Direct answer
CRI (usually reported as Ra) is a legacy measure of colour fidelity: how similarly a set of test colours appears under a light source compared with a reference source of similar colour temperature. It is useful, but incomplete. The CIE now recommends wider adoption of its newer fidelity index Rf, and IES TM-30 adds both fidelity (Rf) and gamut (Rg) information. Colour temperature such as 3000K or 4000K describes the appearance of the light itself; it does not tell you how faithfully objects will look under it.
Key facts
- CCT (Kelvin) and colour rendering measure different things. Two 3000K lamps can render objects differently.
- NIST explains that general CRI Ra averages the first eight test-colour samples; R9 is a separate special index and is not included in the Ra average.
- CIE PS 002:2025 recommends that the lighting community begin adopting CIE Rf and replacing Ra in relevant standards/specifications, with both reported in parallel during transition.
- ANSI/IES TM-30-24 reports average fidelity, gamut and hue-specific colour shifts rather than reducing colour rendition to one number.
- No single colour metric captures every preference. Fidelity and visual preference are related but not identical.
Start with the biggest misconception: Kelvin is not CRI
A 2700K lamp looks warmer than a 4000K lamp because correlated colour temperature describes the chromatic appearance of the source. CRI/Ra and newer metrics describe how the spectrum affects the appearance of coloured objects. You can therefore have two lamps at the same 3000K CCT where one makes skin, food, timber or fabric look noticeably more natural than the other.
This distinction is valuable in ecommerce because product pages often list colour temperature but omit colour-rendering data. Where colour quality matters, CCT alone is insufficient.
What CRI / Ra actually measures
NIST explains that the traditional CRI method compares test-colour samples under the test source and a reference illuminant. General CRI Ra is the average of the first eight sample scores.[1] A higher Ra therefore indicates greater average fidelity for that test set relative to the reference; it does not guarantee perfect rendering of every saturated colour.
Because those first eight samples are relatively moderate in saturation, Ra can hide weaknesses in particular colours. This limitation is one reason the lighting science community has developed newer approaches.
Why R9 is discussed separately
R9 is the special index associated with a strong red test colour. It is not part of the standard Ra average.[1] That makes it useful supplementary information for applications where reds and skin/food tones matter, but R9 is still one sample rather than a complete description of a spectrum.
Retailers should avoid the simplistic claim that “high R9 = perfect colour”. A better buying statement is: when colour-sensitive materials matter, request fuller colour-rendition data rather than relying on Ra alone.
What changed with CIE Rf in 2025
The CIE’s 2025 position statement says its General Colour Fidelity Index Rf was developed to address technical problems identified with Ra. After years of use without evidence contradicting its scientific use, the CIE now recommends that the lighting community begin adopting Rf and replacing Ra in relevant regulations, specifications and standards, while reporting both during the transition.[2]
This is a genuine current-content opportunity for Whitehouse: most consumer lighting guides still stop at “CRI 80 vs CRI 90”. Explaining the transition makes the Knowledge Hub more technically current than generic retail content.
What TM-30 adds
ANSI/IES TM-30-24 evaluates colour rendition using an objective statistical method and reports average properties plus hue-specific information. The system includes a fidelity index (Rf), gamut index (Rg) and graphical/hue-bin measures showing where colours become more or less saturated or shift in hue.[3]
For ordinary home shoppers, this does not mean every bulb needs a TM-30 report. It means professionals specifying restaurants, retail, hospitality, art, premium residential interiors or colour-critical spaces have better tools than one CRI number when manufacturers provide the data.
How to buy lighting when colour matters
First choose the right CCT for the atmosphere and task. Then compare documented colour-fidelity information. For general utility spaces, a mainstream CRI specification may be adequate. For kitchens, bathrooms, dressing areas, food displays, retail merchandise, art and material showrooms, it is reasonable to prioritise products with stronger verified colour-fidelity performance and, where available, R9 or TM-30 data.
Finally, use real samples. Paint, timber, stone and fabric can respond differently to spectra that share the same CCT and nominal CRI. For a high-value interior, viewing the actual material under the intended fitting is more reliable than choosing from one number alone.
Do not confuse these lighting metrics
| Metric | What it tells you | What it does not tell you |
|---|---|---|
| CCT (K) | Warm/cool appearance of nominal white light | How accurately objects are rendered |
| CRI / Ra | Legacy average colour-fidelity score for eight samples | Saturated-red performance or colour preference by itself |
| R9 | Special red-rendering index in the legacy CRI family | Overall fidelity across all colours |
| CIE Rf | Modern general colour-fidelity index | Colour preference or gamut by itself |
| TM-30 Rg | Average relative gamut / saturation tendency | Fidelity by itself |
Frequently asked questions
Is CRI 80 “good”?
It can be adequate for many general-lighting applications, but “good” depends on what you are lighting. Colour-critical spaces benefit from stronger, better-documented rendition data.
Is 3000K better quality than 4000K?
No. CCT describes colour appearance, not colour fidelity. Either CCT can be produced with high or low colour-rendition quality.
What is R9 and why is it missing from some product pages?
R9 is a special strong-red rendering index and is not part of the standard Ra average. Many consumer datasheets report only Ra, so R9 may simply not have been published.
Is TM-30 replacing CRI?
TM-30 is a modern, richer evaluation method used by lighting professionals. Separately, the CIE’s 2025 position statement specifically recommends transition toward CIE Rf for general colour fidelity.
Should Whitehouse claim a CRI value when the manufacturer does not publish one?
No. Treat colour-rendering metrics as verified specifications. If the datasheet does not provide the value, do not infer it from price, colour temperature or product type.
Technical references and further reading
[1] NIST — Color Rendering of Light Sources — Technical explanation of CRI calculation and limitations; Ra averages the first eight test samples. Source link
[2] CIE PS 002:2025 — Position Statement on Colour Quality Metrics, 2nd Edition — Current CIE recommendation to begin adopting Rf and replace Ra, with parallel reporting during transition. DOI: 10.25039/PS.6qdfenyc. Source link
[3] ANSI/IES TM-30-24 — Method for Evaluating Light Source Color Rendition — Current 2024 IES method covering fidelity, gamut and hue-specific colour rendition. Source link
