How to Read Photopic V(λ) and Luminous Efficacy

How do you read a photopic V(λ) curve and calculate lumens from a spectral power distribution? The practical answer is to weight the radiant power at each wavelength by the standard photopic luminous-efficiency function, add those weighted contributions over wavelength, and apply the photometric scaling constant. The result describes light in relation to a standardized daytime visual response—not simply the total optical watts emitted.

Learning how to interpret luminous efficacy from a light spectrum also requires distinguishing three quantities that sound similar: luminous flux in lumens, luminous efficacy of radiation in lumens per optical watt, and the electrical luminous efficacy of a lamp in lumens per input watt. Those denominators are different. A plotted spectrum, a photometer reading, and a manufacturer’s efficacy figure may each answer a different question, even when they describe the same lamp.

This guide explains the graph axes, wavelength weighting, integration, worked calculations, measurement uncertainty, comparison of light sources, and the language needed to report results accurately. It follows the preceding guide to spectral power distribution plots but addresses the separate search question: how can a spectrum be converted into a defensible photometric quantity?

Start with the question the graph can actually answer

A spectral power distribution, or SPD, shows how an optical quantity is distributed across wavelengths. It may be plotted in watts per nanometre, irradiance per nanometre, arbitrary detector counts, or values rescaled so the largest point equals one. The appearance of the curve alone does not establish its physical normalization. Before calculating anything, read the vertical-axis label and the measurement description. If they do not tell you the absolute scale, you cannot obtain absolute lumens by integration without another calibration.

A photopic luminous-efficiency curve answers a different question. It assigns a standardized visual weighting to each wavelength. It is not a measured lamp spectrum, and it does not tell you how much power a particular source emits. Combining the measured SPD with this weighting makes a visual-flux calculation possible. When you explain the process, keep the emitter, the observer weighting, and the resulting photometric quantity separate. That simple separation prevents most interpretation errors in introductory lighting reports.

What the V(λ) curve represents

V(λ), usually pronounced “V of lambda,” is a dimensionless relative spectral luminous-efficiency function for photopic vision. Photopic means the light-adapted visual regime associated primarily with cone-mediated vision. The standard curve peaks close to 555 nanometres, in the green region, and falls toward both shorter and longer visible wavelengths. It is an agreed measurement convention, not a claim that every viewer sees every scene exactly alike.

The word “relative” matters. A value of 0.2 at one wavelength means that, within this standardized function, a watt of monochromatic radiation there receives one-fifth the luminous weighting of a watt near the peak. It does not mean that a person necessarily experiences one-fifth the subjective brightness of a complex scene. Perceived brightness depends on adaptation, surrounds, spatial pattern, observer variation, and other conditions. V(λ) is designed for defined photometric quantities; using it as an all-purpose model of human visual appearance overstates its purpose.

Read the horizontal axis before touching the data

Most source spectra place wavelength on the horizontal axis, expressed in nanometres. A longer wavelength is toward the red end of the visible spectrum, while shorter wavelengths move toward violet and blue. A spectral luminous-efficiency plot uses the same horizontal coordinate. The simplest integration workflow aligns both curves at matching wavelength samples or interpolates one onto the grid of the other. Misalignment changes the answer because the weighting varies sharply across the spectrum.

If one table lists wavelengths in micrometres and another in nanometres, convert carefully: one micrometre equals one thousand nanometres. If the optical instrument reports frequency or photon energy instead of wavelength, a change of variable is required; values expressed per unit frequency are not numerically identical to values expressed per unit wavelength. Merely relabelling the horizontal axis destroys the meaning of a spectral density. A careful reader treats the axis, unit, and bin width as part of the evidence rather than decorative labels.

Radiant power and luminous flux are not interchangeable

Radiant flux measures optical power in watts. Luminous flux measures the same radiation after photopic spectral weighting and is reported in lumens. The distinction is not cosmetic: two sources can deliver identical optical wattage but very different luminous flux. A near-infrared emitter can produce substantial radiant power while contributing little to a conventional visible-light photometric result. A green narrowband source can have a high number of lumens per optical watt because its energy sits close to the V(λ) maximum.

