How to Translate Easily | How to Read CRI Ra and R9 Reports, Colour Samples and Rendering Limits

How do you read a CRI report, and what do Ra and R9 really tell you about a light source? A colour rendering index report compares how selected colour samples appear under a test lamp with how they would appear under a defined reference. Reading it accurately requires more than finding a large number on a product label. The headline Ra, the individual test-colour scores, and the conditions under which the measurements were calculated each answer different questions.

This guide explains how to interpret a CRI Ra and R9 report step by step: how the standard test samples work, why a high general colour rendering index can coexist with a weak saturated-red score, and how to translate a numerical lighting report into useful, qualified language. It also shows how to check whether two reports are genuinely comparable and where a fuller rendition assessment, such as TM-30, adds information.

The aim is not to select a particular lamp or prescribe a universal acceptable score. It is to help anyone reading lighting specifications, laboratory reports, architectural schedules, product comparisons or educational charts say what the data supports and avoid claims it does not support.

Start with the question the report can answer

A CRI report addresses colour fidelity relative to a specified reference illuminant. The ordinary question is: if the objects represented by standard test colours are illuminated by this source, how different are their calculated colour appearances from the reference? This differs from brightness, colour temperature, energy efficiency, glare, lifetime or the subjective attractiveness of a room. Those matters may influence a design decision, but they are not directly measured by the Ra number. When a description says that a source has ‘CRI 90’, the cautious translation is that the calculated general colour-rendering score is approximately 90 under the specified method and conditions. It is not that every colour is reproduced with ninety-percent accuracy.

A useful reading habit is to write the report’s question in a full sentence before discussing its numbers. For instance, ‘The eight standard general-test samples are rendered with relatively small aggregate colour differences from the reference’ is closer to the method than ‘colours are ninety percent correct’. The first statement identifies a comparison, a limited sample set and an aggregate. The second invents a percentage interpretation that the index does not carry. Careful translation begins by protecting those distinctions.

Identify the general index Ra before looking at individual values

The general colour rendering index, conventionally written Ra, is the arithmetic mean of the first eight CIE special colour rendering indices, R1 through R8, in the traditional method. These eight test-colour samples have comparatively moderate saturation. Each special index expresses the calculated colour difference for that sample relative to a method-defined reference. The calculation is not a vote from eight people and is not an observed percentage of satisfied viewers. It is a constructed colorimetric comparison. That construction makes it useful for repeatable specifications but also explains why one average cannot describe every colour or every visual task.

Suppose eight individual scores are 91, 93, 95, 89, 94, 92, 90 and 96. Their arithmetic mean is 92.5, which may be displayed as approximately Ra 93 depending on reporting precision. The average tells a reader something real about this particular group, but it hides the variation between members and says nothing directly about sample R9. A careful report summary should retain that scope: ‘Ra is about 93 based on the eight general test samples.’ If a reader wants to know whether a specific red pigment will look faithful, the general average is only a starting point.

Locate R9 and understand why it is separate

R9 is a special colour-rendering index for an additional saturated-red test sample. It is not included in the conventional Ra average. A lighting datasheet may therefore display Ra prominently while omitting R9 entirely, even when the latter would be informative for a colour-sensitive application. That omission is not proof that a lamp is inadequate; it is a missing piece of evidence. The sensible response is to request the individual score or a fuller report, rather than guessing it from Ra.

A weak or negative R9 can occur alongside a fairly high Ra because the eight general samples do not challenge exactly the same part of the colour space as the deep-red sample. This makes R9 a diagnostic addition, not a replacement for Ra. R9 also does not predict every shade of skin, every red textile, or every food surface. Each object has its own reflectance spectrum, and many practical outcomes depend on adaptation, surrounding colours and illumination levels. The right translation is: ‘The report shows how the specified saturated-red test sample compares with the reference,’ followed by the actual result if it is known.

