1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

What is the real D800 sunlight white balance?

Discussion in 'Nikon Chat' started by ChrisNewman, Jun 6, 2020.

  1. ChrisNewman

    ChrisNewman Well-Known Member

    According to my D800 manual, the ‘Direct sunlight’ white balance setting has a colour temperature of 5,200 K. But if I load a RAW file shot with that white balance setting, Capture NX-D shows the ‘Recorded Value’ as 5215 K. If I investigate the adjustments available in Capture NX-D, the drop-down menu includes ‘Direct Sunlight’ at 5200 K.

    I appreciate the difference is trivial, but I’m investigating the amount of headroom before RAW NEF files become over-exposed compared to JPEGs, and I’d like to be able to edit NEF files to match JPEGs as well as possible. If I set ‘Direct sunlight’ white balance, does my D800 apply a value of 5200 or 5215 K when creating JPEGs in camera? Is the ‘Direct Sunlight’ drop-down setting in Capture NX-D likely to give the best possible match to JPEGs when editing NEF files in Capture NX-D, or would I do better to set a value of 5215 K?


    Chris
     
  2. GeoffR

    GeoffR Well-Known Member

    I don't have a D800 so I can't be sure but I do know that the D2 was able to determine the colour temperature of the incident light, the white thing on the front of the prism housing is a sensor. This sensor was dropped on subsequent cameras because, I recall reading, they were able to measure the ambient light by another means. If the D800 can do the same then it is possible that it is applying a small correction to your pre-set white balance. You might find out more in the handbook.
     
  3. ChrisNewman

    ChrisNewman Well-Known Member

    Thanks for the suggestion, Geoff. The camera offers a couple of options for Auto White balance, which I used to use until a few years ago I took a shot of some primroses under a green woodland canopy, and got a shot of white flowers with yellow centres! Since then I’ve tried to choose the appropriate option from the camera’s list of light sources. It’s straightforward on the D800, with a dedicated button that allows settings to be selected and adjusted with the command dials (I assume your more recent D4s are similar). (I do save RAW files, so can correct things when I forget to adjust white balance for the current conditions). There is also the opportunity to set a chosen colour temperature, or a Preset based on a reference, but I’ve never tried either of these.

    It is possible to fine-tune the camera’s options, but that’s also something I’ve never done, so ViewNX-i displays White Balance: Direct sunlight, 0, 0.

    There’s no mention of any feedback from the available light to adjust the camera’s presets, and I wouldn’t expect that to happen, given the availability of two Auto settings. I assume most cameras uses information from the metering system to set colour temperature when using the Auto settings.

    If I see some sunshine tomorrow, I should try choosing colour temperatures. These can be set at intervals of 10°K, so if I try 5200 K, 5210 K and 5220 K I could check whether Capture NX-D reports these same values. But I think I’d need a very clear sky and a carefully chosen target for sufficiently stable conditions to stand a chance of seeing which of those Presets was most similar to the ‘Direct Sunlight’ setting.


    Chris
     
  4. PeteRob

    PeteRob Well-Known Member

    I don’t use Nikon but I’m very surprised that the camera WB setting isn’t respected. I guess your experiment will tell. The setting should make no difference to the raw file data other than in the embedded jpg, it is only used in the colour conversion and can be freely changed in post.
     
  5. GeoffR

    GeoffR Well-Known Member

    It may not be a case of "not respecting" the WB setting more one of the preset being slightly different from a manual setting 2115K as opposed to 2110K or 2120K though why Nikon would chose to do things that way is beyond my imagination.
     
  6. PeteRob

    PeteRob Well-Known Member

    I misread the post - it was late! I thought the raw file was reporting a different colour temperature to that which was set in camera and to the default preset in the Nikon softwares.

    It has never occurred to me to check presets for colour temperature. I suppose, in software it is possible that the values used for presets might change in time but I can't say that I have compared software versions over the last decade. I mostly find AWB satisfactory but my judgement in altering the green-magenta shift is not good and I'll sometimes use the daylight preset in post to steer my adjustments. Woodland light, with the green of the canopy, is very difficult to deal with. That is the only circumstance that I have tried a custom white balance and was surprised at how much difference it can make.

    Silly question, born of ignorance, doesn't the Nikon raw software do this? For Canon the supplied software will replicate the in-camera jpg processing. The main reason I stopped saving jpgs in the camera was that I could always make them in post using "camera" settings if ever I so wanted. This can be done as a batch process so there is nothing really to do, other than have a cup of tea while the computer chirrs away.
     
  7. ChrisNewman

    ChrisNewman Well-Known Member

    Before making this post I’d checked several shots taken over the last few years ‘Direct sunlight’ white balance preset, and I’ve now checked some shots taken with my SB-910 flash, where I’ve been setting the ‘Flash’ white balance for five years, some where I’d set ‘Cloudy’, and one I found where I’d accidentally set ‘Cool-white fluorescent’ for an outdoor shot!

