Lens Over-Correction?

I have found a strange issue that seems to be occurring when processing Fujifilm RAF files. This was shot at 200mm with the Fuji 55-200 lens. I have seen the same issue with other images shot with this lens. I am using PR2 in Lightroom and PR2 downloaded the lens correction files before I processed the image. I reset all LR edits before processing with PR2. It almost seems like PR2 is over-compensating for lens vignetting or something and lightening the left and right sides. However, the following is with both lens corrections set to OFF and using DeepPrime. If I check lens distortion correction to ON, the lightening is slightly less but still very visible.

Here is the unedited file followed by the result after PR2:

Here is the RAF file: Transfer - Dropbox

The RAF file on Dropbox is not the one shown here. It is difficult to understand where the problem would be.
(There is obviously no problem with the transmitted file)

I have on occasion seen differences in color with uploaded images in Dropbox. I have suspected that Dropbox may alter the display rendering, but I do not know that for certain. Of course for things like uploading raw files and .dop files it’s not an issue.


My apologies. I have corrected the link for the RAF file.

Just to clarify, the 2 images above have not been through Dropbox, they are just jpgs of the RAF file before and after PR2. Comparing the images, the PR2 image (bottom) is overall a bit brighter. But the real concern is how the left and right have been bleached out.

Am I correct in thinking that removing all the LR edits should allow PR2 to process a “clean” file without LR affecting the output?

This issue is related to the PureRaw lens correction profile. I have already seen a few similar cases, always with telephoto lenses and generally at the greatest focal length. But it’s hard to say if it’s a general problem related to this profile, or to the specific case of your lens.
The result is effectively an overcorrection of the vignetting, and it is not possible to deactivate this correction in PureRAW, to use instead the correction of Lightroom with the integrated profile.
Distortion correction normally has no effect on vignetting (except at the margins with lenses with a lot of distortion).
The only possibility I see is when returning to LR, to switch to manual lens profile and enter a negative value for the vignetting slider. But that will probably only partially solve the problem.
It would be interesting to see if this issue affects all shots from this lens at 200mm, or just a few.

With PhotoLab, we can easily solve this kind of problem since we have additional settings to modulate the correction of vignetting.

Thank you for the explanation. That lens correction (of all things) should be an issue from DxO is certainly disappointing. Since posting I have come across a similar issue with an image from the Fuji 16-80 shot at 27 mm. It seems as it stands that PR2 is not quite sorted, at least for Fujifilm lenses. I have also noted color changes with pretty much every file I have put through PR2 where a green color cast is added. Although easily fixed in Lightroom it is still a pain to have readjust color with every file. It is a real shame, as I have been very impressed with its sharpening/noise reduction. The workflow integration with Lightroom also seems excellent. But the hit and miss performance with the files I’ve tried so far is concerning. I don’t think in its current state that PR2 justifies a purchase.

1 Like

As I said above, the problem comes from the DxO profile for this lens… and probably for this focal length. But it’s not clear who is involved::

  • the profile itself,
  • the information transmitted by the camera in the exif values ​​(in particular the exact focal length) that PR needs to define the vignetting correction.
  • or the lens that would have a particular problem (compared to the rest of the series).
    If the profile is in question (and this can obviously happen), generally there are several reports from users and DxO corrects the profile within a period that can be more or less long depending on their workload.
    If the lens does not return the exact focal length values ​​to the camera (the firmware may for example return 200mm when it is really 190mm), this may have a more or less significant impact. Generally this poses little or no problem, but in the case where the vignetting is important, and this is often the case with zooms at the end of the range, this can be a problem.
    And of course it’s possible (although it’s rare) that your lens has a problem such as lens shift. Here, I do not think that is the case on first analysis.
    But it would take more examples to have an idea of ​​the origin of the problem Knowing that DxO’s objective profiles are normally the best.

For the “problem” of green tint, more details are needed: is the file transmitted to LR by PR a dng, or a jpeg? And there again, examples would be needed for us to be able to judge.

Here is an example of the vignetting issue with the Fuji 16-80 lens shot at 27 mm. All LR settings were zeroed before processing with PR2. This one isn’t as bad as the last one but is still obvious when looking at the clouds at the top right and the river at the bottom right. The issue really pops out with further editing.

RAF file: Transfer - Dropbox

Another example with the 55-200 lens shot at 128 mm. Vignetting issue is visible especially along the right hand edge.

RAF file: Transfer - Dropbox

Since the RAW file in Dropbox for the last example is the same as the previous one, I can’t pass judgment on that.

But for the second example, I’m not sure there’s a problem.
Yes, the water is clearer along the right edge, but wasn’t it clear when shooting? Ditto for the cloud above. It’s hard to believe that there could be a problem with a vertical strip and only on the right. The vignetting correction being done from the center of the image towards the edges “in a circle”, there is no reason for only the right edge to be affected, the left should be, and the angles much more than the edges in the middle of the image.

Thanks for your time. Sorry about the raw file. I won’t make any more attempts.

I have the same problem with Sony files, except it is much, much worse than the OP. It is NOT a lens correction problem, because I turned it all off, and it did not solve the problem. I have uploaded my sample images in another post on this forum. Yes, this program does not seem to be sorted.

So far I’ve seen a couple of images in this thread and I have to admit, I struggle to see the super big differences - if an image already contains sort of mist, low contrast and especially no single evenly lit area (I mean the first comparsion if the TO), the “overcorrection” is simply the “properly exposed” mist which is more accentuated because the vignetting doesn’t raise contrats anymore. I’d like to see from any of the persons with “overcorrected lens profile” a simple, evenly lit plane with a uniform colour.

First, in the lab there’s no possibility to focus a lens close to infinity and then measure the amount of vignetting. Typically, lenses are tested (by opticallimits, i.e.) with 25-50× fl as focusdistance (6.25 - 12.5 m for a 250 mm lens). I don’t know how long DxO’s lab is, but at home it’s not that easy to “make way”.

I don’t want to defend DxO’s profiles, I just don’t see the sample images as good examples to find out about “too much vignetting corretion or not”.

PureRAW is supposed to process raw files and then, Lightroom edits are ignored anyway…but might be applied to the files exported by DPR.

Maybe it’s a good idea to process the images in DPR first and catalog the files in LrC afterwards.