Blog Image

DAFOS Photo World


To catalogue my thoughts and experiences on things photographic and related topics, aiming to promote open-minded creativity and respect that hopefully contribute in some way to general peace and well-living.

White Balance: Were it comes from, what it is and what to do about it

Things Digital Posted on Sun, February 27, 2022 22:19:19

By Alun Foster

Where it comes from

Objects in a scene reflect light (so we can see them) in a way that depends on their own surface colour and the colours in the illuminating light. Your eyes expect this to be “white” (though it often isn’t) and correct for this in “real life”: light must be a very strong colour before we stop seeing objects and start seeing art.

But our eyes don’t do a good job when looking at a photograph (which, of course, is always art). If the illuminating light source when the photograph was taken has stronger (or weaker) colours, this will affect the colours we see in the photograph, and this must sometimes be compensated for the photo to look “right”. The same is true if it is viewed as a print or on a screen of some sort.

Most natural sources of light have a colour that depends upon their temperature. In the ideal world of physics, they are “black body radiators” of light. (Paradoxically, an ideal radiator of light is also a very good absorber of light, hence ‘black’). Like this:

Take a piece of metal and heat it up

It starts to glow…

First dull red

Then brighter, yellowish, whiter

Then blue-white

So, the colour of the light emitted is related to its temperature. We can say that light has a “colour temperature”, therefore. In reality, the only things to get that “blisteringly hot” are some very distant stars. Our own sun is quite a cool 5600 K, by comparison. This is why 5600 K (or close) is often used as the reference colour temperature.

I’ve made a table of some common light sources’ colour temperatures here, to get a feel for what this means in some scenes we can relate to.

Light SourceColour Temperature
Candle1000-2000 K
Tungsten Lamp2500-3500 K
Morning / Evening sun (clear sky)3000-4000 K
Fluorescent Lamps4000-5000 K
Electronic Flash5000-5500 K
Direct sunlight (noon)5000-6500 K
Cloudy sky6500-8000 K
Shade9000-10000 K 

A couple of things to note about this.

  • Colours that we associate with “warm” actually have a lower colour temperature than the blue tints we associate with things being cool.
  • Cloudy sky and shade show colour temperatures that are higher than that of the light source itself – our sun. The reason for this is the same as why the sky is blue. The air (and clouds) above us tend to reflect away the red tones, leaving the light looking more blue except when it comes at us head-on.

So, to correct for the colour of these light sources, to make them look as if they are close to our “standard “ 5600 K sunlight, we would need to “add some blue” (or take away some red) if they are too low, or “add some red” (or take away some blue) if they are too high on our colour temperature scale.

The effect of colour temperature (simulated)

In these photos, I’ve simulated the effect of various light sources on what we see in the scene.

Flash (indirect)

Window (blue sky)

Incandescent lamp (tungsten)

Fluorescent light (“TL”)

Electronic flash is designed to be close to our natural view of ‘white’. Though it can look quite blue if used directly, diffused flash is probably your best bet for getting natural-looking colours when indoors.

A shaded area (or indoors near a window pointing away from the sun) is basically being illuminated by a huge blue light – the (hopefully blue) sky, so tends to look quite blue as a result. Giving your subjects a blue complexion is not recommended.

Incandescent (tungsten) light, on the other hand, is quite definitely orange coloured, which makes our photos look correspondingly so.

Gas discharge light of the “TL” variety actually emit a spectrum of light that is far from ideal (see later), but we often see the result as having quite a ‘green’ cast (so, we have to “add purple” to compensate for the missing red and blue). More on this type of light source later.

Adjusting for White Balance – in camera

So, what can we do about all this? Especially if you shoot JPEG (though, also useful to get a good reference if shooting raw), your camera will more than likely offer you a number of “White balance” options.

Auto White Balance (AWB). This is by far the best to use for all general-purpose shots. Here, the camera analyses the scene and tries to compensate for any colour cast. It generally works well, except at extremes (tungsten, shade) or when there is no “white” in the scene to guess from.

Presets. Like the name suggests, these are pre-cooked compensation for a variety of fairly standard lighting conditions (tungsten, fluorescent, flash, daylight, shade, cloudy). Simply select a WB preset as appropriate when taking the shot. (Pro tip – you can select the “wrong” white balance preset to get some creative effects. Try “shade” or “cloudy” on autumn colours, for example. Or “tungsten” during “blue hour” for some spooky effects).

Other Options. Many cameras will allow a “custom white balance” to be selected. This usually requires taking a photo of a white object (paper, …) in the prevailing light, which the camera will then use as a reference for adjusting the white balance of pictures taken with the “custom” mode. Some cameras may also offer a feature where you can “dial in” a desired white balance colour temperature (though this is often harder to use well).

