The Photobiology of Light:
What Every Lumen Measurement Misses
Standard lighting metrics — lumens, CRI, CCT — were built to measure what the human visual system perceives. They were not built to measure what your nervous system responds to at night. This hub covers the parallel photoreceptor system that regulates melatonin, cortisol, and circadian phase: the ipRGC pathway, the melanopic action spectrum, and the engineering implications for residential lighting.
The Third Photoreceptor — Why Rods and Cones Are Only Half the Story
Before 2002, lighting science operated on a two-receptor model: rods for scotopic (dim) vision, cones for photopic (color) vision. A third class of retinal photoreceptors — discovered by Berson, Dunn, and Takao — changed everything we thought we knew about what light does to the human body after dark.
Intrinsically Photosensitive Retinal Ganglion Cells (ipRGCs) are a distinct class of light-sensing neurons that project directly to the suprachiasmatic nucleus (SCN) — the master pacemaker of the human circadian system. Unlike rods and cones, whose output is primarily visual, ipRGCs serve a non-image-forming function: they measure ambient light levels and relay that signal to the hypothalamus to govern the timing of melatonin secretion, body temperature rhythms, cortisol onset, and sleep-wake architecture.
The photopigment in ipRGCs is melanopsin — an opsin with peak spectral sensitivity near 480nm, well into the short-wavelength blue range. This is why blue-shifted light sources are disproportionately disruptive at night relative to what standard photometric measurements predict. A 5000K cool-white LED and a 590nm amber emitter may register similar lux values on a standard meter, but their biological impact at night differs by over 40-fold when measured in the melanopic-relevant metric.
The practical consequence: any lighting design that uses lux alone as its nighttime safety metric is using the wrong instrument. Lux measures the response of the V(λ) photopic luminous efficiency function — a curve whose peak is at 555nm, optimized for cone-based daylight vision, and which is nearly blind to the 480nm stimulation that dominates ipRGC activation.
Berson, Dunn, and Takao published in Science the first direct demonstration that ipRGCs are autonomously photosensitive — they generate action potentials in response to light even when isolated from all rod and cone input. The peak sensitivity of the melanopsin photopigment in these cells was measured at approximately 480nm, establishing the foundational basis for all subsequent circadian photometry standards including CIE S 026/E:2018.
The Melanopsin Action Spectrum — Mapping Circadian Sensitivity Across the Visible Range
The melanopsin action spectrum is not a flat curve. It has a defined peak, a rapid falloff toward longer wavelengths, and a broader shoulder toward shorter wavelengths than the standard luminous efficiency function. Understanding its shape is the prerequisite to understanding why wavelength — not just intensity — determines biological impact at night.
380nm
400–420
420–490nm
490–500nm
500–565nm
565–590nm
590–620nm
620–780nm
The two curves above illustrate the core problem with standard photometry as a night-safety tool. The melanopic sensitivity function sc(λ) peaks at 480nm and falls to near-zero by 590nm. The photopic V(λ) function peaks at 555nm. These peaks are 75nm apart — a gap that spans the difference between cyan-teal and yellow-amber light.
A standard lux meter weighted to V(λ) will read a 590nm amber source as moderately bright — perhaps 80–120 lux for a typical residential application. But that same source registers essentially zero on the melanopic scale. Conversely, a 480nm-dominant blue source at the same lux reading carries full circadian load. The meter says "equal light." The brain says "day" vs. "night."
This mismatch is not a minor rounding error. It is the reason that every residential lighting decision made on lux alone is biologically uninformed — and why the CIE developed melanopic EDI as a mandatory parallel metric for circadian-relevant lighting design.
Melanopic EDI — The Metric Standard Photometers Were Never Built to Measure
Melanopic Equivalent Daylight Illuminance (melanopic EDI) is the CIE-standardized quantity for expressing the circadian potency of a light source in illuminance units that are comparable to D65 daylight. It is the only field-deployable metric that correctly weights light exposure for ipRGC stimulation.
A melanopic EDI reading of 250 lux means the source produces the same ipRGC stimulation as 250 lux of standard D65 daylight. A 590nm amber source at 100 photopic lux will typically register 2–4 melanopic EDI lux. A 5000K cool LED at 100 photopic lux will register 90–100 melanopic EDI lux. Same photometer reading. Biologically, they are not the same stimulus.