Read the report as a comparison, not a direct photograph

A CRI calculation uses the measured or specified spectral power distribution of the test source, defined test samples and a selected reference illuminant. It predicts shifts in colour appearance within the method’s model. It is not a photograph of a shop, a museum gallery or a person’s face. A promotional swatch picture may visually suggest the report’s direction, but photographs bring camera response, white balance, display calibration, viewing environment and editing into the chain. Those factors can change what a viewer sees without changing the source’s underlying CRI calculation.

If a report presents a grid of patches, ask whether it shows the standard sample colours, a calculated comparison against the reference, or an illustrative rendering. Those are not the same. The accompanying numerical table and stated method normally carry more traceable evidence than a dramatic before-and-after image. When publishing an explanation, describe the basis of the illustration plainly. No image is needed to understand the principle: the evidence begins in the spectral measurement and its documented conversion into colour differences.

Understand the reference illuminant

The reference source matters because fidelity is relational. Traditional CRI methods compare a test source with an appropriately chosen reference of similar correlated colour temperature, using reference spectra associated with thermal radiators at warmer temperatures and daylight-type references at cooler temperatures. The precise reference follows the standard’s rules. The reference is not necessarily the lighting someone will prefer in a given setting, and matching it does not automatically make a scene desirable. A source can agree closely with its reference while the overall atmosphere remains too bright, too dim, too cool, or otherwise unsuitable for the task.

This is also why blindly comparing two Ra numbers without checking the measurement basis can be unhelpful. They may have been evaluated at different colour temperatures and therefore against different reference spectra. A comparison is still possible, but its language must respect what is held constant and what is not. Record CCT, reference choice where given, standard version, spectral measurement conditions and reported precision. A strong technical summary names those conditions before making a performance claim.

Separate colour fidelity from colour preference

Fidelity describes closeness to a reference; preference concerns what an observer finds attractive or useful. A light source can make certain colours more saturated than its reference, and some viewers may prefer that presentation for some activities. Traditional Ra does not independently report the direction of saturation shifts or provide a universal score for preference. A high general index should therefore not be translated as ‘best-looking colours’, ‘most comfortable light’ or ‘healthiest illumination’. Those claims need different forms of evidence.

Imagine two lamps intended for a textile display. One may have a slightly higher Ra while the other creates a colour appearance favoured for a particular set of fabrics. Without a controlled visual evaluation and context, neither observation defeats the other. The proper question is whether fidelity to a defined reference is the principal requirement, or whether appearance, consistency, camera response and task performance also matter. Translating a CRI number into a purchasing decision without that intermediate reasoning is where many explanations become misleading.

See how spectral gaps produce selective rendering differences

Objects look coloured because they reflect, transmit or emit wavelength-dependent light. For a reflecting object, the incident source spectrum interacts with the object’s reflectance spectrum. Two sources with similar white appearance can contain different distributions of radiant power across wavelengths. A source with comparatively little power in a region important to a particular object’s reflectance may reproduce that object’s colour differently from a reference, even if the source’s overall white point looks similar.

The mechanism helps explain why a single global index cannot tell the whole story. Ra compresses calculated differences into one mean across selected samples. The individual R values reveal more of the distribution of differences, while the source’s spectral power distribution offers an upstream clue as to why they might occur. However, a peak or dip in an SPD plot does not by itself specify how any particular real object will appear; that prediction requires the object’s reflectance data and the relevant colorimetric transformation. Read a spectral graph as physical input, not a ready-made colour verdict.

Translate scores without converting them into percentages

A value of Ra 85 is not an assertion that colours are reproduced at eighty-five percent of their true appearance. Nor does R9 equal the percentage of red light in a lamp. The indices arise from specific differences in a colorimetric space and a scoring transformation. A negative special index can be possible under the method, so a simplistic percentage metaphor is especially unsuitable. When explaining a report to a non-specialist, avoid the words ‘accuracy percentage’ unless a separate study has actually defined and measured an appropriate percentage quantity.

