A high color rendering index does not promise that every red object will look the same under two lamps. The familiar CRI score summarizes a set of color comparisons, and an average can conceal the behavior of one important hue. That matters when the object you care about is a red cushion, a painting or a plate of food rather than the average of a test collection.
This is a more specific problem than two warm-white bulbs disagreeing in tint. Two lamps can look similarly white while changing the colors of illuminated surfaces differently. The useful distinction is between the lamp's appearance, its rendering of objects and your preference for the resulting scene.
What the number on the box summarizes
CRI commonly means the CIE general color rendering index, Ra. It compares object colors under a test source with their colors under a specified reference and averages results from eight samples. Those samples are relatively unsaturated. The separate R9 index concerns a saturated red sample and is not part of that eight-sample general average.
Think of the score as a summary with a defined question, rather than a complete portrait of the lamp. A good summary can be useful for filtering options while still omitting the detail you need. If your main concern is red artwork, the average across several other samples is not a substitute for asking about red. This is an interpretation of the scoring structure, not evidence that every high-CRI lamp performs badly.
Nor does CRI measure brightness, energy consumption or the warmth of the light. Those specifications answer other questions. A bulb can satisfy a desired appearance of white and still disappoint when it lights a particular fabric. Comparing all of these qualities under the single word “quality” makes buying harder because there is no single number that means all of them.
R9 fills one gap, with limits of its own
A published R9 result provides additional information about fidelity for saturated red. It is worth requesting when reds matter and a product lists only general CRI. Missing R9 information does not prove the result is poor; it means the advertised average alone does not answer that particular question.
R9 is still a fidelity score. Fidelity describes the magnitude of departure from a reference, rather than whether the departure makes a red more vivid, less vivid or shifted toward another hue. Two deviations of similar size can point in different directions. The TM-30 tutorial hosted by the Department of Energy explicitly cautions that equal red fidelity can accompany quite different renderings of red objects.
That distinction is easy to miss in a shopping comparison. “I like this red better” and “this red is closer to the reference” are both meaningful judgments, but they need not agree. A slightly boosted red may appeal in one room while misrepresenting a textile sample in another. The intended use decides which question deserves priority.
A richer report explains the direction of change
The IES TM-30 framework offers information beyond a general average, including hue-specific changes and a color vector graphic. Rather than forcing every color into one headline score, it can show whether groups of hues tend to increase or decrease in chroma, the strength of color relative to a neutral appearance, or shift in hue.
Research on gamut shape explains why average fidelity and overall gamut area cannot describe every local change. A source may alter different hues in different ways while keeping the same overall summaries as another source. This is why a detailed report can be more useful to a lighting professional than a claim that one bulb has a universally better number.
You need not calculate those metrics yourself to ask a better purchasing question. Ask whether the manufacturer publishes red-specific rendering information or a TM-30 report for the actual model. Compare like-for-like documents and pay attention to the test source and stated conditions. A report for a product family should not automatically be treated as a measurement of every bulb carrying the same brand.
Bring the comparison back to your own red objects
A practical viewing comparison is an application of these distinctions, not a laboratory measurement. Place the actual red object beside a few other familiar colors. Compare candidate lamps in a similar fixture and position, with outside light kept as consistent as possible. Judge both the red and its relationship to the surrounding objects.
Do not let a brighter lamp win merely because it makes the entire scene more conspicuous. Likewise, avoid judging one bulb beside a window at noon and the other at night. Those comparisons change multiple variables at once. If the choice is for a room used under mixed daylight, repeat the observation under that normal condition rather than assuming a night-only result settles every use.
The Color Matching controls offer a useful vocabulary exercise: changing hue and changing saturation are different operations. The browser game does not test a lamp's CRI, R9 or spectrum. It simply makes the difference between directions of color change easier to describe when discussing the real object.
Finally, keep product decisions separate from judgments that color cannot support. The appearance of food under a pleasing lamp is not a freshness or safety test. A convincing red photograph is not a calibrated report of the lamp either. Each additional camera or display introduces another interpretation into the comparison.
The better question at the shelf is not just “How high is the CRI?” It is “What does this score cover, what do I need the light to show, and what happens to those objects under this source?” The headline average remains useful. It becomes more useful when you know where its summary ends.
Editorial backfill for October 7, 2026. Prepared on October 8, 2026.