Ad
Ad
Ad
Pages: [1]   Bottom of Page
Print
Author Topic: What is the correct WB for a glow-worm shot . .  (Read 6062 times)
xpatUSA
Sr. Member
****
Offline Offline

Posts: 304



WWW
« on: March 17, 2013, 03:53:47 AM »
ReplyReply

 . . or a luminous paint shot, or any self-luminous object? With reflected light, it's easy - out with the neutral-colored card for a custom in-camera setting or later, click on a neutral color to fix color cast. Suppose you shot in RAW but accidentally had the cam set to tungsten WB - what would you correctly reset your raw converter WB to?

The glow-worm in JPEG format (NEF file long gone):




Here's a Glycine Airman watch in the dark. Same question. It has Luminova luminous paint on the hands:



Here's the paint spectral emission curve, if that helps:



I still had the raw file for the watch shot. I found that varying the WB from shade to daylight (ACR 5.4) gave quite a large shift in hue from 113 degs to 137 degs - quite a color shift, IMHO.

Could go on to mention traffic lights, car back lights, etc. . . .

I'm sure there's a simple answer: just haven't found it yet  Embarrassed

Logged

best regards,

Ted
stamper
Sr. Member
****
Offline Offline

Posts: 2804


« Reply #1 on: March 17, 2013, 04:12:20 AM »
ReplyReply

what would you correctly reset your raw converter WB to?

I don't think .... correctly...... is appropriate? Whatever suits your idea of good colour rendition would be suitable?
Logged

xpatUSA
Sr. Member
****
Offline Offline

Posts: 304



WWW
« Reply #2 on: March 17, 2013, 05:00:20 AM »
ReplyReply

what would you correctly reset your raw converter WB to?

I don't think .... correctly...... is appropriate? Whatever suits your idea of good colour rendition would be suitable?
Yes, perhaps I should explain the use of that word in this post. By 'correctly' was meant a setting that produce the same color in the output image as was seen in the object. That is to say, if a perfect color measuring device pointed at the object gave some value in, for example, XYZ tri-stimulus values then the same device would give the same values when pointed at the output image, all other things being equal.

Putting that more simply, a correct setting would make the color of the object and the image the same. In the RAW image of the luminous paint, the green color looked OK at 5500K, 6500K and 7500K WB settings but the actual hues, as measured by PhotoShop were quite different - which goes to show how much the brain compensates for such things.

Logged

best regards,

Ted
hjulenissen
Sr. Member
****
Offline Offline

Posts: 1683


« Reply #3 on: March 17, 2013, 07:25:29 AM »
ReplyReply

I assume that if you put a picture-frame of a general image inside the same room that the image was shot in, there would be no need for WB. I.e., WB is about removing/reducing the influence of the scene lighting color temperature so that the image will appear as if it is lighted by the light source wherever you are viewing the image. This might only be true up to a point, you may want to keep the warm glow of a candle-lit scene even when viewing the print in daylight.

If this is a reasonable laymans description of the motivation behind WB, then I dont see how to apply it to a self-luminous object.

-h
Logged
xpatUSA
Sr. Member
****
Offline Offline

Posts: 304



WWW
« Reply #4 on: March 17, 2013, 10:42:06 AM »
ReplyReply

I assume that if you put a picture-frame of a general image inside the same room that the image was shot in, there would be no need for WB. I.e., WB is about removing/reducing the influence of the scene lighting color temperature so that the image will appear as if it is lighted by the light source wherever you are viewing the image. This might only be true up to a point, you may want to keep the warm glow of a candle-lit scene even when viewing the print in daylight.

If this is a reasonable laymans description of the motivation behind WB, then I don't see how to apply it to a self-luminous object.

-h

Thanks Hju (Hjule?)

I fully agree with your "laymans description". The problem is that cameras, at least mine, always apply a WB i.e. you can't not apply it!

Is there perhaps a 'neutral' value of WB that can be deduced from the literature that abounds on the subject of color perception, I wonder?
Logged

best regards,

Ted
hjulenissen
Sr. Member
****
Offline Offline

Posts: 1683


« Reply #5 on: March 18, 2013, 03:05:38 AM »
ReplyReply

Thanks Hju (Hjule?)

I fully agree with your "laymans description". The problem is that cameras, at least mine, always apply a WB i.e. you can't not apply it!

