This content has been marked as final. Show 14 replies
>First of all I don't care if the temperature reported by ACR agrees with the preset temperature quoted by Nikon. What puzzles me is the discontinuity in the values reported by ACR when going from 'flash' to 'cloudy'. Is this a problem with ACR or with the Nikon preset values for these two settings?
A full answer to your question would require a response by Thomas Knoll or some other ACR expert with knowledge of how Nikon records the white balance and the algorithm used to apply the white balance. However, I can bring up a few points which may shed some light on the matter.
One should first consider what white balance information Nikon cameras write to the NEF file. The following table refers to the D70 and similar Nikon cameras before encryption of the WB.
The settings of daylight, flash, tungsten, etc are written as descriptive strings, not actual degrees Kelvin. Nikon considers daylight to be 5200K, but Thom Hogan writes that this value often produces a blue result and 5400K might be better.
The bias value refers to fine tuning between the above descriptive settings, approximately 100 Mired per increment. Finally, the red and blue coefficients are recorded (green would be redundant). Some raw converters such as DCRaw use these coefficients for the "as set" white balance. There is nothing corresponding to the ACR tint, which adjusts in the magenta green axis, whereas the color temperature slider adjusts in the yellow-blue axis. The degrees Kelvin figure specifies the white balance only for a black body radiator.
Thus there several ways to interpret the WB data. Mr. Knoll has written previously that ACR attempts to produce a neutral white balance appearance rather than the WB numbers. An interesting experiment would be to determine which converter (ACR or Nikon Capture) gives a more neutral white balance on your color checker with the above settings.
Thanks for the response, Bill.
I understand that the numbers themselves don't tell the whole story. I should have mentioned that there is a *visual* discontinuity. That is, if I arrange the images so that what I see is a gradual progression from less to more yellow, then the order would be (using the names chosen by Nikon) "daylight -> cloudy -> flash". That simply doesn't seem correct.
I too, would like to know what Mr. Knoll has to say about this. I would also like to hear what a user of Capture NX or Bibble sees as the 'as-shot' temperature when doing this crude 'test'.
What to you see in a similar 'test', Bill?
>What to you see in a similar 'test', Bill?
I haven't yet done such a test, but would appreciate hearing from anyone who has done so.
Personally, I don't usually use the presets but rather take a custom WB from a gray card or include a neutral reference in one of the pictures and use the ACR eyedropper. I've had good results with both methods, but the second is usually slightly more accurate.
I admit that I also sometimes use the D200 auto WB for daylight shots. :)
I normally include a gray card in a text image or use an ExpoDisc. I'd just like to know what is going on with the presets or ACR.
Well, the response from Nikon is about what I expected: They can't say why a third-party NEF converter would give white balance values other than those for which the D200 is calibrated. Unfortunately, they didn't respond to my specifric question so I submitted it again :)
As I indicated before, I don't really care about the specific temperature/tint values. What I'm trying to determine is WHY there is a discontinuity in the temperature/tint values when going from daylight -> flash -> cloudy. There is also a visual discontinuity in that I don't see a progression from less to more yellow when the file is displayed using the as-shot values from ACR.
So, one more time: Will someone with a D200 please repeat the simple test described in my original post and report the temperature/tint values from ACR and any temperature information avaiable from Capture, Capture NX, Bibble or any other converter.
I do not have my D200 camera or results with me. I originally left my WB on flash and all from a morning of shooting came back w/ 6150, -9 in ACR. No flash was used.
That got me curious and I did a crude test of going through all the settings in indoor light. Flash at 0 was always 6150, -9.
I can also confirm daylight at 0 setting giving me 4950, -3 consistently under indoor lighting. I plan on trying it in daylight but I suspect it will stay at 4950, -3 based on my flash experience. I believe my tungsten was different since I don't have it at 0.
I don't remember about the others.
Since my current workflow is ACR/Bridge I didn't look at Capture or Bibble.
