Three questions about white balance

[William W]

. . . "Dead on" white balance is rarely used and, more importantly, is very rarely necessary.

I just found this, I made it a while ago to show (another member here) how "dead on" White Balance, for most of our Photography, is not really ever used.

These are three frames made of the same area of garden in Sydney AUS in dappled light, (but the flowers are mainly in open shade) made in wintertime at around midday:

Frame 01 is the JPEG SOOC with Auto White Balance initiated: the image would be pleasing to most Viewer's Eyes, it is nice and warm and the flowers appear rich and healthy.

Frame 02, I dumped a Standard Reference Card in the centre of the frame.

Frame 03 is an A/B comparison and contrast between the AWB shot and the manually corrected to very close to 'dead-on' White Balance using the reference card at source.

Most people who have seen this set (and there have been many) prefer the AWB version, which of course has a totally "wrong" White Balance.

The interesting thing is, I have done similar experiments with Students, (i.e. we go shoot flowers in the winter time) and one reason for doing that is, the flowers, in situ, appear quite lush and rich, but they are not rendered so when there is a reference card used and standardized PP is employed. In situ, the brain interprets that the flowers should be nice to look at, so that's how we 'see' them.

WW

[Cantab]

Bill, when I click on the link in your post, I get a message that the folder is "private". Is there another way you could make the three images available? I'd very much like to see the set.

P.S. I found my way here from your post in the "Natural Beauty" thread.

[TonyW]

Bill, I wasn't able to view your reference here. There was a message that said it was marked as private.

It is an interesting question as to how much we should alter an image to make it like we thought we saw it or to make it like we would like to see it, rather than how it really was, if that can actually be defined. As you suggest, the way we see things in real life is influenced by how we expect to see them. But how do we know what it really was and does that question make sense? This is relevant to one of my original questions of how can we process an image in a standard way uninfluenced by our or the camera's expectations. The concept of a 'canonical' light source could be useful here.

Also, in very low light, I understand that the eyes tend to use the rods rather than cones and these have only a black and white sensitivity, if I understand the physiology correctly (a big if). The other day I was trying to photograph the trees outside my window at night time which were lit from the light in the next room. The scene looked pretty much black and white to me but when I photographed it (with a long exposure) it was quite colourful, so that the camera did not see it as I did.

[William W]

^ Bruce, Tony,

My error.
Sorry for the inconvenience.
I simply just forgot that I had to 'unlock' it. There is nothing 'private' about the folio.
It should be OK now.
Please confirm its OK.
Thanks,
Bill

[Cantab]

j

^ Bruce, Tony,

My error.
Sorry for the inconvenience.
I simply just forgot that I had to 'unlock' it. There is nothing 'private' about the folio.
It should be OK now.
Please confirm its OK.
Thanks,
Bill

Bill, yes, it works fine now. It's an interesting illustration. Like everyone else, I prefer the AWB version.

As an aside, when I saw your foldout colour grid in a metal spiral bound book, it reminded me of a Kodak book I acquired in the 1970's. I'll have to dig out my book and see if in fact it's the same thing.

[chauncey]

It's Okay for me...I want dead on white balance when I shoot mixed race portraits.

[TonyW]

Yes, I can see it now. I would prefer the AWB version too. This is interesting because two of my bêtes noires are getting skin tones right, or even recognising what is right, and photographing flowers - two things I will have to work on.

[William W]

. . . It is an interesting question as to how much we should alter an image to make it like we thought we saw it or to make it like we would like to see it, rather than how it really was, if that can actually be defined. As you suggest, the way we see things in real life is influenced by how we expect to see them. But how do we know what it really was and does that question make sense? This is relevant to one of my original questions of how can we process an image in a standard way uninfluenced by our or the camera's expectations. The concept of a 'canonical' light source could be useful here. . .

Yes, the topic is very interesting to me and I thought that you’d be in interested.
Yes, your question “how do we know what it really was” makes complete sense to me.
Yes, I agree the concept of ‘canonical’ could be useful to quantify and reference and explain: I very much appreciated the link that Ted provided.

