Compared to tungsten halogen lighting, white LEDs offer improved efficiency and much longer life. But can LEDs compete with the excellent colour quality of halogen lighting? An experiment funded by the BRE Trust aimed to find out. It involved two sets of two booths containing everyday coloured objects, lit by different lamps. One pair of booths contained lamps equivalent in illuminance to 35W halogen, and the other lamps equivalent to 50W halogen. 36 subjects were invited to compare the brightnesses of the two booths, colour of the light, and correctness or familiarity of the colours of the items.
The 50W equivalent lamps were all warm white in colour. The 35W equivalent lamps included two cool white LEDs, and a warm white LED, as well as a tungsten halogen lamp.
What we learned was that the subjects preferred the warm white to the cool white LEDs, but preferred the cool white LEDs to the tungsten halogen lamp, which had been dimmed to give a similar illuminance to the warm white LED. This gave it a low Correlated Colour Temperature of 2550 K (see Tech Briefing, page 96) which may have resulted in the colour of that light being less preferred.
The experiment also gave us interesting information on the proposed American measure of colour rendering, the Colour Quality Scale versus the standard Colour Rendering Index. Initially all the LEDs chosen for the experiment had similar CRI and CQS values. However, one of the 50W equivalents was a tuneable LED, whose settings were then specially altered so that it had a higher CQS rating (87) compared to its CRI (77). The colour of this lamp and the items lit by it were preferred to that of another LED lamp with a CQS of 83 and a CRI of 84. The difference in preference was statistically significant. The tuneable LED was even preferred to a tungsten halogen lamp with a measured CRI of 99 and a CQS of 96. The results suggested that for some LED lamps CRI can be a poor indicator of colour preference.
The tuneable LED had a higher colour gamut (see Tech Briefing, page 96) than the other LED lamps and the tungsten lamp with which it was compared. This may be another reason why the colour of the tuneable LED was preferred. Future lamps for display lighting could be specially engineered to have a high colour gamut.
The results showed that LEDs with the right spectral mix may be preferred to halogen lamps. For this type of task, viewing a display of coloured objects, subjects preferred lamps with higher colour gamut.
A BRE Information Paper ‘Specifying LED Lighting’ (available from www.brebookshop.com) gives more advice on LEDs and how to use them.
This work was funded by the BRE Trust as part of its research programme. Gareth Howlett and Stephanie King carried out the experimental work, and Lorna Hamilton performed the statistical analysis.