TM-30: Insufficient with whites — and the question of ultraviolet light
The art of differentiated white rendering without ultraviolet radiation...
In various articles here in the danholt lighting BLOG, I have explained the importance of the different TM-30 metrics and how lighting designers and users can use them as a guide.
However, TM-30 is also useful for other players in the lighting world, including manufacturers themselves. Developing a light source often means making compromises — for example, deciding whether to prioritise lumens per watt or colour rendering. It is crucial that the metrics we use to evaluate these trade-offs are as accurate as possible, so that valuable product properties are optimised rather than abstract numbers.
Today, using the SORAA Vivid full-spectrum LED as an example, I would like to show how such compromises can determine the design of a product.
The SORAA Vivid series aims to render white and all colours of the visible spectrum equally naturally — exactly as they would appear under a natural light source with the same colour temperature (CCT), such as halogen or sunlight.
SORAA’s full-spectrum technology combines a violet boost with three additional phosphors, naturally creating a smooth and continuous spectrum. By carefully balancing the individual spectra, it is possible to come as close as possible to the full spectrum of natural visible light, as shown in Fig. 1.
This leads to natural colours, measured by a high value of the TM-30 colour fidelity index Rf: with the right tuning, for example, Rf = 95 can be achieved.
Fig. 1 Natural sunlight can be very closely approximated with SORAA’s full-spectrum approach, resulting in a high colour fidelity index Rf. Here, the individual peaks of the violet booster and the three phosphors (blue, green, red) are superimposed to create the full spectrum.
So far, however, one crucial question remains unanswered:
How do we achieve a clear and differentiated rendering of whites?
Why are whites different from colours in the first place?
The reason is that many white materials contain so-called fluorescent whitening substances, also known as optical brighteners.
These substances absorb invisible ultraviolet light and emit bluish visible light in return. Our eyes, together with the brain, register this as an increase in whiteness.
Optical brighteners are everywhere: they are found in many manufactured white materials such as fabrics, paper and plastics, and they are also naturally present in our teeth. They are responsible for absorbing ultraviolet radiation and converting it into blue light.
Even though we all like bright, clean white, in most applications we do not want the harmful UV radiation that first makes these whites perceptible.
Conventional LED products naturally avoid UV radiation — but this means that they do not excite the brighteners, making white objects always appear yellowish and dull.
SORAA has found an elegant solution to this problem:
It turns out that optical brighteners can also be excited by harmless violet light — not ultraviolet light — at a carefully selected wavelength. The trick SORAA uses to render natural white is to replace the UV light typical of sunlight with the right amount of simple violet light:
The whitening agents, or optical brighteners, are then excited exactly as they would be under natural light with a UV component.
Fig. 2 illustrates this relationship:
Fig. 2 Halogen and incandescent light stimulate white colours due to their “UV tail”. With intelligent LED design, this can be mimicked by a peak of violet light, which also excites the optical brighteners. In this way, the harmful effects of UV radiation are completely avoided.
How can this effect be expressed in numbers? Surprisingly, there is no strict answer. This is partly because most of the lighting industry, using LEDs with a “blue” source and having no way to determine whiteness, is not even aware of this issue.
And furthermore, to be absolutely clear: colour rendering metrics, including CRI and TM-30, say nothing about white rendering!
This means that a source can have a very high Rf value or CRI and still have terrible white rendering.
Fortunately, the underlying colour science is quite well understood, and it is possible to derive a metric that measures the rendering of white objects in analogy to the colour rendering index Rf.
Based on internal research and academic collaborations, SORAA has done exactly that and developed the white rendering metric Rw. As expected, the Rw value for natural light is around 100.
And here the compromise I promised becomes visible. To achieve the best colour fidelity, we should approximate the shape of the natural spectrum as closely as possible — but to achieve the best white fidelity without UV, we need to add some violet light, not ultraviolet light, to the spectrum, which then deviates from its natural shape!
In short, we are facing an antagonism between homogeneous colour and full-spectrum rendering on the one hand, and the rendering of whites on the other. Accurate measurements are crucial here, because this compromise must be made as well as possible.
Using Rf and Rw, the spectrum of the SORAA Vivid lamps was designed to get the best of both worlds.
This was achieved by optimally tuning the wavelength and intensity of the violet boost in relation to the spectra of the three phosphors. To achieve better performance, particular attention was also paid to another TM-30 index, Rfh1. This parameter measures red rendering and is the modern, better equivalent of CRI R9. In fact, the rendering of red tones is very important for our perception — Rfh1 is therefore at least as important as Rf, perhaps even more important.
As a result, SORAA Vivid achieves very high values across all metrics: Rf = 91, Rfh1 = 95 and Rw = 100. It is fortunate that this specific compromise could be found and optimised without losing the essential aspects of a full-spectrum light source.
Fig. 3 TM-30 colour fields with colour distortion metrics. On the left: a SORAA Vivid with high colour fidelity and high whiteness fidelity (Rw). On the right: seemingly almost identical in TM-30, a standard LED with high CRI also has a high Rf value, but does not render white at all, Rw = 0 — not visible in TM-30.
For comparison, let us now consider an LED source (Fig. 3, right) that has been “naively” optimised only for colours: colour fidelity could reach Rf = 95, but white fidelity would drop to Rw = 0. If this example sounds too dramatic, let me remind you that this is exactly what applies to every blue high-CRI booster LED, which represents the overwhelming majority of products on the market!
Anyone who has compared a white shirt under a SORAA Vivid LED and under a competitor LED can confirm it: the difference is striking!
I hope these insights give you a few useful decision criteria:
1. The design of a light source is a complex and demanding task in which technical and physical aspects mix with subjective perception.
2. Accurate measurements are important because they are the design tool that ultimately allows us to make decisions and then validate them.
Conclusion: Do not focus on just one single metric, but on a range of indicators relevant to your application, such as overall colour fidelity, red rendering and differentiation of whites.
The complete range: SORAA full-spectrum LED
Leave a comment