When a chart is labelled “intensity,” stop to identify the intended quantity. In everyday English, intensity may mean visible brightness, a detector’s counts, irradiance, luminous intensity in candela, or radiant intensity in watts per steradian. Scientific reporting cannot safely treat these as synonyms. Translate the symbol together with its unit and geometry. If a report says a lamp became “more intense,” ask whether flux increased, the beam narrowed, measurement distance changed, or only the plotting scale was adjusted.

How the weighted integral works

For absolute spectral radiant flux Φe,λ expressed in watts per nanometre, photopic luminous flux is approximately Φv = 683 lm/W multiplied by the integral of Φe,λ times V(λ) over wavelength. Read the structure before memorising the equation. At every wavelength, take the amount of optical power present, assign the appropriate visual weighting, and add all the contributions. Integration is the mathematical version of that continuous sum.

The constant is tied to the SI realization of photometric units. The exact definition is based on radiation at 540 terahertz, corresponding to approximately 555.016 nanometres in standard air. In most explanatory calculations, 683 is a sufficient rounded factor. The units work because the spectral density multiplied by a wavelength interval gives watts, the dimensionless V(λ) changes only the weighting, and lumens per watt converts that weighted optical power to lumens. A result that still contains “per nanometre” after integration indicates that the calculation omitted a width or mixed conventions.

Why 683 is not an electrical-efficiency rating

The factor around 683 lumens per watt describes the photometric scaling near the peak of photopic sensitivity. It is not a universal ceiling for the lumens produced per electrical watt by a light source in an energy-conversion model, although the highest luminous efficacy of radiation under this conventional definition is close to that figure. Electrical luminous efficacy includes losses between wall socket and emitted optical radiation, including driver conversion, heat, and package losses.

Consider an imaginary light source that consumes ten electrical watts and emits three optical watts. Even if all three optical watts lay at the photopic peak, the luminous flux would be approximately 2,049 lumens and the electrical luminous efficacy about 205 lumens per watt. The radiation-weighting step and the electrical-input step are separate. A spec sheet that places “lm/W” beside a number must say which watt is the denominator. If it does not, responsible comparison requires clarification before treating one device as superior to another.

Worked comparison: equal optical watts, unequal photopic output

Suppose two idealized, narrowband sources each emit one watt of radiant flux. Source A emits very close to the photopic maximum, where V(λ) is near one. Its luminous flux approaches 683 lumens. Source B emits at a wavelength where the tabulated V(λ) is 0.1. Its flux is about 68.3 lumens, although both sources still radiate one optical watt. This example is deliberately simple: actual LEDs have finite bandwidth, so their energy must be integrated across a wavelength interval.

The comparison also shows why visual weighting is not a statement about a source’s physical energy. The second source did not lose nine-tenths of its radiation. It received a smaller photopic weighting. A translator or science tutor should phrase the result as “lower photopically weighted flux per optical watt,” not “ninety percent of the light disappeared.” Precise language helps readers avoid mistaking a convention for a change in energy conservation.

Use spectral bins correctly

Instrument exports usually provide discrete wavelength samples rather than continuous mathematical functions. When wavelength spacing is uniform, an approximate calculation multiplies each spectral radiant-flux density by V(λ) and by the wavelength interval, sums the products, and applies the lumen conversion. If the instrument samples every five nanometres and the y-axis is watts per nanometre, the width factor is five nanometres for each appropriate rectangular bin.

A common spreadsheet error is to sum all the plotted y-values directly. That yields a number whose magnitude depends on how many samples were taken. A spectrum sampled every one nanometre might appear five times larger than the same spectrum sampled every five nanometres if the interval is omitted. The correct approach respects the physical meaning of density. Where bin widths vary, use each actual interval or a defensible numerical integration method such as the trapezoidal rule. Check units at every stage, especially when import software automatically changes decimal separators or formats scientific notation.