Adjusting White Balance “in post”

All image editors will offer some means of correcting for white balance, and most work in more-or-less the same way. While post-processing corrections for WB can be done on JPEG captures, the range of correction you can achieve before weird things start to happen is more limited than when using raw. (See insert to get a technical background for this).

JPEG and raw JPEG format uses 8 digital bits per channel, whereas raw uses the full range of bits available from the sensor – usually 12, or these days commonly 14 bits per colour channel. This not only gives better precision in setting the level of each colour, it also gives more “overhead” for the arithmetic to work with before digital artefacts start appearing. Each extra bit effectively doubles the available range for adjustment, so going from 8 to 12 bits is already a 16-fold theoretical improvement. JPEG compresses the image data by throwing out a lot of colour detail (more than the luminance, or “B&W” part) – it is “lossy” compression. This lost information can be vital in colour-related corrections such as white-balance. Raw file formats use “lossless” (or sometimes, no-) compression, so all information for making colour-related adjustments is to hand.

In all cases, the first step is to call up a preset (most programmes have them). Like the presets in the camera, they use a standard set of adjustments for each of the commonly encountered modes. From these, small manual adjustments to “get it just right” are also possible.

In more difficult cases, you can usually use a ‘dropper’ to pick a “white” colour reference in the image. This could be a white shirt, or paper item in the shot, or even a colour reference card that you include in a test shot. Pro-tip: “white” objects could potentially be “blown out”, so is best  to choose a neutral grey as a reference. White reference cards usually offer a white or a grey surface.

A more awkward “pro tip” when really desperate is to use the white of a person’s eye as a reference.

When in post-processing mode, you can also use local adjustments for mixed lighting cases (defined by masks or layers, … the terminology may be tool-dependent).

Other light sources

All of the above assumes a fairly “uniform”, ideal light source, like a “black body” radiator. Such sources have smooth, predictable curves that describe the wavelengths (colours) of light they emit. No gaps or major spikes. Nice and smooth, like that in this diagram.

Many man-made lights are far from this ideal.

Domestic LED light
    Fluorescent “eco” light

Such lights are far from ideal for photographing subjects where the colour rendering is of any importance. Certain colours are over-represented, or even completely missing (an object of such a colour would appear very dark in the photo). The better-quality lights, especially of LED lamps, are carefully designed and measured to have a “Colour Rendering Index” (CRI), and this should be higher than 95 (or, even better, 98).

For most applications, understanding the light source in a scene and using white balance correction is quite good enough for most applications. For best colour rendering, use flash if you can, and/or a high quality continuous light source (CRI>>95). If all else fails and the colours don’t work for you at all (for example, when there is a mix of illuinating light, this can be very tricky indeed), there’s always conversion to Black and White…

If your particular application requires very good colour rendering, you can go the whole mile and use a colour calibration chart. Such charts allow a complete colour profile of the light source and the camera sensor to be built, to compensate all of these physical side-effects. This is in fact quite simple to do, and is covered in the next section.

The ultimate solution – custom ICC profiles

Colour management standards for digital imaging are established by the “International Color Consortium” – ICC. “ICC profiles” are essentially correction tables that can be applied to different parts of the flow of image information as it moves through the processing chain – coming in from camera sensor through your favourite image editor and going out to the computer screen or printer.

Manufacturers of the software and hardware we use provide profiles for “the best possible average” for their specific products. These standard profiles are certainly adequate or most applications (with the notable exception of monitor/screens, though some manufacturers do now offer calibrated screens – it’s still worth while checking this though). For applications that require better colour rendering, a custom ICC profile can be created for each lighting condition, and for each camera’s unique response to colour.

This may sound complicated, but manufacturers of such tools do make it rather simple to do. This article describes the use of X-Rite tools and Capture One Pro: other options are of course available, but all work in more-or-less the same manner. A brief step-by step guide follows.

Step One – Create a reference image using the calibration card.

Calibrating the light source and camera starts by taking a reference image using a calibration card. While a grey card can be used to detect non-optimal white balance, a more detailed reference for many different colours is needed to compensate in more demanding situations. The colours are not random: In addition to grey at different levels, they include skin-tone colours, a variety of “nature” colours and some prime references.

When taking the reference shot, it’s best to bracket the exposure to be sure you have one where none of colour channels are “burned out” (which would confuse the later calibration steps).

This example shows a calibration set up for a studio environment (recording a painting, using flash), but the same principle applies at an on-location shoot of any sort.

