Circadian Science · Deep Dive · CIE S 026/E:2018
Melanopic EDI:
The Metric Your Lux Meter Cannot Measure
Every lighting recommendation you've ever received — "use warm bulbs," "keep it dim," "avoid blue light" — was built on the wrong measurement. Standard lux measures what your eyes use for vision. Melanopic EDI measures what a completely separate system in your retina uses to set your biological clock. The two numbers can point in completely opposite directions. This guide explains what melanopic EDI is, how it's calculated, and what the numbers actually mean for your health.
Why Lux Gets It Wrong — Two Systems, One Light Source
Lux is a perfectly good measurement — for vision. It was designed to quantify how bright a light appears to the human visual system. The problem is that "how bright it looks" and "how much it disrupts your sleep" are two completely different questions with different answers.
Here's the simplest way to understand the gap. Imagine two light sources, both measuring exactly 100 lux on a standard meter:
Source A is a 5000K cool-white LED — the kind found in offices and kitchens. At 100 lux it looks bright and crisp. Your melatonin? Almost completely suppressed. Your circadian clock? Convinced it's mid-afternoon.
Source B is a 590nm narrow-band amber LED. At 100 lux it looks noticeably orange, like a dim campfire. Your melatonin? Flowing normally. Your circadian clock? Completely undisturbed.
The lux meter reads 100 for both. The biological effect differs by roughly 45-fold. That gap is precisely what melanopic EDI was invented to measure.
The reason lux misses this is structural. The lux measurement uses a sensitivity curve called V(λ) — the standard luminous efficiency function — which peaks at 555nm (yellow-green) and was designed to match the combined response of cone cells in the human eye. It is an excellent model of how bright things look. It has nothing to do with the separate photoreceptor system — intrinsically photosensitive retinal ganglion cells (ipRGCs) — that regulates melatonin and circadian timing. That system peaks at 480nm, deep in the blue range, and is nearly silent by 590nm.
A lux meter at 590nm registers a significant reading because the amber light stimulates cones. The ipRGC system at 590nm registers almost nothing, because melanopsin barely responds there. One reading says "bright." The other says "dark." Both are correct for their own purpose. Only one tells you what's happening to your sleep.
The melanopsin peak at 480nm and the lux peak at 555nm are 75nm apart. That gap spans the difference between blue-cyan and yellow light. A standard lux meter is essentially calibrated to the wrong wavelength range for measuring circadian impact. It's like using a UV meter to measure audible sound — the tool is real and accurate, but it's measuring something completely different from what you need to know.
What Melanopic EDI Actually Measures — A Plain-Language Explanation
Melanopic Equivalent Daylight Illuminance — melanopic EDI for short — tells you how much melatonin-suppression and circadian clock stimulation a light source delivers, expressed in a unit that's easy to compare: lux-equivalent of standard daylight.
The full name is intimidating, but the concept is simple. Melanopic EDI answers one question: "If I replaced this light source with standard D65 daylight (the reference used in all CIE color science), how many lux of that daylight would produce the same circadian stimulation?"
So if a 590nm amber lamp at 100 photopic lux has a melanopic EDI of 2 lux, that means it delivers the same circadian signal to your ipRGC system as if you were sitting in just 2 lux of standard daylight. That's biological near-darkness — the equivalent of a moonlit room. Your melatonin will flow normally.
If a 5000K cool-white LED at 100 photopic lux has a melanopic EDI of 90 lux, that's the circadian equivalent of standing in bright daylight. Your brain reads it as "midday." Melatonin suppressed. Sleep delayed.
Same lux reading. Wildly different melanopic EDI. Completely different biology.
The "Equivalent Daylight Illuminance" part of the name is important: it means the number is always expressed relative to D65 daylight, making it comparable across different researchers, labs, and studies worldwide. When a clinical paper says "melatonin suppression begins at 10 melanopic EDI lux," that threshold applies universally — it doesn't depend on what light source was used in the study, because melanopic EDI accounts for the source's spectral content.
Melanopic EDI is one of five "α-opic EDI" quantities defined by CIE S 026/E:2018 — one for each type of photoreceptor in the human eye. The other four (S-cone-opic, M-cone-opic, L-cone-opic, rhodopic) are important for specific research applications. For circadian and sleep health, the melanopic quantity is the one that matters most, because ipRGCs are the primary driver of melatonin suppression and circadian phase shifting.
