[JohnBee]

Im only using manual lenses. I dont even own one with AF.

Out of curiosity, what is your fastest lens?

[opiedog]

This might be due to the DOF limitations that smeg is talking about, which may well exist in the AF system. It would account for a lot of the AF inaccuracy complaints that I've had over the years, where the AF just doesn't quite hit the nail on the head.
Seriously, learn how to manual focus. People did it for decades prior to AF cameras, and we did it with fast lenses.

it's not about not knowing to manual focus. i've done that since 1979 , but it's the manual AF "assist" mechanism of the pentax DSLR's that seems to be klunky and inaccurate.

i was hoping from the *ist DL to the kx things would improve, but sadly, it did not, at least not in this regard. i don't know how much better the kr/k5 would be, but it seems folks are commenting on the SAME issue.

[aliasant]

Out of curiosity, what is your fastest lens?

Nothing fancy at all Im afraid.
A bunch of 50mm 1.7's

edit: The problems do get worse with wider lenses.
The one I have that was the worst was my 28mm f2.5
From about 1 meters distance the focus would look perfect in the viewfinder but the actual shot would be 10+ cm infront of were I placed that perfect focus. Very odd.

I have a couple of 85mm 1.8 and with those I have no problems what so ever nailing focus wide open.
My 50's all behave slightly different from eachother. Was a while since I did that test but I remember 1 of my 50/1.7 was almost perfect while the others FF'd by varying amounts.

I think it might have something to do with the focal length and the way the sensor sits/hangs.
It is as if the sensor moves towards the back when you take the shot.
Focusing with the LW is never a problem with any of these lenses.

[opiedog]

Nothing fancy at all Im afraid.
A bunch of 50mm 1.7's

edit: The problems do get worse with wider lenses.
The one I have that was the worst was my 28mm f2.5
From about 1 meters distance the focus would look perfect in the viewfinder but the actual shot would be 10+ cm infront of were I placed that perfect focus. Very odd.

i had the same issue. the wider the lens, the worse the problem.

some folks explained that it's the sensor size and that it's quite BIG and the wider the lens, the more possibility that the sensor will pickup and lock on something that you didn't want.

focusing from close focus to infinity will lock at a different object compared to focusing from infinity to close focus.

this probably also relates to why sometimes i look at my lens's distance scale and it will say something like 7 feet when i'm trying to focus on a subject at infinity.

whatever the issue is, the end result is a mis-focused image.

[aliasant]

Hmm. My problems arnt quite gone. Just tried a nice Nikkor-O C 35mm f2.0 and it does the same.

Im about 50cm from a white can with small black text.
Im focusing on that text and it looks perfectly in focus.
The actual shot has focus about 6cm in front of that text.
6cm is a lot!
How the heck am I supposed to compensate for that? I would have to place focus way behind my subject.... Not even sure I could use the focus compensate function for that.

Ill see if I can run a new test thru my most used lenses tomorrow.
This quick test was using tungsten in a pretty dark room.

UDATE:
The same 35mm f2.0 Nikkor works perfect when I leave this dark studyroom and use it in my livingroom instead. Just came back here and took a new testshot of that white can with black text and again its about 6cm FF. All the shots I took in my livingroom were spot on. All shot at f2.0 50 to 100 cm distance.

[aliasant]

As I mentioned earlier in the thread, the AF assist is tied to the AF module. If it is whacky, then the confirmation light is also going to be whacky.
Hence, my advice to trust your eyes.

That is exactly what I am doing. Ill say it again.
It LOOKS like a perfect focus and Im not looking at the green indicator or the red one but the actual shot is way FF'd

I had my LCD screen pretty close to the subject Im testing on so I tried switching the lcd off and that seemed to help a little bit. Only FFing about 2cm now. Switching it back on made it worse again.

[jstevewhite]

That is exactly what I am doing. Ill say it again.
It LOOKS like a perfect focus and Im not looking at the green indicator or the red one but the actual shot is way FF'd

I had my LCD screen pretty close to the subject Im testing on so I tried switching the lcd off and that seemed to help a little bit. Only FFing about 2cm now. Switching it back on made it worse again.