Step two – import the test image into your image processors

Due to the huge variety of raw image formats, colour calibration tools prefer to take in a standardised image format, such a TIFF or DNG. Therefore, use your favourite raw processor to create a TIFF or DNG image. (JPEG from a camera has already been processed and much of the colour information compressed, so is not directly suitable for this type of work). When converting the calibration image, it is important the no additional processing is done when creating the TIFF (or DNG) reference. This means:

  1. Turning off any existing or default ICC profile. In Capture One Pro, this is done in the “Base Characteristics” tab.

    Base Characteristics in normal use.

    Settings required for the calibration image.

    By selecting “No colour correction” and “Linear response”, the data from the camera sensor is in effect being imported complexly unmodified.

  2. Colour correcting parameters in your editor must all be “neutral”. In particular, “White balance” control should be set to OFF or “None” if possible. The image can/should be cropped to include just the calibration card.

Step three – export the test image

Export the image as 16-bit TIFF (or DNG). For this, ensure that no profile or corrections are included with the file. A trick to do this in Capture One Pro is to “include camera profile”, which has in fact been set to “none” in step two above.

The output recipe format for Capture One Pro

You end up with a TIFF file that, unsurprisingly, looks like this:

Step four – create the ICC profile

The example here uses the software delivered with the X-Rite Colorchecker Passport. Start the programme and drop your newly created reference image into its window.

The programme will automatically detect the corner markers and place its monitoring points appropriately on each colour patch. It may need little help; you can manually adjust the corners to get the best match.

Test image in place within the Colorchecker software.

Simply clicking on “Create Profile” does just that. Give it a name of choice and it’s ready for use!

Step Five – using your new colour calibration profile

The newly created profile is used in place of the standard profile you may normally use in “Base Characteristics”, like this:

This profile can now be used when processing all images taken under the same conditions as the calibration shots.

Pixels and Print sizes. Views from a distance.

Things Digital Posted on Mon, February 21, 2022 21:57:16

When printing a digital image, it is important to know and understand how the size of the print, the viewing distance and the number of pixels in the digital image file interact to determine the perceived quality of the resulting print .

A standard reference is that the human eye can resolve about 300 dots per inch (or DPI – which is 118 dots per centimetre) on a 20×30 cm pages (A4-ish) viewed from a distance of 30 cms. Having more dots per inch than this is wasteful: you can’t see them: having fewer dots per inch and the image may start to appear “soft focus”. To assure quality, a little margin is taken – typically 330 or 360 DPI. For simplicity, like most people, I equate “dots per inch” to “pixels per inch”, though a printer will put down very many more dots in the space of each pixel due to the way they work.

If you do the sums, an A4-sized print needs a little over 10 million pixels to achieve 330 DPI resolution (which is roughly 3800 pixels on its longest side: 330 pixels per inch over 12 inches).

At this same resolution, a print at 40 x 60 cms needs 33 million pixels, and an A0-size (84 x 119 cms) needs 140 million pixels!

Hang on… Such huge files is getting to be a bit silly. Indeed, this observation was mentioned to me in a discussion with an artistic photographer who gets excellent, exhibition quality large prints with less than half the size of file this theory would indicate. Investigating further on the Interwebs I came across this useful little table.

Viewing Distance (m)Resolution (DPI)

In other words, the further away from the print you stand, the lower the resolution needs to be for the same subjective quality. In fact, looking at the table you can see that the suggested DPI is simply 180 divided by the viewing distance in metres.

Elsewhere on the same Interwebs you can read that the best viewing distance is about 1.5 times the diagonal of the picture. Which makes sense, as you typically view larger things from further away. Put those together and you find that the DPI you need decreases as the print size increases, actually quite exactly, such that the number of pixels you need stays roughly the same, no matter how big a print you want to make!

Now, this is of course a theoretical case. Some people do go up close to exhibition prints to look at fine detail, so this guideline should definitely be regarded as the “minimum case”. In practice, a file with about 20 million pixels is quite big enough for the size of larger prints most of us need to make.

 If the size of print you want to make seems to call for rather more megapixels than your camera gave you in the first place, it is tempting to use an image editor programme to “upscale” the number of pixels. Indeed, most do an excellent job of this. However, and this is again due to the way printers actually work (the difference between “dots per inch” and “pixels per inch”), my preference is to let the printer’s software do this upscaling: it will be better adapted to the exact requirements.

So, bottom line is – at least, if you fully trust the printing service – to get the best possible print, give the printer all of the pixels you have without upsizing.

(A complete table of paper sizes, required DPI and the resulting sizes is given here. Not for the NL, but for an article in “Resources).

