• Please use real names.

    Greetings to all who have registered to OPF and those guests taking a look around. Please use real names. Registrations with fictitious names will not be processed. REAL NAMES ONLY will be processed

    Firstname Lastname

    Register

    We are a courteous and supportive community. No need to hide behind an alia. If you have a genuine need for privacy/secrecy then let me know!
  • Welcome to the new site. Here's a thread about the update where you can post your feedback, ask questions or spot those nasty bugs!

Colour Accuracy

Dierk Haasis

pro member
Here are two images of the same subject, taken at the same time with the same equipment. Optimisation was done with Nikon Capture NX:

WB259-01.JPG


WB259-02


Let's talk about which is more accurate, which looks better and what actually happens here with regard to our eyes-brain vs. camera-picture.
 

Michael Fontana

pro member
Dierk,
I don't think one is less accurate than the other. It's just a different mood of the same thing. Disclaimer: I haven't seen the building in reality.

Personally, I prefer the greenish.
A tecnically perfect WB isn't alwith the good one - I suppose the first is so.
And heck - every image is a interpretation, too:

The green °layers° of the 5 floors are better bound to the green turf in the foreground.
 

Asher Kelman

OPF Owner/Editor-in-Chief
Dierk,

I love the green one too! The yellow light in the first version is ordinary. We would say, yes, the lights are on. The second is unusual and has magic. Part of this might be the relationships between the colors Green and Blue in our brains wiring as opposed to yellow and blue.

Now what happens to this impressive picture when the building is corrected and made orthogonal, Will it lose life?

Asher
 

Michael Fontana

pro member
Dierk,
Now what happens to this impressive picture when the building is corrected and made orthogonal, Will it lose life? Asher

I doubt it; the round shapes of the building will be more accentuated with a orthogonal cam-position.

Dierk - BTW, who is the architect ? - It looks netherlandish ;-)
 

Dierk Haasis

pro member
Yes, it is the new office building An der Alster 1. Before I write something on the topic - colour rendiction and accuracy - some more facts about technique and intend:

The photo was taken with a D200 set to Auto White Balance, using the 10.5 mm fish eye, corrected to rectilinear mapping [believe me, this was intended beforehand and it it is necessary].

I used Capture NX to get one of the renditions to reflect what my eye-brain saw.

Let me add that the photo I use here as an illustration, while not bad, is far from optimal. The original idea was to get a composition with a slightly outlandish angle [metaphorically speaking] based upon the porthole look of the building's windows. Something like a Sci-Fi movie look at a space ship. Sadly the area around the building is not really helpful - flat to downward sloping, a body of water with adjacent main streets, lots of trees around. Doesn't much help that Hamburg is a large, dog-turd-smelly construction site. At least the sky and the moon were nice to me.

At least my take looks much more interesting than the boring, over-processed, edge-on at the blog referenced.

As the title of the topic says, this is about colour accuracy, not my feeble attempts at interesting architectural photography*. Lately colour accuracy and white balancing has taken over the place of megapixels when it comes to photographic discussions - not the least because yet another WB tool is currently marketed heavily. Curiously only very few voices put white balancing into perspective - and even less people know the basic physial and physiological [hence creative] principles.

Most of the time discussions centre on some abstract ideals, almost never using real-world examples, disguising themselves as "scientific" by using numbers and measurements regardless of their usefulness. Isn't it interesting that this thread hasn't started a debate on colour accuracy as it should have? At the same time one thread on this forum is absorbed by it. Another prominent forum runs hot with [meaningless] discussions on colour accuracy. Admittedly the latter is more concerned with the accuracy of different tools, which is even more disingenious.

Accuracy is a synonym for 'right', thus I own you the answer to the question which of the two renditions is right. And that is the one answer I will not give, yet. I'd like to see more people giving their opinions and arguments.




*Though the photo can be used to illustrate a point I made about documentary vs. interpretative photography on the 'skewed horizon' thread.
 
Accuracy is a synonym for 'right', thus I own you the answer to the question which of the two renditions is right. And that is the one answer I will not give, yet. I'd like to see more people giving their opinions and arguments.

Hi Dierk,

Since I don't know for sure whether the walls inside are green or not, all I can say is that one of the samples is presumed to be correctly color balanced and the other is not.

My guess would be that the green is the result of fluorescent lighting, and thus the first sample is closer to being perceptually correct, or 'accurate', and for me more pleasing.