*

. . . Also, in very low light, I understand that the eyes tend to use the rods rather than cones and these have only a black and white sensitivity, if I understand the physiology correctly (a big if). The other day I was trying to photograph the trees outside my window at night time which were lit from the light in the next room. The scene looked pretty much black and white to me but when I photographed it (with a long exposure) it was quite colourful, so that the camera did not see it as I did.

Yes.

And using that information - if I want to better conceive my Vision for a given Shooting Scenario which will ultimately be a Black and White Image, I sometimes use an ND Filter over my EYE, that makes the scene very dark and, as you mention, more like 'seeing it" in B&W.

Using an ND Filter to assess a scene was especially very useful for me, for B&W Portraiture when I didn’t have the skill levels that I have now: especially for me to provide best advice and instruction upon the CLOTHES SELECTION for the Subjects.

WW

[xpatUSA]

Frame 03 is an A/B comparison and contrast between the AWB shot and the manually corrected to very close to 'dead-on' White Balance using the reference card at source.

Most people who have seen this set (and there have been many) prefer the AWB version, which of course has a totally "wrong" White Balance.
WW

Bill, are you sure that the difference is just due to the WB method?

Your comp image hacked (so much for their "protection", eh?):



Reason I ask is because not only are the in-camera JPEG reds "redder" (approx -10 degrees of hue, HSV model) but they are approx 5% more saturated and somewhat brighter too. (I am assuming that the camera JPEG was set to neutral and that the color profiles are the same, of course).

So, bearing in mind that the in-camera JPEG converter is arguably less fancy than the one in the proprietary converter, do you still believe hand-on-heart that the difference in color is due to the WB method alone?

Your comments themselves about WB are not being challenged, by the way, I'm just a tad concerned about the example.

[William W]

I want dead on white balance when I shoot mixed race portraits.

You probably don't.

OR - it is very likely that we two are defining "dead on", differently.

In my response (Post #5) to the first mention of the term “dead on” White Balance (Post #2), I made the assertion that ‘dead on’ implied that there would be no variation to the Colour (of the skin).

To achieve this, the lighting at the source would need to be as close as possible to an absolute consistent Colour Temperature. This could be achieved using a (well designed) Photographic Studio and (well maintained) Studio Flash.

Then the shoot would need a reference card.

Then the Post Production would need to be carried out using a Calibrated Monitor, in a lighting calibrated working environment.

Then any variance of Colour Perception in the Technician’s eyesight would need to be accommodated – try this quick quiz: How's you color vision? (test)

Then the PRINT or the screen image would need to be VIEWED by the CLIENT in a similarly standardized and controlled lighting room . . . and after all that hullaballoo, there would still be slight variations.

*

On the other hand, the correct use of reference devices such as X-Rite, WhiBal, etc will afford good batch consistency and make easy work flow and get near to ‘accurate’ White Balance/Colour Balance results; arguably consistency of results across the batch is usually more important.

I think that we are just defining what is “dead on” differently. Using my definition –

“DEAD ON” White Balance I have rarely used in any of the more common photography genres: and only for forensic archival records of ARTWORK.

For everything else, including Mixed Race Portraiture, a bit of ‘tweaking to taste’ comes into the mix for the final image, but, using my definition, you might strive for ‘dead on’ white balance I don’t know, but my guess is that realistically you don’t: either by design or by omission of one of the necessary steps to get it.

WW

[William W]

Your comp image hacked (so much for their "protection", eh?)

Haha! Anything on line can be hacked, you’re welcome to it!

*

Your comments themselves about WB are not being challenged, by the way, I'm just a tad concerned about the example.

No need to mention that.
I would not have thought that.
I've always expected that you make the purpose of your comments quite clear.

*

Bill, are you sure that the difference is just due to the WB method?

OK.

