder. But it is *not* normal for something to be in focus in the shot that was *not* in focus in the viewfinder. Tests like the OP's don't tend to touch on that - he probably wasn't paying attention to whether the background also looked in focus in the viewfinder. But testing with text as I've describe will tell the story well.

[Mikhail_Kriviniouk]

The only *possible* problem would be if the focus screen were not positioned correctly in the camera, but this would have a *tiny* effect on your picture - like moving the focus zone a few millimeters in the posted picture. Whereas you actually missed focus by a lot more than that.

I have to disagree. 2 inches error at a distance to subject of 2ft - that is far from tiny. Clearly visible projection of the subject not in focus on the screen while ok in the image file if the AF confirm was lit. And conversely if the subject is in focus on screen - it will be out of focus in the image file and some subject in front of it - far from "tiny" error, half foot, a foot at a distance of a few yards, just like the situation discussed here - would be in focus. It was impossible for me to get any sharp shots while focusing manually with f numbers under 4.0 with any lens if the AF confirm light did not do the job, and it often fajls due to design limitations as it has been pointed out.

And all it was due to the wrong thickness of the shim under the focusing screen of my K-x. Exactly the same problem resolved - https://www.pentaxforums.com/forums/pentax-dslr-discussion/114013-k-x-focusin...ml#post1174957

[Mikhail_Kriviniouk]

Start with two 1 mm small strips of postit, put you're cam on its back without a lens mounted, and carefully push the release tab of you're focus screen. Flip the focus screen in the direction of the mirror (if it didn't already is in that position), stick the two post-it strips on right and left edge of the focus screen and carefully push the screen and screen plate back to its position and start the test all over.

I would suggest making a thin card shim and stick postits or other tape to it while it is out of the camera, or making a new "shim" out of the initial sheet of paper/card but with tape stuck to it (if you want a thicker shim). And then more tape, then thicker card, and so on, to avoid sticking anything to the actual screen. Or, if you need thicker screen than the one your camera came with - stick the tape to the shim you had installed in your camera originally.

As careful as you may be the technique above is pregnant with the sticky stuff being dropped onto the matte surface of the screen or into the pentamirror well. You don't want to do that ;-)

Shame on Pentax - allowing badly calibrated SLRs out of the factory. And the fit as standard focusing screens so bad that they mask the problem of bad calibration. Leading brand indeed. Well, I suppose there are so few of us users bothered with the accuracy of manual focusing.... And then there are all these compact superzooms with abominable tiny ccd sensors, but which provide almost infinite DOF thus not needing a particularly precise AF system. So people who want the availability of the precision of focus and narrow DOF are by definition a minority - either professional shooters with expensive gear, or annoying weirdos who know a thing or two, like myself and, I am sure, quite a few people on this forum, who are perceived by Pentax aftersales like some kind of perverts. So Pentax, probably, is treating K-x as something which people in the majority would be using the same way as PS cameras, on total auto mode with the kit lens and built in flash, so why bother with the focusing screen calibration? And they, it looks like, don't. Russian Pentax repair shop does not have the shims for K-x in stock, I found out. I don't believe that all K-x sold in Russia are calibrated perfectly, and that only mine was out. It is just that nobody complains, probably, nobody makes demands as it is quite hard to get blood (service) out of the stone (pentar.ru) ;-) You try telling them - and if a human perception is involved (your eye looking at the screen) in the fault you are experiencing - they'll tell you it is a "user technique error" and will do bugger all and that your camera is within the specified tolerances, while, by the way, never specifying what these tolerances they refer to actually are.

Best of luck with the focus screen calibration!

[eccs19]

I know it's not applicable in this case, but if someone was having trouble with front or rear focusing, and they had live view, would the live view be accurate vs using the focusing screen?

[Mikhail_Kriviniouk]

I know it's not applicable in this case, but if someone was having trouble with front or rear focusing, and they had live view, would the live view be accurate vs using the focusing screen?

