blp shabash 430x45
Inspiring and Supporting Photographers of Australian Birds

Welcome, Guest
Username: Password: Remember me
When posting a new topic, please ensure that you select the correct category for your post in the top drop-down box of the edit window. The default entry is the first category shown on the All Categories page; this is unlikely to be the category that you want. The Category drop-down box will be present if you click the New Topic tab in the Forum menu; if you are viewing a particular category of the Forum and you use the New Topic button in the Category Header section, the drop-down box will not be present, and your new post topic will automatically appear in the category that you are viewing.
Discussions about cameras, lenses, accessories, and image-processing.

TOPIC:

Accurate colour management 4 years 1 week ago #2369

  • Simon Pelling
  • Simon Pelling's Avatar Topic Author
  • Offline
  • Elite Member
  • Elite Member
  • Posts: 241
  • Thank you received: 260
Perhaps we are rephrasing the original question about how to achieve 'accurate' colour to how do we achieve 'desired' colour - which is fine.

If I understand correctly it seems that posts have expressed perhaps three different ways of doing this (apologies if I have misinterpreted anyone):
- using DPP Neutral profile, or the equivalent for Nikon etc, as a reasonably approximation of daylight, and then adjusting white balance to compensate for the different K values of different environments so as to achieve a consistent colour regardless of the light
- using DPP Neutral profile or equivalent, and setting the white balance to 'daylight' or an equivalent K value (5000), to attempt to capture 'observed' colours
- perhaps, if I understanding David correctly, attempting to reproduce some 'memory of the scene' in terms of colour.

I think Bob is perhaps also asking whether the second option is not just as valid as the first.

I think the point of my original post is that I have observed colours in submitted images that are significantly different from what I would expect from my own observations of species, particularly variations which shift towards the 'vivid' end of the scale. This might be because people are just using camera defaults, or raw processor defaults, or have poorly adjusted screens, or it might be a deliberate attempt to achieve impact. If critiquing such images I would be inclined to call the colour out as being 'incorrect'. I expect there would not be a lot of differences between Bob and Ian's approaches in practice as both start from Neutral presets and therefore are likely to result in relatively conservative colour palettes even if tonally a bit different - I expect they would be within the boundaries of differences between different raw processors, different monitors etc. I would think that both might be acceptable in terms of what BirdLife Photography is trying to achieve. However, I still think we should call out the more extreme use of colour for our main image databases.

Incidentally, on reading Canon's own explanation of its Picture Styles, they state that the Camera Faithful profile is intended to achieve accurate, true to life colours. It seems to have slightly higher saturation than Neutral, although to my eye the differences are not huge. Neutral is described as being particularly good as a starting point for post processing given it starts a lower contrast and saturation position.

Simon
The following user(s) said Thank You: David Seymour, Ian Wilson, Glenn Pure

Please Log in to join the conversation.

Accurate colour management 4 years 1 week ago #2370

  • Glenn Pure
  • Glenn Pure's Avatar
  • Offline
  • Moderator
  • Moderator
  • Posts: 252
  • Thank you received: 204
Simon, thanks for your summary. Just a point of clarification though: Canon's picture styles (and that of other camera makers) does not relate to white balance. It is purely a way to change aspects like saturation and sharpness without altering the starting white balance. So picture style is not related to an approximation of daylight or any other white balance setting.

Like you, I would call out a shot if the colour looked really wrong. By that I mean really over-saturated, strong cast or white balance way off. My standard would be both my own experience of observing a species (limited) but mostly sources like bird guide illustrations. Having said that, I think there is room for artistic licence in applying some saturation or rendering in a way that may not be fully true to the bird's colours; examples might be dawn and dusk shots where I think most would accept a warm cast on the colours. Not every shot has to be submitted with bird ID in mind. However, I'm aware this started off as s discussion of accurate colour, not artistic licence.
The following user(s) said Thank You: David Seymour

Please Log in to join the conversation.

Last edit: by Glenn Pure.

Accurate colour management 4 years 1 week ago #2371

  • Simon Pelling
  • Simon Pelling's Avatar Topic Author
  • Offline
  • Elite Member
  • Elite Member
  • Posts: 241
  • Thank you received: 260
Thanks Glenn. I'm not too fussed about where the discussion goes and it has been useful in clarifying my thinking. I think there are two aspects to my original question (on which I would welcome any further comments, of course):

- the more technical aspects of managing colour - is there a preferable target in terms of accuracy, and how to achieve this including in images where there are few if any reference points (eg 'mid-grey) equivalents. What should we as photographers be doing when we look at the file on the screen in DPP, Lightroom etc to know we are within a desirable range of colours? I think this has been answered in a few ways.