A clearer explanation is relational and plain: ‘A higher Ra generally indicates closer colour matching for the eight specified test colours relative to the chosen reference.’ For R9: ‘The R9 number describes the method’s rendering comparison for an additional saturated-red sample.’ If you need to compare two scores, say exactly which metric has increased, by how much, and under which calculation. The numeric difference between scores should not automatically be equated with a proportional improvement in human satisfaction.

Recognise what an individual R1–R15 chart can reveal

Some reports show R1 through R15, or a subset of the special indices. The R1–R8 values make up Ra. R9 adds saturated red; other individual indices represent additional defined samples. A bar chart can help readers identify whether the headline average conceals a particular weak spot. The correct reading order is: find the legend, confirm the standard and metric, identify individual labels, compare each bar with the reference framework, and then decide which differences matter for the intended task.

Do not infer that fifteen displayed bars mean the general CRI has been computed from fifteen samples. Under the conventional definition, Ra remains an average of eight. Also avoid assuming that a sample’s common descriptive name perfectly represents all real objects carrying that colour word. A ‘skin’ test sample is a specific reflectance target, not a biological measurement of all complexions. If a chart only labels bars by number, the most defensible prose is ‘R9 is lower than R1 through R8’ rather than an invented account of exactly how every red object will look.

Check the vertical axis before interpreting a bar chart

A bar chart can make a modest numerical gap look enormous if the vertical axis starts near the plotted values rather than at zero. A report showing two bars at 92 and 95 may appear dramatic if the axis begins at 90. The opposite problem occurs when a plot compresses a broad range so much that important differences are visually muted. The axis is a design choice, not a second measurement. Read printed values where available and inspect ticks, truncation, negative ranges and sorting.

Pay attention to whether the chart mixes general Ra, special Ri scores, an Rf value from another methodology or colour-gamut indices. Numerically similar values from distinct scales are not automatically interchangeable. If multiple systems appear together, describe each by its proper name and avoid a single ranking that treats them as one statistic. For a reproducible interpretation, copy the metric label, score and standard as reported; then explain the decision relevant to each metric.

A worked example: high Ra and a low R9

Consider a hypothetical lamp report that lists Ra 91 and R9 8. This example is invented for instruction, not a real test record. The correct observation is that the general index across the first eight test samples is relatively high while the special index for saturated red is much lower. This provides a reason to investigate red rendition more carefully in applications where red surfaces or detailed colour differentiation matter. It does not prove that all reds will appear wrong or that the lamp is unsafe.

A useful explanation might read: ‘Although the lamp’s general CRI is high, its reported R9 result is much lower, indicating that the additional saturated-red test colour is rendered less faithfully relative to the method’s reference. Inspect a fuller colour-rendering report and, where important, test actual objects under representative conditions.’ This statement keeps the quantitative observation, limitation and proposed verification distinct. The unhelpful alternative would be ‘CRI 91 means perfect colour except red’, which overstates both the overall performance and the diagnosis.

A worked example: similar Ra, different distribution of scores

Imagine two hypothetical test sources, A and B, each with Ra 90. Source A has relatively even R1–R8 values near that average, whereas B combines some higher and lower values that average to the same Ra. If the report includes only the mean, the two sources look identical on the headline metric. With individual values, a reader can see that the pattern is different. Add a further difference in R9 and the case becomes a lesson in why an average compresses evidence.

This distinction matters beyond lighting. A class average, clinical average or mean error can conceal variation across cases. In a CRI report, the relevant cases are defined test colours. The appropriate translation is not ‘the sources perform identically’ but ‘the sources share this aggregate Ra value; their per-sample behaviour must be checked separately’. A reader does not need to know every detail of the scoring algorithm to understand why distributions cannot always be recovered from their averages.

