White balance in Lightroom different than set in camera?

When I shoot in my studio, I set my custom white balance in my camera to 6000K. However, when I import the photos in Lightroom, they are all 5500K. Am I missing something or does this have a purpose?

It's because the Canon in-camera settings are proprietary, in that they don't translate into exact values in the Metadata. So, the Raw processor has to make an educated guess. They have done a pretty good job of "reverse engineering" the Canon settings but not perfect -- use your own judgement as to how accurate the WB is and tweak when needed. It's always good to use a WB target in a studio shoot in a test shot anayway, I figure.

I also use the Canon software Digital Photo Professional (DPP) as a reference because of course it's Canon and knows how to read the camera values and it applies those things with more or less accuracy. I'd spend some time using DPP along Lightroom to get a feel of how they compare with your studio shots!

I explain this somewhat differently than Tony.

But the WB values, the multipliers applied to each color channel, are clearly stated together with the deduced color temperature in the Maker's Notes section of the metadata. They can be seen with ExifTool and if Phil Harvey was able to decipher the code, I'm sure Adobe engineers can do the same. (Or they could use Exiftool since it is Open Source and the code is freely available.) Below are three sets of data from three CR2 files - the top one is from a shot taken with AWB in sunlit conditions, the middle with the WB set to 5200K (although the light was actually colder) and the bottom with a custom WB set. In each case there are listed the actual multipliers used by the camera for the embedded jpg as well as the camera's measurement of the reflected light (note that they are not the same when AWB is set).

So what happens in Lightroom? The temperature listed is useless information as far as LR is concerned. What it uses when doing the "As Shot" WB is the actual numerical multipliers. But what is being multiplied? Not the RAW data, because before WB is calculated the image colors have been converted into the camera's native space as defined by the camera profile and (probably) from that space to LR's working space Melissa RGB. And it's that camera profile that Adobe has had to "reverse engineer", because Canon isn't sharing. So the same multipliers on different base values produce slightly different results.

And finally, in regard to the temperature displayed in LR's GUI. The camera has calculated the multipliers needed to neutralize the effect of a light source with an unknown color temperature. LR does the same when doing "Eyedropper" or "Auto" WB, calculating the multipliers from the image data. Knowing what measures are necessary to neutralize the color of the light, its temperature can be deduced and this is what LR displays because only nerds like me are interested in multipliers, but most photographers are interested in (and understand) temperatures. However, in the case where LR is asked to use "As Shot" the multipliers are dictated to it by the metadata and as we've seen, the post WB colors are different from Canon's, so the deduced temperature is also different.

HOSTED PHOTO
please log in to view hosted photos in full size.

An alternative to the studio approach of setting CCT in camera (as the OP noted) is to shoot a reference gray target in the studio for that session, and for each time the lighting is changed in that session. Then, number differences between Canon and the rest of the world are irrelevant, and click-WB'ing on the target in whatever application you use becomes the standard.

Kirk

Not sure what the issue really is, for this thread. The WB scale is defined by Lord Kelvin to be based upon the light output of a black body when heated to a certain temperature. I know that 'daylight' really is a range of temperatures (different at different times of day). I assume that at one specific tempature, the R-G-B values are balanced (neutral), but I do not know at what K value this happens to occur.

Let me relate an experiment which I just conducted, which indicates 'no issue at all exists'...

I just shot a known neutral 18% gray card (Douglas Gray Card, which I have previously verified as neutral against aother standard references like the Macbeth ColorCheker). This was shot under CFL at 1/40 f/2.8. I set Custom White Balance using this reference shot, and shot the same gray card again with the CWB setting. I imported into LR both the reference shot and the shot made with CWB, with no presets at all and with no other adjustments during import.

• The as-shot-in-camera with CWB had a Temp value of 2750, and a Tint value of +18 as shot.
  
• The same shot, when using the eyedropper white balance selector of LR to set WB would use 2750 and +18, as well.
  
• The reference shot, when using the eyedropper white balance selector of LR to set WB would use 2700 and +16.
  

Wilt - If you have an Exif reader that shows you the Maker's Notes section of the metadata, check what the camera wrote there - it will be different from the LR display. Or look at my screen shot above: In the first instance, an AWB shot in sunlight the camera wrote 4982K and LR shows 4600K. In the second, with the camera set to 5200K, the camera wrote 5224K (essentially the same) and LR displays 4850K. And in the third example, a CWB, the camera has 4690K and LR is at 4500K.

The OP was a question about why LR's displayed temperature is different from what is set in the camera when using a user-set Kelvin number WB.