If I understand what you're saying... You're focussing "by eye", right? Not paying any attention to the focus indicators at all, just the image on the screen?

With a matte screen like the one supplied with the K-5, it's *very* difficult to focus accurately with wide aperture lenses. If your focus is accurate in one light level, and inaccurate in another, and all you're using is your eyes, then the light level or color is fooling your eyes. This is exactly the reason split rangefinders and microprism screens exist (e.g. Katz-eye).

My K-5 was front-focusing visually in every lighting situation. I discovered when I replaced the factory screen with a Katz-Eye that the factory shim was way off, and once I swapped it (a .40) out for a .25, everything is good. I still cannot focus accurately on the matte part of the screen in low light, but I *can* using the microprism ring and the split rangefinder.

[aliasant]

If I understand what you're saying... You're focussing "by eye", right? Not paying any attention to the focus indicators at all, just the image on the screen?

With a matte screen like the one supplied with the K-5, it's *very* difficult to focus accurately with wide aperture lenses. If your focus is accurate in one light level, and inaccurate in another, and all you're using is your eyes, then the light level or color is fooling your eyes. This is exactly the reason split rangefinders and microprism screens exist (e.g. Katz-eye).

My K-5 was front-focusing visually in every lighting situation. I discovered when I replaced the factory screen with a Katz-Eye that the factory shim was way off, and once I swapped it (a .40) out for a .25, everything is good. I still cannot focus accurately on the matte part of the screen in low light, but I *can* using the microprism ring and the split rangefinder.

I have a Katzeye that I removed. It does make it a bit simpler but I have the same problem using that one. The focus LOOKS to be perfect with the katzeye ( +optibright) but the final image turns out way FF'd

Just now I fitted my K5 with the x1.2 viewfinder thing from pentax but that doesnt really help much.

Ill have to give this new K5 a couple of days first. The weird FFing is mostly in bad/low light with wide lenses and I mostly use one of my 85's or macro lenses..... but Im thinking about getting a 17mm.... lol That might not work that well...

Thanx Wheatfield but I just got this brand new K5 today and it does the exact same as my old one.
I just hate not understanding what the problem is. Im a programmer so Im used to hunting and finding bugs but this one just eludes me.

Maybe Ill try with the katzeye again...

Edit:
Were do you get new thinner shims??

[jstevewhite]

I have a Katzeye that I removed. It does make it a bit simpler but I have the same problem using that one. The focus LOOKS to be perfect with the katzeye ( +optibright) but the final image turns out way FF'd

Just now I fitted my K5 with the x1.2 viewfinder thing from pentax but that doesnt really help much.

<snip>

Edit:
Were do you get new thinner shims??

DEFINITELY sounds like the screen needs the shims changed - which, according to Rachel Katz, is typical of K-5s. So here's the information she sent me:

To order parts from Pentax, you can simply call them at
(720) 988-2653;

77240.M02200A - 0.10mm
77240.M02200B - 0.15mm
77240.M02200C - 0.20mm
77240.M02200D - 0.25mm
77240.M02200E - 0.30mm
77240.M02200F - 0.35mm
77240.M02200G - 0.40mm
77240.M02200H - 0.45mm
77240.M02200I - 0.50mm

When I ordered mine, it took three minutes to order and I had them in hand in five days; they resolved all my problems. I ended up using a .1 + .15, because they didn't have an .25 when I ordered.

-Steve

[aliasant]

DEFINITELY sounds like the screen needs the shims changed - which, according to Rachel Katz, is typical of K-5s. So here's the information she sent me:

To order parts from Pentax, you can simply call them at
(720) 988-2653;

77240.M02200A - 0.10mm
77240.M02200B - 0.15mm
77240.M02200C - 0.20mm
77240.M02200D - 0.25mm
77240.M02200E - 0.30mm
77240.M02200F - 0.35mm
77240.M02200G - 0.40mm
77240.M02200H - 0.45mm
77240.M02200I - 0.50mm

When I ordered mine, it took three minutes to order and I had them in hand in five days; they resolved all my problems. I ended up using a .1 + .15, because they didn't have an .25 when I ordered.