Bart
 

Cem_Usakligil

Well-known member
..My guess would be that the green is the result of fluorescent lighting, and thus the first sample is closer to being perceptually correct, or 'accurate', and for me more pleasing...
Hi Bart,

I agree with your logic on the guessing part (i.e. the first sample being more accurate). Strangely enough I find the second one more pleasing, but I cannot say which one is the real McCoy :). If I am pushed I'd say the second one against the logic, eventually.

Cheers,

Cem
 

Michael Fontana

pro member
In night- or eveningshots of buildings a photographer automatically runs into these WB-questions, because of different light sources.

Another recent example is here
Its situation is different, as the emphasis is on the interior, meanwhile Dierks shots is clearly a exterior.

As for the green, my guess is Fls, as well - with 4'000 K. That one just needs a little magenta-shift, CC 30 m.... still remembering the insane filtertesting fom the analog days :-(

Theory:
But why do we prefer the °tecnically° wrong one? (= green)

Because our visual experience, or visual memory knows, that Fls do have colorshifts, and a tecnically accurate WB looks more generic. Plus we like colors....

And as we have seen it on photos many times, we accept it to look °natural°.
On the very first photo that manifested these cast, probably the architect told the photographer: hey, that's wrong, it's not a green light....
 
Theory:
But why do we prefer the °tecnically° wrong one? (= green)

Because our visual experience, or visual memory knows, that Fls do have colorshifts, and a tecnically accurate WB looks more generic. Plus we like colors....

Whether we prefer one over the other, depends on what our goal is. Do we want to record the architect's creative intent (shapes/light&shadow/space/...), or do we want to create an artistic image (which basically means that anything goes). How about selectively changing the interior green cast to blue (like the sky) to create an sensation of 'coolness', or to orange to create a color contrast (making the building jump forwards).

I do not entirely agree about the visual memory of green fluorescents. The Fls come in "warm" or ''white" flavors, the human visual system (HVS) will always correct to see white objects (e.g. a piece of paper) as white. We may have developed some memory of perceptually incorrect white balanced photographs, but that is more a flaw (because the light isn't green) which stems from tri-chromatic recording of a mismatched spectrum. It also kills human skin color, which may be artistic, but not even close to correct.

I'm not saying that we shouldn't play with color balance/white balance. In fact we can use it to get our intentions across in a more or less subtle manner, but then we intentionally divert from perceptually correct. Again, it's all about our goal, our intent.

Bart
 

Michael Fontana

pro member
Bart: "Again, it's all about our goal, our intent."

I can't more than agree.
Some personal preferences interact as well, as the circumstances of the shots do.

Talking about °artistic vs documentation°, the rigid border between these two has gone:
A archi-shot is - today, more than ever - a documentation and a interpretation at the same time.

In the 40's, we had that rigid idea of archi-photography beeing purely documentary; kinda neutralising all the °exterior° influences of a building-image - but later on we became aware, that this is just another °style°, only.

Today, I'm much more allowed to have people, etc in the archi-shots, than, lets say, some 15 years ago. The image of the architecture has changed itself as well, due to the discussion of postmodernism in architecture itself, in the 90ties. Not beeing advocat of °anything goes° I feel more free to integrate some °exterior influences° - as long as they aren't anecdotic - meanwhile they had strictly to be banned, some years ago.

Back to topic:
>We may have developed some memory of perceptually incorrect white balanced photographs, but that is more a flaw (because the light isn't green) which stems f<

Not a flaw anymore, if it makes the image greater, or more interesting looking. Most of the architects will today accept that!

Personally, I know about the tecnically accurate WB, but I' ve no problems to break this laws, if it enhances the interpretation to the intended side.
 

Asher Kelman

OPF Owner/Editor-in-Chief
Accurate?

For shooting at a target one can measure the distance by which the shot misses or not. Here dierk we have no basis. For matching colors, accuracy means different things in different circumstances.

For clothes, separates, ie jackets and skirts sold separately to accomodate different figures, can seem the same in the store but outside in sunlight the differences become obvious. So there's a practical use-specific sense of accuracy.

For showing fashion in a magazine, one has to be able to look at the pictures in different lights and the colors have to be seemingly the same and not garish and ugly.

So for this building, what is the intent?

Obviously they are not using this to build and extention for matching the colors. So why would one use the term "accurate"?

Asher
 

John_Nevill

New member
Sorry to come in late on this one, but I actually prefer the colour rendering of the first one.

Perhaps I prefer to see the lights white rather than green. Strange!
 

StuartRae

New member
If you open up the shadows it becomes even more interesting, and may give a clue to Cem who knows the building in question.