Paper size nameSize (in.)Size (mm)XcmYcmViewing Distance (m)DPIPxXPxyMinimum MPxNote
A101 ” x 1½”37,125 x 52,5 mm3,71255,250,1060087612401,1Size in inches is approximate
A91½” x 2 “37,125 x 52,5 mm3,71255,250,1060087612401,1Size in inches is approximate
A82 ” x 3 “52,5 x 74,25 mm5,257,4250,14600124017532,2Size in inches is approximate
A73 ” x 4¼”74,25 x 105 mm7,42510,50,19600175324804,3Size in inches is approximate
 3¼” x 4½82,5 x 120 mm8,25120,22600194828345,5Size in inches is approximate
3R3½” x 5″89 x 127 mm8,912,70,23600210230006,3Called “9 × 13 cm” worldwide.
4R4″ x 6″102 x 152 mm10,215,20,27600240935908,6Standard 135 film & print size in US, Canada, Australia and India. Called “10 × 15 cm” worldwide.
A64¼” x 6¾”105 x 148,5 mm10,514,850,27600248035078,7Size in inches is approximate
4D4½” x 6″114 x 152 mm11,415,20,29600269235909,7New size for most consumer level digital cameras and Micro 4/3 cameras. Also known as “6D”.[3]
5R5″ x 7″127 x 178 mm12,717,80,335492743384510,5Twice the size of a 3R print. Called “13 × 18 cm” worldwide.
6R6″ x 8″152 x 203 mm15,220,30,384732831378110,7Twice the size of a 4R print. Called “15 x 20 cm” worldwide.
A56¾” x 8¼”148,5 x 210 mm14,85210,394672727385710,5Size in inches is approximate
8R8″ x 10″203 x 254 mm20,325,40,493692949369010,9Can be used for contact prints from 8×10 film. Called “20 × 25 cm” worldwide.
S8R8″ x 12″203 x 305 mm20,330,50,553282617393210,3Closest approximation to A4 (210×297mm), twice the size of a 6R print. Called “20 × 30 cm” worldwide.
A48¼” x 12¾”210 x 297 mm2129,70,553302727385710,5Size in inches is approximate
10R10″ x 12″254 x 305 mm25,430,50,603023023363011,0 
S10R10″ x 15″254 x 381 mm25,438,10,692622620393010,3 
11R11″ x 14″279 x 356 mm27,935,60,682652914371810,8Called “28 × 36 cm” worldwide.
12R12″ x 15″305 x 381 mm30,538,10,732462952368810,9 
S11R11″ x 17″279 x 431 mm27,943,10,772342567396510,2 
A312¾” x 17½”297 x 420 mm29,7420,772332727385710,5Size in inches is approximate
S12R12″ x 18″305 x 457 mm30,545,70,822182622392910,3“30 x 45”
“40 x 60”16¾” x 24½”400 x 600 mm40601,081662620393010,3“40 x 60”
A217½” x 23½”420 x 594 mm4259,41,091652727385710,5Size in inches is approximate
A123½” x 33 “594 x 840 mm59,4841,541172727385710,5Size in inches is approximate
“60 x 90”24½” x 35½”600 x 900 mm60901,621112620393010,3“60 x 90”
A033 ” x 47¾”840 x 1188 mm84118,82,18822727385710,5Size in inches is approximate

Colour management from front to back.

Things Digital Posted on Thu, July 22, 2021 23:54:24

Well, the front part today at least. Regarding the “digital” parts of photography, many articles have already covered the importance of checking that your computer output device (monitor, and eventually printer) are properly calibrated/profiled etc…, so that the colours represented “in the digital domain” (i.e. as numbers in your computer) come out looking like we’d expect them to. Up to now, not so much attention has been paid to what happens before we get into the “digital domain” in the first place – i.e. how the camera we use perceives colours, and also how the lighting we have may be affecting the way it “sees” the colours.

There are basically two things that can affect the way a camera will record the colours in a scene:

  • The light illuminating the scene. For most sources, its “colour temperature” – though other sources like LEDs and low-energy lights have special characteristics too. (These can be weird: they emit light with very irregular spectra, with some colours missing altogether so they could appear as black in your picture!) In general, your camera flash is the best bet where practical, but if you must use continuous lighting, lamps (including LEDs) have a “Colour Rendering Index” (CRI) which must be above 90, preferably 95, if you need to render colours correctly.
  • The camera sensor. Here, problems could be caused by a variety things, but the combined effect of these is that the cameras sensor is a good, but not perfect, converter of light to digits.

Both sources of colour variation can be addressed in one go, by using a colour calibration chart. By photographing this chart alongside the subject of your picture, a “profile” can be built that can be used to cancel out pretty well all of the affects mentioned above.

