Post processing preferences

[DanK]

I've gotten into a debate with a more experienced night photographer about white balance, and I would be interested in your opinions.

Starting in late afternoon, the color balance in remaining light shifts dramatically as night falls. During "blue hour," for example, the light is very blue.

Normally, a "pure" white balance adjustment fully removes a color cast of this sort. If you shoot a neutral target under these varying conditions and apply the right WB, the color of the image of the neutral target will stay neutral.

The question is how much of the evening color casts one should remove. I recently posted a blue-hour photo that only partially removed the color cast:



Elsewhere, I posted a more neutral version that more fully removes the blue hour color cast (although what counts as "neutral") depends a bit on which boat transom one uses as a target):



Which do you prefer? Can you explain why? or would you prefer something in between?

thanks

Dan

[Thlayle]

I've gotten into a debate with a more experienced night photographer about white balance, and I would be interested in your opinions.

Starting in late afternoon, the color balance in remaining light shifts dramatically as night falls. During "blue hour," for example, the light is very blue.

Normally, a "pure" white balance adjustment fully removes a color cast of this sort. If you shoot a neutral target under these varying conditions and apply the right WB, the color of the image of the neutral target will stay neutral.

The question is how much of the evening color casts one should remove. I recently posted a blue-hour photo that only partially removed the color cast:



Elsewhere, I posted a more neutral version that more fully removes the blue hour color cast (although what counts as "neutral") depends a bit on which boat transom one uses as a target):



Which do you prefer? or would you prefer something in between?

thanks

Dan

Dan,

I can go either way with this specific example. I love blue hour shots but I normally go for a color range closer to the second example, especially where the night sky is more dominant in the framing. I really dislike a blue cast in Milky Way compositions -- they are not how the night sky looks to me.

Otherwise, in shots like this that are not so much about the night sky, I am good with the blue cast. I would say for this shot, I might like something in the "in-between" category.

Randy

[xpatUSA]

As you probably know, Dan, the conventional literature tends to emphasize how our visual system compensates for lighting to a great degree thus causing us to see white as white, i.e. neutral-looking, unlike our dumb cameras. What is less common is specific advice as to when or even if to apply a "color imbalance" to a scene. Such advice is often written in SPV (Standard Photographic Vague).

For your boat shot, perhaps balance-pick each boat transom and average the resulting temperature and tint values. Kind of like a local "gray world" WB analysis. Apply that average and then mess with the color-balance or color level sliders to taste - which is itself subjective and can not be defined.

I sometimes ask this question in these kind of discussions:

What is the correct white reference for a traffic-light illuminated scene with no other lighting in it?

[Manfred M]

One thing that made a significant impression on me during the print making masterclass that I finished earlier this month was a comment made by the instructor; "colour is an opinion".

The example he used is that companies that are in the business of selecting colours for products (he used Crayola the makers of children's wax crayons as an example) will use focus groups / committees rather than a single person in making colour decisions. He mentioned that when he made test prints for Yousef Karsh to review, there would always be 10 slight variants that Karsh would review. He noticed a trend that Karsh favoured prints with a slight yellow bias versus totally neutral or slight magenta bias that Michael favoured.

If I am shooting at blue hour, I tend to process the images on the cooler side, like your first version. If I shoot at golden hour, my bias tends to be toward the warmer tones. In both cases, I make a deliberate effort to look at the scene during the shoot and tend to make my PP direction based on what the light is doing in the scene from my perpective.

[Shadowman]

My preference between the two images is sort of split, the more pleasing image is the second as the boats and dock look more natural in tone, however symbolically and perhaps atmospherically the blue tones of the water and reflections of the sky are normal. I think balancing the two areas would be an interesting alternative.

[DanK]

Thanks. An interesting and helpful set of comments.

I haven't read any scientific studies of this, but I think that the notion that our brains compensate for color differences is overstated. I think we partially adjust. We perceive golden hour and blue hour differently from open sunlight or shade. Ted, do you have any references that actually evaluated the degree to which we compensate?

