Wednesday, 16 November 2016

Colour - Plumage Morphing

In this Birding Frontiers posting the late, and sadly missed Martin Garner describes a phenomenon often attributed to Siberian Chiffchaff (Phylloscopus collybita tristis), referred to as colour or plumage morphing.  Clearly plumage itself doesn't have the ability to morph, or change pigment colour.  So it must have something to do with light.  What is it about this taxon in particular that seems to attract our attention to what is surely a widespread and common problem?  The photographs which Martin uses to illustrate his point give some clues as to what is going on.  It is easy to get side-tracked on various related matters so I am going to break it down here as clearly as I can.  I have already delved into this subject under various postings in the past.  But as this subject is a direct complement to my recent posting on ghostly birds, I thought I might be useful to pull these threads together once more.

A typical 'grey and white' appearance characterises the 'classic' look of a Siberian Chiffchaff.  But many are less obvious than this, and that is part of the confusion.

White Balance
At the heart of this problem lies the concept of white balance.  The colour of light changes all the time due to scattering caused by the atmosphere.  It just so happens that a dramatic shift in the quality of our light coincides with the arrival of Siberian Chiffchaffs here in Western Europe.  No surprise then that we are perhaps more acutely aware of colour morphing now than at any other time of the year.  I never tire of the illustration below which I lovingly and painstakingly compiled in order to satisfy my fascination with this cyclical phenomenon.  This, in a nutshell is I suspect the main underlying cause of plumage morphing.  

Away from the equator the sun's position in the sky is dictated by time of year.  On the winter solstice here in the northern hemisphere the sun is at it's lowest ebb, bumping along just over the horizon before plunging us into a long night.  This means that sunlight is not very pure, even at noon.  And, by early afternoon the sun's rays are already beginning to dim and yellow significantly.  Returning to Martin Garner's posting it's quite easy to distinguish between the brighter, sunlit, and invariably warmer yellow toned images and the colder images.  The yellow cast in the brighter images is almost certainly a natural effect of low winter sunlight.

Colder, bluer light is associated with pre-dawn and dusk, as well as shadow.  In a posting examining white balance (HERE) I demonstrated simply enough that two different white balance's can exist for the same image.  On a sunny day there is one white balance for direct sunlight and another one for shade.  The reason for this is simple - the blue sky dome projects blue light into the shadows, therefore a bird photographed in the shade on a sunny day is bathed in this cold, blue light.

This point may account for some of the colder looking birds in Martin's posting but not all I suspect.  Its difficult to say if conditions were sunny or overcast in some of the images, which leads to the next important point.  On an overcast day shadows are not blue.  This is because the blue sky is obscured by cloud on an overcast day.  Instead of the sky dome projecting blue light it projects diffuse sunlight which is white (or perhaps yellowish or reddish depending on time of day).  Much like a frosted light bulb, or lamp shade, the sunlight is scattered throughout the cloud cover and thus scattered to earth from that massive diffuser in the sky.

Last but not least, I must of course point out that camera white balance correction is prone to significant error.  Without proper white balance calibration we are always left discussing these birds in a bit of a vacuum.   For instance, in a number of the images posted in Martin's blog I can detect a fairly obvious green cast.  Green is not one of the colours of sunlight as the sun traverses the sky, though a momentary green flash may occur just as the sun dips below the horizon.  As can be seen from the animation above, sunlight passes from blue light to slightly magenta then reddish at dawn through yellow to white and back again.  Thus the green cast to these images has nothing to do with the sun at all.  It is a camera white balance error.  This is not uncommon when a bird is shot against a green background.  As discussed in a posting (HERE) simple white balance correction tools only cater for the normal sunlight colour shift in the yellow-blue axis.  To correct for errors in the magenta-green axis one needs a more advanced white balance correction tool.  

Returning to our subject, the green cast created by an in-camera white balance error in this case in turn adds a false olive tone to the bird, which for another species might go unnoticed.  But as the subject matter is the subtle shade of a dingy warbler, white balance correction is really important.  This leads me to the next point.

The Bold Versus The Bland
So, having established that white balance error is the main cause for plumage morphing why are we not aware of this problem more often?  The problem lies in the nature of the colours we are trying to analyse.  Effectively we are looking at subtle, very low-saturation colours, or the ultra-bland as I like to call them.

I have penned a few postings on the differences between bold and bland field marks.  The conclusion I reached while analysing various parameters was that bold field marks including plumage colours were more 'resilient' to image quality deterioration than bland features.  What I mean by this is it is still possible to accurately gauge bold features in an image of almost any quality, whereas an image must be of higher quality in order for bland features to be so readily assessed.

The Camera Versus The Eye
The points which apply to the camera's ability to capture accurate colour is similarly true of our observations of birds in life.  To make an accurate assessment of the subtle colours of a Chiffchaff for instance, we need good, neutral light.  It must be said however that we do appear to possess a surprisingly effective ability to discern subtle colour differences, even in challenging circumstances, particularly with a bit of training.  As a former quality manager for a photofinishing company I was always amazed by the keen eyes of the more experienced printing operators, who could accurately gauge individual colour corrections to obtain neutral, white balanced prints, based often on very limited visual cues.

The concept of white balance arose because of the need to emulate this adaptation of human vision to work around the challenges of ever-changing light quality.  Sometimes however our vision fails us.  Optical illusions like 'The Dress' viral phenomenon and Beau Lotto's colour cube experiments are a reminder how our eyes can deceive us.  This must surely account for at least part of the explanation for colour plumage morphing.

Colour Quality
Having already devoted an entire section of this blog to colour I won't rehash it all again here.  Suffice to say that accurate colour capture is dependent on many variables.  Key among these is the actual calibration of the camera sensor itself.  Surprisingly in this day and age camera sensors are not calibrated for colour to any recognised standard.  No two cameras, even of the same model will display colours exactly the same.  We resolve this by using a professional tool like the X-rite colorchecker passport.  Because few if any birders calibrate their camera sensors we start out with a certain level of colour bias which we cannot measure or rectify.

Secondly, as outlined above we must then calibrate white balance to account for variations in both natural lighting colour and in-camera white-balance error.  Provided we have obtained a reasonable exposure we now have done the very best we can to 'approach' accurate, representative colours.  I stress the word 'approach' because of course a camera's colour pallet only accounts for a fraction of all the colours we can actually appreciate with our eyes.  So even with the best calibration in place a photograph will never match exactly the colours we see in nature.

I developed the concept of colour profiling as a pinnacle of the study of colour capture in digital photography.  If we can obtain accurate colour captures we can begin to accurately name and profile the subtle colours of the birds we capture.

I think it's clear that colour plumage morphing is a real phenomenon.  When it comes to observation there is a natural lighting explanation, coupled perhaps with a slight optical illusion at work as the brain's white-balance and brightness settings try to make sense of what it is seeing.  When we add the additional quality parameters needed to accurately capture a bird with a camera we must consider other layers of complexity.  I have summarised these with the aid of a quality control tool I have developed (HERE).  These can be grouped as parameters related to image capture, those related to the quality of the light and those related to accurate colour calibration.

This turns out to be a pretty good image of a tristis-type Chiffchaff, made possible by a set of ideal circumstances - a confiding bird, an ideal, overcast day, good neutral mid-day light, a very lucky camera exposure, and finally and most importantly, both sensor and white-balance calibration.  Where there is a will, there is a way and plumage morphing need not be feared!