Melanopic EDI = "how much this light source acts like daylight, as far as your biological clock is concerned." A melanopic EDI of 2 lux means your clock thinks it's practically dark. A melanopic EDI of 90 lux means it thinks it's a bright afternoon. The photopic lux reading is irrelevant to this question.
D65 is the CIE standard illuminant representing average northern European daylight at a colour temperature of approximately 6500K. It's the universal reference illuminant in colour science, used everywhere from camera calibration to paint matching. Using D65 as the melanopic EDI reference means all published measurements share a common baseline — a researcher in Tokyo and a researcher in Toronto get directly comparable results. Without this normalisation, melanopic EDI readings from different labs would be internally valid but mutually incomparable.
These two terms are related but not the same. Melanopic EDI is an absolute number in lux-equivalent — it depends on both the spectrum and the intensity of the source. M/P ratio is the melanopic EDI divided by the photopic lux — a dimensionless ratio that describes the source's spectral character alone, independent of brightness. M/P ratio is useful for comparing fixture types. Melanopic EDI is what you use to compare against clinical thresholds. You need both. Learn more in the M/P Ratio section of the Circadian Science Hub.
The CIE S 026 Formula — Plain-Language Walk-Through
You do not need to be a mathematician to understand how melanopic EDI is calculated. The formula has three ingredients: the spectral power of the light source, the melanopsin sensitivity curve, and a normalisation factor that converts the result to daylight-equivalent units. Here is exactly what each piece does.
Measure or obtain the spectral irradiance of your light source — how much power it emits at each wavelength from 380nm to 780nm. This comes from a spectroradiometer measurement or a manufacturer's datasheet. Without this, you cannot calculate melanopic EDI accurately — CCT alone is not sufficient.
Units: W/m²/nm at each wavelength step (typically 1nm or 5nm intervals)At each wavelength, multiply the source's power by the melanopsin sensitivity value sc(λ) from CIE S 026 Table 2. At 480nm, sc(λ) = 1.0, so 100% of the power there counts. At 590nm, sc(λ) ≈ 0.003, so less than 0.3% of the power there counts. This weighting step is what makes the calculation biologically meaningful — it gives full weight to the wavelengths that matter to ipRGCs and nearly zero weight to the ones that don't.
sc(λ) table: CIE S 026/E:2018, Table 2, 380–780nm at 1nm intervals — download free from CIEAdd up all those weighted power values across the full 380–780nm range. This gives you the total melanopically-weighted irradiance — how much biological signal the source delivers to your ipRGC system. In a spreadsheet, this is simply the SUM of (power × sc(λ)) at each wavelength row.
Result: melanopic irradiance in W/m² (before normalisation)Multiply by Kmel,D65 to convert from raw W/m² to the "daylight-equivalent lux" unit that makes the result comparable to clinical thresholds. This step is where most non-standard melanopic calculations go wrong — many tools skip it or use an approximation, making their numbers incomparable to peer-reviewed research.
Kmel,D65 ≈ 1/(∫ ED65,λ(λ) · sc(λ) dλ) — pre-calculated in the CIE S 026 ToolboxMany circadian lighting calculators and apps calculate what they call "melanopic lux" without applying the D65 normalisation correctly. The result looks like a melanopic EDI number but sits on a different numerical scale. When a research paper says "melatonin suppression onset at 10 melanopic EDI lux," that threshold only applies if your measurement used the same D65-normalised methodology. An unnormalised number might read 6 for the same source — and you'd incorrectly conclude you're below the threshold when you're not.
Always verify that any melanopic EDI tool you use applies the CIE S 026/E:2018 normalisation. The free CIE S 026 Toolbox (available at cie.co.at) is the reference implementation.
If you know the M/P ratio of a source and you know the photopic lux at your measurement point, melanopic EDI is simply: Melanopic EDI = M/P ratio × Photopic lux. A 2700K warm-white LED (M/P ≈ 0.45) at 80 photopic lux delivers approximately 0.45 × 80 = 36 melanopic EDI lux — above the clinical melatonin threshold for most adults. A 590nm amber source (M/P ≤ 0.02) at 80 photopic lux delivers ≤0.02 × 80 = ≤1.6 melanopic EDI lux — biologically dark. The M/P ratio shortcut gives you the right order of magnitude without a spectroradiometer, as long as you use verified M/P values.
