[roguephysicist]

I'm not entirely sure that is correct. Exiftool has the following metadata fields which reference focuspoints - the formatting here is rather ugly but you can look up the Pentax makernotes on the exiftool website. Is that in any way helpful or am I merely misinterpreting what you are trying to do?

Those are the some of the many "names" that the different camera models use for the focus points. The actual position of each point on the image file (in pixels) is definitely not specified in the metadata. Hence, I'm asking for people to map their camera AF points by providing me with the pixel values that represent each location.

Other manufacturers do provide these positions in the metadata, so this mapping is not necessary for Canon, Olympus, etc.

[newmikey]

Those are the some of the many "names" that the different camera models use for the focus points. The actual position of each point on the image file (in pixels) is definitely not specified in the metadata. Hence, I'm asking for people to map their camera AF points by providing me with the pixel values that represent each location.

Other manufacturers do provide these positions in the metadata, so this mapping is not necessary for Canon, Olympus, etc.

Aha, now I get the point - I wasn't aware of that. But isn't the number and naming of the focuspoints (11 for "older" Pentax AF systems, 27 for newer) somehow mappable to the focuspoint layout in the viewfinder? With other words, wouldn't it be possible to take a physical layout of the focuspoints in the viewfinder and determine what the tag-names Top-left, Top Near-left, Top, Top Near-right etc. would correlate to?

I'm struggling to understand why the number of focuspoints in the makernote tags correlate so strongly to the SAFOX 11 or 27 focuspoints if they weren't directly related. But I'm not technical or sftware-savy enough to challenge anyone on that besides my simple logic which may turn out terribly faulty - apologies for that.

[beholder3]

I'll call the rear button "AF2" for the purpose of this discussion. It turns out that if you press-and-hold AF2 while taking the picture, it works just like using the regular shutter-button activation and records both AFPointSelected (when available) and AFPointsInFocus will contain the correct information. However, this is not the case if you press AF2 to focus and then release it before pressing the shutter button to take the photo. Any subsequent photos taken without holding down the button will obviously also register "None" for AFPointsInFocus.

Which I would consider a behavior which makes a lot of sense. In that case you did configure the camera to have its autofocusing system passive unless you press the backside "AF" button. You briefly pressed it earlier, but in the moment of shutter release and image file storage you did not, so the AF system was not active then and probably there was no AF sensor readout in that moment just because of that. Seems logical that nothing is stored then, because exif could only store the last/historical data, not any current.

Imagine you would use that exact same usage pattern to frame a portrait and focus on the eye in the center of the frame and afterwards pan the camera to the left and release the shutter. Basically you are asking the metadata to record that the center AF point had something in focus where the image might show you that at the point in time of capture the AF point is on the far away background which definitely is not in focus.
Anyone being unhappy about this way of filling metadata would be very right in my book.

---------- Post added 11th Apr 2017 at 23:49 ----------

But isn't the number and naming of the focuspoints (11 for "older" Pentax AF systems, 27 for newer) somehow mappable to the focuspoint layout in the viewfinder?

For single points selected the numbering scheme for all Pentax DSLRs is the same and extremely easy. Whatever number or position of the AF points, they are numbered just as people read: from top to bottom, from left to right. Row by row and column by column. So number "1" is always the uppermost, leftmost AF point and the bottommost, rightmost AF point has the number which is the total of AF points available (typically 11, 27 or 33).

[roguephysicist]

Aha, now I get the point - I wasn't aware of that. But isn't the number and naming of the focuspoints (11 for "older" Pentax AF systems, 27 for newer) somehow mappable to the focuspoint layout in the viewfinder? With other words, wouldn't it be possible to take a physical layout of the focuspoints in the viewfinder and determine what the tag-names Top-left, Top Near-left, Top, Top Near-right etc. would correlate to?

I'm struggling to understand why the number of focuspoints in the makernote tags correlate so strongly to the SAFOX 11 or 27 focuspoints if they weren't directly related. But I'm not technical or sftware-savy enough to challenge anyone on that besides my simple logic which may turn out terribly faulty - apologies for that.

Sure, there is a direct correlation between the naming and the positions.

However, there's another important issue at play here: not all Pentax cameras spit out the same names in the metadata to indicate which points are selected/in-focus. For instance, the K-50 records names like "Center," "Upper-left," etc., but the K-1 stores numbers 1-33 to indicate each point. Obviously, this needs to be accounted for since each camera has its own metadata quirks and I have found that the documentation is often inconsistent. This is exactly I was asking for the RAW files -- to figure out what each camera records in the metadata over a variety of situations.

I'm struggling to understand why the number of focuspoints in the makernote tags correlate so strongly to the SAFOX 11 or 27 focuspoints if they weren't directly related. But I'm not technical or sftware-savy enough to challenge anyone on that besides my simple logic which may turn out terribly faulty - apologies for that.

This actually brings up an interesting point. Are there really only three Pentax autofocus systems with 11, 27, and 33 points? I already have the mapping for the 11 point system that I made for the K-50. I'll need to scale it to accommodate the pixel dimensions of the different cameras (the K-50 is 16 MP), but that's trivial. We also already have the 33 point map for the K-1. That means that we're only missing the map for the 27 point system.

---------- Post added 04-12-17 at 12:59 AM ----------

Which I would consider a behavior which makes a lot of sense. In that case you did configure the camera to have its autofocusing system passive unless you press the backside "AF" button. You briefly pressed it earlier, but in the moment of shutter release and image file storage you did not, so the AF system was not active then and probably there was no AF sensor readout in that moment just because of that. Seems logical that nothing is stored then, because exif could only store the last/historical data, not any current.