- is there a preferred way of doing this for BLP given our role? As a moderator, or as someone critiquing images (eg in the critique gallery, assessing competition entries, etc) what are the things I should be looking for. I'm not suggesting that BLP define a 'policy' on this at all, or even reach an agreed conclusion, but just looking to provoke some thoughtful responses to help me continue to frame/refine my own perspective. Your second para goes to this, I think.

Simoon
The following user(s) said Thank You: David Seymour, Ian Wilson

Please Log in to join the conversation.

Accurate colour management 4 years 1 week ago #2372

  • Ian Wilson
  • Ian Wilson's Avatar
  • Offline
  • Platinum Member
  • Platinum Member
  • Posts: 431
  • Thank you received: 496
A very interesting discussion going here. My take on the subject is that the colour management used to render a bird photograph depends on the purpose for which the image was created. If it is your intention to create an image with the colours of the bird looking something like the reference images we find in ABG, then you will usually need to make some lighting adjustments like those suggested by a number of contributors. These adjustments are in effect trying to produce an image of the bird like it would appear with ‘standard’ illumination. Historically, the standard illumination used by ornithologists and scientific illustrators when describing and recording the colours of birds is what we loosely call daylight or its equivalent. An important function of the BLP image library is to provide reference images enabling users to identify birds, ideally to the subspecies level, to show any sexual dimorphism, breeding plumage, plumage morphs, moult stages and age classes. To meet this objective we need to produce images with reasonably accurate colour rendering. What I have to say in this post is all to do with trying to achieve accurate colour rendering in the sense just described. In writing this I am well aware that we create bird photos for other purposes which are just as worthy such as giving priority to what some photographers call ‘capturing the light’; an obvious example of this is warm images captured during golden hour.

The aim of this post is to providing more practical advice and examples of how to accurately render a bird. One of the main difficulties is that we are dealing with multiple light sources. Even on a clear sunny day we have one primary source and one or more secondary sources making a contribution to the illumination of the bird. The different sources have different spectral characteristics from warm to cool (colour temperature). We can adjust for this using the amber-blue white balance slider or colour temperature slider. Contrary to common belief, these two adjustments are doing the same job, you will get the same result whether you adjust the amber-blue slider or the colour temperature slider. A difficulty arises when part of the bird receives more illumination from one source than another part of the bird. The white balance adjustment is a global adjustment so we must choose which part of the bird has adjustment priority. We can end up with part of the bird having the correct white balance adjustment and part of the bird with a colour cast. To deal with this situation we will need to make colour adjustment on the part of the bird having the colour cast. A common example of this problem occurs in avian flight photography on a clear day when photographing birds with a dark top side, for example, most raptors and many petrel species. The main illumination is the direct sunlight and we use the appropriate daylight or sunny white balance pre-set. This usually gives a reasonably accurate rendering of most of the bird. However, the blue sky hemisphere above the bird is a secondary source of illumination and there will be a component of this light reflected off parts of the top side of the bird to the camera (specular reflection). If the bird has a dark top side, the blue sky reflection will appear on those parts as a magenta colour cast. When I encounter this problem I select the part with the colour cast and decrease the magenta and or blue saturation. A hue, saturation, luminance adjustment will achieve the same result.

A second example of where a single global white balance adjustment needs to be followed by selective area colour adjustment is regularly encountered in pelagic sea bird photography. Many sea birds have a convenient neutral white part on their top side, for example rump or crown, that is an effective spot to use for white balance adjustment. Some species have good neutral grey upper-parts that can also be used. They will usually be photographed flying over the sea and on a sunny day there will be a significant amount of light scattered up from the top few metres of water. This upwelling illumination is usually blue or blue-green and if the under-side of the bird is white, as often with seabirds, then a pale blue colour cast will be observed on these parts. Once again, the remedy for this problem is to make a selection of the affected parts and apply some desaturation of the blue channel.

There are other examples where multiple sources of illumination can cause difficulties similar to those already described. Without going into specific details, another common example is the green or yellow-green colour cast that can affect parts of birds having white under-parts when they are perched over green vegetation. Another example is seabirds having brightly coloured feet; light reflected up off their feet when the birds are standing can cause a colour cast if they have a white belly. A good examples of this are Gentoo Penguin and Red-footed Booby (white morph); what do you do in this case? Leave as is and make an appropriate comment would be perfectly acceptable.

