Overshooting Vignetting correction

i was editing some snapshots and noticed some negative vignetting (brighter corners then centre.
(don’t mind the image i was beating myself i didn’t bring my 300mm… :stuck_out_tongue_winking_eye:)

i have this on auto mode most of the time but now it’s overshooting.
about 50%

never noticed this before, was i blind?
(meaning does anyome noticed this earlier also on his images?)

I’ve seen the problem before, but not to this extent. As the help (?) text in the Vignetting palette says, in auto mode 100% intensity means full correction. So what you’re seeing is either a bug in the lens correction module or a mismatch of the module with your camera/lens combination. Another possible cause is the use of filters or other add-ons that modify the lens optics.

No addons or filters.
Only new thing i did was running them through FRV as test if this program is something i like.
Used the check buttons like usm, d, s, e, and s h.

Doesn’t touched adjustments in FRV or safed xmp’s.
It’s odd.

I have to check if all images that day has this problem.
Yep, not only the point in the sky shots. at this folder.

test older image around the same type same lens:

no negative vignetting lowerleft is a cloud.
copied it, opend in DPL clean no edit done so no DOP-file or viewed by FRV.
took both to be sure and the 351 has clear visual issue of negative vignetting(overblown corners)

@Egregius: could you test my file on your computer?
See if it’s a system thing or a software thing? (i proccessed this file on a lower version of V3 earlier so it could be a update error.)
P1050351.RW2 (18,5 MB)

I never looked at the vignetting. But with a camera/lens module and the setting vignetting on auto all pictures are 100 and 50 as with you. Makes me wondering what auto means.


As far as i understood from a dxo staff member it’s mostly not image intelligent. It’s a preset they think as correct for the given iso, aperture, focallenght and such. They use exif data and the optical module data to preset this magic wands. That’s why they do change shifting through images. I broke my head also on the denoise magic wands why it wasn’t iso depending but just 40%. And almosr never changed on underexposed overexposed or low /high iso. They explained this a bit more. I think microcontrastauto setting is depending on edgecontrast or micro detail like the two filters in FRV edge and detail are detecting.
So maybe they step in monday to explain. :blush:
Maybe the “deadpixels” removal is based on image scanning and some others.

What’s the advance of 100 and 50 over 0 and 0?


It’s just the amount or percentage of the correction you want to apply.
Every lens has some vignetting and if you don’t like the effect turn on correction.
Until resesently it’s was in automodes fine now i see overcorrection.

I just tried PL1, the same. Auto means fixed numbers of 100 and 50. So once again what is the advance of 100 and 50 above 0 and 0? It’s done on white images as well as on dark images.


Did you run my image through pl1?
Same overcorrection?

About your question, i hope dxo staff can answer that. :slightly_smiling_face:

Just some raw’s of my own. Auto is a fixed value :neutral_face:


I have had this type of phenomenon in the past - so with earlier versions (Optics Pro).
It was always in similar situations: blue sky and use of a long focal length lens.
I don’t know why … I had fixed the problem with a partial preset decreasing the intensity of the vignetting and increasing the preservation value.
I just tested with this file, it’s good with intensity 30 and preservation 80.
It may be related to the module that overcorrects, but it was not systematic depending on the image

1 Like

I don’t see why vignetting in a lens would change with the subject being photographed. It changes with focal length and aperture. If DxO’s camera/lens correction modules account for the amount and gradations of lens vignette as they should, then Intensity=100 always fully corrects the vignette so that the light gathering is even across the frame. I don’t see why automatic correction would be anything other than 100. That the vignette isn’t corrected properly at 100 tells me something is going wrong - I would guess either with the lens or with the correction module. Is the problem happening only with one lens?

The description of the Preservation slider is, “Prevents highlights from being clipped and noise from appearing in the shadows.” 50 seems to be a middle setting that is calibrated for the lens. If I lower or raise this setting, the effect is dramatic. However, if in PL3.2 I grab the slider using the mouse and drag it through the range 0-100, something goes wrong. Settings below 50 become indistinguishable from 50, while settings above continue to alter vignette correction. This is obviously a bug.

EDIT: My mistake, I accidentally had the correction off when I first clicked on it. The Preservation slider seems consistent in only having a noticeable effect above 50 with my image files. It looks terrible, though - the radial gradients it creates aren’t smooth.

Vignetting is also visual, not only mathematical. It can be caused by lens design but also by the use of a lens hood. It’s more visible on lighter corners as on darker corners.
From what I understood PL is trying to correct vignetting based on a mathematical approach based on the optical modules. It doesn’t look to the need of a correction. Automatic is a fixed setting of 100 and 50. Also with dark images.


Maybe the optical modules are modified or the cameradepending profile.
Which could be effecting this vignetting correction strenght.
I labeld it as “Bug” see what come’s up.
(it isn’t the lens because it also happens with a older image which had before no issue.)

Hello Peter (@OXiDant),

looking at your image (thanks for the RAW file) it seems we have a vignetting overcorrection, we will look at it and correct the optics module.



@Marie, do you need more rawfiles? I can deliver some from one day or more days. Different focallenghts is also possible.


Hello Peter (@OXiDant),

yes please, it will help to secure the correction.
And if you can tell us on what config (Focal + FNumber) you have that issue it will also help to be sure we don’t miss something.

best regards,

I will select a row when i am home.
And to be sure i wil do a full test wide open til f11 at wa, midrange fulltele.
Around 20? Rawfiles.
I think that will cover it nicely. :grin:

yes, that would be very great :slight_smile:
Thanks a lot !

Hello Peter (@OXiDant),

for you overcorrection issue we will have to refine our calibration of Panasonic 14-140mm F3.5-5.6 and Panasonic G Vario 100-300mm F4.0-5.6.
Due to the confinement we will have to wait a bit till we can do necessary images but we whould be able to release new optics modules in June or July at the latest.