-Steve

Galactical Thank Yous !

I will call my Pentax dealer tomorrow

[Smeggypants]

Hi Smeggypants,

Did you happen to noticed if/when AF confirm began to jive along the aperture scale? I don't have a 1.2 but I do own two 1.4 lenses and my test confirm that they begin to enter accuracy at/or around 2.8.

However, on an AF lens, I can MF perfectly at 1.4! - And so I'm guessing that the AF system implements some "sort of" focus correction measure to cope with wider apertures. Though this doesn't happen with none Pentax or... lenses that can't be identified by the system.

Would be very cool to get more feedback on this hypothesis based on experiences from others.

I'm not sure how you mean ????

Regardless of what aperture is set to take the photo, the lens is always held at widest aperture while focussing. focus confirm is happening.

Of course if you take a picture with a smaller aperture then the greater DOF will hide any inaccuracies of the focussing. This alos accounts for why a few people claimed their K-5 wasn't front focussing as they were using slow lenses which don't exhibit the razor thin DOF a 1.4 or 1.2 has

And take no notice of people who claim they can focus a 1.2 accurately on a non split screen focus screen. The wider apparent DOF caused by the screen means you can turn the focus barrel on a 1.2 with no noticeable change in the viewfinder. This is why I use LV magnified to focus my 1.2.

One method I used before Live View days was to rock the focus barrel back and forth between the same FF and BF and guess the centre.

[jstevewhite]

And take no notice of people who claim they can focus a 1.2 accurately on a non split screen focus screen. The wider apparent DOF caused by the screen means you can turn the focus barrel on a 1.2 with no noticeable change in the viewfinder. This is why I use LV magnified to focus my 1.2.

One method I used before Live View days was to rock the focus barrel back and forth between the same FF and BF and guess the centre.

I still use that technique when I'm using a fast MF lens, trying to focus on something not in the split rangefinder area.

Since apparent DOF depends on magnification, people who don't pixel-peep routinely believe images are sharp that are not. There is frequently a certain... "sparkle" to a sharp image, though, even at lower magnification, I think.

[Quicksand]

To order parts from Pentax, you can simply call them at
(720) 988-2653;

77240.M02200A - 0.10mm
77240.M02200B - 0.15mm
77240.M02200C - 0.20mm
77240.M02200D - 0.25mm
77240.M02200E - 0.30mm
77240.M02200F - 0.35mm
77240.M02200G - 0.40mm
77240.M02200H - 0.45mm
77240.M02200I - 0.50mm

When I ordered mine, it took three minutes to order and I had them in hand in five days; they resolved all my problems. I ended up using a .1 + .15, because they didn't have an .25 when I ordered.

Oh, they are stackable? That's good to know!

I'd just order a 0.10, 0.15, 0.20, and 0.30 then to cover everything from 0.10 to 0.65!

I'll be ordering a split prism screen shortly, and I want to have my shims ready to go. I don't want to have to repeat the masking tape solution from my K20D -- it works, but it was tedious.

[Smeggypants]

Take even less notice of people who snipe without knowing what they are talking about.
If smegma knew what he was talking about, he would know that the people he was directing that barb at were using split image screens.
It was even mentioned earlier in this thread.

I knew you couldn't ignore me for long. When you said you putting me back on ignore my finely honed Olfactory machinary rang the alarm bells

Anyway I apologise for exposing your complete lack of knowledge regarding focussing screens giving a wider apparent DOF than is reality. I can see it caused you much embarrassment.

It was entertaining reading your claim of a superior skill set regarding focussing though. When in reality it was a split prism screen that was the enabler rather than any skills on your part. Exposed again.

[rhodopsin]

from the horse's other end