Earlier in this thread I mentioned that colour checkers can work well as colour standards in some situations. In my experience, colour checkers are most useful at set-ups, for example at a bird bath. This kind of set-up can be regarded as a kind of outdoor studio but with one important difference, the ambient light usually changes during a shoot. This can be ameliorated if the colour checker can be mounted close to the action in the same light and colour calibration shots taken a few times during the session. If this is not possible, the colour checker can be temporarily placed in the centre of the field of view at the beginning and end of the session to enable calibration shots to be taken. An average of the before and after colour checker values can be used to generate the working colour profile. The illumination of the bird at a set-up will usually be a mix of light from one of more flashes and the ambient light which will usually have a primary source and a number of secondary sources. I usually aim for about half the light to come from the flashes and use the white balance pre-set for flash and neutral picture style as the starting point for generating the color profile. The first step is to fine-tune the white balance using the neutral density panels on the colour checker and adjusting the brightness so that the measured neutral panel RGB values match the calibration RGB values provided with the colour checker. The next step is to use Hue, Saturation and Luminance (HSL) adjustments in your photo-editing software to bring the measured results for the RGB and CMY colour panels into agreement with the calibration values. When this is achieved to your satisfaction, the colour profile is saved as a recipe for use with the images captured during the session. RGB colour space and HSL adjustments for the colour calibration work well for modelling the output of physical devices like monitors and printers. An alternative colour space which also works is CIELAB which is designed to approximate human visual perception. The X-Rite Color Checker provides calibration data for both sRGB and CIELAB colour spaces.
The following user(s) said Thank You: David Seymour, Simon Pelling

Please Log in to join the conversation.

Last edit: by Ian Wilson.

Accurate colour management 4 years 4 days ago #2376

  • David Seymour
  • David Seymour's Avatar
  • Offline
  • Premium Member
  • Premium Member
  • Posts: 93
  • Thank you received: 65
Apologies for a somewhat tardy return to this topic to add a few more thoughts. My strong impression of the majority opinion in the submitted discussions is that for bird photographs submitted to the BLP galleries, the favoured apporoach is to ensure that the colours in most images are calibrated to a standard, particularly the D65 daylight standard. There are various technical methods for ensuring this, and the result is that species colours can be reliably compared between images.

If I could refer back to Simon's last post for a moment, I firmly believe that what is being sought here is best described as 'calibrated colour', and I still have a problem with the use of the term 'accurate colour' to describe the favoured approach. In my view, 'accurate colour' would be achieved when the colours in the final image closely match the appearance of the subject colours at the time of capture - an approach often used in landscape photography, but not as often favoured in bird photography, for the reason given above. I do note in passing that some of the discussion in this topic referred to allowing the warm light of 'magic hour' to remain un-recalibrated in the final image - but I add that there can be a continuous spectrum of such warmness in colour, whereby depending on atmospheric conditions, latitude etc., it can still be present slightly even at midday (try photographing in Tasmania when autumn forestry burnoffs are in full swing!). Conversely, colours can be cool all day when the weather is heavily overcast.

In regard to the image database represented by the BLP galleries, I see an issue with the viewer not knowing whether the image being viewed was calibrated to the common standard at the time of capture or during PP. One suggestion springs to mind:

An additional field on the image submission page, entitled 'Colour calibration method' with a drop-down selection which might include:

Unknown
Uncalibrated
Faithful to time of capture
In-camera AWB with no further adjustment
In-camera Daylight setting with no further adjustment
In-camera Cloudy setting with no further adjustment
In-camera AWB with grey-card or white point adjustment in PP
In-camera Daylight setting with grey-card or white point adjustment in PP
In-camera Cloudy setting with grey-card or white point adjustment in PP
In-camera Custom White Balance
In-camera Custom White Balance with grey-card or white point adjustment in PP

Incidentally, I am aware that some of the above information can be viewed in the EXIF data if it is all there, but the information is incomplete - the EXIF may show that the camera was set to AWB, but if the colour was further calibrated in PP that typically doesn't show up in the EXIF.

Cheers, David
The following user(s) said Thank You: Ian Wilson

Please Log in to join the conversation.

Last edit: by David Seymour. Reason: Additional words

Accurate colour management 4 years 4 days ago #2377

  • David Seymour
  • David Seymour's Avatar
  • Offline
  • Premium Member
  • Premium Member
  • Posts: 93
  • Thank you received: 65
Further to my last post, a revised suggested drop-down list for 'Colour calibration method':

Unknown
Uncalibrated
Faithful to original scene
In-camera AWB with no further adjustment
In-camera Daylight setting with no further adjustment
In-camera Shade setting with no further adjustment
In-camera Cloudy setting with no further adjustment
In-camera AWB with grey-card or white point adjustment in PP
In-camera Daylight setting with grey-card or white point adjustment in PP
In-camera Shade setting with grey-card or white point adjustment in PP
In-camera Cloudy setting with grey-card or white point adjustment in PP
In-camera Custom White Balance
In-camera Custom White Balance with grey-card or white point adjustment in PP
In-camera preset Colour Temperature
In-camera preset Colour Temperature with grey-card or white point adjustment in PP
Camera calibration applied in PP (e.g. X-Rite ColorChecker or similar)

Cheers, David
The following user(s) said Thank You: Ian Wilson

Please Log in to join the conversation.

CONTACT US

The easiest way to contact us is by emailing us at This email address is being protected from spambots. You need JavaScript enabled to view it.

The Our People page, in the About Us section, contains email links to each of the committee members.