CRI, R9 and TM-30: how to read a datasheet properly
CRI is the most quoted and least sufficient figure on a datasheet. Two luminaires with the same CRI 90 can produce completely different results on a red.
- Reference standard
- CIE 13.3 (CRI) · IES TM-30-20
- Scale
- CRI 0–100 · Rf 0–100 · Rg ≈ 80–120
- Where it applies
- Retail y museo >90 · moda y arte >96
- How it is measured
- Espectrómetro sobre la luminaria real
CRI (Ra) is the average fidelity across 8 pastel colour samples (R1–R8). Its problem is twofold: it is an average, so a poor value can hide inside a good mean; and it excludes saturated colours. That is why CRI 90 does not guarantee that reds will look right.
R9 measures saturated red and is the decisive figure for meat, wood, leather, textiles and human skin. Many CRI 90 sources have an R9 of 20 or 30: on paper they comply, at the butcher counter the meat looks grey. Always ask for R9 separately — we publish R9 ≈ 96.
TM-30-20 is the IES modern system and gives two numbers: Rf (fidelity), an improved CRI equivalent using 99 samples, and Rg (gamut), showing whether colours become saturated or dull. Rg ≈ 100 is neutral; above, colours look more vivid; below, washed out. Rf 93 with Rg 103 — our case — means high fidelity with slightly positive saturation: faithful colours, but alive.
For video and photography there is TLCI, which evaluates how the camera responds, not the eye. It is the relevant figure in studios, filmed rooms and showrooms producing content.
- Do not compare luminaires by CRI alone: ask for R9 and, if possible, TM-30 Rf/Rg.
- Be wary of a datasheet that publishes CRI but hides R9.
- CCT and CRI are independent: a 3000 K source can have CRI 70 or 98.
- The measurement that counts is the spectrometer reading on the finished luminaire, not the bare LED chip.