WB259-03.jpg


The sky and the foreground colours don't change much, but the building is dramatically different.

My guess, based on the colour shift middle right on the second image, is that the first one has the 'correct' WB, but I'm curiously drawn to the second one.

Regards,

Stuart
 

Dierk Haasis

pro member
Yes, Stuart, I changed the WB only for the building - note to JPEG shooters: go RAW! - since that was the one object [apart from the street lamp to the left] showing up completely different from what I saw on scene.

I will later on chime in with more, including my take and another interesting colour comparison.
 

Dierk Haasis

pro member
My stance: There is no 'accurate' colour [reproduction]. At least not in any meaningful way.

In theory you can measure colours in the scene and adjust the photo file until the values are as close a fit as possible - for monitor as well as printing. That is what physicists do all the time. But that's not how we perceive colours. It's even less how we use colour [or any other characteristic of an image]. We go for effect*.

A few weeks ago I first noticed An der Alster 1, mainly for its form, it looks like a space ship crossed from Moebius' designs with Douglas Adams and an Apple computer. Without researching it further I assume the architects wanted what I called the porthole look simply because many architects designing for Hamburg are desperately trying to fit some seafaring reference into their buildings.

Back then I saw two people in one of those panorama windows on the side, making for quite an interesting picture. Obviously I wanted to revisit that. Driving by one night I saw this slightly outlandish look with the fluorescent light on. Exactly what I wanted! Last week I went with my tripod to explore the building; one result is the photo on top.

The second rendition, with the green house, is what we actually see when standing in front of the house at night. The camera's Auto White Balance of my D200, correctly, corrected towards white, coming up with the first rendition. As Stuart showed, the light is actually white, not magenta tinged.

Interestingly this does not answer the question what 'accurate colour' is. Physically correct is the greenish rendition since the light used is in the green frequencies. Thus the camera is right to correct for this so the white walls look white - which is what we would see touring the offices. If 'accurate colour' means 'that what we perceive' - a viable definition - the green rendition is correct since that is what you see at night from the outside.

Our brain adapts to different lighting situations by comparing what it sees to what it has seen, coming up with corrections. We all know that. Sometimes the system fails [to good effect] because comparisons are missing or the "wrong" colour is overwhelming:

WB-04.jpg
WB-03.jpg


Everyone has probably seen the motive once or twice, most likely with the orange rendition. Like the camera - and the same as with the building - our brain sees the orange version when on the scene! Two exceptions: the lamps themselves and the tarmac. Our brain has models for them, "knowing" that light is white and tarmac some kind of grey.

The second rendition is very close to the correct colours but has major problems with the off-white tiles on the walls, getting a bit too much blue. BTW, I used a WhiBal for the correction. The tarmac, the pavement, the green tiles and the yellow tiles on the ceiling are fine. I had to boost the colour - thus the surplus blue in the off-whites - because otherwise the whole thing would have looked very slightly tinted b/w.

In contrast to the building shot this one does not exhibit mixed lighting. Still, when we are on the scene we do not see the correct colours but the orange one. And again we have no meaningful way to talk about accurate colour.

My personal preference, BTW, lies with the green house because it looks more interesting [and coincidentally is what we see in reality]. In the case of the tunnel I side with the corrected version simply because the other rendition has been used too often.**

Admittedly the genre I used is not the one usually called up in these discussions, disputants talk about portraits. I think I do not need to come up with examples as everyone here should be aware that subjects do decidedly not want accurate colour. The only genre where right colours are important - up to a point - is product photography. Mainly because the customer wants his goods to look in the pic as in reality. That is where proofing comes from.







*A linguistic explanation is in order: 'Effect' as I use it here means the emotion elicited by the viewer [or listener]. In the vernacular 'effect' often means the technique as applied to a photo etc. The latter is probably better called 'special effect'.
**There's some other photos I did which gain from the orange look.
 

Michael Fontana

pro member
.......... The only genre where right colours are important - up to a point - is product photography. Mainly because the customer wants his goods to look in the pic as in reality. That is where proofing comes from.

You forgot the reproductions of paintings, etc.
If the word °accurate° can be used, it'l be in that context, as there's a well defined original.

While the image reproductions are shot most of the time in good light aka studioflash, some museums - even good ones - have unstable (!!) mixes of fluoreszent and daylight, depending on the daylights intensity: if the sun is behind a cloud, some fluoreszent light is added automatically, to get the desired brightness. No chance to shoot a accurate WB, as the light trmprtatur and tint might change in 1 second...

So even in that context, one has to adapt the WB.
 
Top