A worked example: same Ra at different colour temperatures

Suppose a warm-white source and a cool-white source both report Ra 95. That does not imply their emitted light looks like the same white, or that a fabric illuminated under one will look identical under the other. Ra is about fidelity relative to each source’s method-defined reference, not the equality of the sources’ chromaticities or spectral power distributions. The sources may also differ in perceived ambience and in their suitability for the viewer’s adaptation state and task.

If an observer asks which lamp has ‘more accurate colour’, first clarify the reference goal. For maintaining a target appearance across two distinct lighting environments, a comparison may require colour constancy experiments, appropriate adapted-state colorimetry or direct measurement of the relevant samples under both sources. For simple quality screening within each CCT family, the Ra reports may remain informative. The art of translating a score is knowing what question the score answers and refusing to make it answer a different one.

Why Ra cannot substitute for correlated colour temperature

CCT describes a white light’s chromaticity by reference to a nearby Planckian radiator in a particular colorimetric system. It is typically given in kelvin. Ra describes calculated relative colour rendering. CCT is not a quality grade, and Ra is not a label for warm or cool appearance. A lamp described as 3000 K with Ra 90 and another described as 5000 K with Ra 90 share a general rendering score but differ in white appearance. The pairing of values answers two independent questions.

Duv can add information about the direction of tint relative to the Planckian locus, a subtle distinction that CCT alone does not capture. That third measure is also separate from Ra. A competent datasheet explanation keeps the three on different lines of reasoning: CCT for approximate warm–cool appearance, Duv for a tint deviation in the defined space, and rendering metrics for calculated colour differences between test and reference. Collapsing them into a phrase such as ‘ninety percent daylight’ loses the meaning of all three.

Why illuminance and glare require different measurements

A high-Ra lamp can be uncomfortable if it produces glare, is poorly positioned or creates excessive contrast. Conversely, a lamp with modest colour fidelity can still meet a task’s illuminance requirement. Lux describes illuminance at a surface, while glare depends on luminance distribution, viewing geometry and other factors. CRI does not measure these. The decision to choose a light source for reading, assembly, artwork or public circulation requires more than a rendition chart.

When a sales summary says a high CRI lamp is therefore ‘gentle on the eyes’, treat that as an unsupported logical leap unless independent evidence is provided for the relevant conditions. The correct move is not to dismiss Ra but to put it in its compartment. Ask for illuminance measurements, luminance or glare assessment, flicker information where relevant, and the actual environment of use. A carefully translated report helps the next decision-maker request the right missing tests.

Keep camera performance separate from human-observer rendering

A camera’s spectral sensitivities and colour-processing pipeline differ from the standard human-observer assumptions in traditional colour-rendering calculations. Two sources that have similar CRI values can produce different camera colours, and a source attractive to the eye may introduce difficulties for video or still photography. White balance can correct some neutral shifts but cannot generally reconstruct colour information absent from the recorded signal.

Broadcast and production contexts may therefore refer to camera-oriented metrics or conduct controlled camera tests. When writing about CRI in a photographic context, state that Ra and R9 are human-observer colour-rendering indicators, not comprehensive camera-match scores. Record the camera, profile, exposure, white-balance setting and test targets if a photographic comparison is made. This avoids treating a consumer photograph of a display as laboratory evidence of CRI, or attributing a camera colour error solely to the light source without checking the whole image chain.

Understand measurement uncertainty and rounding

A report is produced from measurements and calculations, and every link carries conditions. Spectroradiometer calibration, wavelength accuracy, stray-light correction, measurement geometry, source stabilisation and the algorithm used to calculate the final indices can affect results. Two laboratories may report small differences even when assessing nominally similar devices, particularly if the sources are unstable or conditions are not aligned. A product datasheet may then round those results to a convenient integer.

If two reports give Ra 92 and Ra 93, do not automatically present the second as visibly superior. First check whether the results are within the stated measurement uncertainty, whether they were obtained under comparable operating conditions, and whether the difference is meaningful for the selection task. If no uncertainty is documented, the careful phrase is ‘the reported values differ by one point’ rather than ‘one lamp produces measurably better colours’. Rounding and confidence are part of translating data honestly.