Looking into what is available, it became apparent that, while support for Adobe products seemed fairly prevalent, other platforms such a Capture One (my photo processor of choice) seemed less well served. X-Rite Pantone® seemed to be the only reputable source of the necessary tools (my ColorMunki Photo monitor calibrator comes from there too). On the basis of this I purchased an X-Rite “ColorChecker passport”, and downloaded its accompanying software.

The colour temperature of the light can, to a great extent, be compensated by selecting the appropriate “white balance” setting, either when you capture the image or in post processing. (If using JPEG, you really must do this when shooting). The camera’s default “profile” can also be used of course, but is this precise enough?

To be honest, for the vast majority of pictures, the answer is yes. But for the specific application I have (reproducing paintings etc…), it may not be ideal. Thankfully, companies the likes of X-Rite, Adobe and Phase One make it very easy to set up the necessary “camera profile” that provides compensation for all these non-ideal phenomena (at least when using raw format), by including the colourful patches in a test shot or two, using RAW format. Process the test shots, assuring that the default profile is turned off (in Capture One Pro, this is under “Base Characteristics” – see the screen grabs: they are similar in Lightroom) and setting the response curve to “linear”, we then need to isolate the colour checker by cropping to it, to make an image suitable for the X-Rite software to analyse.

  Select these for camera calibration

Having done that, we then need to export an image that the X-Rite programme can read. This must be 16-bit TIFF, and not be influenced by any output profile. (For Adobe users, a DNG file can be used instead of TIFF). So, we select “embed camera profile”, which through the above step is now set to be fully neutral.

  Output process

We can then simply “drag and drop” the resulting TIFF image into the X-Rite programme, tweak the automatically detected reference points and let it do its magic. Give the resulting profile a name to remember, selecting “ICC” for Capture One Pro or “DNG” for Adobe, and save it.

Back in Capture One (or Lr / Ps), select the camera ICC profile just created (and select “auto” Curve) in the “Base Characteristics”

and you’re ready to roll with a very nicely calibrated front end!

Turn turn turn…

Things Digital Posted on Sun, August 23, 2020 15:50:02

Not about photography this time, but another web programming exercise. I wanted to make a “spinner” to keep people busy while the new photo carrousel I put in the into section of my website was loading.

As my site is already quite “clunky (*)” I thought to try something using only the built-in features of HTML and CSS (no javascript for this one).

It’s quite hypnotic, and I’m rather proud of it, actually!

(* I’m not a professional web programmer and learned my computer programming skills during my career as a product marketing engineer for microprocessors, and while building this site over the last 8 years or so. Some of the original code I wrote early on is not so easy to maintain, therefore. When I changed jobs a long time ago, I promised myself to not go and learn another programming language – I broke my promise… 🙂 )

<!DOCTYPE html>

.container {
border: black solid 1px;
padding: 20px;
background-color: grey;
height: 500px;

.spincontainer {
padding: 10px;
text-align: center;
background-color: lightgrey;

.spinpaper {
display: inline-block;
position: relative;
box-sizing: border-box;
text-align: center;
width: 250px; height: 250px;
background-color: rgba(255,255,255,1);
border-radius: 100%;

.spintext::after {
width: 100%;
position: absolute;
top: 50%; left: 0;
margin-top: -11%;
text-align: center;
font-size: 2.2rem;
content: "\00231B";

.turning1, .turning2, .turning3
display: inline-block;
position: absolute;
box-sizing: border-box;
margin: auto;
width: inherit; max-width: 90%;
height: inherit; max-height: 90%;
top: 0; bottom: 0; left: 0; right: 0; /* Force centering by pulling on all four corners */
border-radius: 100%;

.turning1 {
border-top:    20px solid rgba(0,0,127,1);
border-right:  20px solid rgba(0,0,127,1);
border-bottom: 20px double rgba(0,0,127,0);
border-left:   20px double rgba(0,0,127,0.5);
animation: turner linear 3s infinite normal;
.turning2 {
border-top:    20px solid rgba(127,0,0,1);
border-right:  20px solid rgba(127,0,0,1);
border-bottom: 20px double rgba(127,0,0,0.5);
border-left:   20px double rgba(127,0,0,0);
animation: turner linear 1.8s infinite reverse;
.turning3 {
border-top:    20px solid rgba(0,127,0,1);
border-right:  20px solid rgba(0,127,0,1);
border-bottom: 20px double rgba(0,127,0,0);
border-left:   20px double rgba(0,127,0,0.5);
animation: turner linear 1.2s infinite normal;

@keyframes turner{
from {transform:rotate(0)}
to {transform:rotate(360deg)}

<div class="container" id="background" >
 <div class="spincontainer">
  <div class="spinpaper" id="spinner">
     <div class="spintext"></div>
     <div class="turning1"><div class="turning2"><div class="turning3" ></div></div></div>


Levels or Curves?

