paper A literature review was conducted to examine the influence of light source spectrum on the visual needs of pedestrians including spatial brightness. The results of the reviewed studies suggest that metal halide lighting is brighter than sodium lighting (whether HPS or LPS) of equal illuminance.

Fotios SA, Cheal C, Boyce PR. Light Source Spectrum, Brightness Perception and Visual Performance in Pedestrian Environments: A Review. Lighting Research and Technology 2005; 37(4); 271-294.

paper An experiment was conducted to compare brightness under lighting of different spectrum, specifically to test the then-permitted (in BS5489-1:2003) illuminance reduction when using lighting of high colour rendering index. While CRI is not expected to be a robust predictor for spatial brightness, it was the metric used in the standard and hence this validation exercise was carried out. The experiment used three experimental procedures (matching, discrimination and category rating) and five types of lamp (LPS, HPS, CFL and two types of MH) with evaluations carried out using horizontal illuminances within the recommended range (2.0, 7.5 and 15 lux).

The results suggested an effect of spectrum on spatial brightness, with the CFL and MH lamps creating brighter environments than HPS and LPS, and of sufficient magnitude to support the one-lighting-class reduction permitted in BS5489-1:2003. It was also stated that CRI is not the right metric to characterise this effect.

Fotios SA, Cheal C. Lighting for subsidiary streets: investigation of lamps of different SPD. Part 2 – Brightness, Lighting Research and Technology, 2007; 39(3); 233-252.

See also

Fotios S, Cheal C. Correspondence. Correction: Lighting for subsidiary streets: Investigation of lamps of different SPD. Part 2 – Brightness, Lighting Research and Technology, 2009; 41(4); 381-383.

paper This second series of experiments was intended to test different metrics for characterising spatial brightness at mesopic light levels. The metrics were colour rendering index, correlated colour temperature, gamut area, the S/P ratio and the SWS/P ratio – this latter simply replaced the rod (S) part of the S/P ratio with the s-cone response. Five different light sources were used, in an all-possible-pairs approach, these chosen to enable the predictions to be tested. The experiment used a side-by-side brightness matching procedure, supported by discrimination at equal illuminance, with the reference light level being 5 lux.

The results suggested that the S/P ratio gave better prediction of spatial brightness than did CRI, CCT, SWS/P and gamut area.

Fotios SA, Cheal C. Predicting Lamp Spectrum Effects At Mesopic Levels. Part 1: Spatial Brightness, Lighting Research and Technology, 2011; 43(2); 143-157.

paper The Digest series in Lighting Research & Technology was initiated to summarise a topic when it appears to have reached a consensus. The first digest focussed on brightness at mesopic levels.

Fotios S. LRT Digest 1. Maintaining brightness whilst saving energy in residential roads. Lighting Research and Technology research digest, 2013; 45(1); 7-21.