For that reason, I'm reluctant in some cases to remove the color cast entirely. Like Manfred, I tend to process golden hour shots on the warm side and blue hour on the cold side. However, looking at other photos from the past, I realize that I have been inconsistent. For example, in the case of some blue hour photos, I have left the sky or reflections in the water blue but come closer to balancing the rest.

In the case of this photo, there were two color casts: the sky overhead was blue, but the sky in front of the boats was magenta after sunset.

so at this point, my inclination would be to remove only some of the color cast, perhaps a little more than I have removed in the first photo, but not to make it truly neutral, as in the second. I don't think the second mirrors how I perceived the scene at the time.

I'm thinking of going out in blue hour and taking a series of shots at different fixed Kelvin values to see which comes closest to how I actually perceive the scene at the time. (I would have to view the LCD while shading it from the blue light in the environment.)

[xpatUSA]

Thanks. An interesting and helpful set of comments.

I haven't read any scientific studies of this, but I think that the notion that our brains compensate for color differences is overstated. I think we partially adjust. We perceive golden hour and blue hour differently from open sunlight or shade. Ted, do you have any references that actually evaluated the degree to which we compensate?

I don't recall any such, Dan, and am struggling to think of what measure might be used for "the degree".

Perhaps a good indication of the range over which we compensate is the range of CIE standard white references as shown here in the LUV color space:



I can image a Kodak R27 white card looking stubbornly "white" for most of us under the above standard illuminants.

I wonder what CCTs represent the 'blue' and 'golden' hours ... anybody?

Another teaser: what WB is correct for Sodium street lighting?

[Antonio Correia]

Just a few lines about this matter.

I usually tend to warm my photographs as they look more pleasant to my eyes.
However, I agree with John on a third alternative of the two posted images.

It is indeed a question of taste and feeling...

[DanK]

I started doing a bit of reading on color constancy. Not surprising: it's very complex, and I don't think one can find an answer as simple as mine. I've picked up a few interesting things already. One is that physiologically, color constancy requires a broad spectrum of light. That may be germane in night photography. Another is that some of what experiments in the past had identified as color constancy was actually something entirely different: relative color constancy, that is, the ability to recognize a constant difference between colors.

I'm going to have to go back to some blue hour scenes to try to sort this out for myself.

[Manfred M]

I started doing a bit of reading on color constancy. Not surprising: it's very complex, and I don't think one can find an answer as simple as mine. I've picked up a few interesting things already. One is that physiologically, color constancy requires a broad spectrum of light. That may be germane in night photography. Another is that some of what experiments in the past had identified as color constancy was actually something entirely different: relative color constancy, that is, the ability to recognize a constant difference between colors.

Dan - I feel this is a very solid summary. While there is a basic scientific basis for what we are trying to achieve as photographers, it plays a very limited role in what we end up with as a final image. Physiology, psychology and creativity are probably even more important.

[xpatUSA]

... While there is a basic scientific basis for what we are trying to achieve as photographers, it plays a very limited role in what we end up with as a final image.

Sorry Manfred, I don't understand this statement. Take a simple example: exposure, for which the scientific basis is aperture size and shutter period. I'm wondering why those would only play a "very limited" rôle in the final image. To my simple mind, exposure is a major factor in the final image.

Physiology, psychology and creativity are probably even more important.

Seems to be a variation on the oft-stated theme of "the camera maketh not the photographer" ...

[Manfred M]

Sorry Manfred, I don't understand this statement. Take a simple example: exposure, for which the scientific basis is aperture size and shutter period. I'm wondering why those would only play a "very limited" rôle in the final image. To my simple mind, exposure is a major factor in the final image.

The discussion has been around colour in these shots and my comments have been directed at that subject. I don't recall making any comments regarding exposure.

Getting a "proper" exposure is always going to be critical to getting a good base image to work from, so ideally no crushed shadows or clipped highlights.

That being said, had we been discussing portraiture in these lighting conditions, I would have said that getting the skin tone to look "right" is critical and the rest of the colours don't really matter in that situation. I know one photographer who will use heavily gelled flash in mixed lighting situations to shoot a model and will then colour correct to get appropriate skin tone and have really strange background colours. The work can be quite effective.