Things Digital Posted on Fri, May 31, 2019 20:06:41

Again, based on discussions in the Viewfinders club…

Once you get to the stage where you get your digital photos onto a computer and start working with “image processors” to make adjustments, a question one often hears is “what should I use – Levels or Curves?”. The Questionner is referring to different types of tool that can be used to adjust the rightness and/or contrast of an image. The Questionee’s standard answer is whatever suits you best”. Which is correct though not necessarily very helpful. I’ll try to elaborate on the difference here, without getting too technical.

All photo processing software tools are equal (though some are more equal then others), but they do tend to use their own names for things or operate in slightly different ways to what’s described below. However, the basic principles are the same, and some experimentation will help along the way (remember, digital bits are very cheap, but keep a safe copy of your original photo in case something goes totally west).

Brightness and Contrast sliders.

Most photo editors have two simple sliders called like this; usually right in your face so you can’t miss them. For a first approach they do what it says on the label (some have “Exposure” sliders as well). It is a bit of a blunt instrument though. You can increase the general “brightness” of the image, a bit like the volume control of a radio – you can make it louder (brighter), or quieter (darker), to taste.But then everything gets louder (or quieter if you turn the knob the other way). A good start but lacking some subtlety sometimes. If the music is quiet you may be tempted to turn the knob up a little more, at which point some hissing may become audible. Much in the same way, a “too dark” (under-exposed) image can be boosted up to the point that “noise” can become visible in the picture too.

To continue the analogy with a radio, the Contrast control is a little like a basic “tone” control found on some simple radios. Moving it changes the general tone of the sound (how the highlights sound), and by analogy, the Contrast slider adjusts how the blacks, greys and highlights are related.

The simplest of adjustments, and usually the first (and, to be honest, quite often the only) “tweak” on a photo.


Moving up from a simple radio to a classic home “hi-fi” system, you may well find some extra controls – Bass” and “Treble”, and sometimes even “Mid” knobs to twiddle. Bass adjusts the low-end, darker sounds, treble the high-pitched brighter sounds, and Mid (if present) kinda gets freaky with the middle tones (it’s funny that way).

From our analogy again, the “levels” in your photo software will have one control for the “darks” end, one for the bright end and usually at least one for the mid-greys. And these too do what it says on the label. The Black control will control the level (yes) of the blackest parts of the picture, White the top end, and the mid-point control will let you shift the middle grey tones up (brighter) or down (you got it).

Still not the most accurate of instruments, but it’s certainly very usable.


This is where we go from the normal household hi-fi audio system to something more resembling a professional recording studio. At least as far as the “tone control” goes, this has now become a set of sliders controlling quite well-defined pitches of sound, giving very precise control of the amount of each one in the sound you hear. The knobs are usually sliders which, when lined up next to each other, can look like a nice, graceful curve (so they call them “graphic equalisers”). Yes, you get where we’re going here now, right?

The “Curves” tool in photo processing would probably be rather cumbersome if it had a slider for each level of grey to be adjusted. Instead, they allow you to define the nice-shaped curve (from which they calculate all the intermediate values required), by adding points along the curve that can be pulled up or down as desired.

Normally, curves are used to give fine control of the appearance of the different tonal ranges in the source image. You can boost (push the curve up a bit) or reduce (pull the curve down a bit) at different places along the histogram at will.

Pulling the lows down a bit and pushing the highlights up (or vice-versa) produces a slight “S”-shaped curve, sometimes referred to as a “sigmoid” (just to be geeky, I think), but which is very effective and intuitive for correcting many images taken in “difficult” lighting conditions.

You can also get quite extreme: pushing the “Black” end all the way up to “White” and pulling the White end all the way to Black produces a negative image. Keeping the “White” end at white and pulling down the middle of  the curve gives an interesting “solarisation” effect. Experiment, and observe what happens. It’s the only way!

Local adjustment

The above description applies changes to the whole of the image. However, most photo processing tools will allow you to apply any of the above adjustments to specific areas only. Often, a “mask” can be drawn by hand (usually shown as a contrasting colour, so you can see what you’ve done – though this colour itself doesn’t appear in the final image of course). More advanced software will also give you all sorts of tools to help draw very precise masks, fill large areas (e.g. the sky,…), etc. etc… Where the mask is painted, the adjustment will be applied, elsewhere not. This allows you to pick out one problematic part of your picture and correct it (face too much in shadow? Just brighten that bit. Here, I’ve darkened the building facades just to show the point).

