We have a brand new look! Take a tour with us and explore the latest updates on Adobe Support Community.
Copy link to clipboard
I set a custom white balance in camera (using kelvin settings) to 5600, but LR develop module shows it as 6000. I'm using a Nikon D810, shooting RAW, of course. This happens whether I import as NEF or DNG. When I set the camera to "Flash" mode for white balance, LR says I used 7100K. The camera manual says "Flash" is 5400K. One of my colleagues, who also shoots with a Nikon D810, sees exactly the same results.
Is this a problem with Lightroom? With all Nikon D810s? Is there a setting I'm missing? I need to submit RAW files for a certification I'm applying for, and they need to verify the WB setting in the files (as shot). I'm concerned that they'll see the wrong values if they use LR to read the data. Beyond that, I'm concerned that my images are ending up too cool because of this issue. Sure, I can adjust the WB to taste for client images, but would much prefer to be able to rely on the specific temperature values.
Has anyone else noticed this? Is it a bug? Is there a better way to inspect the RAW files to see what settings were actually used?
Is this a problem with Lightroom? With all Nikon D810s? Is there a setting I'm missing?d?
Neither. Not a problem with LR or the camera or any setting. LR interprets that WB setting you are using in the camera. If you loaded that image into some other RAW developer is would be different than the camera or LR as that other RAW developer will interpret the WB differently.
Copy link to clipboard
Copy link to clipboard
CCT values define a rather large variation of possible colors. IOW, CCT 5600 isn't one specification of a color, it's a large possibility of colors as they fall on the line(s) of correlated color temperature. So that they don't match isn't at all a problem or something you should expect to match.
ANY color between line e and f below can be called 5000K.
So this isn't a bug. It's simply a different number being interpreted and it has zero effect or practicality on editing your images.
Copy link to clipboard
This happens whether I import as NEF or DNG. When I set the camera to "Flash" mode for white balance, LR says I used 7100K. The camera manual says "Flash" is 5400K. One of my colleagues, who also shoots with a Nikon D810, sees exactly the same results.
Nikon SB and Canon EX series flashes ( and others?) communicate the white balance setting back to the camera based on the actual flash settings and and light output. With shorter duration (i.e. lower light output) the flash white balance is higher than with a full-duration 100% flash output. You will see a different 'As Shot' WB inside LR because of this variation in flash color temperature at different light output.
So you have a small offset due to the way LR interprets the 'As Shot' WB setting coupled with the camera's WB setting changing due to variations in the flash output. If you have LR's WB set to 'As Shot' then it will apply the correct settings as communicated to the camera by the flash. As long as you keep your camera's WB set on AUTO and LR WB on As Shot there's nothing to worry about!
Thanks, Todd. Your comments are certainly correct for certain situations. The thing is, I never said I was using speedlights. In fact, there's zero communication from the flash back to the camera in my case because I'm using studio strobes with everything in manual mode. No TTL, no auto WB, no auto anything. The camera doesn't even know the strobes are firing, as they're triggered by the "dumb" center pin in the hotshoe (using a radio controller to trigger the strobes).
That's not to say that the setup is devoid of variances! The strobes could still have slight inconsistencies from one firing to the next, and the light modifiers can certainly shift things depending on modifier. I'm just saying there's no feedback loop back to the camera, so the RAW file's metadata remains constant, set to whatever I set things to in the camera. In this case, I'm setting a kelvin value explicitly in the camera, so that doesn't change at all from shot to shot. Even when I set it to "Flash WB", that's just a fixed WB value in the camera. I could put on the lens cap, and crawl into a deep hole in the dead of night, producing a totally black image, and the WB set to Flash would still give the identical WB setting in the file's metadata.
I do agree that this all has very little effect on how I adjust my images in post. The only impact, really, is on the "as shot" settings. Unfortunately, my current project requires that I submit unedited RAW files for evaluation by a certification committee, and they need to know exactly what settings I used in camera - not what settings make the image look good in LR. To this end, I seem to be out of luck.
