White balance offset in DNG with PureRaw 4 and Ligthroom


I already posted something 3 years ago about this issue with Photolab:

The problem is still there, and as PureRaw is mainly designed to export DNG, maybe there will be more reactions here.

This problem happened with my Olympus E-M5 mk III and is the same with my OM-1.

Any image will get a white balance offset when exported in DNG.
This offset is integrated in the “as shot values”, so for an image that has no setting in Lightroom, both the RAW and the exported DNG will show the same colors, but with different values.

For instance, the RAW will show with T=5000 K / Tint+2. The DNG T=5300 K / Tint-4

However, if the image as already been processed, and the white balance is changed, let’s say, to T=5600 K / Tint+5, the exported DNG will show the same values, and the color balance will be wrong.
I would have to apply the offset (300 K/-6) to the values to get it right.

Does anybody else have this issue?

I see differences in WB temperature and tint values…but I don’t consider them to be an issue and simply adjust the two until the image looks like I want it to look.

Yes, I understand.
Most of the time I decide if an image goes to denoise only close to the final stage, so my settings are already there, especially white balance…

So having to redo the white balance adjustments is a pain to me.

:exclamation: Please observe that PureRAW acts on your original RAW files.

DPR does NOT understand what Lightroom wrote as edits to the XMP sidecar. If you use DPR as one of the last steps, the things you did in Lr are ignored by DPR. Lr can later apply the settings from XMP to the DNG files you get from DPR, but some of these settings don’t work or do so in unexpected ways, e.g. applying CA corrections again does not make sense.

:exclamation: DPR should be one of your first steps in order to eliminate such irritations.

If I think I need DPR denoising and optical corrections, I use DPR standalone and rename the exported files with Finder before syncing the source folder again in Lr. All other edits (image and metadata) are then done with Lightroom.

It’s not correct.
PureRaw (or Photolab) leaves most of the XMP settings unchanged, but changes some of it, in order to make to avoid double corrections in Lightroom.

For instance:

  • optical corrections are set to zero
  • noise reduction is set to zero
  • sharpening is set to zero

So, it understand what it needs to understand, not more, and leaves the rest unchanged.
That means that apart local correction that may be inaccurate because of the distorsion correction, there is nothing wrong when opening the DNG file in Lightroom.

It reads the modified XMP data, and the unchanged data (other settings, keywords, GPS etc…) are still there.

We’re basically seeing the same things and use different words to describe it. But does it explain differences in white balance values as mentioned in the first post herein?

According to what I’ve read in several places, WT temp and tint values are ballpark rather than exact in photography and photo apps. And while we can establish a colour temperature value from the balance of red and blue light from a black body radiator, real life lighting in combination with all the technical properties of sensors and applications can produce values that differ by several (tens or hundreds of) degrees Kelvin. Nevertheless, differences can seem to be too big and that might qualify as bug or sample variation to be corrected by DxO.

That’s not the issue here.

I don’t expect Dxo, Lightroom, C1, Rawtherapee or whatever to show the same White Balance values.
It’s not important, as each software will have its own color rendering anyway so I don’t expect the colors to be the same in all software.

However, here I’m looking at the values in the same software: Lightroom.
What I like with DxO product is that the DNG that they generate are not developped in term of colors.
I can get the exact same colors in Lightroom with a DNG coming from DxO, than with the original RAW, or with the DNG denoised by Lightroom.

This is not the case with some other products like Topaz, and that’s a reason why I’m not considering using their products.
I don’t want Topaz colors, or DxO colors, or whatever.
I want my denoise file to render exactly the same as the other pictures in the same series that don’t necesarilly need denoising.

That was the case with my old camera (Olympus E-M10) and it worked perfectly.
It’s probably working like this with many other camera bodies.
But with my new ones, I get this annoying White Balance offset that means that all my images coming from DxO products need additional work.

I can still get the exact same colors, but I need an additional step.
I don’t get this issue in Rawtherapee for instance: the white balance values are the same in the RAW and the DxO DNG.

So, Lightroom has an issue reading DNG from DxO, maybe it’s a bug in Lightroom, maybe it’s a bug in DxO, I can’t be sure.
I don’t know which other bodies encounter the same issue.

Currently, this issue is blocking for me: I won’t update my DxO software if I need to manually correct every single image coming from them.
I won’t change my workflow either → I’m not going to get through DxO first before knowing if the image is worth getting denoise.

The WB vs. Temp and Tint values intrigued me and I therefore ran another test:

  1. Processed a series of images in Adobe DNG Converter, in PhotoLab versions 4 to 7 and PureRAW versions 3 and 4
  2. Synced the folder in Lightroom → Images looked different due to Lr applying different presets. Temp and Tint values were different too
  3. Standardised the images in Lr using one preset and WB setting for all images → Images looked fairly similar, DxO output is darker than Adobe’s rendering as a rule, some images are fairly similar too, results seem to depend on image characteristics. Temp and Tint were mostly different and Temp was 300 to 650K lower for DxO output DNGs.

    (FLTR: Original, Adobe DNG, PhotoLab DNGs, PureRAW DNGs; check filenames for details. First row looks like most output, lower rows show examples of obviously darker DxO output)

When I reset all images, I found Temp values to differ by as much as 800K.

Be it as it may, I’ll keep using DxO SW as first step after culling and adjust images (OOC RAWs and DxO DNGs) in Lightroom. I’ll stick to ignoring WB Temp and Tint values too.

Your conclusion seems to be in alignment with your choice and/of procedures … and saves you some money for not getting updates. No need for them either, specially if the things you have do what you need them to do. :+1:

I’ll read your answer when I have a bit of time…
Not sure to understand all of it today… Probably too tired after a work day :slight_smile:

I think most people to the denoise first, and that’s probably why I seem to be one of the only ones having an issue here.
And it make sense to do so, just that I want to keep the pictures with additional DNG to the strict minimum so I want to denoise only the ones that I want to export, and this I know after having set some basic parameters.

I’ll retry it anyway when Photolab 8 will be out.
I still own a PL6 licence so it will depend how much the update is and what are the other news in PL8.

For PureRaw, I’m not fully convinced & the renaming scheme is a blocker, so I’ll pass for now.

Advances have bee incremental rather than revolutionary in the last few years. Skipping an update seems perfectly reasonable, also in view of the current upgrade price discounts. Check your shop account for details.