Exercise 10: Colour Cast and White Balance

One of my early memories from when I first became interested in photography and was experimenting with my first film SLR was being shocked by the extreme colour cast from a some pictures I took under tungsten lighting. The abiding thing this experience taught me was to be wary of how our eyes can compensate for different lighting situations and how the camera will not do this. Since the advent of digital photography however, white/colour balance has become less of an issue, especially if photographs are shot in RAW as the colour balance can be altered post capture. (I wonder however if this causes any problems - for example, is there a degradation in quality doing this versus setting white balance at the time of capture.)

As I always shoot in RAW mode I only ever have my camera set to automatic white balance. I find this gives a neutral result most of the time, however, I always experiment with white balance as this can significantly alter the mood of a photograph - higher colour temperatures giving a cooler/bluer result and lower settings giving a warmer/red/orange effect. White balance can also significantly effect colour saturation in an image and perception of exposure.

I skipped the first part of this exercise which called for photographs to be taken using different white balance settings in sunny, cloudy and shady situations as this is something I am familiar with already as described above. The second part of the project interested me however as this called for photographs to be taken in mixed lighting situations. The pictures I have chosen are taken outside my house at dusk with a streetlamp in shot. The different colour temperatures of the sky and the lamp give interesting results

All of these images were shot in RAW format and the Kelvin reading here is taken from Lightroom without any processing, interestingly the amount indicated was different to the one my camera manual gave.

Automatic White Balance:

7350 Kelvin (-23 tint)

Firstly I wanted to compare how my camera rendered the scene with the automatic white balance setting. This is pretty close to how I remember the scene looking with the sky being grey/blue and the orange of the streetlamp illuminating the scene. The streetlamp itself is rendered white.

Incandescent:

2950 Kelvin (-3 tint)

This setting has given the light source the appearance of white light, the orange cast from the street lamp is virtually eliminated and the sky is a dramatic, saturated blue. The centre of the light appears pure white as does the 'halo' around it. This also makes the light and the sky contrast more and the light appears to be brighter while the fence and grass appears much darker.

Fluorescent:

3950 Kelvin (+30 tint)

A much redder element is added here although the blue of the sky is maintained. The light has a more pink than orange tint.

Daylight:

5000K (-6 tint)

This image appears much brighter than the previous 2, the sky is much more muted and grey rather than blue. The cast from the lamp has a warm orange feel.

Cloudy:

5700K (-8 tint)

Very similar to the daylight setting, the main difference is an apparent increase in brightness.

Flash:

6100K (-6 tint)

There is a slight difference in the perceived brightness here in the sky and grass. The grass appears slightly less orange and more green.

Exercise 9: Scene Dynamic Range

Having not enjoyed the previous exercise I decided to give it another go to see whether I could get any better results - I achieved a 4 stop difference which was even worse than my previous result! With this in mind I decided not to complete this exercise. I did do some more reading about dynamic range however.

In his book 'Perfect Exposure', (Ilex 2009) Michael Freeman discusses dynamic range and how this relates to exposure extensively. He defines dynamic range as "the ratio between the maximum and minimum luminance values in a scene or in an image." He also concludes that "the highest image quality in digital photography comes from using the full dynamic range of the sensor to capture the full range of the scene." I am unsure about this statement. "highest image quality" seems a particularly loaded phrase and while I think I understand what Freeman is asserting I do not think I share his valuation of what makes an image of the highest quality. As I became more interested in photography I become less worried about making images that comply with an accepted notion of aesthetic beauty and more concerned about how a picture can be more interesting because it breaks convention. For example, previously I would always select the lowest ISO setting I could to achieve the smoothest tonal values in my pictures (and what I thought was the highest quality.) Now, I find myself often using high ISO settings and have come to appreciate the immediacy noise can add to a picture.

