PL4 requires manual focus distance entry while PL3 didn't

Hi all,

I am puzzled by the fact that PL4 now requires to manually enter the focus distance for my Sony 24-105 shots while PL3 didn’t. I suspect a bug, too, because the PL4 optical corrections are far off from the Lightroom ones (regardless of the focus distance) while the PL3 corrections are identical to those of Lightroom.

Would be nice if the team could look into this…
Thanks
Julien

Hi Julien. I’m not experiencing the behavior that you describe, but I can tell you that PL both v3 and v4 require manual entry of the focusing distance when it is not clear from the EXIF what that distance is. Perhaps the EXIF has somehow become corrupted? DXO does not use any info stored in the EXIF to provide lens corrections however, so the lens corrections provided by DXO camera/lens modules will not be the same as in Lightroom and are often much better.

I have the Sony 24-105 lens too - and I can confirm that PL4 does NOT require me to manually enter a focus distance. As Mark suggests, you may have an issue with the EXIF data stored within the ARW file … Is it possible you edited/changed the EXIF data with some other tool ?

Also suggest you click on the DxO Optics Module menu item, and check that a module is correctly installed for this lens.

John M

Thanks for the feedback,

That’s weird. So I also tried with an A7III+24-105 raw file and on my system (16 inch macbook pro, latest catalina) it showed the same behaviour, opposite to what is reported above.

This seems to point towards a problem in my own install. So I did the usual “user problem” stuff i.e. removing/resinstalling the lens module, removing/reinstalling PL4, logging in under a fresh user account. None of these did solve the problem.

Below are screenshots showing the problem with the distance scale on the exact same raw, where 3.3.4 interprets the distance encoded in the exif while 4.1.1 does not. Also, once the correct distance (from the exif) is manually entered into 4.1.1 the correction is still different from that applied by 3.3.4.

Doing a little bit more testing, it is difficult to assess which correction is the most accurate. Sometimes the 4.1.1 correction is closer to Lightroom, sometimes it is the 3.3.4. I know LR can’t be taken as a reference point for optical corrections but that’s what I have right now. The problem is, DxO corrections can’t be taken as a reference either if they change from version to version.

I also believe that early versions of PL4 (before 4.1) did not show the problem, so it seems to have been introduced in a subversion of PL4. But I can’t confirm that as I have not found a way to access to earlier versions of the software on DxO’s website.

3.3.4

4.1.1

Mark, But the info must be there for a reason, otherwise it shouldn’t be there?

I am using M43 cameras and lenses as well as some manual lenses. And I am also confused. Some lenses show 128m, some other nothing. And unless there is a value, I can not change.

My operating system is MacOS, and I am wondering if this is another bug DxO PhotoLab have in the MacOS version that does not exist in the Windows version?

Hi Lars. I wish I could help you but I must admit that when I responded to this thread I didn’t notice that it was in the Photolab Mac section. I think that all of what I wrote in my earlier post is applicable to Mac as well as Win or I would remove it, as I know nothing about the Mac version of PL. Maybe someone who is familiar with the Mac version will come along to clarify. It may very well be a bug in the Mac version. Perhaps you should submit a ticket to DXO support to let them know.

I’ll be very glad to submit a ticket but have never done this before - TBH I thought that starting a thread here was a bug report in itself. Is there a specific procedure?

Thanks
J

Hi Julien. Just go to https://support.dxo.com/hc/en-us/requests/new
and follow the instructions. Good luck!

Thanks. I guess the devs also scan this forum as there is a “bug” tag that can be attached to threads.

Yes, the staff participate in this forum but they don’t work on holidays or weekends. You have to get their attention by “pinging” them, which I’m not going to do since it’s an holiday. :innocent: Sometimes their advice is “Create a support ticket” anyway so you’ve already gotten the ball rolling, just don’t expect any response until Monday. :grin:

The help text is not clear on this topic, but … I’m pretty sure that focus distance is required only when an Optics Module does NOT exist for the {body+lens} combo … in which case, it’s used to assist with distortion corrections, etc.

Regards, John M

Not really - without an optics module the software does not require focus distance, as there is nothing it can do with this info. It is when an optics module is available, and when the EXIF does not report the focus distance, that a manual entering is required to fine-tune the optical corrections.

The Sony 24-105 f4 gives an accurate reading of the focus distance in the EXIF, so the software should use it. This is what PL 3.3.4 does (focus distance is greyed out as it should) but PL 4.1.1 doesn’t, and produces corrections that appear a bit random. So this definitely looks like a bug in PL4 Mac (the windows version appears OK, see above).

I’m now really intrigued, Julien !

I’ve never paid any attention to this correction before, as I didn’t think it had any relevance when an Optics Module is available - but now I’m not so sure (?)