Two lamps with identical 2700K colour temperature can have M/P ratios ranging from 0.30 to 0.55 depending on their phosphor formulation and blue pump amplitude. A 2700K rating tells you how the light looks — it tells you nothing reliable about its melanopic content. M/P ratios estimated from CCT alone carry ±30% error or more. For accurate melanopic EDI, you need the spectral power distribution — the SPD — not the CCT. See The Hidden Blue Spike in Warm LEDs for why.
D65 Normalisation — The Step That Makes Results Globally Comparable
The D65 normalisation is what transforms a raw biological measurement into a number you can compare against clinical research worldwide. Without it, melanopic EDI readings from different labs, tools, and studies are internally valid but mutually incomparable — like measuring temperature in Celsius and Fahrenheit without a conversion table.
D65 is the CIE Standard Illuminant that represents average northern hemisphere daylight — the sunlight most humans have historically spent their days under. It has a colour temperature of approximately 6500K and a well-defined spectral power distribution published by the CIE.
The normalisation constant Kmel,D65 is calculated once from the D65 SPD and the sc(λ) function. It represents the relationship between melanopic irradiance and photopic illuminance specifically for D65 daylight. By multiplying every melanopic EDI calculation by this constant, the result is expressed in "how many lux of D65 daylight would produce this same circadian signal" — a universally understood unit.
The practical result: when a clinical study says that 10 melanopic EDI lux is the approximate threshold for measurable melatonin suppression in the general population, that threshold applies directly to your installation's melanopic EDI reading — provided you used the correct D65-normalised CIE S 026 methodology. No further conversion needed. The number means the same thing everywhere.
This is why the CIE S 026 Toolbox includes the pre-calculated Kmel,D65 value as a fixed constant. You do not need to re-derive it — just apply it. The toolbox also includes worked example calculations that verify your implementation against CIE reference values, so you can confirm your calculation chain is correct before applying it to real measurements.
Without the D65 normalisation, a tool using its own reference illuminant (or no reference at all) would still output a number — but that number would be offset from the CIE-standard result by an unknown factor. It might read 7 for a source that the CIE toolbox reads as 10. You'd conclude you're under the melatonin threshold when you're actually above it. The error is systematic, invisible, and potentially large enough to matter clinically.
Real-World Melanopic EDI Numbers — What Common Light Sources Actually Deliver
Here are melanopic EDI values for the most common residential and nighttime light sources, calculated using CIE S 026 methodology at two representative illuminance levels. These numbers are what your ipRGC system actually receives — not what the lux meter says.
| Light Source | M/P Ratio | Mel. EDI @ 50 lux | Mel. EDI @ 100 lux | vs. Threshold | Night Safety |
|---|---|---|---|---|---|
| 5000K cool-white LED | ~0.88 | ~44 mel. lux | ~88 mel. lux | 4.4–8.8× above threshold | Day use only |
| 4000K neutral LED | ~0.65 | ~33 mel. lux | ~65 mel. lux | 3.3–6.5× above | Day use only |
| 3000K warm LED | ~0.52 | ~26 mel. lux | ~52 mel. lux | 2.6–5.2× above | Day use only |
| 2700K "warm" LED | ~0.45 | ~23 mel. lux | ~45 mel. lux | 2.3–4.5× above | Inadequate |
| 2200K filament LED | ~0.21 | ~11 mel. lux | ~21 mel. lux | Above threshold | Poor |
| Candle flame (~1800K) | ~0.10 | ~5 mel. lux | ~10 mel. lux | Near threshold | Marginal |
| 590nm+ amber (LumeCircadian) | ≤0.02 | ≤1 mel. lux | ≤2 mel. lux | 5–10× below threshold | ✓ Verified safe |
| 625nm deep red LED | <0.01 | <0.5 mel. lux | <0.5 mel. lux | 20×+ below threshold | ✓ Maximum safety |
A 2700K warm-white LED at 100 photopic lux — a perfectly ordinary bedroom lamp at a comfortable brightness — delivers approximately 45 melanopic EDI lux to the retina. The widely cited clinical threshold for meaningful melatonin suppression in healthy adults is around 10 melanopic EDI lux. At 100 lux, a 2700K LED delivers 4.5 times that threshold. Dimming it to 50 lux helps — but only brings it to 2.3× the threshold. Even at 20 photopic lux (a very dim room), a 2700K source delivers about 9 melanopic EDI lux — still near the suppression threshold for sensitive individuals. "Warm and dim" is not a solution. It's a partial mitigation that still keeps you above the biological boundary.