    In all cases the drop-down menu in Nikon’s Capture NX-D matches what is given in my D800 manual, but the settings that Capture NX-D displays as the ‘Recorded Value’ are different from these, but consistent for every shot taken with that preset. All of the presets, as applied by Capture NX-D, have a ‘Tint’ value of 0, but the ‘Recorded Values’ for the presets I’ve investigated, other than ‘Direct sunshine’, have a negative ‘Tint’ value.

    The manual and Capture NX-D have a colour temperature for ‘Cool-white fluorescent’ of 4,200 K, but Capture NX-D’s ‘Recorded Value’ for a shot taken with this preset is 3738 K, ‘Tint’ -3.40.

    The manual and Capture NX-D have a colour temperature for ‘Flash’ of 5,400 K, but Capture NX-D’s ‘Recorded Value’ for all shots I’ve checked taken with this preset is 6516 K, ‘Tint’ -0.06.

    The manual and Capture NX-D have a colour temperature for ‘Cloudy’ of 6,000 K, but Capture NX-D’s ‘Recorded Value’ for all shots I’ve checked taken with this preset is 5985 K, ‘Tint’ -0.01.

    So the difference between the colour temperatures of the preset and recorded ‘Cool-white fluorescent’ and ‘Flash’ values are far greater than the differences for ‘Direct Sunlight’ and ‘Cloudy’. I’m particularly puzzled that the ‘Recorded Value’ for the shot taken with the ‘Cool-white fluorescent’ preset has a very strong ‘Tint’ value (understandable, given the uneven spectra from fluorescent lights), yet the Capture NX-D drop-down setting has a zero ‘Tint’ value.


    Chris
     
  8. ChrisNewman

    ChrisNewman Well-Known Member

    I was assuming that a JPEG taken with a particular preset on the camera would be the same as a JPEG from a RAW file with that preset applied by the appropriate Nikon software (Capture NX-D). That’s why I’m surprised and puzzled that Capture NX-D is suggesting that the values used when recording the JPEG are different to those in the D800 manual and the matching Capture NV-D drop-down menu settings. I want to find out whether the colour temperature (and tint setting) used by the camera when creating JPEGs are those listed in the manual, or those reported by Capture NX-D.

    Chris
     
  9. GeoffR

    GeoffR Well-Known Member

    In capture NX-D you can change the white balance in 1K steps so if you want a very specific setting it is possible. However, when you look at the settings you can apply a preset or the "recorded value" or, obviously make corrections as you see fit using the colour temperature and tint sliders. If you click the WB you will see the sliders and, as you know, that will allow you to see the recorded values.

    You can also see the values in View NX-I as I am sure you know. If you look at a NEF these values can be adjusted but looking at a JPEG made from a NEF there is no adjustment available and you cannot read the recorded values.

    I have just opened an image from last week and in View NX-I the tint shows as 0 but in Capture NX-D it shows as -0.39 the camera is set at 0. I cannot explain why Capture should be displaying a different value from that set in the camera. I checked with an image from the other body and that has a -1.65 tint. I wonder whether it has anything to do with the lens in use?
     
    Last edited: Jun 7, 2020
  10. ChrisNewman

    ChrisNewman Well-Known Member

    Thanks Geoff for that interesting suggestion, but the shots I’ve mentioned checking so far have been taken with a total of five different lenses, and the ‘Recorded Values’ for colour temperature and tint don’t vary between lenses. In any case, Capture NX-D should have access to the same data on the lens as the camera used when creating its JPEG, so ought to offer to produce a similar JPEG (although I don’t yet know whether this would need the preset white balance or the one Capture NX-D displays as the ‘Recorded Value’!)

    I notice that the disparity with the ‘Flash’ white balance is so large (it claims 5,400 K, little different to ‘Direct sunshine’ but reports 6516 K, appreciably higher than ‘Cloudy’) that when we get back to some reliable continuous sunshine, I should be able to take some test shots, in sunshine but using the ‘Flash’ preset, that will show which of these values set in Capture NX-D best matches what the camera is using for JPEGs. But I would really like to know just what colour temperatures my camera is using, and why, rather than simply finding out by trial and error how to get similar results from Capture NX-D.


    Chris
     
  11. ChrisNewman

    ChrisNewman Well-Known Member

    After asking the same questions on another forum, again without enlightenment, I asked Nikon. After I pointed out why an initial answer was fatuous, they referred my query back to Japan, and I finally got a definitive response just before Christmas. I’m told Nikon use different "color adjustment axes” for different light sources, and the "color adjustment axis suited to sunlight", applied when white balance is set to a “Recorded Value” of 6516 K gives a similar appearance to their "color adjustment axis suited to flash lighting" with white balance set to 5400 K in Capture NX-D. I confirmed this by switching an image between a “Recorded Value” setting of 6516 K and a “Flash” setting of 5400 K, and failing to notice any change. Yet if I set “Recorded Value” at 5400 K or “Flash” at 6516 K, the colour balance of the image changed substantially.

    Nikon also confirmed that the colour temperature value for the white balance used is saved in the NEF file; given their use of different "color adjustment axes”, some description of the adjustment axis selected must also be recorded.

