What the number is actually measuring
The general colour rendering index, written Ra, is calculated by lighting a set of standard test colours with the source under examination and with a reference illuminant at the same correlated colour temperature, then measuring how far each sample shifts. Ra is the average of the first eight test colours. A perfect match to the reference scores 100.
Two things follow from that definition, and they explain most of the confusion around CRI. First, the comparison is always against a reference at the same colour temperature, so a 2700 K source and a 6500 K source with the same Ra are being judged against different references. Second, an average of eight samples can be high while an individual colour renders poorly.
The R9 problem
The eight samples behind Ra are moderate, unsaturated colours. Strong red is sample nine — R9 — and it sits outside the average. A source can post a respectable Ra and still render saturated reds weakly, which is exactly the case where the difference is most visible: skin tones, red fabric, timber with a warm grain, meat and produce in a retail display.
If red matters in your space, ask for R9 as a separate figure rather than reading it into Ra. More recent method, IES TM-30, replaces the single average with a fidelity index and a gamut index across 99 samples, which describes the same behaviour in more detail.
| Space | What is being judged | What to ask for |
|---|---|---|
| Living room and bedroom | Skin tone, timber, textiles | Ra together with a stated R9 |
| Kitchen | Food, produce, worktop finish | Ra together with a stated R9 |
| Retail display | Product colour against competitors | Ra, R9, and consistency between fittings |
| Wardrobe and dressing area | Matching two garments | High Ra at a consistent colour temperature |
| Corridor, plant room, store | Little or no colour judgement | Ra is rarely the deciding factor |
A guide to which question to ask in each space. No product, brand or numeric threshold is recommended here.
Why the same paint looks wrong in two rooms
Colour is not a property of a surface alone; it is what happens when a light source, a surface and an eye meet. Change the spectrum of the source and the reflected spectrum changes with it. Two lamps that both read as neutral white can contain very different amounts of energy at each wavelength, so a paint sample chosen under one can genuinely look like a different colour under the other.
This is also why mixing sources in one room is risky. A downlight from one batch, a pendant from another supplier and a strip light chosen for its price will rarely agree, and the eye is very good at noticing the disagreement even when it cannot name it.
Common mistakes
- Treating Ra as a complete description of colour quality when it is an average of eight moderate samples.
- Comparing the CRI of two lamps at different colour temperatures as though they were judged against the same reference.
- Choosing paint, stone or fabric under showroom or site lighting that is nothing like the lighting the room will finally have.
- Mixing fittings from different sources in one room and expecting their colour to agree.