[westmill]

For flash work... it works perfectly only when you set your red eye in motion.
It uses the preflash to work out the correct exposure.
This is the only method I have found to work with the K5.
I use the metz 50. try this... you will see a vast improvement.

[Ray Pulley]

I saw your posts on the other forum, and as someone who has tested several K5 bodies with many lenses, both before and after the firmware upgrade, my first question is what is the true light level in the shots you took with flash?

If the light level in the shots was out of the design limits of the camera (-1ev), and unless the body or flash assist light came on, the camera will not focus properly. However, in my experience there are some situations where it should just give up because it is out of range, but instead it indicates focus lock. The K5 seems also a bit less than logical as to when it decides to use the AF assist.

Based upon my personal tests of 4 bodies (two of which I still own) using many lenses, I have no doubt that the low light FF issue was fixed by firmware 1.03, but that does not rule out other causes, including a defective body that needs adjustment.

You should try and eliminate other possible causes (like the size of the af point and so on) but also try and determine the true light level of the scene without flash to make sure you are not expecting the auto-focus system to work outside of the designed range.

Ray

[RKKS08]

This is rather a question than a suggestion:

With different colour temperatures of the light sources, I always thought there will be some shift in the focus plane. Can this get big enough to create a kind of FF/BF effect when focusing with ambient light, but shooting with flash (and with aperture wide open)? If this is the case, pre-flash from pTTL may or may not help (mixed light may or may not improve the situation).

Just an idea.

[westmill]

Front and back focus issues are genraly nothing more than people not realising the size of the huge focus points.
The central focus point for example stretches out to the semi circle spot metering markings. its not possable to focus accurately on small targets such as eyes.
Otherwise you can miss altogether since it may latch onto the ear or nose... hence front or back focus thoughts.
In most situations flash overpowers all light sources anyway. All I can say is that using the pre flash is the only way I have been able to get good reliable exposures.
Colour temprature should no way effect foccus.

[RKKS08]

@westmill

Colour temprature should no way effect foccus.

Then please can you explain what sense is in the correction marks on the distance scales of most high quality lenses of the film era?

[westmill]

Ive used film cameras for years and mostly medium format and Im not sure what you are talking about to be honnest.
Im pretty certain colour temprature has certainly not caused any back or front focussing issues though.
You have got me racking my brains trying to remember though lol.
The only colour correction mark I seem to remember on a lens was for the use of infra red, and Im not 100% certain about that, but it rings a bell lol.
Other than that, they were colour coded to help you find the hyperfocal distance points.
Ive certainly never seen a lens that you focus at different spot under differing light conditions ?

[twitch]

Let's say you find that your lens AF's accurately in in good light, and consistently FF in poor light. If you can test the amount of FF you can set up a user mode, say "low light" where you set a global AF adjustment for your lenses in that mode, say -5 or whatever adjustment. AF adjustments are one of the settings that can be adjusted between user modes (I think, I haven't tested it though).

I admit it is a bad thing that this would be required, but it is one way to solve the problem.

[Ray Pulley]

Front and back focus issues are genraly nothing more than people not realising the size of the huge focus points.
The central focus point for example stretches out to the semi circle spot metering markings. its not possable to focus accurately on small targets such as eyes.
Otherwise you can miss altogether since it may latch onto the ear or nose... hence front or back focus thoughts.
In most situations flash overpowers all light sources anyway. All I can say is that using the pre flash is the only way I have been able to get good reliable exposures.
Colour temprature should no way effect foccus.

With all due respect, I must take exception to some of what you say above,

The K5 absolutely had a low light/low kelvin AF problem, which was admitted by Pentax and a firmware fix was issued for the problem. I daresay that Pentax knows more about AF than any of us ever will, and based upon my extensive testing with several bodies and many lenses under well-controlled conditions, the problem was fixed on the K5 after this firmware was installed on my two bodies. I shot before and after tests that conclusively prove this, not to mention many real world examples before and after.

As for flash, yes it over-powers ambient light in the foreground in dark scenes, but it has nothing to do with focus. The camera begins focusing as soon as you half-press the shutter release button or press the AF button. It either locks or does not lock, and it does so in the ambient light before the pre-flash goes off and before the shutter is tripped. Unless the flash turns on the AF assist light, it is not involved in focusing in any way.

As for color temp not affecting focus, well this is not the case either. Different wavelengths focus at different planes, and although well-designed optics attempt to correct for this, it is not always possible to completely correct for the problem, especially in relatively simple (cheap?) optics used in AF modules. Pentax designed a color-sensing element into the AF module of the K5 to attempt to correct for this issue, which, in my experience, has been more or less of a problem with every Pentax AF camera i have used.

Looking at the pics posted by the OP, I have a strong suspicion that the ambient light the AF module had to work with was below the lower limit of the PDAF module (-1ev), and if the assist light did not come on, the camera will not focus correctly because it was simply too dark.

There are certainly other possibilities as to the cause of the problem, but if you picture the conditions in most bars at night, it is very likely that it is too dark for proper focusing without the assist light. The OP said he used focus/re-compose (which has it's own problems) so unless he released the AF button between focus and re-compose, had AF-C on, or was using auto-select multi-point AF mode, it is not likely that focusing on a face with just the center point caused a large and consistent front focus due solely to the size of the AF point.

Ray

[RKKS08]

@westmill

Of course I was speaking of the IR mark.
When focusing, you are using the light you can see, or the distance scale which is calibrated for roughly the mean of the light frequencies visible for film or sensor. With the extreme low pass filter for IR, you will get ONLY light for which a shifted focus plane is valid. I just do not know - and this was my question - how much of such an effect is still there with well corrected lenses, and whether this could have any visible impact on Pentax AF.