    I’m very relieved that the different white balance values reported for the same lighting under a different source name, which seemed anomalous to me, are intentional, rather than an oversight by Nikon. I would still like to understand how such different white balance values can be considered as equivalent for different light sources. I had assumed the colour temperature used to describe a light source would approximate to the mean of its visible light spectrum. But Nikon didn’t respond positively to my request as to whether they could refer me to a general optical information source that might enlighten me on the topic. If anyone is aware of an explanation I could study, I’d be very grateful.


    Chris
     
  12. Learning

    Learning Ethelred the Ill-Named

    You might try Prof Bob via AP, or Thom Hogan on this. They are very different people I would think and might need approaching in different ways.
    Personally I don't care what Nikon do to generate their jpegs. I either shoot Raw or Raw plus jpeg. In the latter case if the jpegs look ok I use them, if not I process the raws to taste.
    I don't get hung up on correct white balance. I don't do scientific documentary. I process photographs to suit my mood.
     
  13. Learning

    Learning Ethelred the Ill-Named

    The setting makes a difference if you process the raw in ViewNX-i or Capture NX-D. By default the Nikon raw processing engine takes account of camera settings. Other raw processors may treat the raw files differently.
     
  14. ChrisNewman

    ChrisNewman Well-Known Member

    Thanks for the suggestion, but now I’m reassured that Nikon’s software is handling my photos properly rather than getting mixed up when assigning colour temperature, it’s only to satisfy my curiosity that I hoped Nikon could point me at some information that would explain why different "color adjustment axes” would need very different colour temperatures to match colours. So I don't want to plead with experts to explain to me something that no longer seems to have any practical consequence. However, when editing RAW files in future I’ll aim to set the light source that I think was appropriate for the shot, rather than leaving the setting on “Recorded value”, in case there is some slight effect on tints (otherwise I don’t understand why they would go to the trouble of using different "color adjustment axes”.
    I always shoot RAW + JPEG, and have much the same view regarding editing, except that sometimes, for example when editing shots of a group walk to contribute to the group’s website, I’ll take the middle route of looking at the effect of quick and dirty edits to JPEGs, such as using “Auto Smart Fix”. But I don’t find the time to edit nearly as many RAW files as I’d like to.

    Chris
     
  15. cliveva

    cliveva Well-Known Member

    I leave my D800 on auto W/b, shooting a grey card, then get on with what I want to shoot. When setting w/b in PS, it tends to be warmer than the camera chose.
     
  16. PeteRob

    PeteRob Well-Known Member

    It seems not. White balance controls let you adjust red-blue (coded temperature) and green-magenta (coded tint). It seems credible to me that the end-points in these adjustments could be tailored to different illuminant sources or colour profile mappings.

    I’ve been digitising my film-stock and found some pictures of my late brother’s wedding. Having gotten the results together I realised I’d got his suit varying in colour from brown (overcast outdoor, no flash) to blue (flash dominant lighting). It isn’t credible that he wed in a brown suit, unlikely to be blue, most likely grey so I normalised the suit across the pictures and it had remarkably little effect on other colours - the dominant other feature in each frame being the (several bright shades of grey), white wedding dress.

    I’ve not looked to see if LR changes colour/tint for WB (on camera “as shot” if the colour preset is changed e,g from portrait to landscape. It might well do.

    I sometimes “warm” or “cool” a shot using the temperature slider. Occasionally it looks wrong. Setting a daylight preset, which changes tint, then changing temperature, can give a more pleasing effect.
     
  17. ChrisNewman

    ChrisNewman Well-Known Member

    I like the idea of nailing colours by shooting a grey, or better still colour checker card. But the majority of my shots are taken when I’m out walking. In Britain there are often rapid changes as the sun goes in and out of gaps in the cloud, and I might move into shade. I’m not prepared to update a reference frequently enough to try to keep up with that.

    Chris
     
  18. ChrisNewman

    ChrisNewman Well-Known Member

    I realise that different light sources will have different spectra, and so I can believe that setting "color adjustment axes” for various light sources might allow more precise tweaking than simply adjusting colour temperature and tint. But I don’t see an obvious way to ascribe a colour temperature to a light source other than to average the visible spectrum in some way. So I still don’t understand why flash light with a colour temperature of 5400 K should be more similar to outdoor light with a colour temperature of 6516 K than to outdoor light at 5400 K!

    Chris
     
  19. PeteRob

    PeteRob Well-Known Member

    Me neither. I think something has been lost in translation somewhere. I suspect a clash between colour temperature notation and white balance adjustment notation. When I calibrate my monitor I have white point options. I use D50 - the option for pre-press whereas there is also an option for “outdoor photography” of D65 which two, as you say, you’d think of as worlds apart (5000K vs 6500K) in terms of white balance adjustment.
     
  20. ChrisNewman

    ChrisNewman Well-Known Member

    I’m afraid I’m not sure what you mean by “a clash between colour temperature notation and white balance adjustment notation.” But the effect is definitely real, rather than being due to misleading terminology. As I stated previously, I don’t notice any change if I switch an image between a “Recorded Value” setting of 6516 K and a “Flash” setting of 5400 K. Yet if I set “Recorded Value” at 5400 K or “Flash” at 6516 K, the colour balance of the image changes substantially.

    Chris
     

Share This Page