[Smeggypants]

How about forgetting about trying to find fault with your camera, and just go take some pictures?

How about reading the members original post before making those kind of comments?

[Smeggypants]

Unfortunately, the faults found me. I've already previously posted a couple of shots with horrendous focus:

https://www.pentaxforums.com/forums/pentax-k-5-forum/133556-what-%25%24-*-no-1-a.html

I have many more such examples, taken with various lenses. Af adjustment and/or FW 1.03 has helped a lot, but neither perfectly nor consistently.

Hi there bxf

I have found that V1.03 has pretty much fixed the FF problems, BUT .. it has introduced an inconsistent AF accuracy. FW before 1,03 ALWAYS front focussed in low light. with 1.03 it focusses in low light most of the time, but occasionally does not. I would say 7 times out of 10 it is accurate.

[bxf]

For those with af problems i have a really good test chart...


Beacuse of the fact that the issue occurs also on reddish objects, a red testchart reveals af issues...


so maybe some of you can post their results here using the added charts...

RED CHART

WHITE CHART


For quick tests you can use your screen! For me it worked the same way as the printed version... but maybe it depends on the type of your computer screen....

Thanks, these could be useful. My problem is that I get different results with different tests

[bxf]

I think something we are just going to have to live with regarding the K5 is some AF inconsistency in low light/ some colours of light. It looks like they managed to get something not quite right in the AF, and this is being magnified tremendously by the low light performance of the imaging sensor allowing people to work in light levels that were unphotographable previously.

I would debate that EV6 would be a problem, and would wonder exactly what the original file looks like. The performance of the sensor is such that an underexposure of several stops can be salvaged in post, and this can make the EV level look a lot higher than it actually is.
It is not really possible to accurately measure EV with a reflected light meter, one needs an incident meter that is colour calibrated to do it properly.

Please see samples in subsequent post. Note that the quoted LV are as shown by PhotoMe.

[bxf]

As for definitive conclusions: no. My impression is that some still suffer from low (tungsten) light AF problems which FW 1.03 mitigated; otoh some reported no problems even before the update. An explanation for this would be that the problem occurs with some copies of lenses with some copies of bodies as is the case with FF/BF in general. My guess for a root cause would be chromatic aberrations produced by the combination of the lens optics and those in the AF sensor. Pentax lenses could be better matched than 3rd party ones in this respect, that is, with them there is a fair change that such matching has been a design consideration.

Each of my lenses appear to require a different level of AF adjustment. Daylight tests show the following:

Pentax 12-24 +8 at all FL (whew!)
Pentax 18-55WR +5 wide, -9 mid & tele
Pentax 50-200 +9
Sigma 50mm f1.4 -4 (though -9 with a different test setup using laptop screen)
Tamron 28-75 f2.8 -5 (though -9 with a different test setup using laptop screen

The Sigma produces horrendous fringing at wide apertures - purple in one shot and green in the next, making it difficult to assess the quality of the AF at times.

As you say, the Pentax lenses do exhibit somewhat more consistent behaviour.

[bxf]

My first K-5 ahd severe FF problems in low light. V1.03 did NOT fix it.

I got the camera replaced. The replacemetn camera exhibited the same FF problems in low light but V1.03 DID fix it. Two days after that I bought a 2nd K-5 and it's low Light FF problems were fixed with 1.03.

The original K-5 was built in December 2010, the replacement and second K-5 were built in Jan 2011 and Feb 2011 respectively.

I've been following your saga quite attentively all along, Smeggy, and I'm happy for you that your replacements perform better than your first body. The inconsistency, even in light levels that are not that low, is bothersome, however. Let's see what comments I get on the images I'm about to post - perhaps I'm being overly critical and expecting too much from some of my lenses.

Mt K5 was built in December, a day after the date I'd seen as being the cutoff date for the stain problem

[bxf]

How about reading the members original post before making those kind of comments?

Thanks

Hi there bxf

I have found that V1.03 has pretty much fixed the FF problems, BUT .. it has introduced an inconsistent AF accuracy. FW before 1,03 ALWAYS front focussed in low light. with 1.03 it focusses in low light most of the time, but occasionally does not. I would say 7 times out of 10 it is accurate.

Yes, it appears to be inconsistent. I'm curious as to what comments I'll get on the images I'm about to post.

[bxf]

Please judge for yourself.

I usually prefer to use my 28-75mm Tamron for this type of shooting, but here I switched to the 18-55WR, just to see if the focusing is any better. AF adjustment is set to -7, though my tests with this lens indicate that -9 is optimal at 35 & 55mm, and around +5 at 18mm. I'd say that this lens is focusing better than the Tamron (which was terrible before I did some AF adjustment and which prompted my looking into this AF problem), but the wider FL and smaller aperture may be responsible for the apparent improvement. Flash was (mostly) bounced off the corner of the ceiling behind me. Please do not comment on the lack of artistic merit here - these are just snapshots, and no effort at all has gone into making these examples of good portraiture. ISO is 1600, SR ON.