[DanK]

Some additional visual data:

I should have looked more carefully in response to Sharon's post in the earlier thread. It turns out that setting white balance off one of the transoms caused a substantial magenta shift. I reduced that and exported the file with the reduced magenta shift and three color temperatures. The first and third are the same temperatures as the two in the OP, 4000 and 4900K. The second is a compromise, 4300 K.

4000K:




4300K




4900K

[xpatUSA]

The discussion has been around colour in these shots and my comments have been directed at that subject.

Pardon me. I read the statement as general.

I don't recall making any comments regarding exposure.<>

Never said that you did.

[Rufus]

I won't even try to be scientific about this. I can only offer my opinion in qualitative terms.

I prefer the middle shot (4300K). This I feel conveys the most realistic feeling, albeit of a well lit night rather than blue hour. I can identify with this version.

The first is just too blue for me, but I can't help feeling that the colour would be more "acceptable" to my eye/brain if it was in a cityscape rather than water. Irrational but true (for me only, of course).

The third image still has quite a "cold"/metallic feel to it and is quite attractive but not as realistic as the middle one. I also notice that the colours in the reflections and on the sides of the boats are a little less vibrant in the third image.

[xpatUSA]

I won't even try to be scientific about this. I can only offer my opinion in qualitative terms.

I prefer the middle shot (4300K). This I feel conveys the most realistic feeling, albeit of a well lit night rather than blue hour. I can identify with this version.

The first is just too blue for me, but I can't help feeling that the colour would be more "acceptable" to my eye/brain if it was in a cityscape rather than water. Irrational but true (for me only, of course).

The third image still has quite a "cold"/metallic feel to it and is quite attractive but not as realistic as the middle one. I also notice that the colours in the reflections and on the sides of the boats are a little less vibrant in the third image.

In the world of color there is a property called the "just noticeable difference" (JND). In over-simplified terms, any difference smaller than that can't be seen by "most of us". The difference is measured on a chromaticity diagram - usually CIELAB but sometimes CIELUV like the one I posted earlier - literally as the distance between two colors or, in this context, the distance between two white references.

The difference between 4000K and 4300K can be calculated in u'v' terms by referring to para 3.1, where it also says that 1 JND is 0.0013 :

http://files.cie.co.at/738_CIE_TN_001-2014.pdf

I estimate that the difference between Dan's 4000K and 4300K on the diagram is about 0.007. Being quite lot more than the JND, the difference is, well . . . . noticeable.

LOL

[Rufus]

The difference between 4000K and 4300K can be calculated in u'v' terms by referring to para 3.1, where it also says that 1 JND is 0.0013...I estimate that the difference between Dan's 4000K and 4300K on the diagram is about 0.007. Being quite lot more than the JND, the difference is, well . . . . noticeable.

LOL

Fair enough

But to be fair to me, too, I was attempting to answer Dan's original questions which were "How much of the evening color casts should one remove" and "Which do you prefer? Can you explain why? or would you prefer something in between?" To these there can be no definitive or scientific answers as they ask about our taste and our reactions to and perceptions of the images and the real world which Manfred refers to such as psychological, physiological, etc.

[xpatUSA]

Fair enough

But to be fair to me, too, I was attempting to answer Dan's original questions which were "How much of the evening color casts should one remove" and "Which do you prefer? Can you explain why? or would you prefer something in between?" To these there can be no definitive or scientific answers as they ask about our taste and our reactions to and perceptions of the images and the real world which Manfred refers to such as psychological, physiological, etc.

Yes, I understood what was meant by "I can only offer my opinion in qualitative terms. I prefer the middle shot (4300K).".

Wasn't disagreeing with that.

[billtils]

Enjoyed reading this thread, but when I go back to the original one my personal view remains that to me this is a lovely image. The various edits in this thread may give a more accurate representation but the original is the one that passes the "would I hang it on my wall" test.

[DanK]

Thanks to everyone. This has been an interesting and useful thread.

It's clear that we filter out much of the blue cast during blue hour, but I don't know how much remains in our perception at the time. I have to read more and experiment more. And I actually can't decide which of these many versions I prefer, except that I don't like the truly neutral one.