Know when a test sample’s colour name can mislead

A defined colour sample is a material or reflectance description within a calculation. Its everyday label—yellow-green, light skin, red, blue or another term—helps readers remember its approximate region, but it does not capture all possible object spectra. Two real objects that seem similarly red under one lamp may reflect light differently and diverge under another. A score for one red test sample is evidence about that sample, not a universal theorem about all red objects.

When presenting sample charts to a general audience, prefer language such as ‘the R9 saturated-red test colour’ to ‘all saturated reds’. A similarly cautious note applies when a sample is used as shorthand for a class of human appearances. Avoid claiming that any one prescribed swatch adequately represents diverse skin tones or practical design contexts. Detailed decisions should involve representative materials, suitable viewers and controlled lighting conditions in addition to general indices.

Compare CRI with TM-30 without treating them as the same statistic

TM-30 reports different complementary information about colour rendition using a substantially larger set of colour evaluation samples. Its fidelity index Rf and gamut index Rg answer related but distinct questions. Rf summarizes fidelity within that system, while Rg reports a form of average gamut change relative to its reference. Colour vector graphics can also show directional effects across hue regions. Traditional Ra and R9 remain useful familiar quantities, but they do not exhaust what these later tools can show.

Do not convert a CRI value directly into an Rf value through a universal fixed offset, and do not compare a TM-30 Rg number of 100 to an Ra number of 100 as if both were percentages of the same property. Keep index names explicit. A sound comparison might say: ‘Traditional CRI gives a general eight-sample score and selected special scores; TM-30 offers a broader account of fidelity and colour-gamut changes.’ Which measures belong in a project specification depends on the project’s goals and requirements.

Build a fair comparison table for two products

A useful comparison starts with identifiers: product model, test mode, drive setting, nominal or measured CCT, the method version and measurement date where supplied. Then place Ra, R9 and any additional special indices in clearly labelled columns. Add TM-30 Rf and Rg only with unambiguous headers. Finally record the data source and any laboratory or instrument information. Without these identifiers, a tidy table can combine results that were measured under unrelated operating states.

Keep omissions visible. If product A discloses R9 and product B does not, write ‘not reported’ rather than zero. If a manufacturer describes a value as ‘typical’ instead of a tested minimum, preserve that word because it changes the nature of the claim. If the datasheet reports a range, retain the range. A transparent table is not merely aesthetically tidy; it shows precisely where evidence ends and assumptions begin.

Read performance changes across dimming and drive current

LED output spectra can change with operating conditions, depending on device architecture and control approach. A score measured at a stated drive current does not automatically apply at every dimming level, operating temperature or ageing state. For colour-sensitive installations, obtain data at the settings that actually matter. If a report compares several conditions, label them consistently and inspect both changes in white point and changes in rendering metrics.

For instance, imagine a luminaire tested at full output and at a low dimming level. If its CCT changes and its R9 value moves, two separate effects may be present: the light appears a different white and it renders the selected red sample differently relative to the relevant reference. Do not sum those effects into an invented single ‘colour loss’ quantity. Instead report both measured changes, check uncertainty and ask whether those changes matter in the intended use conditions.

A responsible procurement interpretation

The procurement task is not to buy the largest number available. It is to establish a performance requirement and verify that the evidence matches it. Colour-sensitive inspection, galleries, workplaces, homes, video production and public environments have different practical goals. Some may prioritise predictable reproduction of specified materials; others may balance efficiency, glare control, visual comfort, affordability and a preferred appearance. A CRI report informs the colour-fidelity component and should not replace the whole specification.

A robust decision might require actual object tests, a current report from the intended operating configuration, documented tolerances across units and a complementary gamut assessment. It should also consider whether the procurement language refers to ‘minimum’, ‘nominal’ or ‘typical’ values. These terms matter: a typical Ra 90 claim is not necessarily a guarantee that every delivered unit meets 90. Translate the commercial claim accurately, then ask for acceptance criteria and measurement methods if the decision depends on a strict threshold.

