SPECTRAL QUALITY: WHY CRI DOESN'T TELL THE WHOLE STORY

For decades, the Colour Rendering Index (CRI, or Ra) has been the main reference used to describe a light source's ability to reproduce colours. A high value — 90, 95 or 97 — has become almost synonymous with good colour quality.

But anyone working with lighting knows that it is not always that simple.

Two light sources with the same CRI 90 can make the same materials, objects or skin tones appear different. This becomes particularly relevant in spaces where colour perception is critical, such as retail environments, museums, healthcare facilities, offices or other settings where materials and finishes play an important role.

The explanation lies in the limitations of CRI itself and, above all, in what is happening within the spectrum of the light source.
 

WHY CAN TWO CRI 90 LIGHT SOURCES LOOK DIFFERENT?

CRI was developed by the CIE long before LED technology became widespread. The traditional Ra calculation is based on comparing just eight colour samples, predominantly low in saturation, illuminated by the light source being tested and by a reference source.

The result is an average value that provides a useful indication of colour rendering, but it does not tell the whole story.

One of its best-known limitations is that saturated red, represented by the R9 index, is not included in the calculation of CRI Ra.

This means that a light source can have a high CRI and still perform relatively poorly when rendering reds.

In practice, this difference can be clearly visible. Skin tones may appear less natural, certain foods may look less appealing, and materials containing red or warm tones may appear duller.

This is why, when colour rendering is important, looking only at the CRI value stated on a technical datasheet may not be enough.
 

WHAT DOES CRI NOT SHOW US?

The problem is not that CRI is useless. On the contrary, it remains a simple, widely recognised and useful reference for an initial assessment.

The difficulty arises when a single number is interpreted as a complete description of light quality.

CRI does not tell us, for example, whether certain colour regions are slightly more saturated or more muted. Nor does it make it easy to identify specific areas of the spectrum where significant differences exist compared with the reference source.

With the development of LED technology, it has become possible to create very different spectra for light sources that may display exactly the same colour temperature and CRI on a technical datasheet.

It was precisely the need for a more comprehensive assessment that led to the development of metrics such as TM-30.
 

TM-30: A MORE COMPLETE VIEW OF COLOUR

TM-30, developed by the Illuminating Engineering Society (IES), significantly expands the amount of information available about colour rendering.

Instead of the eight samples used in the traditional CRI Ra calculation, the method analyses 99 colour evaluation samples covering a much broader range of hues and saturation levels.

This analysis provides two particularly important indicators: Rf and Rg.

Rf: How Faithful Are the Colours?

Rf, or the Fidelity Index, assesses how closely colours observed under the light source being analysed match the same colours under a reference source.

The closer the value is to 100, the greater the overall colour fidelity.

It can be understood as a more comprehensive assessment of colour fidelity than that provided by traditional CRI.

Rg: How Saturated Do Colours Appear?

Rg, or the Gamut Index, adds another dimension to the analysis: it indicates the average variation in colour saturation.

An Rg close to 100 generally represents a level of saturation similar to the reference. Higher values may indicate an average increase in saturation, while lower values may indicate a reduction.

This information is particularly useful because two light sources can provide similar colour fidelity while still creating different perceptions of the same colours.

In certain retail environments, for example, a controlled change in saturation may be desirable. In a museum, gallery or environment where accurate reproduction of materials is essential, the objective may be different.

TM-30 also provides graphical representations that make it possible to identify the regions of the colour gamut where changes in hue or saturation occur. Instead of reducing the entire behaviour of a light source to a single number, it allows us to see in greater detail where the differences are.
 

THE SPECTRUM IS THE CAUSE, THE INDICES ARE THE SYMPTOM

CRI, R9, Rf and Rg are different ways of interpreting a reality that exists before any of these values: the Spectral Power Distribution, or SPD.

SPD represents the amount of energy emitted by a light source at different wavelengths across the visible spectrum.

This is where we find the reason why two light sources with 3000 K and CRI 90 may not produce exactly the same visual result.

Despite displaying similar values on a technical datasheet, their spectral distributions may be different. One source may emit more energy in certain regions of the spectrum and less in others, changing the way some colours are perceived.

This is also where the phenomenon of metamerism comes into play: two colours or materials that appear identical under one type of lighting may reveal differences when the light source changes.

For this reason, in projects where colour accuracy is particularly important, analysing the SPD curve can provide information that no single index can fully communicate.

Instead of looking only at a result summarised by a number, we can directly observe how energy is distributed across the spectrum.
 

THE SPECTRUM INFLUENCES MORE THAN COLOUR

The importance of the spectrum does not end with the way we see objects, materials or finishes.

We now know that light is also associated with non-visual responses in the human body and plays a role in regulating the circadian rhythm.

CIE S 026 established specific metrics to assess the influence of different regions of the spectrum on human photoreceptors. These include melanopic EDI (m-EDI), which is used to quantify the melanopic component of light in relation to reference daylight.

This approach has become particularly relevant in spaces where people spend long periods of time, such as offices, schools and healthcare facilities, and is also considered in circadian lighting strategies and building certification systems such as WELL.

This introduces a new dimension to lighting specification.

Two light sources may have similar CRI and Rf values and provide comparable colour rendering while still having different spectral distributions.

Good colour rendering alone does not guarantee that the spectrum will meet the objectives defined for a circadian lighting strategy.
 

WHAT SHOULD WE ANALYSE WHEN SPECIFYING LIGHTING?

There is no single indicator capable of describing every characteristic of a light source. The information required will always depend on the specific demands of the project.

However, when spectral quality is particularly important, there are several parameters worth considering:

CRI (Ra)
It remains a useful and universally recognised reference for an initial assessment of colour rendering. However, it should be interpreted as a starting point rather than a complete description of the spectrum.

R9
It specifically evaluates the rendering of saturated red, a particularly important region for skin tones, food, fabrics and a wide range of materials.

Rf
It provides a more comprehensive assessment of colour fidelity through TM-30.

Rg
It helps determine whether colours tend, overall, to appear more or less saturated compared with the reference.

SPD
It allows the spectral distribution of the light source to be observed directly and can reveal differences that may not be evident from summary indices alone.

m-EDI
It becomes relevant when a project includes objectives related to the non-visual effects of light and circadian lighting strategies.

Not every project requires this level of analysis. For many applications, CCT, CRI and other traditional parameters continue to provide sufficient information for an appropriate lighting specification.

However, when colour rendering, consistency between light sources or the relationship between light and well-being are determining factors, it is important to go further.
 

BEYOND A SINGLE NUMBER

CRI has not stopped being useful, nor is it likely to disappear from technical datasheets. Its simplicity continues to make it an important reference when comparing light sources.

What has changed is our ability to analyse light in greater depth.

Today, we know that two light sources with the same CRI can have different spectral characteristics and produce different results within a space. Metrics such as R9, Rf and Rg, together with spectral distribution analysis, allow us to understand these differences much more comprehensively.

The quality of light cannot be reduced to a single number.

When a project demands greater accuracy, understanding the spectrum means understanding not only how much light we have or its colour temperature, but what kind of light we are actually bringing into the space.