Freeman continues to describe (what he considers) to be the 12 types of exposure situation and how recognising these in combination with an understanding of the cameras dynamic range can enable the photographer to select the desired exposure settings. He says that he arrived at this analysis through experience:

"the way we assess types of scenes subconsciously has been built up over time and is embedded somewhere in our brains"

Analysing a scene relies on understanding key tones (one or more areas of a scene that have commanding importance) and average tone (most people expect most subjects to be average in tone.)

1st Group - The Range Fits:

1. Range Fits - Average

The dynamic range of the scene fits the range of the sensor - histogram within 5-10% of the limits.

Good practice - take a test shot (especially for action.)

Technique - spot maximum highlight and minimum shadow in a scene and make a decision which to set exposure for.

2. Range Fits - Bright

Examples - snow, white walls, white clothing.

Average meter readings would not give the right result, use centre weighted on the key tone and positive exposure compensation.

3. Range Fits - Dark

Opposite approach to 2 to keep dark details black and not grey.

2nd Group - Low Range:

Atmospheric conditions diffuse light and act as a filter that even out tones in a scene e.g.. Haze, mist, fog, dust.

4. Low - Average

Histogram will be cantered with room either side, often 'bell' shaped.

Sometimes called flat lighting

5. Low - Bright

No important shadow areas.

Will often need positive exposure compensation to keep bright but danger of highlight clipping.

6. Low - Dark

Not common - mainly due to taste - people prefer light rather than dark images. Most low key images tend to have small bright tones which raise the dynamic range e.g.. Dusk, dawn, night, deep colours e.g.. Purple.

3rd Group - High Range:

Also referred to as over scaled or high contrast. Important to make decisions about clipping.

7. High Key - Average

Extreme mixture of tones, some clipping inevitable. Most common lighting situation when range is high.

8. High - Large Brighter

Subject matter brighter than average and the surroundings.

Technique - centre weighted and positive exposure compensation.

9. High - Small Brighter

Difficult to measure - spot metering best. Histogram of little use because small area of brightness displayed as relatively few pixels. When key tone small danger of over exposure - care needed to avoid clipping but high enough to show detail.

10. High - Edge Lit

Most specific, rarest and difficult lighting situation. Effect depends on situation e.g.. Thin edge may blow highlights, broader expose as light, not clipped.

Bright edge usually key tone meaning loss of shadow detail is often acceptable. (An example of a low key image)

11. High - Large Darker

Background will suffer from over exposure unless contrast dealt with by adding light or in software.

Very common e.g.. Outdoor scene with a band of sky.

Choices - expose for average tone or darker or silhouette.

12. High - Small Darker

Background always dominates here and is where exposure decisions should be taken.

Freeman brings his analysis of exposure situations with a description of Envision - being able to imagine how a scene will look if you give it a certain exposure:

1. Learn to discount your eyes efficiency at seeing detail in deep shadows and bright highlights as your camera cannot see both.
2. Be able to imagine how a scene will reproduce of given 'normal' or 'average' exposure.
3. Decide how you would like it to look.
4. Anticipate how it might look under different lighting conditions that are practically possible, such as change in the angle of sunlight or under controlled conditions by changing lighting.

While I find the information here interesting, even invaluable, it remains difficult to translate into practical application. The notes imply the ability to consider a scene carefully and make adjustments as necessary without time pressure. Unfortunately, in practice this is not often possible. I guess this is where experience comes to the fore and becomes key. I imagine accomplished photographers like Freeman are able to analyse a scene quickly and come make exposure decisions almost intuitively. I have a tendency not to think too much about the exposure settings I make, partly this is because experience means I have a clear idea of the result that will be achieved but also it I because I always shoot raw and know I have a certain latitude to be to alter exposure in processing. I have never worried about this but increasingly I am beginning to think this is laziness on my part - should I be striving to have a clearer understanding of what a photograph will look like when and attempting to get my camera settings as close as possible to this or is it okay to make these decisions at the computer screen? I do not think there is a right or wrong answer to this and experimentation is probably the best course of action. I will attempt to apply the knowledge gained here to my shooting however and see if the results are different.