Back in the days of “analog” printing, you’d use your hands or bits of card with holes cut out to “dodge and burn” the print (lighten or darken respectively) in certain areas, simply by casting a shadow on part of the print, or only exposing one area of the print for part of the exposure time. Tipping the hat to the pioneers of photography, the terms “dodge and burn” are still in common use in digital photo processing today. Some really good tricks like this will never die, I guess…

Picture Window Pro is now available free

Things Digital Posted on Wed, April 24, 2019 22:10:52

A recent discussion on our lively Facebook group highlighted a trend that seems to be growing – getting away from processed images andgetting things as right as possible “in-camera”. Still, the need for minor trimming of a few technical parameters, and possibly some local adjustments, is there, even in the purest of approaches: even with this trend, there exists a need for basic, no-fuss, cost-effective software, therefore.

A very long time ago, when I was looking for an affordable and usable alternative to Adobe’s already Gargantuan Photoshop (I talk of before the days of Lightroom – yes, the Earth had cooled sufficiently by then…), I trawled the interwebs and found the highly-acclaimed “Picture Window” programme, from the company Digital Light and Color. Back then (at version 3) it cost about €70 and proved to be very good value for money. DL-C is a small company, set up by photographers to make a photographer-friendly programme for quality processing of their images: I dutifully bought the upgrades until Version 5, which is still installed and frequently used on all the PC and laptops I possess. Now, at Version 7, it has officially been released as a free programme. I’ll describe my experiences with it below, which on the whole are Very Good, and now it’s free, you could hardly go wrong to just try it out.

While it offers the ability to handle RAW files, PWP is best used for editing JPEG, TIFF or other standard formats. Though offering an extensive pallet of quite advanced processing possibilities, it really shines for just the basic “tweaking” that may be needed on most images, which can be done very quickly indeed – it doesn’t require files to be “imported” in any way: you just open the file, directly from the memory card if you want. Common things like cropping, exposure, contrast, colour saturation, white balance are immediately to hand, work intuitively and give very good results.

A “mask” feature is available, which is very neat and useful. All adjustments can be applied to the whole image (no mask) or by using such a mask to locally control how much effect the adjustment has. It works like a paintbrush with soft edges, to “paint” an adjustment just where you need it to be, with some very handy tools to make creating these masks very quick and easy. I have often used it, for example, to lighten a face that’s a little dark from being in shadow, or even using teeny-tiny masks to accentuate highlights in the eyes…

It also has cloning tools, which are useful for removing bits and bobs that you may not want in the picture. It can also do composites of different pictures, or create layouts of several on a page. It also does something simply that many high-end programmes seem to find hard – make a nice, properly calibrated print without any fuss.

It’s no Photoshop, so don’t expect wonders of retouching or major manipulation, but I find that it is in fact a VERY handy programme to have around (my secret – most of the pictures that you have ever seen from me in the Viewfinders Newsletters have gone through it, if only just for cropping).

My Digital Workflow

Things Digital Posted on Wed, April 24, 2019 21:23:29

To try and launch a kind of learning experience around how to make photographs, I thought to describe my own “workflow” and see if this could identify some items for interesting and educational discussion.

Step 1. Take the pictures.

You could say that the processes are different for every single shot you’ve ever taken. For me, as probably for most, it means setting the camera to a basic set-up, from which any special needs can be selected without too much trouble. This may change during a day, depending on the lighting, indoor/outdoor, etc… but it’s a pretty standard way to work. A trick I use which may not be so standard is as follows. I generally favour Aperture Priority mode and set up the camera accordingly. I do take the trouble of  programming a set of Manual settings I can quickly switch to, that can be useful to “save the day”. Typically, settings for using a flash when the rest of the shoot is available light.

Step 2: Get the pictures off the camera.

For this, I connect the camera to my main computer and use the Canon “EOS Utility” to download the photos. You can of course take the memory card out of the camera and use a card-reader to copy the files, but I find that the Canon utility saves me a lot of bother: all orientation information about the shot is kept (portrait / landscape – something I had trouble with in the past), and the photos all end up in a logically named folder in a place I can easily find, without the need for any input from me (I’m lazy).

My computer is set up to automatically take a backup copy of any new files every day on a local drive, and also to my network connected server. This backup also saves the pictures I have worked on. Very handy. I don’t delete the pictures from the card until I next need to use it though – just in case!

Step 3: Selecting and Basic Adjustment.

I power up Capture One Pro, which uses my single, standard “Session” as a default. I don’t have the discipline to tag all my photos, making a Catalogue quite redundant for me. Using this one Session means I don’t have to  “Import” the images or any such step – I can just navigate over to the newly created folder with images and go.

Stepping through the pictures chronologically, I can quickly identify the one’s I’ll probably keep. I do the basic adjustments at the same time (exposure, colour balance, cropping, and usually add some “Clarity”), then give it one “*” if I’m happy with it before quickly moving on to the next.