A museum example: fidelity is necessary context, not the whole conservation problem

Consider a museum team comparing two light sources for an exhibition. CRI can contribute to discussion of colour rendition, but conservation decisions may also depend on radiant exposure, spectral content outside the visible region, exposure duration, sensitive materials and environmental conditions. A high Ra does not certify that artwork will avoid photochemical damage. Nor does a particular R9 determine whether all red pigments will look as intended under a display illuminant.

An appropriate summary would say: ‘The candidate source has the stated general and saturated-red rendering scores. The team should separately evaluate the exhibition’s colour appearance under representative viewing conditions and the conservation requirements of the materials.’ This is a useful example of scope control. One report supports one part of a complex decision; it must neither be ignored nor inflated into a comprehensive certificate.

A healthcare example: distinguish colour fidelity from clinical validation

Clinical tasks involving colour observation can make lighting performance important, but a manufacturer’s Ra or R9 number alone does not establish diagnostic suitability. Skin, tissue and diagnostic targets vary in spectral reflectance, and the relevant visual tasks require task-specific evaluation. A generic ‘CRI above 90’ phrase does not document the conditions, observer factors or clinical performance measures needed for a stronger claim. It can be one screening requirement, not the entire validation method.

When translating lighting specifications for healthcare audiences, be precise about evidence levels. You may state that a lamp has the specified rendering index, and you may explain why selected colour differences could matter to visual inspection. Do not state that it improves diagnosis or patient outcomes unless such effects have been tested in an appropriate context. The distinction protects readers from extending a lighting calculation into a medical conclusion.

A workplace example: choose evidence for the task

A workplace can include inspection stations, office desks and general circulation. The same colour-rendering threshold may not be equally informative for every activity. A surface-inspection task involving small colour differences may call for well-controlled illumination and real sample comparisons. General circulation may focus on other performance requirements as well. Translating the report into a site plan therefore means connecting the measured quantity to what workers actually need to see.

Imagine a team that has already selected two lamps with similar Ra values. The next step could be to examine individual indices, TM-30 information or the spectral reflectance of actual products under each lamp, depending on the task. An effective report summary should specify the unanswered operational question instead of declaring a winner solely from a marginal index difference. That makes the translated information actionable without overstating certainty.

A design example: visual appearance can differ without an index error

Two luminaires with similar Ra, CCT and even R9 can produce somewhat different impressions in a furnished space. Their spectral distributions, optical distributions, intensity patterns and surrounding finishes may differ. Observers adapt to the overall field, and the same material appears in relation to neighbouring colours. None of these observations implies that the reported Ra was false. It means the index was never intended to be a full description of lived visual appearance.

A designer can use the report for initial screening, then examine physical samples in a controlled mock-up. Written commentary should separate instrument-derived evidence, controlled subjective evaluation and personal preference. For example: ‘Both candidates meet the specified Ra requirement, while the mock-up revealed different appearances on the selected materials.’ Such wording preserves the authority of each observation without pretending that the metric predicts all design outcomes.

Recognise common misleading phrases

Watch for descriptions that say ‘CRI is the percentage of visible colours reproduced’, ‘R9 measures the percentage of red wavelengths’, or ‘a CRI of 100 recreates sunlight exactly’. None accurately states the method. A reference source receives a perfect match by definition within the relevant calculation; that is not the same as replicating daylight in every spectral detail. Likewise, a large Ra number does not establish that a lamp is flicker-free, energy-efficient, biologically beneficial or ideal for photography.

Replace vague superlatives with traceable comparisons. Say ‘reported Ra 95 under the stated method’ rather than ‘ninety-five-percent natural light’. Say ‘R9 not supplied’ rather than ‘R9 probably fine’. Say ‘the sources were compared using different reported test settings’ rather than claiming a precise performance order. Technical language need not be intimidating. It can be concise, but it must retain the variables that prevent a sentence from becoming false.