It depends on the focusing mode you use, if you use phase AF, it will be exactly the same as with normal shooting mode (the camera drops the mirror back, does the focusing the same way as with normal shooting, lifts the mirror back up and shows you the AF area in focus)

If you use contrast focusing - it'll be something completely different, the camera will focus the way PS cameras focus, by achieving the highest contrast in the selected area of the frame. this will not depend on the phase detection AF mechanism or adjustments done to it. Regarding accuracy of contrast-type AF - it just has different underlying principles so comparing it to phase AF would be like asking what's better - apple or a tennis ball? it depend on the subject you are shooting that much I would say....

I tried manual focusing in liveview, even with magnification it is not easy as the LCD on my K-x is only 230000 pixels. It would have been of considerably more use for live view manual focusing if it was 900000 pixels. Waiting for the K-r to hit the shops, if it is as good as K-x at high ISOs, good bye K-x with its outdated LCD and lack of AF assist lamp and the useless SAFOX 8. Anyone wants a K-x with still 1 year warranty on it? I wonder if SAFOX 9 is actually any better - they claim the total revision of AF algorithms. But then with K-x they claim compatibility with manual lenses and sell cameras with focusing screens totally useless for manual focusing...

---

Last edited by Mikhail_Kriviniouk; 09-10-2010 at 05:07 PM.

[Ron Kruger]

I'm not absolutely sure what your problem is, but it sounds technical.
However, I developed a MF technique decades ago that still works for me just as well in this digital age. (Even though all but one of my lenses are AF, I still use MF the majority of the time, because I find it more reliable.)
I believe most look for focus and as soon as they think they find it, they hit the button. I go from one side, all the way through to where focus starts to blurr, then back through again to the other side where focus begins to blurr. I do this rather quickly just to find the edges. Then I slowly go back until the image sort of pops.
This may sound time consuming, but with practice you can complete it and get the shot faster than if AF hunts even a little, and I've found that AF doesn't always lock on the pop. For aunt Betty standing in an open field, AF is fine, but for any cluttered on complicated composure a practiced MF technique is better.
Don't look for clear focus--look for the pop.

[Laurentiu Cristofor]

I focus like Ron described above. I would just add that I'm usually looking for some fine detail to pop, like Ron said, during the procedure. If I don't see detail, then I can't focus accurately. I'll miss focusing on a white wall, for example, if I can't see texture details in the viewfinder. But if I focus on a face and I can make the eyelashes pop, I'll nail focus most of the time.

My impression from the shots I've seen earlier is that this is a technique issue - I cannot imagine a camera so poorly calibrated that you can miss focus by as much as it was missed in those shots.

[Marc Sabatella]

I have to disagree. 2 inches error at a distance to subject of 2ft - that is far from tiny.

I agree that is not tiny, but I also rather doubt it would be possible for a focus screen to be so far out of alignment that it would have that effect. And that was my point.

[RolloR]

guys, is there a generally accepted step-by-step diopter adjustment method?

when i adjust mine, my eyes just "autofocus" to any setting, and everything seems clear, except when it's all the way to the left.

i've had so much success with using LiveView instead, but i don't like it for the batteries.

[labnut]

I know it's not applicable in this case, but if someone was having trouble with front or rear focusing, and they had live view, would the live view be accurate vs using the focusing screen?

Live View together with manual focusing is probably the most accurate way of focusing.
Peter

[labnut]

I go from one side, all the way through to where focus starts to blurr, then back through again to the other side where focus begins to blurr. I do this rather quickly just to find the edges. Then I slowly go back until the image sort of pops.
Don't look for clear focus--look for the pop.

I agree with Ron. It is all to do with the apparent DOF problem that Marc mentioned.
Because of the eye's inability to resolve fine detail in the viewfinder the depth of field appears to be rather greater than it really is, with the possibility of focusing in the wrong place. Ron's technique minimises this problem.
Peter

[Marc Sabatella]

guys, is there a generally accepted step-by-step diopter adjustment method?

when i adjust mine, my eyes just "autofocus" to any setting, and everything seems clear, except when it's all the way to the left.

Sounds like maybe you're looking at the image in the viewfinder. Don't - look at the lines etched into the viewfinder or the info display at the bottom. Easiest while pointing at something like a plain white wall, and deliberately focusing the lens incorrectly.

[stevebrot]

I agree that is not tiny, but I also rather doubt it would be possible for a focus screen to be so far out of alignment that it would have that effect. And that was my point.

That is my thought as well Marc. I suspect that something else is going on here.


Steve