[rawr]

Full EXIF (not stripped, as it is in the sample images) for one of the OOF shots would be useful diagnostic information, to show:

(a) which AF points were active at the time the shot was taken;
(b) what the per metering segment AE Light Value readings were; and
(c) what the flash setting were

as well as other useful stuff about what the camera was up to. PhotoMe provides a visualization of (a) and (b) too, which is handy.

[sanjeevdas]

Front and back focus issues are genraly nothing more than people not realising the size of the huge focus points.
The central focus point for example stretches out to the semi circle spot metering markings. its not possable to focus accurately on small targets such as eyes.
Otherwise you can miss altogether since it may latch onto the ear or nose... hence front or back focus thoughts.
In most situations flash overpowers all light sources anyway. All I can say is that using the pre flash is the only way I have been able to get good reliable exposures.
Colour temprature should no way effect foccus.

For small targets I would agree. But in a lot of my shots the subject is fairly larger than the semi circle surrounding the focus point.

[sanjeevdas]

I saw your posts on the other forum, and as someone who has tested several K5 bodies with many lenses, both before and after the firmware upgrade, my first question is what is the true light level in the shots you took with flash?

If the light level in the shots was out of the design limits of the camera (-1ev), and unless the body or flash assist light came on, the camera will not focus properly. However, in my experience there are some situations where it should just give up because it is out of range, but instead it indicates focus lock. The K5 seems also a bit less than logical as to when it decides to use the AF assist.

Based upon my personal tests of 4 bodies (two of which I still own) using many lenses, I have no doubt that the low light FF issue was fixed by firmware 1.03, but that does not rule out other causes, including a defective body that needs adjustment.

You should try and eliminate other possible causes (like the size of the af point and so on) but also try and determine the true light level of the scene without flash to make sure you are not expecting the auto-focus system to work outside of the designed range.

Ray

It is true that the levels were fairly low. However, the camera promptly locked focus with no hunting. It also didn't use the AF assist lamp. I don't know what the low levels light limits for K5 are before it will not focus. I was comparing it to my D7000 where I have shot hundreds of similar pictures of people in my backyard during parties with very few OOF. I do think one session is not enough and I need to do some more controlled tests, once I get some time.

[cali92rs]

You may indeed have some focussing problems that should be diagnosed before jumping to any conclusions...but even once you diagnose and solve your issues, don't expect the K-5 to focus as accurately or as quickly as the D7000, even in good light. My D90's AF was better than my K-5.

[Ray Pulley]

It is true that the levels were fairly low. However, the camera promptly locked focus with no hunting. It also didn't use the AF assist lamp. I don't know what the low levels light limits for K5 are before it will not focus. I was comparing it to my D7000 where I have shot hundreds of similar pictures of people in my backyard during parties with very few OOF. I do think one session is not enough and I need to do some more controlled tests, once I get some time.

Hi Sanjeevdas,

As I explained (and I do not know why) the k5 seems very capable of finding something to lock on to even if it is too dark to do so reliably, and in my experience, when it does this, it is always a FF result.

The K5 AF system is specified to focus down to -1EV with the FA 50mm f1.4. Note that the f1.4 part of this equation is critical to understand.

What it means is that if you are using a slower lens, say one that is f4.0 max aperture, the lower light limit for reliable focus is actually 3 stops higher than -1EV. This is true because the AF operation takes place with the lens wide open, and f1.4 lets more light into the AF system than f4.0. This is true of all cameras, BTW.

As for your Nikon, my experience with Nikon (we use them at work) is that it turns on the AF assist light very consistently in light that is actually not very dim (at least the models I have used at work do so). I am not sure why the AF assist on the K5 body does not always come on when it seems like it should, but it often does not. When I have the AF540 flash on the camera, it seems to turn on the AF assist on the flash much more consistently than the one on the body.

The only way to understand what is really going on is to get some data. By that I mean get some exposure data for the dim scenes, note the lens used and max aperture and so on. It is also useful to try and determine if the AF assist came on or not (I have not looked for this in EXIF, so I do not know if the assist status is reported or not?). This exposure data should be without flash as the camera sets certain values when using flash no matter what the non-flash exposure would have been (depending upon the mode and settings, of course).

Here's a flash shot from a reunion in a bar. Note that the ceiling is not white, so there was little bounce, it was full dark, and as can be seen in the background, the lights were tungsten. This was shot with the K5 and the AF540 on camera using a Demb bounce/diffuser, but I do not know if the AF assist came on or not (K5 w/ DA* 16-50 f2.8):



The K5 is capable of properly focused images like this in all of the normal low light and dimly lit scenes I have used it in, but it would not produce such images in the same conditions before firmware 1.03. If you cannot get similarly focused images, something is wrong, and one way or another, it can be corrected.

Ray

[rawr]

By that I mean get some exposure data for the dim scenes, note the lens used and max aperture and so on. It is also useful to try and determine if the AF assist came on or not (I have not looked for this in EXIF, so I do not know if the assist status is reported or not?).

Not only are the AE metering segments reported in the EXIF, the flash metering segments are reported in the EXIF too, along with the LV per segment. (Once again, PhotoMe can do a graphic display of the flash metering as well).

The AF assist light status may be in the EXIF, but I couldn't locate it in EXIFtool's current list of pentax EXIF tags. There is, however, a wealth of other flash related data in there, including flash power etc.