In recent years, more and more terms that sound highly technical have emerged in the lighting industry: CAF, CS, EML, m-EDI, EDI, DER…
Many manufacturers, designers, consultants, and system providers have heard of them—and to some extent, used them. But if we push one step further and ask:
- What exactly does each of these metrics describe?
- Are they actually talking about the same thing?
- Can they be used interchangeably?
The answers are often far less clear than people assume. This reflects a very typical situation in today’s lighting industry: There are more and more terms, but a true common language has not yet been established.
If the industry genuinely wants to move from simply “lighting up spaces” to accurately understanding how light affects people, then the first step is not to invent yet another new term.
It is to put these commonly used metrics back into their proper context.
Are these metrics really describing the same thing?
Let’s start with the conclusion: CAF, CS, EML, and EDI/DER are not different names within the same framework.
They originate from:
- different stages
- different objectives
- different modeling approaches
Some function more like spectral efficacy ratios.
Some behave more like physiological response models.
Some are closer to application-level compromise metrics.
And others are more like standardized, computable, and transferable baseline coordinates.
So the issue is not whether these terms should exist
The real issue is this: If the industry treats all of them as interchangeable “healthy lighting metrics,” confusion becomes inevitable.
But if each metric is placed back into its proper role, many of the current ambiguities start to resolve themselves.
Why is traditional lighting language no longer sufficient?
In the past, the most familiar language of lighting was built around:
- illuminance
- luminance
- correlated color temperature (CCT)
- color rendering (CRI)
- light distribution
- glare
These metrics are still essential. They primarily serve visual tasks and spatial quality:
- Can we see clearly?
- Is it comfortable?
- Are colors accurate?
- Is the space bright enough?
But the scope of lighting has expanded
A growing body of research now shows that light also affects:
- circadian rhythms
- alertness
- emotional experience
- even certain behaviors and physiological responses
Standards and position statements from the International Commission on Illumination have clearly indicated that: the traditional photopic system is not sufficient to fully describe human responses related to ipRGCs (intrinsically photosensitive retinal ganglion cells).
This changes the fundamental questions
The industry can no longer stop at asking:
- “How many lux is this space?”
- “Is it 3000K or 4000K?”
Instead, it needs to ask:
- Which photoreceptive channels is this light stimulating?
- What does this imply for vision, circadian regulation, and emotional response?
This is the real context behind new metrics
This shift is precisely why metrics like:
- CAF
- CS
- EML
- EDI / DER
have emerged.
They are not just “new terminology,” but attempts to extend lighting language from: visual description → human biological interaction
In other words: from “how the space looks” to “how light actually affects people.”
The human eye doesn’t just “see” — it also “feels” light.
In the lighting industry, the most commonly discussed elements are rods and cones. That’s not wrong.
Rods are mainly associated with low-light (scotopic) vision. Cones are responsible for color, detail, central vision, and typical daytime visual functions.
But today we know that, beyond rods and cones, there is another critically important photoreceptive pathway in the human eye: ipRGCs (intrinsically photosensitive retinal ganglion cells).
These are associated with melanopsin, are more sensitive to short wavelengths, and are closely related to non-visual responses such as circadian rhythms, pupil response, and alertness.
However, to be more precise, what truly needs to be considered is not three systems, but five classes of photoreceptor channels: S-cone, M-cone, L-cone, Rod, and Melanopsin / ipRGC.
What CIE S 026:2018 establishes is a standardized metrology framework based exactly on these five photoreceptors.
In other words, for the first time, the industry has a shared language that is not only about “what can be seen,” but about “how light stimulates the five types of receptors.”
This is a critical shift. Because it means the lighting industry is moving from “spatial output” to “human input.”
What do CAF, CS, and EML actually represent?
1. CAF: closer to a “spectral efficiency ratio” mindset
CAF (Circadian Action Factor) has long been used to compare the potential of different spectra to stimulate circadian-related responses.