Clinical Thresholds — What Numbers Actually Protect Your Sleep
Multiple independent research groups have measured the melanopic EDI levels at which melatonin suppression begins, sleep onset is delayed, and circadian phase shifts occur. These are not theoretical numbers — they come from controlled human studies published in peer-reviewed journals.
The clinical evidence tells a clear story: the residential lighting most people consider "warm and comfortable" operates well inside the active range of the melatonin suppression dose-response curve. Standard bedroom illuminance of 50–200 lux with white LEDs of any colour temperature delivers melanopic EDI values of 20–90 lux — from 2× to 9× above the threshold where measurable suppression begins.
The dose-response relationship is not a cliff edge — there is no single "safe" threshold below which exposure is harmless and above which it is damaging. Rather, it's a smooth sigmoidal curve: small suppression effects start around 5–10 melanopic EDI lux, grow through the 10–100 range, and approach saturation above ~500 melanopic EDI lux. This is why the LumeCircadian specification targets ≤2 melanopic EDI lux rather than simply "below 10" — building in a buffer below the onset region rather than operating at its edge.
There is also an important duration dimension that single-number thresholds don't fully capture. Zeitzer et al.'s dose-response data was measured after 6.5 hours of exposure. Shorter exposures (a 5-minute bathroom trip at 3am) carry proportionally lower risk than sustained evening exposure. This is why the practical recommendation is to transition to amber lighting in the 2 hours before bed — not to avoid all non-amber light forever. Duration matters, especially in the context of the nighttime infant care environment, where parents may be present for extended periods.
The direction, however, is unambiguous: lower is better, and there is no biological benefit to any melanopic EDI above approximately zero during intended sleep hours. The only reason to have any light at night is functional — to see well enough to navigate or perform care tasks. A 590nm+ amber source at 10–20 photopic lux provides that functional illumination while delivering less than 0.4 melanopic EDI lux to the retina.
How to Measure Melanopic EDI — From Smartphone to Spectroradiometer
You cannot measure melanopic EDI with a standard lux meter — full stop. The only way to get a valid CIE S 026 melanopic EDI reading is with a detector that captures the spectral content of the light, not just its total output. Here's what each measurement approach can and cannot tell you.
| Method | What It Measures | CIE S 026 Valid? | Approximate Cost | Best Used For |
|---|---|---|---|---|
| Standard lux meter | Photopic illuminance only — V(λ) weighted | No | $20–200 | Verifying photopic illuminance only. Useless for melanopic EDI. |
| "Circadian-aware" lux meter (uncalibrated) | Approximate blue/white ratio — not CIE S 026 | No | $50–300 | Rough screening only. Cannot produce comparable numbers to clinical literature. |
| Manufacturer SPD data + CIE Toolbox | Full spectral calculation — from published SPD | ✓ Yes | Free | Pre-installation fixture qualification. Requires manufacturer SPD data. |
| α-opic validated meter (e.g. Gigahertz-Optik) | Direct melanopic EDI and all five α-opic quantities | ✓ Yes | $800–3,000 | Field verification of installed systems. Best option for residential commissioning. |
| Spectroradiometer + software | Full SPD measurement → all CIE S 026 quantities | ✓ Yes — full | $3,000–15,000 | Research-grade. Required for product SPD verification and clinical-standard measurement. |
For most residential HCL retrofit applications, the most practical measurement strategy combines two approaches:
Before installation: Obtain the SPD datasheet for your chosen emitter and run the CIE S 026 Toolbox calculation to verify M/P ratio and projected melanopic EDI. This is free and requires no measurement equipment — just the SPD data and the Excel toolbox available from cie.co.at. This step qualifies the emitter before you install it. If the manufacturer won't provide an SPD, don't use the product for a night-safe application.
After installation: Verify the installed system with either an α-opic validated meter (field-deployable, gives direct melanopic EDI readings) or a standard lux meter combined with the known M/P ratio of your verified emitter (melanopic EDI ≈ M/P × photopic lux). The latter has slightly more uncertainty but is acceptable for residential applications where the emitter SPD has been pre-verified.