It’s taken me a minute or two to understand the meaning of question - but maybe I still don't get it

(I am assuming that the camera JPEG was set to neutral and that the color profiles are the same, of course).

Yes confirmed, your assumption is correct. The in-camera JPEG settings were ‘neutral’.

*

Reason I ask is because not only are the in-camera JPEG reds "redder" (approx -10 degrees of hue, HSV model) but they are approx 5% more saturated and somewhat brighter too.

So, bearing in mind that the in-camera JPEG converter is arguably less fancy than the one in the proprietary converter, do you still believe hand-on-heart that the difference in color is due to the WB method alone.

The raw image was processed in Photoshop.

Yes PS is more sophisticated than Canon in camera JPEG converter.

Yes I agree that there could be, probably are, other influences of non controlled variables.

Though I don’t quite understand that those non controlled variables are particularly relevant to the use of the example, because the purpose of the example is simply to show how AWB can (and often does) make the scene less to be way than “dead on”, apropos White Balance.

So whilst other technical elements of the image might be different, the colour is definitely different, and that's the point of the example.

Does that make sense?

Maybe I am missing your point?

WW

[William W]

. . . when I saw your foldout colour grid in a metal spiral bound book, it reminded me of a Kodak book I acquired in the 1970's. I'll have to dig out my book and see if in fact it's the same thing.

Someone else asked me about that, too.

I'll get to digging mine out, It is one of the Kodak Professional Guides, I have a few others too, such as the Darkroom Guide.

WW

[Cantab]

Someone else asked me about that, too.

I'll get to digging mine out, It is one of the Kodak Professional Guides, I have a few others too, such as the Darkroom Guide.

WW

Yup, I just dug mine out: Kodak Professional Guide, with wonderful detail, inter alia, about various then current films....

[xpatUSA]

Though I don’t quite understand that those non controlled variables are particularly relevant to the use of the example, because the purpose of the example is simply to show how AWB can (and often does) make the scene less to be way than “dead on”, apropos White Balance.

So whilst other technical elements of the image might be different, the colour is definitely different, and that's the point of the example.

Does that make sense?

Maybe I am missing your point?

WW

It does make sense if your point was only that a particular SOOC JPEG taken with AWB had better colors than a converted raw shot with properly corrected WB and then saved as a JPEG - then, yes Bill, I must agree that the colors are different and that the SOOC looks better to the eye.

If, on the other hand, your further point was that the different white balance processing is the only cause of those color differences then I am less able to agree, sorry. I can see that WB can affect a color hue significantly and even luminosity to an extent - but perhaps others here can explain how WB can affect saturation that much.

It's perhaps a bit moot. When I saw the processed raw image I was struck by it's washed-out appearance compared to the SOOC and that's what got me thinking.

[William W]

Ted,

Thanks for taking the time to explain further. I now believe that I understand your point and the reason that you raised it:

It does make sense if your point was only that a particular SOOC JPEG taken with AWB had better colors than a converted raw shot with properly corrected WB and then saved as a JPEG - then, yes Bill, I must agree that the colors are different and that the SOOC looks better to the eye.

Yes, that was sort of the reason for posting the example, but really it was even simpler than that: my point was that AWB may (will often) make the image’s colours NOT the same, compared to if the scene and imaging process were Colour Balanced (white balanced) throughout the whole image process, so that the WB would be ‘dead on’: – particularly this is so if the scene is shot under ‘unfavourable’ lighting.

My general point being: - that we rarely want ‘dead on’ Colour Balance or ‘dead on’ White Balance: the example images were mainly prompted by: wm. c boyer’s comments about ‘dead on’ and the general interest in AWB scribed Tony’s Opening Post.

Please note that my comments regarding wm.’s ‘dead on’ are using MY definition of ‘dead-on’ and what that phrase means to me. I have already commented that I think that wm. and I are using different definitions for the phrase 'dead-on'

*

If, on the other hand, your further point was that the different white balance processing is the only cause of those color differences then I am less able to agree, sorry. I can see that WB can affect a color hue significantly and even luminosity to an extent - but perhaps others here can explain how WB can affect saturation that much.