Do not invent lumens from a normalized graph

Many papers publish a relative SPD with its highest peak scaled to one hundred percent. That graph is excellent for comparing spectral shape, finding phosphor bands, and studying changes between operating conditions. It is insufficient, by itself, for calculating absolute luminous flux. The missing quantity is the actual radiant-power scale of the source. Multiplying a normalized curve by 683 produces a scaled relative result, not a measured lumen output.

You can still calculate a meaningful luminous efficacy of radiation from a consistently normalized shape. The normalization factor cancels when weighted optical power is divided by total optical power, provided that the spectrum covers the relevant wavelength region and is correctly represented as a density. State this condition explicitly. A defensible explanation reads: “The relative spectrum supports an estimate of spectral luminous efficacy but cannot establish the absolute lumen output without calibrated radiometric data.” That sentence is both technically accurate and easy to translate without losing the distinction.

What spectral luminous efficacy of radiation means

Spectral luminous efficacy of radiation, often abbreviated LER, is luminous flux divided by the radiant flux of the relevant emission. The numerator is in lumens and the denominator in optical watts. In terms of a spectrum, LER equals about 683 times the ratio of the V(λ)-weighted spectral integral to the unweighted spectral integral. Because the same overall normalization multiplies both integrals, a relative SPD can often support the ratio even when it cannot support absolute lumens.

LER answers how well the shape of the emitted spectrum aligns with photopic weighting. It does not incorporate electrical conversion losses, optical losses in a luminaire, or actual delivered illuminance at a work surface. When a report says a source has excellent “efficiency,” ask which efficiency. Wall-plug efficiency, photon conversion efficiency, spectral luminous efficacy, and task efficacy belong to related but non-identical chains of measurement. Naming the denominator is the fastest way to resolve a confusing comparison.

Why wall-plug lumens per watt can tell another story

Electrical luminous efficacy is luminous flux divided by input electrical power. Two sources can have equal LER and very different electrical lm/W because one wastes more input as heat. Conversely, a highly efficient electric-to-optical converter can still have modest photopic lm/W if much of its light lies in spectral regions that receive low V(λ) weights. Real device performance is a combination of spectrum, electrical conversion, packaging, optics, and operating conditions.

A clean comparison therefore keeps at least three lines of evidence. First, inspect the SPD to understand wavelength allocation. Second, assess total radiant and photometric output under stated measurement geometry. Third, obtain measured electrical input under matching current, voltage, temperature, and dimming settings. Mixing a spectrum taken at one operating point with power consumption recorded at another can yield a compelling but fictitious efficacy figure. In reports intended for buyers, researchers, or learners, transparent matching conditions are more useful than a single large number with no measurement notes.

How bandwidth changes the answer

A narrow peak and a broad peak may have the same maximum y-value yet very different area under the curve. Luminous flux depends on spectral area after weighting, not merely the tallest point. Full width at half maximum is a helpful descriptor of peak breadth, but it is not a substitute for integrating the actual emission. Two LEDs can share a nominal peak wavelength and still have different results because their tails, phosphor components, or secondary emissions fall in regions with different V(λ) values.

If you are comparing plots with different wavelength scales, do not estimate total flux by the apparent visual height of their coloured shapes. Examine the axis range and whether the plotting software clips low-amplitude tails. A wavelength window that excludes near-red or violet emissions changes the unweighted integral even when it has a smaller effect on photopic lumens. Report the integration bounds so another reader can reproduce the calculation and understand how much of the source was included.

A spectrum with two peaks needs both peaks counted

A typical white LED may feature a blue excitation peak and a broader phosphor band extending through green, yellow, and red wavelengths. The visual weighting of that broad band can dominate the luminous-flux calculation even if the blue peak looks sharper. On a normalized plot, a sharp maximum may attract attention while a wide, lower band contributes more integrated power. The difference between peak height and area is central to accurate interpretation.