I just set the WB to a specific K value of 6000, and when I imported it into LR with no presets or any other adjustment, it initially came in with a value of 5700. If you choose 'Daylight' for WB value, LR sets it to 5500, but 'As Shot' is 5700 (not the 6000 set in camera).

I see why the question now. I've read your earlier response a couple of times, but I must not have eaten enough brain food because it is pretty hard for me to comprehend !

Wilt wrote in post #12006781
I just set the WB to a specific K value of 6000, and when I imported it into LR with no presets or any other adjustment, it initially came in with a value of 5700. If you choose 'Daylight' for WB value, LR sets it to 5500, but 'As Shot' is 5700 (not the 6000 set in camera).

I see why the question now. I've read your earlier response a couple of times, but I must not have eaten enough brain food because it is pretty hard for me to comprehend !

Let me try again, but without any guarantee of doing a better job.
I just set my camera to 6000K and took a shot. The metadata says that the As Shot settings were R=2362, G=1024, B=1538. This means that the camera thought to itself, "The Master wants me to do a WB for 6000K. Now I know what the sensitivity of my sensor is to R,G and B light (because my Creator in Tokyo endowed me with that knowledge) so to get the color channels balanced when I make a jpg I will have to boost the R by 2.362 and the B by 1.538." But instead of letting the camera make a jpg I gave the file to LR to play with and LR looked at the metadata and thought, "Hm, it says here that the As Shot was R=2362 and B=1538. I have a profile for this camera (that The Lord Adobe gave me) that embodies what TLA thinks he knows about how the sensor reacts to light, so I'll use that as a base and apply the As Shot corrections to it." But because LR's UI displays the WB in terms of temperature, it has to do another operation also. It has to do a calculation that is the reverse of the one the camera did (from user-requested temperature to multipliers) and figure out what temperature would require that those particular multipliers be applied to the profile. The result is that because the profiles are different, LR comes up with a different value for the temperature.

But it really doesn't make any difference, because that temperature display is really just window dressing. The important thing is that the WB is obtained and will be correct for 6000K light, even though LR thinks it is balancing for 5500K light.

tzalman wrote in post #12007149
Let me try again, but without any guarantee of doing a better job.
I just set my camera to 6000K and took a shot. The metadata says that the As Shot settings were R=2362, G=1024, B=1538. This means that the camera thought to itself, "The Master wants me to do a WB for 6000K. Now I know what the sensitivity of my sensor is to R,G and B light (because my Creator in Tokyo endowed me with that knowledge) so to get the color channels balanced when I make a jpg I will have to boost the R by 2.362 and the B by 1.538." But instead of letting the camera make a jpg I gave the file to LR to play with and LR looked at the metadata and thought, "Hm, it says here that the As Shot was R=2362 and B=1538. I have a profile for this camera (that The Lord Adobe gave me) that embodies what TLA thinks he knows about how the sensor reacts to light, so I'll use that as a base and apply the As Shot corrections to it." But because LR's UI displays the WB in terms of temperature, it has to do another operation also. It has to do a calculation that is the reverse of the one the camera did (from user-requested temperature to multipliers) and figure out what temperature would require that those particular multipliers be applied to the profile. The result is that because the profiles are different, LR comes up with a different value for the temperature.

But it really doesn't make any difference, because that temperature display is really just window dressing. The important thing is that the WB is obtained and will be correct for 6000K light, even though LR thinks it is balancing for 5500K light.

Now THAT made sense! So the POTN simplified explanation morphs into:
So given that Canon knows the way its sensor responds to different colors, it knows how to bias the RAW data. Adobe, on the other hand, had to reverse engineer (intelligent guess) the bias values, and in the process its computation comes out with a somewhat different WB temp number than what Canon's did, even though both render a neutral gray value to a gray card exposure.

So, I'm glad that we all agree that our brains mush together!

I know this is an old thread but say you set the wb to 6000k and when imported into LR it shows up as 5700k. What if you adjusted the wb in LR back to 6000k and exported to jpg. If you compare the two jpg's (one SOOC and the other adjusted to 6000k in LR) will their wb be slightly off? Would you have been better to keep the "as shot" wb value in LR (in this,case 5700k) to match wb of SOOC camera shot?

The LR version would not be neutral. It would have a slight red/amber bias. Unless you could get a copy of the Canon profile, convert it to a .dcp file and install it in LR.

But you might discover that you like the warmer rendition. The technically correct WB is not always the right WB.

Gotcha. I agree sometimes WB is a personal choice