Regression on a wrongly filled optical module

Hi

Here a photo taken with 5D + 35mm f/2 IS.
ExifTool partially recognizes it but all other software (DPP, Lr, C1) recognize it without any problem.

PhotoLab 3 asks to choose between 9 proposals.
PhotoLab 6 doesn’t recognize it at all !

Pascal


PL3

image
PL6

IMG_0476.CR2 (12.9 MB)

Lens recognition has changed indeed. Combinations of older Canon cameras with newer lenses used to produce module ambiguities that could be resolved manually. Now, the ambiguities are gone, together with the ability to recognize certain lenses.

I’ve reported this a while ago and uploaded example files. But nothing has been done after that.

1 Like

However, LensType is correctly filled in

Pascal

I don’t know, which tags are read by DPL in order to determine the lens model/specification and suitable module. But I found that if I add the “Lens Model” tag and lens specification with ExifTool, DPL proposes the module that fits.

Modification:


Lens model entry (left, line 54) in an exiftool dump (exiftool -G -a -u) compared to the dump of the unaltered file. All other differences are either timestamps or offsets.

@Cecile-C

And?

Can we expect an official answer on the causes and remedies?
It worked on one leg with PL3 and not at all since.

Pascal

We have a task to fix the issue with the Canon modules and it should be done for the summer at the latest. Thank you four patience and understanding

2 Likes

The issue seems to have been fixed with updates of June 7th, 2023 for

  • Tested: PhotoLab (Mac) version 5.12 build 93
  • Tested: PhotoLab (Mac) version 6.5 build 52
  • Untested: Other apps: DVP, DFP, Nik

Thank you DxO!

Not so sure @platypus and @Cecile-C

Case half settled, PhotoLab 6.7 now behaves like PL3.
We have a choice of 9 different versions.
Pascal

Imo, proposing a selection of modules is far better than adding a badge saying “no module available”.

I like the implementation with the orange markers (although they might look better in a column like on Mac, instead of at the end of the name) and the possibility to make a selection even before the selected module is downloaded.

PureRaw had the “decide-before-downloading” for a while already, but the updated module database (and possibly built-in procedures) have resolved the lens/module recognition issues in my case.

1 Like

This is what users needed, since it seems metadatas are difficult to interpret since there is no standard and some datas can be the same for different optics from brand to brand.
If this choice is back again (let’s hope not for canon users only), this seems to resolve the issu, no ?

…only time will tell, because there are way too many camera-lens combinations and detection depends on what the respective firmware versions have written to the files.

If the detection algorithm finds a wrong module, but is dead sure about it, why bother proposing options? And how should the algorithm know that it’s gone astray? In view of all the variations, DPL does a fairly good job and in some cases, this is not good enough. Adding manual module selection or a way to limit e.g. brands could improve the situation.