Translate an unfamiliar report in five passes

First, identify what is being reported: a general Ra index, a special Ri index or a different rendition metric. Second, note the test source and reference context, including CCT and measurement or calculation method if disclosed. Third, read the individual scores, units and chart axes without assuming that the number is a percentage. Fourth, compare only like with like, and state any missing values or uncertainties. Fifth, express the practical meaning in a short sentence that includes the limit of the evidence.

This order is useful because it postpones judgement until after identification. A reader who begins with ‘Which one is best?’ is likely to force the report into a ranking it cannot reliably support. A reader who begins with ‘Which property was measured?’ can build a justified comparison. The same discipline transfers to laboratory graphs, translation tasks, educational assessments and other evidence-rich writing: preserve the original meaning first, then interpret.

A reporting template that preserves the facts

An accessible technical description can take the following form: ‘At the stated operating condition, the source has a reported general colour rendering index Ra of [value]. Its special saturated-red index R9 is [value or not reported]. These indices compare specified test colours with a defined reference under the cited method; they do not measure every possible object or predict lighting preference. For the intended application, check [additional samples, full spectral data, uncertainty, or visual mock-up].’ This is a template for documenting evidence, not a formula for making a product sound good.

Notice the structure. The first sentence reports a measurement or manufacturer claim. The second specifies its diagnostic scope. The third prevents a false universal inference. The final sentence proposes a test appropriate to the unresolved question. If a reader can reproduce that reasoning from the disclosed numbers, the translation has done its work.

Frequently asked questions: is Ra 100 perfect light?

Ra 100 means the test source matches the selected reference in the method’s general eight-sample colour-rendering calculation. It does not say that every object’s appearance is identical under all light sources, that the light has ideal intensity or that it produces no visual discomfort. It is a relative statement about a particular family of standard test colours. Even a light source with an excellent Ra may need other measurements for a complete specification.

A related question is whether a modest difference such as Ra 92 versus Ra 94 will always be visible. It will not necessarily be perceptible in a given environment, and no responsible conclusion follows without considering samples, adaptation, viewing conditions and measurement uncertainty. The index is a screening and comparative tool, not an automatic predictor of a viewer’s experience.

Frequently asked questions: does a negative R9 mean a broken measurement?

Not necessarily. Special indices can have negative values because of the score’s calculation structure. A negative R9 should prompt attention to the saturated-red test sample and verification of the report’s validity, but it is not by itself evidence that the instrument malfunctioned. Confirm the source identity, reference choice, method and testing conditions before drawing stronger conclusions.

Another frequent question is whether two lamps with equal Ra must have equal R9. No. Ra excludes R9, and two sources can have quite different individual-score patterns while yielding similar averages. Ask for the special indices if the distinction matters.

Frequently asked questions: does CRI tell me how a camera will render colours?

Not fully. Traditional CRI uses standardized colorimetric assumptions for human vision rather than a specific camera’s sensor and processing. A camera test may expose differences that the Ra average does not resolve. If the application is broadcasting, photography or video, request camera-appropriate evidence, carry out a representative shooting test and document white balance and recording settings.

If a product advertises both Ra and TM-30 values, record both without blending them into one percentage. Each framework supplies particular information, and a careful explanation can describe why they complement one another.

Conclusion: report the measured comparison, then explain the limitation

The most reliable way to read a CRI report is to keep each value attached to its definition. Ra summarises a traditional eight-sample general rendering comparison. R9 examines one additional saturated-red test colour. Reference illuminants, CCT, spectral power distributions and test conditions shape the result. Individual values, fuller methods and representative object tests may be necessary for a decision beyond basic screening.

A good translation is not a bigger claim than the report. It is a clearer claim: what was compared, what was found, what remains unknown and what should be checked next. That makes colour-rendering numbers useful to designers, researchers, educators and readers without turning them into promises they never made.