The CIE S 026/E:2018 standard was the first to formally define five α-opic quantities corresponding to each of the five human photoreceptor classes: S-cone-opic, M-cone-opic, L-cone-opic, rhodopic (rod), and melanopic. Of these five, the melanopic EDI is the most clinically and architecturally significant for night environment design, because ipRGCs are the primary driver of melatonin suppression and circadian phase shift.
The recommended nighttime target for environments supporting melatonin production is typically cited in the literature as below 10 melanopic EDI lux — though some research protocols for clinical-grade sleep environments push toward 1 melanopic EDI lux or lower. The LumeCircadian standard targets the M/P ratio below 0.02 at the fixture level, which at typical residential illuminance levels (50–80 photopic lux) translates to approximately 1–1.6 melanopic EDI lux.
Measuring melanopic EDI in the field requires either a spectroradiometer with software integration across the sc(λ) function, or a dedicated α-opic meter (such as those calibrated to the Bioptron or Gigahertz-Optik MSC15 class). Standard lux meters, even those marketed as "circadian-aware," may not correctly weight the 480nm region unless explicitly validated against CIE S 026.
The M/P Ratio — Full Technical Methodology
The Melanopic-to-Photopic ratio is the most actionable single-number circadian metric for fixture selection and nighttime environment specification. It is not the same as melanopic EDI — it is a dimensionless ratio that encodes how much of a source's biological impact falls on the circadian pathway per unit of visible output. Here is the complete calculation methodology.
The M/P ratio allows direct comparison between fixtures of different technologies, CCT, and output levels on a common biological scale. It is independent of absolute light level — meaning it is a property of the spectral power distribution of the source, not its intensity. A 590nm amber source at 20 lux and the same source at 200 lux will have the same M/P ratio, while their melanopic EDI values will differ by a factor of 10.
This distinction matters for fixture auditing: you can screen an entire lighting inventory by M/P ratio before a single illuminance measurement is taken, because M/P is calculable from spectral datasheet data alone.
| Source Type | CCT / Peak | M/P Ratio | Melanopic EDI @ 100 lux | Night Use |
|---|---|---|---|---|
| Blue-enriched office LED | 6500K | ~1.05 | ~105 mel. lux | Never |
| Cool white LED (typical) | 5000K | ~0.88–0.92 | ~88–92 mel. lux | Day only |
| Neutral white LED | 4000K | ~0.60–0.70 | ~60–70 mel. lux | Day only |
| Standard "warm" LED | 3000K | ~0.50–0.55 | ~50–55 mel. lux | Day only |
| Warm white LED (2700K) | 2700K | ~0.42–0.48 | ~42–48 mel. lux | Inadequate |
| High-CRI filament LED (2200K) | 2200K | ~0.18–0.24 | ~18–24 mel. lux | Poor |
| Candle flame | ~1800K | ~0.08–0.12 | ~8–12 mel. lux | Marginal |
| Deep red emitter (625nm) | 625nm | <0.01 | <1 mel. lux | ✓ Safe |
| LumeCircadian Amber (590nm+) | 590nm+ | ≤0.02 | ≤2 mel. lux | ✓ Verified |
A 2700K warm-white LED — the default "circadian-friendly" recommendation across mainstream smart lighting platforms — carries an M/P ratio of approximately 0.42 to 0.48. At 100 photopic lux (a standard bedroom lighting level), this source delivers roughly 45 melanopic EDI lux to the retina. The clinically established threshold above which melatonin onset is meaningfully delayed is widely cited in the range of 10 melanopic EDI lux for sensitive individuals, and 30 melanopic EDI lux for healthy adults. A 2700K LED at normal bedroom brightness is operating 1.5 to 4.5× above that biological threshold. This is not a circadian-safe light source. It is simply less bad than a 5000K alternative.
The 500nm Spectral Cutoff — Where Biology Draws the Line
The 500nm threshold is not an arbitrary engineering convention. It is the wavelength below which melanopsin stimulation remains biologically meaningful — derived directly from the sc(λ) spectral sensitivity curve and representing the practical engineering cutoff for night-safe spectral power distributions.