Hi I've noticed the same thing just this past weekend and have been searching for answers since! What triggered my recognition of this issue was doing a portrait shoot under studio strobes (daylight balanced) on the D200 daylight setting. Upon loading them into Lightroom I was looking at the same thing and wondering why Lightroom's daylight setting was far warmer (and slightly incorrect) than the D200 as shot setting. So I opened the nef file in picture project just to get Nikon's rendition and the difference in color saturation and WB really made the image pop! So after a lot of reasearch (taking up the rest of the weekend) it appears that with RAW, Lightroom can read NEF's but it loses all of the in camera settings (save for perhaps WB = As Shot) Lightroom has it's own presets of WB that we can apply at will or tweak... but they don't match up to the D200 because Lightroom is an ACR/DNG editor not a NEF editor. So I am now working on a preset that will "do all the in camera work" so that the NEF's I look at in Lightroom look like what my settings in camera are doing. I am really hopeful that in a future update Nikon will give API access to adobe to "unlock" the NEF files and that Lightroom will automatically apply what the camera did. This is where I'm at so far!
Bob, i did your test and get the exact same numbers.
>Lightroom can read NEF's but it loses all of the in camera settings (save for perhaps WB = As Shot) Lightroom has it's own presets of WB that we can apply at will or tweak... but they don't match up to the D200 because Lightroom is an ACR/DNG editor not a NEF editor. So I am now working on a preset that will "do all the in camera work" so that the NEF's I look at in Lightroom look like what my settings in camera are doing. I am really hopeful that in a future update Nikon will give API access to adobe to "unlock" the NEF files and that Lightroom will automatically apply what the camera did.
The white balance with ACR (Lightroom uses the same raw conversion engine) and Nikon cameras has been discussed before. WB is the only camera setting that ACR reads.
When I first began using ACR, I regarded its inability to read the camera settings as a problem, but now I don't really care about the these settings except the WB. When shooting in the field, I am mainly concerned in capturing the image with proper exposure and composition. The optimum tone curve, contrast, saturation, etc can be difficult to determine in advance. When these parameters are determined in editing, they can be applied en masse to images with the same characteristics.
I'm surprised to see this thread still 'active'.
My question regarding the visual color discontinuity resulting from the D200 temperature presets was ultimately forwarded to Nikon Japan. The answer was (drum roll please): The D200 is operating within specifications. In other words, Deny, Deny Deny! Granted, I don't *really* care. Nevertheless it galls me to get such a response from Nikon.
I ran a similar test on D200 presets and get the same temperature and tint values that Bob listed above. I also recorded values for the six fine tune values (-3, -2, -1, +1,+2,+3). I have not tried Nikon's RAW converter, but if I input the temperature and tint (6150, -9) for images taken with the D200 and SB800 flash using ACR the image appears too red. If I select "flash" in ACR I get 5500K and tint of 0, which looks much better. Does the Nikon RAW converter get correct color with the in camera presets? Why such a large difference between Nikon's flash values and ACR (6150 & -9 vs 5500 & 0)? The same applies for all other presets.
The preset icons don't match anything I know of, the custom white balance function is convoluted, and playing with menus while taking photographs is absolute nonsense. I'm involved in framing and focusing usually.
When you set 5000K, you're just telling the camera that the light source is 5000K, you're not placing a 5000K "gel" over your lens. Quite the opposite!
Take a test shot and dial in the Kelvin. Or is it the anti-Kelvin?
This issue has been discussed a dozen times in these forums recently. There is of course no precise replication of NEF WB in ACR 4 or LR or anywhere else--it is at 'best' an approximation by the Adobe engineers. Samo Samo for any Nikon proprietary settings. If you want to come close to Nikon values, you won't get them by any other than indirect calibration settings that you devise--and in that case you will only come close. Better yet, just use your eye--it is already calibrated to your 'photographic vision' :)
Anyone absolutely needing Nikon camera values will simply have to go to NX and die of impatience using its godawful antiquated gui and slow processing. A Fate 100 times worse than tweaking in ACR IMO :)
Since I don't know the target of some of the recent comments allow me to reiterate:
(1) I don't care about the actual temperature/tint values.
(2) Whatever the temperature/tint values are, I don't care one bit if the agree with the Nikon values or not.
What I DO care about and the reason I contact Nikon over a number of months is the fact that despite the published values in the D200 manual being monotonic in temperature, there is a distinct *visual* discontinuity in images when viewed in 4(!) different NEF converters. And this discontinuity appears at the same point in the test image sequence.
But as I said above, it really doesn't matter.