Its core logic is straightforward: under the same visual lighting conditions, is this spectrum more biased toward “circadian effect” or “visual effect”?
So CAF is essentially a weighted efficiency ratio. It helps compare different SPDs under the same photopic lux to determine which produces stronger circadian-related stimulation.
This approach is not without value. Its advantages are simplicity, intuitiveness, and suitability for early-stage comparisons.
But it also has clear limitations:
- First, it reflects spectral properties rather than actual human exposure dose.
- Second, it does not inherently include time, spatial context, viewing direction, or actual eye exposure.
- Third, it is not the primary shared language in current international standards.
So CAF helps you understand “how biased a spectrum is,” but it is not suitable as a complete coordinate for human response.
CAF is more like a spectral screening tool, not a full human-centric lighting language.
2. CS: closer to a “specific physiological response model”
CS (Circadian Stimulus) has also had significant influence in recent years, especially in North America. Its logic differs fundamentally from CAF.
Rather than being a simple ratio, CS attempts—through a circadian phototransduction model—to map spectral stimuli onto a response scale related to melatonin suppression.
UL’s DG 24480 also proposes design targets based on this type of framework. The strength of CS is that: It goes beyond saying “more or less biased,” and tries to quantify “how strong the circadian system stimulation is.”
But this is also where the challenge lies. Once a metric moves from “describing input” to “predicting response,” it inevitably introduces modeling assumptions:
- What spectral sensitivity functions are used
- How rod, cone, and melanopsin interactions are handled
- How dose-response is defined
- How exposure duration is treated
- How pupil state, timing, and exposure history are incorporated
As a result, CS has been accompanied by considerable methodological debate.
So a fair summary would be: CS is important, but it is better suited as an application-layer or response-layer model, rather than a foundational common coordinate system for the entire industry.
3. EML: closer to a “transitional language for application”
EML (Equivalent Melanopic Lux) has been widely promoted in application contexts such as WELL.
Its key contribution is that it helped many people realize, for the first time: Not all lux are the same.
From a communication and adoption standpoint, EML has played a significant role. It translates complex spectral–receptor relationships into a format that is easier to understand and specify in project requirements.
However, from a stricter standardization perspective, EML is not the ideal end state. The industry has increasingly shifted toward melanopic EDI, and further toward the more comprehensive α-opic EDI / DER framework, because these align better with the standardized structure defined in CIE S 026 and enable consistent use across organizations and systems.
So in one sentence: EML is a bridge toward human-centric lighting—but not the most suitable final coordinate system.
Why are EDI / DER closer to a true coordinate system?
Because they resemble a system that can be recorded, compared, and transmitted—like a “spectrum.”
1. EDI: describing “equivalent stimulus dose”
EDI (Equivalent Daylight Illuminance) can be understood as: How much illuminance from standard daylight (D65) would be required to produce the same level of stimulation for a given photoreceptor?
This allows results from different spectra to be compared within a unified framework.
2. DER: describing “stimulation efficiency”
DER (Daylight Efficacy Ratio) can be understood as: How efficient a given light is, per unit of photopic illuminance, at stimulating a specific photoreceptor.
CIE TN 015:2023 clearly defines the relationship: melanopic EDI = illuminance × melanopic DER
Together, these two quantities are powerful:
- EDI reflects the actual dose reaching the human body
- DER reflects the intrinsic efficiency of the light spectrum
One is exposure-focused, the other is source-focused. This combination is exactly what manufacturers, designers, control systems, and simulation tools need.
Why move from single m-EDI toward a full EDI / DER framework?
This is not a rejection of melanopic metrics. In fact, many recent consensus recommendations are indeed centered on melanopic EDI.
For example, Brown et al. (2022) suggest indoor light exposure guidelines such as:
- At least 250 lx during the day
- Preferably below 10 lx in the evening
- As close to 1 lx as possible at night
These are important. But looking further ahead: Humans do not respond to light through melanopsin alone.
