P: In-camera double exposure colors have changed dramatically

[FeedbackCommunityMember]

Hi Guys,

I have a problem with the RAW file of an in-camera double exposure that has changed colour dramatically and I'd appreciate any advice.

Image 1. - this is the base image for the double exposure.

large?v=1.0&_1.png

This is image 2 of the double exposure:

large?v=1.0&_2.pngExpandIcon.svgExpandIcon.svgExpandIcon.svg

This is image 3: - the in-camera merged double exposure.  The image below is SOOC and exported directly from Lightroom last week and these are the colours I would like to keep.  This image is a low res jpg and is all I have left of the original colours as the RAW version in Lr has changed to nuclear Yellow.

large?v=1.0&_3.jpgExpandIcon.svgExpandIcon.svgExpandIcon.svg

This is the RAW file showing now as funky, nuclear yellow.  Despite resetting all the settings and even reimporting the files I am stuck with the yellow and can't  get the original colouring back.  I can get it to revert to green by using Auto WB in Lr but it's not the same as the original.

large?v=1.0&_4.jpgExpandIcon.svgExpandIcon.svgExpandIcon.svg

There has been a long discussion in The Adobe Support Community and the thread for that can be read here:

https://community.adobe.com/t5/lightroom-classic/colours-have-changed-dramatically-in-lightroom/m-p/...

The RAW file may also be downloaded here: https://we.tl/t-6LcQqFhZnw

I would be very grateful for any help.

Kind Regards

Karen

[Todd Shaner]

I can confirm the issue is present in LR as far back as version 5.7.1 so not a new problem. The OP's multi-exposure CR2 file was checked in numerous other raw converters including Canon's Digital Photo Professional and none of them exhibit the issue. So it appears LrC, LR, and Adobe Camera Raw are misinterpreting the As Shot white balance setting in some multi-exposure raw image files.

[Rikk Flohr: Photography]

Thanks for the report. The Camera Raw team has logged a bug for a fix in an upcoming release!

[Todd Shaner]

Thanks Rikk–Much appreciated!

[FeedbackCommunityMember]

Hi Rikk,

Thank you very much, that's great:)

Kind Regards

Karen

[Todd Shaner]

Rikk, I had a chance to shoot a sunset multi-exposure picture similar to the OP's using my Canon EOS 6D and the issue is not present. The As Shot settings for my multi-exposure CR2 file are very close to the OPs (6350 +11), but the WB looks correct with those settings inside LrC 10.0. The issue appears to be unique to the EOS 5D MKIII camera. Just an FYI for the Camera Raw Team.

[drtonyb]

I would say this is the Canon version of the white balance problem I reported for Nikon raw files produced by combining two or more images using Nikon's overlay function. The problem also afflicts double exposures taken with my Nikon D750.

I investigated this problem myself and discovered that the resulting raw file is produced by applying the white balance to each source file before being combined. LrC doesn't understand this and doesn't know (hasn't been programmed) how to handle it. White balance for these 'combined' raw files from later Nikon models seem to be handled differently, but still have problems.

You can read through my reports and conclusions here:

https://feedback.photoshop.com/conversations/lightroom-classic/lightroom-classic-strange-white-balan...

The problem has been acknowledged and although Adobe has claimed they are working on it, I seriously doubt anything is being done about it.

It would seem that the Canon double exposure does a similar thing in that it applies the white balance to each source file before combining them to produce the raw double exposure. It makes sense to do this so that two source files shot with different white balance settings can be combined.

The best way do deal with the problem is to use the WB Temperature and Tint sliders to get the white balance required or sample a known grey part of the image with the WB dropper.

Below is a screen shot of your double exposure image and a test double exposure from my Nikon D750 (bottom row).

I imported with Embedded & Sidecar Previews, so the images on the left of each row are showing the embedded JPG previews created by the cameras and therefore, the white balance (colours) appear as expected. The middle images are virtual copies without any Develop adjustments as rendered by LrC. Your image is very yellow and obviously wrong, while mine looks normal. The last two images in each row are also virtual copies. For your image I have set the WB Temperature to 3300 and Tint to +30. It looks close to the JPG preview. For my image, I have set the WB to Daylight and as you can see, LrC has got it wildy wrong!

Don't hold your breath waiting for Adobe to fix it. Two years and it's not fixed for my Nikon overlay and double exposure raw files.

[Todd Shaner]

Anthony, thank you for this information...I had completely forgot about your report! In the case here with Canon 5D MKIII multi-exposure CR2 file the OP set the camera to a fixed Flash WB setting. So one would expect the WB results to be the same as the individual shots. Clearly the method used for calculating the actual As Shot WB is incorrect. Here's a comparison of the OP's multi-exposure CR2 file shot witht he EOS 5D MKIII and a similar one I shot with the EOS 6D camera.

The color temperature in the 5D MKIII file is lower than mine (4436 vs 6509), but entering 4436 into LrC WB the image is still shifted heavily toward yellow. It's still a mystery to me!

[drtonyb]

Todd, have a look at RawMeasuredRGGB in the double exposure CR2 and then look at it in a normal shot.

The double exposure has RawMeasuredRGGB = 0 0 0 0 ????

I examined the EXIF data for normal daylight shot CR2 image from a Canon 5D Mrk III and it has RawMeasuredRGGB = 465264 957711 923189 578046

Something to ponder upon.

[Todd Shaner]

I noticed that as well, but was looking for differences between the 5D MKIII and 6D multi-exposure WB data. I agree a normal shot shows values as you've indicated, but my multi-exposure test shots all show the 0 0 0 0 RawMeasuredRGGB values and do NOT exhibit the WB issue. That's why I said "It's still a mystery." Here's a download link to one of my EOS 6D multi-exposure shot similar to the OP's. Maybe you can see something that's different in the 5D MKIII.

https://www.dropbox.com/s/oncdwbf9rgpumtb/IMG_0334.CR2?dl=0

[drtonyb]

Todd, I tried changing the RawMeasuredRGGB value in a normal 5D MKIII shot to 0 0 0 0 and imported into LrC. The photo looked normal, so RawMeasuredRGGB might be a red herring here.

In the end, it isn't up to us to work out what is wrong. It is for the development team to solve it.

Just out of interest, below is a screen shot of opening the OP's file in RawTherapee 5.7. 'As Shot' or Camera WB looks as expected.

However, changing to Flash WB (or any other setting other than Auto) looks very wrong, as you ca see below, but nothing like LrC's yellow cast. I might add LrC's Auto looks much better than RawTherapee's.

Opening your 6D double exposure in LrC is handled as expected, but in RawTherapee, only Camera (As Shot) and Auto give a reasonable result. All the other WB settings result in a magenta cast.

So, like the Nikon case where the D750 and D850 double exposure raw files' white balance seem to be handled differently by LrC, the 5D and 6D double exposure raw files also appear to be handled differently in regards to white balance.

[Todd Shaner]

In the case of the 5D MKIII it was the first EOS camera that had multi-exposure capability. Canon may have made changes to later camera models such as the 6D......and same for Nikon cameras? Most likely this will require version 2 camera profiles for the fix.