Imagine a source whose blue peak is relatively tall but occupies only a narrow wavelength interval, alongside a modest green-yellow plateau extending over many tens of nanometres. The plateau can matter greatly in the photopic sum because it is both broad and well positioned under V(λ). This is a question about numerical integration, not a preference for one lamp colour. A sound report identifies the spectral contributions and then checks the final photometric measurement. It avoids claiming that the tallest spectral line explains the whole lumen rating.

How to compare different white-light spectra fairly

Before comparing two white-light sources, verify the measurement geometry, instrument calibration, current, temperature, input power, and whether both spectra are absolute or relative. If one source is normalized to its own maximum and another is plotted in calibrated watts per nanometre, their y-axes do not share a scale. Simply overlaying them can mislead. Normalize both for spectral-shape comparison, or use calibrated values for flux comparison, but label which question you are answering.

Different spectra can generate similar luminous flux with different colour-rendering characteristics. Therefore the greatest calculated number of lumens is not automatically the best lighting choice. A display, museum, road, operating area, and reading room have different requirements for colour, distribution, glare, temporal modulation, and energy use. Efficacy is an important dimension, not the only one. Connect the calculation to the intended task instead of presenting a high photometric score as a complete quality assessment.

Why colour rendering adds another constraint

A source that concentrates power near the photopic maximum may achieve high LER but reproduce a wide range of object colours poorly. Colour rendering depends on how the source illuminates surfaces across a broader collection of wavelengths. This is why a spectral design optimized only for lumens per optical watt may be unsuitable where faithful colour appearance matters. It is also why photopic flux and a colour-rendering score should not be interchanged in a product comparison.

For a complementary examination of fidelity and hue changes, read eduKateSG’s guide to TM-30 Rf, Rg and colour vector graphics. That guide asks what happens to object colours under a source. The present calculation asks how emitted radiation is weighted by a standard photopic function. Those are different scientific questions even though they use information derived from the same underlying SPD.

Photopic, scotopic, and mesopic language

Photopic weighting is not a universal description of human vision under all illumination conditions. Scotopic vision refers to dark-adapted, rod-dominated conditions and is associated with a different spectral sensitivity function, commonly written V′(λ). Mesopic conditions involve a transition in which both rod and cone contributions matter. A figure that labels a photopic curve simply “eye sensitivity” may conceal this important limitation, especially in discussions of very low-light environments.

The key educational move is to preserve the regime when translating. “Photopic luminous efficiency” should not silently become “visibility in all darkness.” A comparison between light sources in road or exterior lighting may require careful treatment of adaptation and the relevant measurement conventions. At ordinary practical lighting levels, standard photometry remains widely useful, but the interpretation should match the regime for which the quantity is defined. Technical honesty begins with the boundaries of the model, not just its most familiar graph.

Do not equate luminous flux with illuminance

Lumens measure luminous flux; lux measures illuminance, the luminous flux incident per unit area under the photometric definition. A lamp can emit many lumens while delivering little lux to a distant surface because its radiation spreads, misses the target, or is blocked. Optical distribution, geometry, distance, reflections, and fixtures determine the relationship between emitted flux and local illuminance. A spectral conversion alone cannot predict all of those spatial outcomes.

For example, two luminaires with equal lumen output can produce markedly different illuminance maps if one uses a narrow beam and the other distributes light broadly. The beam angle is not encoded in a single wavelength curve. When reading an engineering report, identify whether the measurement is a source-output test, a luminaire-output test, or a field measurement at a specific position. These may be complementary parts of one design, but they should not be merged into one word such as “brightness.”

Luminous intensity is a direction-dependent quantity

Luminous intensity is measured in candela and refers to luminous flux per unit solid angle in a specified direction. A goniophotometric distribution gives directional information that an overall SPD does not. An integrating-sphere result may provide total spectral radiant flux or total luminous flux when properly calibrated, while an angular scan tells a different story about where the light is sent. Neither should be substituted for the other without the necessary geometry.