I never said I was using speedlights. In fact, there's zero communication from the flash back to the camera in my case because I'm using studio strobes with everything in manual mode. No TTL, no auto WB, no auto anything. The camera doesn't even know the strobes are firing, as they're triggered by the "dumb" center pin in the hotshoe (using a radio controller to trigger the strobes).
The below screenshot shows the WB Temp value claimed by Nikon for the D810 when set to 'Flash and the LR Flash WB setting. The readings are very close to each other. If you're seeing 7000 'As Shot' reading with the camera set to 'Flash' WB there may be a problem. Please upload ANY NEF file shot with the camera's WB set to 'Flash' with ANY flash used. I'll take a look at it and see what I can find. If you can shoot one with a colorchecker or white balance card in the scene it would help.
Copy link to clipboard
The underlying reason for the apparent temperature difference is that different processing engines are used. They don't use the same algorithms, and they don't apply the same parameters, and not in the same way. To produce visually similar (never equal) results, different input parameters are used. They can't be equal.
Inversely, to use the same numbers would produce different visual results. So it's a choice. Match the numbers, or match the appearance of neutral color balance?
Lightroom reads the camera white balance metadata, and then applies its own numbers to give a similar visual white balance.
And the camera WB metadata isn't necessarily 'correct'. Easy to capture a raw, measure the illuminant with a spectroradiometer and see different values provided for CCT between the two sets of values. I'd side with the spectroradiometer in terms of a far more accurate measurement of the illuminant. So not only do differing raw converters provide differing CCT values, the data they are using to represent the values are from a device that's not really the best tool to gather such values in the first place. The numbers really don't matter very much in the grand scheme of things; you adjust visually. The numbers hare somewhat of a guide for knowing which direction provides a warmer/cooler (greener/magenta) shift but otherwise, they are not very useful in processing images.
Thank you all for your detailed responses! You've certainly given me something to think about, and I'm still processing the conflict in my mind. I totally agree that this all has very little practical impact on our day-to-day image processing. The only real impact I see may be with how far you can pull the WB into the extremes - e.g., going from 8000 to 10,000 is a smaller adjustment than going from 3000 to 10,000 - but seriously, who does that?
The problem for my specific situation is that I have to submit unedited RAW files to a certification committee, and they need to be able to verify the exact settings I used in camera. It has nothing to do with how I process the images in post after capture. I get that color science is much more complex than simple exposure settings, but I can't shake the idea that there's still a parallel here. Yes, different sensors react differently to the same stimuli. One sensor may see a given color/shade as #ABC (just making up numbers here) while another may see the same color/shade as #AC7. But then doesn't that apply to exposure settings, too? Once sensor may be more sensitive than another, so the raw data would indicate different exposures. And yet, LR accurately reports the exact shutter speed, aperture, and ISO as written to the file by the camera. It doesn't try to interpret the exposure, perhaps changing the reported ISO, to conform to some Adobe standard. Yes, I know, the "exposure" slider always starts at 0, and all changes are deltas from there. I'm making an analogy. It's apples to grapefruit, I know. But the principle still applies - I want to know what settings I used in camera so I can use that as a basis in the future, when confronted with a same/similar lighting challenge, to either duplicate the setup or make proper adjustments. It's metadata. It's useful to know exactly what the camera used during capture.
I do find it interesting, based on all your comments and links, how processing engines, like Camera RAW and others, handle the differences between manufacturers and camera/sensor capabilities. I'd have thought they could map the raw sensor data to a common, universal color space (like ProPhoto RGB), adjusting the mapping but maintaining the camera's metadata. Perhaps that's what's happening, just with the added step of trying to somehow normalize the colors after mapping, by tweaking WB and tint. I don't totally get that. I expect RAW to be RAW. When I shoot video using a LOG profile, I expect a flat, log image, not a colorful natural-looking image. I don't want RAW files to match the embedded JPG images. Still, I'm sure it's all more complicated that that. Smarter people than I have been working on this for much longer, so I'll reluctantly accept that it is what it is. As you all say, it really doesn't have much/any impact on my daily activity as a photographer.