Is there perhaps a 'neutral' value of WB that can be deduced from the literature that abounds on the subject of color perception, I wonder?
If your display is calibrated for a D65 whitepoint, would simply dialing in a 6500K WB mean that end-to-end WB is more or less constant?

I am guessing that self-glow objects can look more realistic on a (self-illuminating) screen that on reflecting paper?

-h
Logged
theguywitha645d
Sr. Member
****
Offline Offline

Posts: 970


« Reply #6 on: March 18, 2013, 09:52:05 AM »
ReplyReply

WB is not fixed until you process, so whatever you set the camera to is just a placeholder and does not effect the image.

If the channel you are imaging has its center wavelength in clearly in the green channel, you really have nothing to worry about. Either set the WB to daylight (5500K) or whatever color temperature of the light source used illuminating the sample.

However, if there is a bleed to the R or B channels, then you have a color problem. But a color in a file does not represent a wavelength at that point. What you do is make the white light image look natural and then use selective color to give the green you want. It is possible that your green is out of gamut or it is clipped.
Logged
xpatUSA
Sr. Member
****
Offline Offline

Posts: 304



WWW
« Reply #7 on: March 18, 2013, 11:22:18 AM »
ReplyReply

Quote from: hjulenissen
If your display is calibrated for a D65 whitepoint, would simply dialing in a 6500K WB mean that end-to-end WB is more or less constant?
Quite possibly so. Playing around with Bruce's calculator mentioned below does indicate that if an image comes into the camera with a color of XYZ then the RAW file gets RGB values according to our, or the camera's, best guess at the illuminant. Which I'm beginning to think makes the question in the OP impossible to answer, there being no illuminant  . . .

Quote
I am guessing that self-glow objects can look more realistic on a (self-illuminating) screen that on reflecting paper?

. . . impossible to answer because, instead of the WB being set for the scene illuminance it is instead set for the output. That it is to say, a colorimeter test would show that my uncalibrated monitor would show a different color to your perfectly calibrated monitor and different again to theguywitha645d's printer output (under whatever lighting he uses)

WB is not fixed until you process, so whatever you set the camera to is just a placeholder and does not effect the image.
Er - thank you.

Quote
If the channel you are imaging has its center wavelength in clearly in the green channel, you really have nothing to worry about.
Although it asked a specific question, the OP was trying to be general and theoretical, not about green as such. Green self-luminous images are all that I had to hand. How about a car amber side light next to a red rear light and a blue cop light, all in the same image?

Quote
Either set the WB to daylight (5500K) or whatever color temperature of the light source used illuminating the sample. However, if there is a bleed to the R or B channels, then you have a color problem. But a color in a file does not represent a wavelength at that point. What you do is make the white light image look natural and then use selective color to give the green you want.
In the OP, I should have said that the objects are not illuminated by another source which is what I meant by 'self-luminous', so there is no light source illuminating the example. Good point about wavelength, though. Bruce Lindbloom uses the term 'dominant wavelength' in his CIE color calculator, accessible from here. Always fun to play with, many choices of illuminant and white point correction matrices, color spaces, etc. Interestingly, if you set an XYZ there and then change the illuminants, the RGB values change but the CCT and dominant wavelength do not. Makes sense - further confirming that WB is only to do with reflected light and that camera makers do not expect accurate color rendition of Christmas tree lights shot in the dark?

Could I ask what was meant by 'a bleed to the R or B channels' although I'm not asking for trouble-shooting of my images.

Quote
It is possible that your green is out of gamut or it is clipped.
Again I was trying to talk theoretically, not really looking for image comments per se. But I agree that out-of-gamut problems can show up. I had one image that puzzled me for weeks - a yellow flower was in-gamut in the camera RAW file but out-of-gamut in Adobe's ACR screen rendition of the working file. Reducing saturation just messed up the image coloration. Other converters e.g. dcraw handled it more gracefully after messing with the clipping options.

I propose that the question does not have a unique answer, in that WB setting varies with the desired output.

All those in favor, say 'aye'.

