CRI: Why Two Bulbs at the Same Brightness Look Different

Colour rendering describes how accurately a light source reveals the colours of the things it lands on. It’s independent of brightness and independent of colour temperature — two lamps can be equally bright, identically warm-white, and still make the same room look completely different. Poor colour rendering is the usual explanation for a space that seems flat, grey, or somehow cheap without anything obviously wrong.

It’s also the spec most often missing from a box, which tells you something.

Why a light can fail to show colour

An object has no colour of its own. It has a reflectance — a preference for reflecting some wavelengths and absorbing others — and what you see is that preference applied to whatever light arrives. A red cushion looks red because it reflects long wavelengths. If the light arriving contains very little energy at those wavelengths, there is nothing for the cushion to reflect, and it looks brown, grey, or muddy.

Sunlight and an incandescent filament both emit smoothly across the whole visible range, so everything they illuminate has something to work with. That’s the standard everything else is measured against.

White LEDs work differently. Most are a blue diode coated with phosphors that absorb some of the blue and re-emit it at longer wavelengths, and the sum reads as white. But the sum is not smooth. It typically has a strong blue peak and a broad yellow-green hump, and — critically — often a deficit in deep red. Cheaper lamps have a larger deficit, because deep-red phosphors reduce efficiency and add cost. A lamp can therefore look perfectly white, measure correctly for colour temperature, and still be missing the wavelengths that make skin, wood, brick, and warm textiles look alive.

What the number measures

The colour rendering index, written CRI or Ra, compares how a set of standard test colours appear under the lamp against how they appear under a reference source of the same colour temperature. The reference scores 100 by definition, so 100 is the maximum and lower is worse.

Reading it:

• Ra in the 90s is high fidelity, and the difference from the 80s is visible in a domestic room.
• Ra in the 80s is the mainstream figure for general-purpose lamps. Acceptable, unremarkable.
• Below 80 is where a room starts to look wrong, and it’s common in very cheap lamps and in decorative strip lighting.

Two caveats that matter more than the number itself.

Ra is an average, and it averages away the failure. The headline figure is the mean across the test colours. A lamp with a large deep-red deficit and good performance everywhere else averages out to a respectable number. That’s why the industry quotes R9 — the deep-red test colour — separately. A lamp with a good Ra and a poor R9 is the classic case: whites and pale colours look fine, while red-toned things go dull. In a room whose warmth comes from wood, brick, terracotta, warm textiles, and skin, R9 is doing most of the work you care about.

Ra is compared against a reference at the same colour temperature. So a high-Ra daylight-white lamp is faithful as daylight, not warm. Rendering and colour temperature are genuinely separate choices, and you need both right — the second one is covered in what colour temperature actually means.

There is also a newer measurement, TM-30, which reports fidelity and saturation separately across a much larger colour sample and gives a far more complete picture. It’s mostly seen in professional specifications rather than on consumer packaging, but if you see it quoted, it’s better information than Ra.

Where it shows up in a home

Colour rendering is invisible until you compare, which is why people live with bad lighting for years. The places it bites:

Anywhere skin tone matters. Bathroom mirrors and dressing areas. A low-R9 lamp makes people look ill, and the effect is strong enough to be unsettling without being identifiable.

Rooms whose character is warm materials. Wood floors, exposed brick, leather, rugs, a red sofa. These are exactly the reflectances that a deep-red deficit flattens. A living room designed around warm materials and lit by low-rendering lamps loses most of what was chosen.

Kitchens, where you’re judging whether food is cooked or fresh.

Anywhere you compare colours — choosing paint, matching clothes, any craft or detail work.

Decorative strip and accent lighting. Colour-changing LED strips are a special case: their output is a mix of separate red, green, and blue emitters, which is a handful of narrow spikes rather than a spectrum. Excellent for washing a wall in a colour, poor at illuminating anything you want to look natural. Many include a separate white channel precisely because the RGB mix renders objects badly. Use them as coloured light, not as light to see by.

How to judge it without instruments

Since the number is often absent, three practical tests:

Compare side by side, with something warm in shot. Put the candidate lamp next to one you trust, and look at wood, skin, and something red. Nothing else reveals the difference as quickly, and you don’t need to know the specification to see it.

Look for the number at all. Manufacturers who have invested in rendering print it. Absence of a CRI figure on the packaging is a soft signal about where the lamp sits.

Buy the whole room together. Rendering varies between product lines and between production runs, and mixing a high- and low-rendering lamp in one room produces the same kind of visible disagreement as mixing colour temperatures.

Why it matters for ambiance specifically

A cosy room is not just a dim room. Every warm-light source people find genuinely pleasant — a fire, a candle, a filament lamp turned low — is simultaneously low in colour temperature, low in intensity, and continuous across the spectrum, with plenty of deep red. That combination is what makes materials look rich rather than merely darkened.

Which is also why a flame effect reads as warm at all: it is deep-orange light at low intensity, working on the same principle. The optics that produce it are described in how does an electric fireplace actually work? Reproducing that quality with electric lighting takes all three properties, and rendering is the one people skip. A dimmed low-rendering lamp gives you darkness without warmth — see also how LED dimming works for the other half of that problem.

What to actually do

1. Look for a stated Ra and prefer the 90s in any room you spend evenings in.
2. Look for R9 separately, and treat a good Ra with an unstated R9 with suspicion.
3. Treat a missing CRI figure as a warning, not as a neutral omission.
4. Compare candidates side by side against wood, skin, and something red before committing to a whole room.
5. Buy all a room’s lamps together from one product line.
6. Don’t use colour-changing strips as general lighting — use the dedicated white channel or a separate fitting for that.
7. Get colour temperature and rendering right together. Either one alone won’t produce the result.