What to measure: Always measure vertical illuminance at eye level — not horizontal illuminance on the floor or desk. Melanopic EDI affects you through light entering the eye, which responds to the vertical component of the light field at your specific eye position. Horizontal lux measurements underestimate retinal exposure from overhead fixtures by 15–35% depending on fixture angle and room geometry.
This is one of the most consistently misapplied aspects of melanopic EDI measurement. Clinical melatonin suppression research measures vertical illuminance at the plane of the eye — approximately 1.2m height for a seated adult, 0.8m for a child, 0.6m for an infant in a crib. If you measure horizontal illuminance on a desk or floor (which is what most lux meters are designed for) and use that number to estimate melanopic EDI, you will underestimate actual retinal exposure. For overhead fixtures, the underestimate can be 20–40%. For infant nursery applications, always measure vertical illuminance at the crib mattress level.
Before spending anything on measurement equipment, work through the CIE S 026 Toolbox calculation for every light source in your target installation. Download the toolbox at cie.co.at, obtain SPD data from the emitter datasheet, and calculate M/P ratio and projected melanopic EDI. This tells you whether your chosen source can meet the target — no equipment required. See the full methodology at the CIE S 026 citation on the References page.
5 Mistakes People Make with Melanopic EDI — And How to Avoid Them
Now that you understand what melanopic EDI is and how it's calculated, here are the five most common errors that undermine otherwise well-intentioned circadian lighting decisions.
Colour temperature (the K number on a bulb box) tells you how warm or cool the light looks. It does not tell you reliably how much melanopsin-stimulating blue light is present. Two lamps from different manufacturers, both labelled 2700K, can have M/P ratios that differ by 50% or more — depending on their phosphor formulation and blue pump design. Never estimate melanopic EDI from CCT alone. Always use the spectral power distribution. If an SPD is not available, the M/P ratio and melanopic EDI are unknown — not estimable.
Correct approach: obtain SPD from manufacturer → calculate M/P from CIE S 026 sc(λ) weighting → verify before specifyingThe word "amber" on a product label means nothing scientifically specific. Many products marketed as "amber," "candle amber," or "fire amber" are phosphor-converted white LEDs with a warm colour appearance and a residual blue pump emission spike at 440–460nm. That spike alone can push the M/P ratio above 0.05, disqualifying the product from the LumeCircadian specification. The only valid test is the SPD. See the Hidden Blue Spike section for why phosphor-converted amber isn't the same as narrow-band InGaAlP amber.
Test: obtain SPD → confirm zero output below 550nm → calculate M/P → if M/P ≤ 0.02, proceedHorizontal illuminance on a desk or floor is not the same as vertical illuminance at eye level, and neither is the same as melanopic EDI. Three distinct conversions are required: (1) lux → vertical lux (geometry correction for your fixture and room), (2) photopic lux → melanopic EDI (M/P ratio correction for your source), and (3) horizontal → vertical plane (fixture angle and position correction). Skipping any step introduces error. The standard clinical measurement plane is vertical illuminance at eye height — approximately 1.2m for seated adults.
Measure: vertical illuminance at eye height → multiply by M/P ratio of verified source → result is approximate melanopic EDIDimming reduces melanopic EDI proportionally — it does not change the M/P ratio of the source. A 2700K LED at 20 lux still has an M/P of ~0.45, delivering ~9 melanopic EDI lux — near the clinical onset threshold. Some individuals, including those with heightened sensitivity and infants with transparent crystalline lenses, may show responses at this level. "Keep it dim" is harm reduction, not harm elimination. Spectral correction (using the right source) is the only complete solution.
Rule: dimming × M/P ratio = still above threshold for white sources at residential lux levelsSeveral consumer devices and apps claim to measure "circadian lux," "melanopic lux," or "biological lux" without disclosing whether they use the CIE S 026/E:2018 sc(λ) function and D65 normalisation. If the output is not labelled as "melanopic EDI (CIE S 026)" and the device documentation does not reference CIE S 026 calibration, the number is on an unknown scale. Comparing it against clinical thresholds derived from CIE S 026 measurements will give wrong answers. Insist on CIE S 026 traceability — or use the free CIE Toolbox with manufacturer SPD data instead. See the full standards chain at the References page.
Requirement for valid measurement: CIE S 026/E:2018 sc(λ) function + D65 normalisation + traceable calibration