But, I do now understand what’s meant by this;


I have 4 different camera bodies, but for only one of them (my excellent little Sony RX100 iv) does the Focusing Distance correction appear (in the Geometry group) … I’m on Win10 - whereas, I understand this always appears on the Mac version … which is probably why you’re noticing this, whereas I’m not.

The fact that it does NOT appear for some bodies (which, presumably, write to the EXIF data the focus info needed by PL) suggests to me that DxO does not consider it necessary for us to fine-tune this info for these cases.

@sgospodarenko/Svetlana: Could you please clarify this for us - That is; in what circumstances, if any, do we need to be concerned about providing Focusing Distance details ?

Also, note Julien’s observation here;

John M

The snippet about the focus distance sliders that you posted makes me even more confused! I have a hard time to figure out why the Mac and Win versions should differ on this point - if the info is there then the software should use it regardless of the platform?

Agreed - but the “interesting” part is that if it’s found in the EXIF data then it’s not displayed at all (in the Win version) … which implies that DxO does not consider it necessary for us to fine-tune this info (?).

Let’s wait for Svetlana to elaborate on this for us.

Meanwhile: I experimented with some images from my Sony RX100 iv - for which the Focus Distance correction is revealed (rather than being hidden - which it is for all my other camera bodies). I manually changed the focus distance - which resulted in different distortion correction results (from the default of assumed infinity).

John M

Hello everybody,

I wish you a happy New Year :slight_smile:

Let me clarify a bit this question of focus.

  • The Sony FE 24-105mm F4: in early December we modified DxO Optics Modules for this lens because some configuration of Focal/Focus were not well corrected in distortion. Before the modification the focusing distance was considered as non inlfuent on the distortion correction (focus min to max would end up with same default for the same focal). The correction showed it was wrong.

  • Focusing distance in Sony cameras: for a long time we didn’t get the focusing distance in Exif on Sony cameras. Now we do but Exif reading Sony A7 III has not been modified yet so Optics modules will still consider we don’t read it. It will be modified later.

  • Focus slider: There is an ambiguous focusing distance when slider “focusing distance” is available in pannel “Geometry”. This slider is present or absent for all images made with same couple camera + lens.
    If there focusing distance isn’t declared as ambiguous in module then slider won’t be present for Windows and will be grayed out for Mac.


    so when focus slider is present and usable it means that the focusing disance you set with it well have an impact on distortion correction (sometimes it wont as maybe focus has influence for some focal but not all). By default focus distance is set to max value, as it is the most common one.

  • Focus values: inifinty value we can have in Exif has evolved through years that’s why you can have 40 or 128 for some lenses we did support some years ago.

I hope it’s clearer now.

Regards,
Marie

4 Likes

Thanks a lot for the clarification Marie and happy new year,

Still I believe there is a problem. Following your algorithm chart for the Sony 24-105, the slider should be grayed out (Mac) or absent (PC) as this lens writes a ‘Focus Distance 2’ entry in the exif which looks very accurate to me. The camera I use is an A7RII (not an A7III).

Thanks,
julien

I have a Panasonic Lumix GX9 and with v4 (on a Macbook running Catalina) I see “Focusing Distance” which I didn’t notice with v3 (but maybe it was there and I just didn’t see it?). Regardless of lens and actual distance it shows me a range of more than 40.00m, the distance sometimes being 60m, sometimes 128m, even if the focus point is 2 meters away from my camera. That looks quite mysterious to me. If I correct the distance by hand I don’t really see a difference in the distortion correction. I should add that all relevant optics modules are loaded.
Using exiftool I don’t see an item “focusing distance” in the output.

Thank you for the explanation, Marie - - that’s much clearer now.

Was this announced anywhere - perhaps in update release notes ? … It’s useful to have such info, so that we can re-process images taken with this lens.

When it is, can this please be announced - so that we can go back and re-process images from this lens.

Yes, I can confirm this with images taken with my little Sony RX100 (iv) … In this case, the Focus Distance is made available for manual entry - and I can see this reflected in the degree of distortion correction that’s applied for different manually entered values.

For my Sony a7 III, I identified FocusDistance2 as the likely EXIF value for Focus Distance (as, I see, Julien did too). It does reflect some “strange” values, tho … for subjects quite close to the lens it seems to be accurate … but, for subjects far away it shows values such as “222”, “248” etc !!

Regards, John M

Yes, I guessed that FocusDistance2 might be subject distance too, Julien (as revealed using EXIFtooGUI) - but it’s showing some strange values for me … See my comment to Marie, above.

It won’t necessarily be labelled that clearly, Rainer - - eg;

  • It’s clear for Olympus: image

  • But I’m just guessing for Sony: image … 248 meters, perhaps ?

John