Short answer
ANSI/IES TM-30 is the Illuminating Engineering Society's method for rating how a light renders colour, currently in its 2024 edition. It compares 99 real-world sample colours under the lamp and under a reference in a modern colour space, reporting average fidelity (Rf), average saturation change (Rg) and hue-by-hue shifts — so it can say not just how far colours move but which way.
The CIE's colour rendering index was first published in 1965, revised as CIE 13.2 in 1974 and republished unchanged as CIE 13.3 in 1995. Its general index Ra averages colour shifts of eight fairly desaturated samples, measured in the CIE 1964 U*V*W* space, which is now regarded as obsolete and is noticeably non-uniform. Three consequences follow. A small sample set invites spectra tuned to score well on those eight colours. The non-uniform space penalises some shifts more than equal-looking shifts elsewhere, notably for reds. And an average hides direction: CIE 13.3 itself notes that the direction of colour shifts is recognised as important but not included. The CIE has since acknowledged that Ra deviated from an accurate fidelity measure.
TM-30 rests on four components: the CIE 1964 10° colour-matching functions, the CAM02-UCS colour space derived from the CIECAM02 appearance model, a set of 99 colour evaluation samples, and a reference illuminant that depends on the lamp's CCT. The 99 samples are measured reflectances of real things — paints, textiles, inks, skin, natural objects — chosen from a pool of more than 100,000 to spread evenly through colour space and across wavelength, which makes gaming the index much harder. The reference is a blackbody below 4,000 K, CIE daylight above 5,000 K, and a blend between the two in that range. Since the 2018 edition the calculation core has been identical to the CIE's own fidelity index, CIE 224:2017.
Rf is the average colour difference across all 99 samples, scaled and log-transformed to run from 0 to 100, where 100 is identical to the reference. Rg describes saturation: the samples are grouped into 16 hue bins of 22.5° each, the average position of each bin forms a 16-sided polygon, and Rg is the area under the lamp divided by the area under the reference, times 100. Above 100 the lamp tends to saturate colours; below, to mute them. The colour vector graphic draws each bin's shift as an arrow. For individual hues there are local chroma shift (Rcs,hj), hue shift and fidelity (Rf,hj) for each bin; bin 1 is the reds, so Rcs,h1 and Rf,h1 are the TM-30 successors to R9. For typical architectural work, Rf, Rg, Rcs,h1 and Rf,h1 together are usually enough.
Rf and Ra are on similar scales but not interchangeable: for lamps with Ra around 80, Rf can differ by more than 20 points in real products.
Annex E of TM-30 recognises that maximum fidelity is not always the goal. It defines three design intents — Preference (a pleasing, natural-looking space: retail, hospitality, offices, homes), Vividness (a vibrant scene, natural or not: entertainment, display) and Fidelity (appearance close to the reference: manufacturing, medical, colour matching) — each at three priority levels, P1 to P3, V1 to V3, F1 to F3. Criteria apply to rooms with a variety of colours lit to roughly 200–700 lux, and can be combined: meeting both P3 and F3 means Rf of at least 85, Rg of at least 89, red chroma shift between −12% and +23% and red fidelity of at least 85. P1, V1 and F1 cannot all be met at once, by design. Programmes including the DesignLights Consortium technical requirements, ASHRAE/IES 189.1 and the WELL Building Standard have written TM-30 criteria into their rules.
IES TM-30-15 appeared in 2015; ANSI/IES TM-30-18 harmonised it with CIE 224:2017 and added the specification annexes; TM-30-20 republished that text without technical change; ANSI/IES TM-30-24 is the current edition. The calculations are published through a reference implementation, and the method describes a light source, not an installation: the light level parameters inside the colour appearance model are fixed so that products can be compared. When buying, ask for Rf, Rg, Rcs,h1 and Rf,h1 — or the full TM-30 report — and treat an Ra-only datasheet as incomplete for any space where colour is judged.
| Measure | What it describes | Range and reading |
|---|---|---|
| Rf | Average colour fidelity over 99 samples | 0–100; 100 = identical to the reference |
| Rg | Average change in saturation (gamut area of 16 hue bins) | About 100 = neutral; above saturates, below mutes |
| Rcs,h1 | Chroma shift of the red hue bin | Percentage; negative = reds muted, positive = reds boosted |
| Rf,h1 | Fidelity of the red hue bin | 0–100; the TM-30 counterpart of R9's purpose |
| Rf,hj / Rcs,hj / Rhs,hj | Fidelity, chroma and hue shift for each of 16 bins | Shown on the colour vector graphic |
Each statement is labelled by kind — established fact, a standard’s requirement, observed market data, a convention, or Colourwise’s own interpretation or analysis — with the strength of the evidence behind it.
StandardStrong evidence
ANSI/IES TM-30-24 is the current edition of the IES method for evaluating light source colour rendition.
Source: ANSI/IES TM-30-24, IES Method for Evaluating Light Source Color Rendition
StandardStrong evidence
The TM-30 framework uses the CIE 1964 10° colour-matching functions, the CAM02-UCS colour space, 99 colour evaluation samples and a CCT-dependent reference illuminant, and shares its core calculation with CIE 224:2017.
Source: Tutorial: Background and Guidance for Using the ANSI/IES TM-30 Method for Evaluating Light Source Color Rendition (LEUKOS 18:2, 191–231); CIE 224:2017 CIE 2017 Colour Fidelity Index for Accurate Scientific Use
FactStrong evidence
For light sources with a CRI of about 80, TM-30 Rf can differ from Ra by more than 20 points among existing products, despite the two measures being well correlated overall.
Measured: About 1,500 measured spectra of real products and about 165,000 theoretical spectra.
StandardStrong evidence
TM-30 Annex E defines Preference, Vividness and Fidelity design intents at three priority levels each, applicable to polychromatic interiors lit to roughly 200–700 lux.
FactStrong evidence
The CIE states that its fidelity index Rf is not intended to take over CRI's unintended role as an overall colour-quality measure for end users.
Source: CIE 224:2017 CIE 2017 Colour Fidelity Index for Accurate Scientific Use
Reviewed 29 September 2026. Colourwise summarises its sources in its own words and does not reproduce standards text or proprietary colour data. Spotted an error? Tell us.