No, confirming that was not my aim. It was a simple and simplistic experiment.

*

It's perhaps a bit moot. When I saw the processed raw image I was struck by it's washed-out appearance compared to the SOOC and that's what got me thinking.

Not moot. I like thinking, it often travels us to different and worthwhile places that we would not have thought about had we not started thinking in a tangent.

Thanks,

Bill

[george013]

@William,

I don't think you can compare 2 JPG created from 2 different converters. They are by definition different.

George

[William W]

. . . I don't think you can compare 2 JPG created from 2 different converters. They are by definition different.

Yes, the two converters were different.

The two processes used to get to the two JPEGS were different.

BUT - point was not about comparing the JPEG converters with each other.

The point was to compare the results of two different processes - and yes those JPEG results are different.

WW

[Manfred M]

Bill - an excellent discussion on white balance.

In my experience, white balance is more related to what the client wants (and the client can be you). If you are shooting a piece of product photography and the client is Coca Cola, then their product logo colours had better be 100% bang on (and they will have the Pantone swatches out to check!), and the other colours in the image can be off a touch. Likewise, if you are shooting a wedding and the mother-of-the-bride's dress does not look right to her, she will hone in on that "issue" and nothing else will matter. I would also be more concerned about accurate colour reproduction if I were taking an image for scientific reasons.

I recently completed a colour correction course given by a long time commercial photographer, and this is one point that he hammered home; 100% "correct" white balance is impossible once you step out of the studio (and in theory there are things that prevent that from being 100% correct in a studio situation too). There are just too many coloured surfaces that reflect light onto subjects and throw off the white balance and effectively introduce a mixed lighting environment.

When it comes to portraits, I find that I really don't like skin tones that are 100% neutral; I prefer them a touch warmer and even when I shoot a gray card, I will use that as a starting point and will warm them up slightly in post.

Again., the only time I get really particular in getting my white balance "right" is when I print. While I use a profiled and calibrated wide gamut screen on my computer. The room I do my editing in has neutral walls and because it is in my basement office, the light I work under is quite consistent. I do quite well on the x-Rite colour vision test.

http://www.xrite.com/online-color-test-challenge

I can't control the conditions other people see my images under when I post my images on the internet, so I really don't get too concerned about white balance comments. Prints are different; when you and I view the image at the same time, we should see more or less the same image (our colour vision differences being the main variable).

So, "perfect" white balance is most photography is a myth; as Bill has demonstrated in his postings. The end purpose of the image should dictate how we prepare them. We need to understand colour theory, how our cameras differ from how we see and of course, our audience. We need to recognize issues in reproducing the images on different devices. We usually should not get hung up on being 100% "correct" as this can result in a less "appealing" image, which likely bears little or no relationship to the conditions we captured it under.

So once again, there is no right or wrong answer. I generally aim at showing colours the way I remember seeing them, not necessarily the way that the technology suggests is "technically correct". With a few limited exceptions, the white balance call is often more of an artistic than technical call.

[William W]

"colour correction course given by a long time commercial photographer, and this is one point that he hammered home; 100% "correct" white balance is impossible once you step out of the studio (and in theory there are things that prevent that from being 100% correct in a studio situation too). There are just too many coloured surfaces that reflect light onto subjects and throw off the white balance and effectively introduce a mixed lighting environment."

"With a few limited exceptions, the white balance call is often more of an artistic than technical call."

Yes.

Those are the two main points which stimulated my contributions to this thread.

WW

[chauncey]

Interesting results last evening when I took a picture at a lake of my son holding a Passport...
it coughed out perfect skin tones but, when I aimed past him to a white gazebo and took a shot,
the colors were a little funky after syncing them with the Passport numbers.
Adjusting WB based on a shaded white post brought it back in line...300 K difference.