Visual performance, color discrimination, adaptation, spatial perception, aspects of emotional experience, and more complex neural responses all involve the combined action of rods, S/M/L cones, and ipRGC pathways.
So if the industry aims to build a future-oriented human-centric lighting coordinate system, focusing only on m-EDI is not enough.
What we need is a more complete EDI / DER framework: Not just melanopic—but incorporating stimulation across all five photoreceptor classes into a unified language.
This does not mean every project must present all five values. It means: The foundational language of the industry should leave room for a complete human model.
From “selling light” to “describing humans”: the industry needs a new staff notation
I like to use an analogy: EDI / DER in human-centric lighting is like musical notation in music.
Musical notation is not the music itself, but it is the foundational language that allows music to be recorded, transmitted, reproduced, and collaboratively created.
EDI / DER is similar.
It is not sleep itself.
Not emotion.
Not comfort.
Not spatial aesthetics.
But it provides a way to more precisely describe: What this light is doing to the five photoreceptive channels of the human body.
With such a coordinate system, many long-standing ambiguities in the industry can finally be addressed collaboratively:
- LED manufacturers can provide more meaningful spectral data
- Luminaire manufacturers can define products in terms of human impact
- Control systems can move beyond brightness and CCT to modulating receptor stimulus
- Simulation tools can evolve from illuminance-based to human-input-based modeling
- Designers can move from “feels healthier” to “designing with coordinates”
Without such a system, the industry easily remains stuck in vague language:
More natural
Closer to daylight
More circadian-friendly
More comfortable
Healthier
These terms are not useless—but without an underlying framework, they struggle to become a shared language across organizations and product chains.
The real value of EDI / DER lies in this: For the first time, “how light affects humans” can be written down—like a score.
What does this mean for LEDs, luminaires, systems, and designers?
For LED and module manufacturers
Future competitive data cannot be limited to lm/W, CCT, and CRI.
SPD and α-opic / EDI / DER information will become increasingly critical.
For luminaire manufacturers
In the future, luminaires won’t just deliver lumens into space.
They will deliver specific receptor-stimulation structures to the human eye.
For control system manufacturers
Control strategies should no longer stop at “what time to switch to what CCT and what dimming level.”
A more advanced control objective should be: To achieve a target balance of stimulus dose and experience
for a given time, space, task, and user group.
For designers
Human-centric lighting design will go beyond “cooler in the morning, warmer in the evening.”
It will require thinking in terms of:
- Which photoreceptors this light primarily stimulates
- What the actual dose at the eye level is
- How to balance visual performance, circadian support, and emotional experience
- How daylight, electric light, reflections, and viewing direction interact
Once designers start thinking this way, lighting design evolves from “placing fixtures” to “modulating human response.”
Final point: the industry doesn’t lack terms—it lacks the ability to read the “score”
CAF, CS, EML, EDI / DER…
These terms often feel confusing not because they lack importance, but because they are frequently discussed at the same level.
In reality, they answer different questions:
- CAF → more like a spectral efficiency ratio
- CS → more like a specific physiological response model
- EML → more like an application-layer transitional language
- EDI / DER → closer to a standardized, computable, and transferable coordinate system
If the industry truly wants to move from “illuminating spaces” to “effectively influencing people,” the next step is not to invent yet another concept— but to learn how to read this system.
Illuminance tells us how bright it is. EDI / DER begins to tell us how light acts on humans. And that may well be the real starting point of the human-centric lighting era.
CTA
If your organization is exploring:
- How to upgrade LED or luminaire data from traditional photometric parameters to a language closer to human-centric lighting
- How to integrate EDI / DER into product definitions, control systems, or design simulations
- How to establish lighting evaluation methods that address circadian rhythms, visual performance, and emotional experience
You’re welcome to get in touch.
English 
Chinese (Taiwan) 
Chinese (China) 
German