The following are JPEGs saved directly from the DNGs (using Faststone Image Viewer 4.4), with no processing at all. Light was a combination of mostly halogen with a bit of tungsten. Light level value in each file name is the Effective LV from PhotoMe, though the individual AE matrix segments are shown as always being lower than the Effective LV (I don't understand this). Metering is CW.

I would have supposed that some of the softness is due to lens limitations (and perhaps even some camera or subject movement?), but nevertheless, it seems to me that the focus is neither accurate not consistent, but perhaps I'm expecting to much from some of my lenses. Your comments about whether or not the lens is soft, is in proper focus, and if not, where it's actually focusing (i.e. FF or BF) would be appreciated. I do realize that, in the absence of any foreground, not all the shots here make it possible to determine the actual point of focus, though I selected some photos specifically because they DO show some foregroung and background.

Once again, thanks for any input.

Edit: I thought the file names would show up and provide appropriate info. Here it is (left to right, top to bottom):

55m F5.6 1/90 LV7.8
23mm F4 1/60 LV5.9
23mm F4 1/60 LV5.9
35mm F4.5 1/60 LV6
37mm F4.5 1/60 LV6

[bxf]

A few more...

55m F5.6 1-90 LV7.8
55m F5.6 1-90 LV7.8
55m F5.6 1-90 LV7.8

[pentmax]

I see slight mis-focus in most of this pictures...

This one´s very hard FF: look at the bedclothes...




Best,
martin

[pentmax]

Thats also what i figured out with my K-5... till i sent it back

I sold all my Pentax stuff, switched to nikon d700 because photography became more commercial and professional to me.... spent also a lot of money on pro lenses and since then i have no AF problems at all...it was a great relief

But for those who can´t afford a more expensive camera, a fully functioning K-5 would be great!

And i´m still interested in this problem... so for me it´s annoying to hear that there is still a problem

[bxf]

I see slight mis-focus in most of this pictures...

This one´s very hard FF: look at the bedclothes...




Best,
martin

That's how I see them as well - nothing really sharp where it should be. And let's not forget that the DOF is not all that shallow here. It does look as if using an AF adjustment of -9 (instead of -7) could be sufficient in most of these, I think. No big deal, except for the fact that I get different results under different lighting, at least with some lenses. Does this even make sense? Should AF consistency be lens dependent?

Thanks for your comment.

[Christine Tham]

Have you considered that perhaps there is something wrong with either your camera or lens? Maybe worth getting Pentax to check out.

My K-5 requires minimal or 0 AF fine adjustment for most of my lenses.

And for reference, a friend of mine (non photographer) took this picture - he wanted to try out my camera, and effectively just pressed the shutter button in Green mode.

FA43Ltd: shooting conditions are f1.9 t=1/80 ISO=800 (so EV is approximately 5) in a restaurant with halogen lighting (JPG, no additional processing):


As you can see, her fingers and chin are out of focus, as well as her hair and jacket, but her eyes are focused, so I would say the AF is pretty spot on in that photo.

[Class A]

There has been far too much rumour spreading about, and makes people think there has to be something wrong.

I feel you are adding insult to injury for those who are affected. The problem has been documented numerous times. Are you saying that falconeye's thorough experiments fall into the the camp of "spreading rumour"?

Even improved after the last firmware update.

Why would a firmware update address a "rumour"?

Pentax lenses could be better matched than 3rd party ones in this respect, that is, with them there is a fair change that such matching has been a design consideration.

I think this is a pretty wild speculation. It would mean that Pentax cameras discriminate third-party lenses (I know that in-camera lens corrections only work for Pentax lenses but this is a different matter). This would be a serious detractor for buying Pentax cameras, AFAIC. Pentax needs the third-party lens manufacturers to offer their customers a reasonable set of lenses to choose from.

Your hypothesis would imply that Pentax can control the chromatic abberrations for each lens copy they produce to a very high level of precision (otherwise an in-camera compensation doesn't make any sense). I think we can dismiss the hypothesis on the basis of this observation. Even "pro" lenses like the DA* 16-50 have too much sample variation to make such an approach even a remote possibility, AFAIC.

As you say, the Pentax lenses do exhibit somewhat more consistent behaviour.

I don't think the difference you are seeing is caused by Pentax vs non-Pentax lenses. The lens that gives you more inconsistent results is the only really fast lens you have used. I believe that this causes the difference. If you tested an FA 50/1.4 or a DA* 55/1.4, I'm sure you'd see very similar focusing behaviour. Forum user robgo2 tried five (5) DA* 55/1.4 and couldn't find one that didn't focus erratically on his K-7.

[Christine Tham]

Here's another example, shot by my friend.

This time, I adjusted the settings to TAv mode f=5.6 t=1/60 (so ISO was calculated at max of 3200 with the camera slightly underexposing)

Again he shot without explicitly focusing or composing, he just pressed the shutter. Mind you, with the above settings, it would have been impossible for the camera to misfocus - I'm just saying that with typical shooting conditions (rather than wide open) there should be no reason why the camera won't do an adequate job so I'm struggling to understand why you are getting the pictures you are posting: