I have spent some time recently trying to improve the lighting conditions in the room where I do colour work. This is an informal report on what I discovered. It shows how far I got with some comment on how much further I could go.

All the light measurements have been done using an Eye-One Display 2 colorimeter which not only allows me to produce reliable monitor calibration but also to accurately measure ambient light levels.

Aims and Initial Observations

Two features control the quality of the light in a room. The colour of the walls and fitments and the strength and colour of the light source.

Walls should be a neutral colour preferably light grey especially in the eyeline and where light will be reflected back towards the screen.

The lighting should not be particularily bright, especially in the user's eyeline when sitting at the screen. It should ideally have a colour temperature which goes neither below 4500°K (a warmer white) nor over 6000°K (a cooler white).

ISO 3664:2000 sets out certain viewing conditions. It gives some targets, D50, effectively 5000°K, for the ambient light colour temperature and a light level below 64 Lux - quite dark.

Overall lighting during the day could be best achieved using daylight as it will get close to the required colour temperature and is free. However it is far too bright.

Reality

I am lucky enough to have a north-facing room. This provides a more even colour temperature than I would get in direct sunlight. I do get quite strong reflections off the neighbour's house opposite in bright sunlight. As I am unlikely to be able to knock it down, I have moved my sitting position and the monitors to reduce the glare.

Daylight readings: colour temperature, 4300°K; luminance 160 Lux.

The wall colours are white above picture rail-level and magnolia below, the magnolia will drop the colour temperature.

The furniture and the flooring in the room are predominantly light-wood coloured, this too will drop the colour temperature.

At night the normal room lighting was a single 100 watt tungsten bulb with a colour temperature which seriously distorted the colours rendered by the screen.

The solution was either to edit only during the hours of daylight or provide some form of daylight-adjusted artificial lighting.

The Theory

Bulbs are assessed using two numbers:

• the Colour Temperature, and;
• the Colour Rendering Index.

Some time on the internet showed that the majority of so-called daylight bulbs had 'issues'.

Colour Temperature

This is a comparison of the bulb's overall light output to that of a black-body source. It is measured in degrees Kelvin. Normal northern mid-day daylight varys between 4500°K and 10000°K. The best bulbs can reach about 5000°K. Colour temperature is a normal measure of any light source and is equally applied to light bulbs, monitors and daylight.

Colour Rendering Index (CRI)

The components of daylight are spread quite evenly across the whole visible spectrum. Most 'Daylight' bulbs, often filtered tungsten or fluorescents, produce most of their light in quite restricted areas of the spectrum.

Consequently, even though, when the light they produce is 'added up', it produces a quite adequate colour temperature reading, the light can be flat and featureless. Properly produced halogen bulbs can have a spectrum very close to that of daylight. CRI is measured as a percentage comparing the bulb's spectrum to the spectrum of the same black-body at the same colour temperature. Fluorescents typically manage 55-85% and a good halogen spot can reach over 99%.

Internet investigation (and personal recommendation) established the Solux brand of halogen spots to be both good and readily available.

The Move to Solux Low Voltage Halogen Lighting

Solux bulbs are all 12 volt units so a low-voltage light fitment is needed. The bulbs are GU5.3 (M16) push fit. They come in 35 watt and 50 watt powers, a range of colour temperatures and a range of angles of beam spread.

The high cost of Solux light fittings meant a search for lower cost ceiling lights from local DIY stores. Most halogen spotlighting now seems to be mains-based so the choice of low voltage (12v) fitments and their supply may be limited.

As the Solux halogen bulbs come in 36 watt and 50 watt powers, any light fitting must support at least the lower of these numbers.

The fitting was a four bulb unit with 35 watt/bulb power handling. The power handling is primarily controlled by the transformer so it may be possible to upgrade to one which would handle the 50 watt bulbs but heat dissipation may become more of an issue. It came with four halogen bulbs of unknown quality but, to the eye, they were still a great improvement on the old tungsten light.

The colorimeter made less happy reading. The colour temperature was still way down around 2500°K and the luminance had risen to over 200 Lux, more than three times too high.

A simple monitor cowl made from a dark charcoal mounting board can be a great viewing aid

After fitting those and adjusting the light spread the measurements rose to a colour temperature of 3200°K and the luminosity dropped to 59 Lux. Disappointing colour temperature but good light level.

Much of the reduction in the light level comes from the cowl now fitted over the monitor, cheap and seemingly effective. Although monitor cowls can be bought commercially, it is quite easy to construct one from a dark-faced mounting board.

Not as good as I had hoped so where to next?

Certain further options exist:

• repaint the room to a grey - maybe even a grey with a bluish tint;
• check the voltage of the light's transformer. If it is running below the rated 12v the colour temperature will be reduced. It may be possible to increase the voltage to 'blue' the light though this would have a bad effect on the longevity of the bulbs, or;
• accept things as they have become which is much improved on the original lighting in the room.

Hmm...

The daytime luminosity has been reduced with a (grey) venetian blind over the window. This has raised the colour temperature of the room to 4300°K and luminosity is now a very presentable 60 Lux. This angling of the vanes provides a surprisingly fine adjustment of the light level.

Conclusion

The daytime figures are almost perfect. The nighttime figures are not as good as had been hoped for but could probably be improved with the expenditure of more time and money.

There is a disadvantage to having the two lighting situations, that being the need to match the artificial lighting to daylight. A single lighting system, useable at any time would be better but this would dictate the constant use of artificial light. Personally I don't feel this is a compromise I want to make so the mix of lights will have to do.

Were a room to be designed from scratch for colour work then much better control could certainly be gained by removing all daylight from the room and just using artificial light in a properly coloured, shaded room with no colour distractions. The use of neutral clothing, no wall hangings.

So...?

Just for the moment I am happy to have improved things as much as I have, and the law of diminishing returns seems to have cut in, I will stop and see how things go.

A few points do stand out, these are:

• the improvements achievable by the use of good halogen bulbs;
• the control of the ambient light during the day by the use of a grey venetian blind on the window;
• the use of a monitor cowl with a dark, neutral coloured interior, and;
• closing the door.