Imagine you would use that exact same usage pattern to frame a portrait and focus on the eye in the center of the frame and afterwards pan the camera to the left and release the shutter. Basically you are asking the metadata to record that the center AF point had something in focus where the image might show you that at the point in time of capture the AF point is on the far away background which definitely is not in focus.
Anyone being unhappy about this way of filling metadata would be very right in my book.

In total agreement! A subtle point indeed. Unfortunately for our purpose, I do think that it is quite common to press AF2 once to achieve focus then take a series of photos without necessarily refocusing.

Anyway, this is just one a few specific scenarios that would have limited success with the plugin. Thanks a lot for clarifying this behavior.

[beholder3]

That means that we're only missing the map for the 27 point system.

I have once written/adjusted a Photoshop script that automatically displays selected AF points of the K-3 as an extra layer upon image loading.
I put in the AF point coordinates as relative percentages of total image width and breadth from top to bottom and left to right (to allow for lower JPG resolutions)
This is the data (obviously the 0/0 is dummy data).
I bet you can understand it.

k3x = new Array(0.000, 0.352, 0.426, 0.500, 0.574, 0.648, 0.352, 0.426, 0.500, 0.574, 0.648, 0.203, 0.352, 0.426, 0.500, 0.574, 0.648, 0.797, 0.352, 0.426, 0.500, 0.574, 0.648, 0.352, 0.426, 0.500, 0.574, 0.648)
k3y = new Array(0.000, 0.338, 0.338, 0.338, 0.338, 0.338, 0.419, 0.419, 0.419, 0.419, 0.419, 0.500, 0.500, 0.500, 0.500, 0.500, 0.500, 0.500, 0.581, 0.581, 0.581, 0.581, 0.581, 0.662, 0.662, 0.662, 0.662, 0.662)

[roguephysicist]

I have once written/adjusted a Photoshop script that automatically displays selected AF points of the K-3 as an extra layer upon image loading.
I put in the AF point coordinates as relative percentages of total image width and breadth from top to bottom and left to right (to allow for lower JPG resolutions)
This is the data (obviously the 0/0 is dummy data).
I bet you can understand it.

Awesome, thank you! I will create these maps in a few minutes, I am currently working on the K-1.

Related question: do you think that the 11 and 27 point systems are identical across the relevant models? I have just created maps for every 11 point camera assuming that this is true.

[roguephysicist]

Thanks to the productive discussion we've been having, I've updated the plugin to now include support for most models. You can see the full list here.

I'm fairly certain that all 11 point models will work correctly with all features enabled. I have done basic tests with the K-3, K-3 II, and K-1 and can confirm that they work as expected, but I know that they have a "Zone Select" mode that I have not encountered before. I would appreciate anyone willing to help me out with this.

Overall, what's needed now is continued testing of the plugin for every model.

[blackest]

Thanks to the productive discussion we've been having, I've updated the plugin to now include support for most models. You can see the full list here.

I'm fairly certain that all 11 point models will work correctly with all features enabled. I have done basic tests with the K-3, K-3 II, and K-1 and can confirm that they work as expected, but I know that they have a "Zone Select" mode that I have not encountered before. I would appreciate anyone willing to help me out with this.

Overall, what's needed now is continued testing of the plugin for every model.

I was just trying it on my k5 and it does something odd it correctly ids the focus point but puts all the focal points in a corner of the photograph.

Is that how it is intended to work?

[Mikesul]

His linked page is dead so maybe it is no longer available. The k-1 shows the focus point in the second information screen.

[blackest]

His linked page is dead so maybe it is no longer available. The k-1 shows the focus point in the second information screen.

I downloaded and installed from the git project running the plugin does id the focus point but it brings up a window with the photo but the focus points are shown as a group in the corner rather than sitting over the focused area on the photo. With the k5 its not too hard to figure where it is on the photo since the focus points are based on the rule of thirds grid one on each intersection one halfway horizontally or vertical one bang centre and 2 half way in the outer thirds on the centre line.

Would be nice if it put the focus points over the points in focus.

[rawr]

Just pinging this thread in the hope that the Lightroom plugin project is still alive.

From the discussion so far it appears that the plugin author did indeed have the K-1 AF points mapped, but it's not clear if that mapping found it's way into a workable plugin.

[texandrews]

GitHub - musselwhizzle/Focus-Points: Plugin for Lightroom to show which focus point was active in the camera when a photo was taken

On GitHub, this fellow has created a Lightroom plugin that allows you "to show which focus point was active when the picture was taken."

With this tool, you can test your lenses for actual dof characteristics, both using a setup with measures or just in field work.

Already quite a few Pentax cameras are supported (click on the link in the short list to see the whole one...). There are instructions as well on how to add your camera to the list if it isn't there or isn't yet "working".


Because it's GitHub, it's free, but also he is calling for any and all contributions. So, Pentaxians: if you have some level of expertise here, please contribute! I may have a try with the 645Z, but first I'll see if someone far more adept than me can do it.

[stevebrot]

This is definitely old news here, but a cool project, none-the-less, at least for people that need to know that sort of thing. The author is a forum member.


Steve

[Sandy Hancock]

I have merged @texandrews recent post with this worthwhile old thread to keep things together.

[texandrews]

I have merged @texandrews recent post with this worthwhile old thread to keep things together.

Thank you, Sandy.