The melanopsin action spectrum sc(λ) does not switch off cleanly at a single wavelength. Like all photopigment response curves, it tapers gradually. However, for engineering purposes, a cutoff threshold must be defined. The 500nm value represents the point at which sc(λ) has declined to approximately 15–20% of its peak value — a meaningful but diminishing biological signal.
Below 500nm, every nanometer step toward the UV edge carries compounding circadian impact because you are moving toward, not away from, the 480nm melanopsin peak. Above 500nm, each step into yellow, amber, and red reduces the circadian signal exponentially — by 500nm itself the sc(λ) function is falling steeply, and by 590nm it is functionally negligible for practical light levels.
The engineering implication: spectral output below 500nm must be treated as circadianly active, full stop. There is no concept of "a little blue light" being tolerable at night in a scientifically rigorous design. Even a narrow residual spike at 440–480nm from an amber phosphor-converted LED can deliver measurable melanopic stimulation depending on its intensity and the duration of exposure.
This is why the LumeCircadian standard uses a 590nm minimum for night emitter peak wavelengths, not 500nm. The 500nm figure identifies the boundary of meaningful activity; 590nm is where the sc(λ) curve has declined to below 1% of peak — effectively zero for any realistic residential lux level.
| Wavelength | sc(λ) Relative Value | Biological Zone |
|---|---|---|
| 440nm | ~0.60 | Active |
| 460nm | ~0.87 | High Active |
| 480nm | 1.00 (peak) | Maximum |
| 500nm | ~0.17 | Threshold |
| 530nm | ~0.04 | Low |
| 555nm | ~0.015 | Negligible |
| 590nm | <0.005 | Near-zero |
| 620nm+ | <0.001 | Zero (practical) |
sc(λ) values: CIE S 026/E:2018 Table 2, normalized relative sensitivity. Values at specific wavelengths interpolated from published tabulated data.
The Hidden Blue Spike in "Warm" LEDs — Why Phosphor Conversion Is Not Spectral Elimination
Every phosphor-converted white LED — regardless of its marketed color temperature — is built on a blue or near-UV pump LED. The warm amber appearance is created by a phosphor coating that down-converts some of that blue energy into longer wavelengths. "Some" is the operative word. The unconverted residual is always there.
The SPD diagram above illustrates the fundamental difference between a phosphor-converted LED and a narrow-band amber emitter. The 2700K warm-white LED has a characteristic blue pump spike near 450nm — residual unconverted pump energy that the phosphor layer did not absorb. This spike sits at the heart of the melanopsin action spectrum. Even if the spike is small in absolute terms (perhaps 8–15% of total radiant output), it carries disproportionate biological weight because it falls directly on the sc(λ) peak.
The phosphor hump — the broad warm emission from 550–650nm — contributes almost no circadian stimulation because it sits primarily above the 500nm safety cutoff. The problem is that you cannot separate the hump from the spike. They are the same LED. When you dim a phosphor-converted warm-white LED, you reduce both components proportionally — but the spike does not disappear; it merely decreases. At no dimming level does a PC-LED become spectrally equivalent to a narrow-band amber source.
This is the spectral argument against "just use warm and dim." Dimming reduces total melanopic stimulation by reducing lux, but it does not change the M/P ratio. A 2700K LED at 10 lux still has an M/P of ~0.45. A 590nm amber source at 10 lux has an M/P of ≤0.02. At equal lux, the warm-white LED delivers over 20× the circadian stimulation of the amber source — regardless of dimming depth.
Warm White
- Blue pump spike at 440–460nm always present
- M/P ratio ~0.42–0.48 regardless of dim level
- Cannot achieve M/P < 0.05 at any output level
- Spike amplitude proportional to drive current — not eliminable
- Marketed as "circadian warm" — misleading
Spectrally Pure Night Source
- No emission below 550nm by spectral design
- M/P ratio ≤0.02 — verified by CIE S 026 methodology
- Biologically dark at any output level above 0 lux
- Melanopic EDI < 2 lux at 100 photopic lux
- Requires purpose-selected emitter, not filtering
CIE S 026 Applied to Real Fixtures — From Standard to Specification
CIE S 026/E:2018 is often cited but rarely applied correctly in residential practice. Most references use it as a naming authority — mentioning "melanopic EDI" without the full weighting function, conversion factors, or illuminance conditions required for valid measurement. Here is what correct application actually requires.