Most of my pictures are delivered in big batches, so this works OK. If there are any I come across that are candidate for some “special treatment” (like B&W, a print or whatever) I may give it two “**”, just to make them easier to find again.

I have on some occasions used the “Auto” feature to set the baseline exposure etc…, but this takes a while to do on all images, and while it sets up a close reference point, it really doesn’t speed me up much, I find.

Notice I didn’t mention “raw” or “JPEG”. I happen to use raw, because I get a lot more leeway in the adjustments (especially in dark areas where noise could become an issue, or in tungsten lighting conditions), but that’s a personal choice. The same method works for either in Capture One Pro.

Step 4: Output the images.

As I use raw format, the images need to be processed to make them into jpegs. The same is true for any jpeg images that have been adjusted (Capture One Pro uses non-destructive editing – it doesn’t touch the original file – so any edited image must be “processed” into a new file to make it available). For this I have a number of “pre-set” recipes that serve mostly to allow me to quickly create different pixel-count images for different needs. I also have some with a watermark pre-programmed. Each “recipe” saves its processed images in a clearly named folder, in the same folder as the images they came from. I do have a few special ones that saves them somewhere more centralised, for “one-off” edits, or those “special” images with two “**”, should I need them. These would also be the candidates for printing, of course.

For saving JPEG images, I have created recipes for different fairly standard needs. They set the image size (in pixels) and compression factor. These can be to fit a box 400pixels each side, to use as thumbnails, to fit a box 1080×1920 pixels (for an HD-sized computer screen), 1600 pixels longest side for “every day” use (including Viewfinders newsletter), or “100%” for full-resolution images. Oh, and one that fits a box 1400 x 1050 pixels, for projection at Viewfinders meetings! 😊

So, I select all the “*” images and activate the batch queue, with the recipes I want to use selected- Capture One Pro will process several recipes in parallel. Don’t go for a coffee yet, though, because it actually happens fairly fast. As the images all end up in a special folder, it’s then relatively easy for me to upload them to my web-site for delivery, send them in to the Newsletter (hint), or copy them to a USB stick. Or whatever.

For printing at home, I tend not to use Capture One Pro directly (except for sporadic contact sheets etc…). It needs a lot of hand-holding to get nice results, I find. Instead, I use a fairly basic photo manager, and stick to using the standard Canon drivers that came with the printer. I usually get very good results that way. For printing at a service lab, I use 100% size (maximum pixels) and 95% compression factor JPEG. Never had a problem!

Deeply dippy … about DPI

Things Digital Posted on Wed, April 24, 2019 21:10:50

The world is unjustly full of misunderstandings about Dots Per Inch, or “dpi”, sometime also referred to as “pixels per inch” or “ppi” (but then my cool title wouldn’t work… 😊 ).

For my work (day-job or as photographer) I often get asked to “send the picture to me in high resolution – 300dpi”: the person asking often doesn’t seem to realise that this is a meaningless request. In order to comply, I then have to ask “how many inches will the picture be printed to?”.


I usually end up just sending the picture with the maximum number of pixels I have in the image, which is often a very big file and certainly “overkill” for most uses (with a file from my Canon 5DmkII, which has 21MPx, a 300dpi print would be about 30×45 cms big – roughly A3 size).

One time, to make a point, I sent a “300dpi” file which had 300px on each side, which is of course the same as 300dpi (ppi) for a 1 inch (2.5cms) square print, but which looks rather poor when printed on an A4 page…

The moral of this sorry tale? There are a few:

  • Digital files don’t really know about “dpi”, as this refers to the physical size of a print. All they know about is pixels, and how many there are. The “dpi” data found in some digital files is calculated from: dpi = (number of pixels along a side of the print) / (length of the side of the print in inches) – literally “dots per inch”.
  • Start thinking about your image in terms of how many pixels it has. Some typical applications have typical numbers of pixels for optimum use. See the table below.
  • Learn how to re-size your digital pictures to give files with a suitable number of pixels. There are some tutorials on the Viewfinders web-site under “Resources” that can help with this.

Typical pixels needed for various uses:

 200 –  400 px highSmall web-page image / thumbnail
 900 – 1200 px highLarger web-page image
1050 x 1980 pxFull HD image (computer monitor / TV)
4200 x 7920 pxQuad HD monitor (but this is getting silly…)
1200 x 1800 px10×15 cms print
2500 x 3500 pxA4 / 20×30 cms print
3500 x 5000 pxA3 / 30 x 45 cms print
 900 px short sideVF newsletter, small illustration
1600 px short sideVF newsletter, large image

Next »