Perhaps this is all Nikon's fault, for not sharing the details on where to find WB values in their NEF files. Or maybe it's on Adobe for ignoring data it can only find in some but not all RAW formats. Or maybe knowing what camera settings were used has absolutely no value - toss it all out! Why should I care to know what lens I used or at what focal length, anyway? Metadata is useless! (deep sarcasm, sorry)
I truly do appreciate all the feedback. Thanks!
Copy link to clipboard
I'm not sure how to handle marking this question as answered, so I'll mark this one that way, as a summary.
Fact is, every response provided here was useful and informative. Every response added value to the topic and is worth reading! If you're here to see the answer, please read all of the responses to gain a better understanding of the issues.
Bottom line, Lightroom does not display the WB setting from the camera. It interprets the value to achieve a certain look (see other responses for details). If you absolutely need to know what WB setting was actually set in the camera, and that camera happens to be a Nikon, you'll have to use one of Nikon's software tools, like ViewNX or Capture NX. Those tools let you see all of the camera settings as recorded in the NEF file. Do note, however, that they only work with NEF, and not with DNG, so you'll have to go back to the original files to use them.
I had a chance to test this with my Canon EOS 6D set to 'Flash' white balance with no flash attached. The LR 'As Shot' reading is for all intents in purposes identical to the camera's white balance setting differing by -145 WB Temp value. I can't duplicate what you're seeing with the Nikon D810.
Concerning the LR 'Flash' WB setting it looks like Adobe uses a generic 5500 value for all cameras. This is obviously not very useful since actual flash output can vary from ~5000 to 7000 Temp.
Canon EOS 6D with Flash WB Setting and no flash attached
Camera Maker Notes
Color \TempAsShot 6345
LR As Shot 6200 +15
Copy link to clipboard
I do have a Nikon D810, and I get the same 7100K reading in Lightroom with the camera set to Flash white balance. It's consistent, and apparently Lightroom's interpretation of that camera setting.
When I attach a Nikon Speedlight SB-700, the resulting image in Lightroom looks perfectly right, with a dead on neutral white balance. And it reads 7100K.
So this is just the way it is. It has no bearing on the raw file itself, so I usually just leave the camera on Auto and never worry about it. What the actual light output from the SB-700 is I have no idea. I also use studio strobes, but haven't noticed any significant inconsistencies when used in combination with the SB-700 (I often use a remote speedlight for backlighting and strobes for main).
Dag, this is very similar to what I'm seeing with my Canon EOS cameras and Speedlights. In the screenshot above you can see the Canon WB settings correlate well. Unfortunately Nikon doesn't place an actual WB Temperature value in the raw file maker notes probably. Here's an example NEF file, which shows just the WB R and B Levels with G assumed as unity (1.0). The bottom line is if the LR 'As Shot' setting provides proper WB it really doesn't matter.
So to answer PhotosByAlejandro concern:
"The problem for my specific situation is that I have to submit unedited RAW files to a certification committee, and they need to be able to verify the exact settings I used in camera. It has nothing to do with how I process the images in post after capture."
No one looking at the NEF file can determine the actual target WB Temp and Tint settings. You can with Canon CR2 files, but not with Nikon NEF files. This 2005 article may be of interest concerning Nikon encrypting the white balance metadata. I don't think this applies to the current Nikon camera models, but clearly Nikon doesn't share verbose WB information in the NEF file format like Canon.
The bottom line is if the LR 'As Shot' setting provides proper WB it really doesn't matter.
Precisely. This is a non-problem, ignore it.
If the certification committee knows what they're doing, they will see it the same way. They know they can't conclude anything based on numbers in Lightroom, which uses a completely different processing engine than Nikon. The numbers aren't supposed to match.
I am having the same problem with Canon Eos R CR3 files set to 5600K and Lr Classic opens the file as 6000K and all the color is totally different from the shot image.
When I use a CR2 file from an old Canon Eos 6D set to custom Kelvin values, Lr reads those values correctly and displays an image similar to the one shown on the camera display.
I'm sorry but it seems like a problem with some particularly raw files and not with normal behavior.
CaptureOne doesn't make these mistakes!