Step 1 — Obtain the spectral power distribution (SPD). Valid melanopic EDI calculation requires the spectral irradiance in W/m²/nm, not just a CCT or CRI value. Many fixture datasheets do not publish full SPD data. When SPD data is unavailable, M/P ratios must be estimated from published reference spectra for comparable CCT classes — which introduces significant error, particularly for the 2700–3000K range where the blue spike amplitude varies widely between manufacturers and phosphor formulations.
Step 2 — Apply the sc(λ) weighting function. The CIE S 026 standard publishes the melanopic spectral sensitivity function sc(λ) in 1nm increments from 380 to 780nm. The melanopic irradiance Emel is the integral of spectral irradiance times sc(λ) across this range. A common error is using a simplified three-band weighting (blue/green/red) instead of the full published function — this can underestimate circadian impact of sources with narrow blue spikes by 30–50%.
Step 3 — Apply the D65 normalisation constant. Converting melanopic irradiance to melanopic EDI requires dividing by the integral of D65 spectral irradiance times sc(λ) — which CIE S 026 defines as the Kmel,D65 constant of approximately 1/kmD65 = 0.9990 (for sc(λ) normalized to unity at 480nm, with D65 as reference illuminant). This step is often skipped when practitioners use raw melanopic irradiance values — making their numbers incomparable to published clinical literature which always uses the D65-referenced EDI.
Step 4 — Specify the measurement plane and distance. Melanopic EDI values are meaningless without specifying the vertical illuminance at eye level — not horizontal task illuminance. At a standard seated eye height of 1.2m, a ceiling-mounted fixture will typically produce 15–35% lower vertical melanopic EDI than its horizontal lux reading at desk level, depending on fixture angle, beam pattern, and room reflectances.
Clinical research on melatonin suppression consistently measures vertical illuminance at eye level — the metric directly relevant to the light entering the pupil. Most architectural illuminance calculations use horizontal illuminance on the work plane. For a 2700K source at 200 lux horizontal, the vertical melanopic EDI at a seated occupant's eye may be 80–90 mel. lux — well above the biological threshold — even in a "dim warm" setting. Correct CIE S 026 application requires vertical measurement. This distinction alone explains much of the discrepancy between industry claims of "low-impact" warm lighting and measured biological outcomes.
The CIE S 026 Toolbox (freely available from the CIE website) provides reference SPDs, the full sc(λ) tabulation, and conversion factor worksheets for calculating all five α-opic EDI quantities. The LumeCircadian methodology uses this toolbox as the primary calculation reference, with all fixture M/P ratios validated against the published tabulated sc(λ) values rather than approximation methods.
Circadian Science — Deep-Dive Articles
Each article below goes deeper into a specific aspect of the science, with full methodology, reference data, and field application guidelines.
- CIE S 026/E:2018. CIE System for Metrology of Optical Radiation for ipRGC-Influenced Responses to Light. Commission Internationale de l'Éclairage. Vienna, 2018.
- Berson DM, Dunn FA, Takao M. Phototransduction by retinal ganglion cells that set the circadian clock. Science. 2002;295(5557):1070–1073. doi:10.1126/science.1067262
- Hattar S, Liao HW, Takao M, Berson DM, Yau KW. Melanopsin-containing retinal ganglion cells: architecture, projections, and intrinsic photosensitivity. Science. 2002;295(5557):1065–1070.
- Lucas RJ, Peirson SN, Berson DM, et al. Measuring and using light in the melanopsin age. Trends Neurosci. 2014;37(1):1–9. doi:10.1016/j.tins.2013.10.004
- Rea MS, Figueiro MG. Light as a circadian stimulus for architectural lighting. Lighting Res Technol. 2018;50(4):497–510.
- Thapan K, Arendt J, Skene DJ. An action spectrum for melatonin suppression: evidence for a novel non-rod, non-cone photoreceptor system in humans. J Physiol. 2001;535(1):261–267.
- ANSI/IES TM-30-18. IES Method for Evaluating Light Source Color Rendition. Illuminating Engineering Society, 2018.