Ted
Logged

best regards,

Ted
bjanes
Sr. Member
****
Offline Offline

Posts: 2825



« Reply #8 on: March 18, 2013, 12:08:42 PM »
ReplyReply


Although it asked a specific question, the OP was trying to be general and theoretical, not about green as such. Green self-luminous images are all that I had to hand. How about a car amber side light next to a red rear light and a blue cop light, all in the same image?
In the OP, I should have said that the objects are not illuminated by another source which is what I meant by 'self-luminous', so there is no light source illuminating the example. Good point about wavelength, though. Bruce Lindbloom uses the term 'dominant wavelength' in his CIE color calculator, accessible from here. Always fun to play with, many choices of illuminant and white point correction matrices, color spaces, etc. Interestingly, if you set an XYZ there and then change the illuminants, the RGB values change but the CCT and dominant wavelength do not. Makes sense - further confirming that WB is only to do with reflected light and that camera makers do not expect accurate color rendition of Christmas tree lights shot in the dark?

Could I ask what was meant by 'a bleed to the R or B channels' although I'm not asking for trouble-shooting of my images.
Again I was trying to talk theoretically, not really looking for image comments per se. But I agree that out-of-gamut problems can show up. I had one image that puzzled me for weeks - a yellow flower was in-gamut in the camera RAW file but out-of-gamut in Adobe's ACR screen rendition of the working file. Reducing saturation just messed up the image coloration. Other converters e.g. dcraw handled it more gracefully after messing with the clipping options.

I propose that the question does not have a unique answer, in that WB setting varies with the desired output.

All those in favor, say 'aye'.

You raise some interesting questions, but in short I would vote 'aye'.

As I understand the question, you want to produce some semblance of the actual wave lengths emitted by the worm or some other emissive source. The camera is not a spectrophometer and can only attempt to produce a tristimulus metametic match for the wavelength(s) in question. Normal photography includes chromatic adaption to the illuminant, but there is no incident light in this case. If you want to reproduce the appearance of the worm, the chromatic adaption of the observer would need to be taken into account (possibly for mesopic adaption).

If you want to reproduce a metameric match to the actual colors of the source, it appears intuitive to me that you should use a WB of approximately 5500K, where the spectrum is flat and equally weighted to red, blue, and green (see Illuminant E). This works more or less for sunsets and candle light where you want to preserve the reddish spectrum. If you tweak the WB manually, then chromatic adaption would have to be taken into account.

Bill
Logged
xpatUSA
Sr. Member
****
Offline Offline

Posts: 304



WWW
« Reply #9 on: March 18, 2013, 12:39:43 PM »
ReplyReply

You raise some interesting questions, but in short I would vote 'aye'.

As I understand the question, you want to produce some semblance of the actual wave lengths emitted by the worm or some other emissive source. The camera is not a spectrophometer and can only attempt to produce a tristimulus metametic match for the wavelength(s) in question. Normal photography includes chromatic adaption to the illuminant, but there is no incident light in this case. If you want to reproduce the appearance of the worm, the chromatic adaption of the observer would need to be taken into account (possibly for mesopic adaption).

If you want to reproduce a metameric match to the actual colors of the source, it appears intuitive to me that you should use a WB of approximately 5500K, where the spectrum is flat and equally weighted to red, blue, and green (see Illuminant E). This works more or less for sunsets and candle light where you want to preserve the reddish spectrum. If you tweak the WB manually, then chromatic adaption would have to be taken into account.
Thanks Bill,

A good point about illuminant E and yes, I was thinking more of metameric rather than how it looks so to speak.

It all started when I was trying to determine the luminance of watch luminous paint from an image similar to the one in the OP. A truly enormous amount of research went into "what is ISO", "what is color", "when is a sensor deemed to be saturated", "who TF is Bradford", etc. ad naus. But the WB was the one thing that I and the well-respected Doug Kerr never quite saw eye-to-eye on.

By coincidence, I usually leave my WB set close-ish to E, namely 5490K which happens to be the measured (custom) value of my table-top lighting.

I would add that, as a refugee from a couple of other fara, it is most refreshing to be greeted by knowledgeable posts such as those in this topic. Oft-times elsewhere, there is quite a tendency to scorn with copious references to the "Real World" and "The Print" . .  Sad


« Last Edit: March 18, 2013, 12:48:14 PM by xpatUSA » Logged

best regards,

Ted
Pages: [1]   Top of Page
Print
Jump to:  

Ad
Ad
Ad