In practical communication, a claim such as “the lamp measures 900 cd” is incomplete without its direction or distribution context. A product’s maximum candela can be high because its beam is focused, not because it emits more total lumens than another product. Retain the words “total” and “directional” when reporting photometric quantities. Readers who learn to connect units to geometry are better prepared to interpret specification sheets and avoid false comparisons across measurement systems.

What an integrating-sphere measurement contributes

An integrating sphere captures light from many directions through a diffuse reflective interior and, with an appropriate measurement method, supports total-flux measurements. A spectrometer connected to a sphere can characterize spectral content for photometric conversion. However, sphere measurements require calibration, correction for the sphere’s response, suitable handling of the test object, and awareness of absorption or spatial-response errors. The presence of a professional-looking sphere does not remove the need for measurement documentation.

If a report lists an absolute lumen output, ask whether that value comes from calibrated photometry or a relative spectral plot later scaled to a known reference. Both may be useful methods when explained correctly, but uncertainty differs. A small, strongly directional emitter and a large, self-absorbing lamp may interact with the measurement setup differently. Reported values are more persuasive when the calibration reference and measurement conditions are visible rather than hidden behind a polished curve.

Instrument spectral response and stray light

A spectrometer does not automatically measure equal physical power as equal detector counts at every wavelength. Its grating, optics, detector, and electronics have wavelength-dependent responses. Calibration converts measured counts toward the relevant spectral quantity. If that correction is absent, a graph’s shape can be distorted, and any subsequent luminous weighting inherits the distortion. The photometric integral may look mathematically correct while resting on inaccurate spectral inputs.

Stray light is another issue. Optical radiation at one wavelength can contribute signal elsewhere through scattering or imperfect spectral isolation. Out-of-band leakage matters when a bright feature sits near a region where the source’s true signal is weak. Before inferring subtle shifts in photopic efficacy from two nearly identical curves, check whether the observed difference exceeds calibration and instrument uncertainty. A clear report distinguishes instrument resolution, wavelength accuracy, stray-light control, and radiometric calibration instead of treating them as one vague statement of “precision.”

Temperature and drive current can change the spectrum

LED spectra depend on operating conditions. Junction temperature, drive current, phosphor behavior, and optical packaging can influence wavelength position, peak intensity, and spectral balance. A curve measured at low current cannot always stand in for performance at the specified electrical input. Likewise, thermal stabilization conditions can affect comparisons between laboratory and installed operation. The lesson is not that published spectra are untrustworthy; it is that spectra are measurements taken under particular circumstances.

When a manufacturer provides several SPD plots at different currents, calculate or compare photometric results within each matching condition. Do not combine the peak position from the hot curve, the radiant flux from the cold curve, and the electrical input from the nominal current line. That hybrid combination describes no actual operating state. Marking each condition clearly is especially important when figures are translated into another language or rearranged into a summary document.

Example: two lamps with equal lumens and unequal electrical input

Suppose Lamp A produces 1,200 lumens while using 10 watts of electrical power and Lamp B produces the same 1,200 lumens while using 15 watts. Their electrical luminous efficacies are 120 lm/W and 80 lm/W respectively. That comparison is straightforward because both numerator and denominator refer to corresponding measured operating points. It says that Lamp A delivers more photopic luminous flux per input electrical watt in those conditions.

It does not, however, tell you whether the lamps have the same beam distribution, colour rendering, glare, lifespan, or spectral impact outside photopic vision. Nor does it identify the spectra that created their lumen outputs. A procurement note might reasonably prefer A for its measured electrical efficacy but still request colour, optical, and lifetime information. This is what a well-translated metric should do: answer the question it was designed to answer without swallowing every other relevant design question.

Example: equal electrical efficacy can hide different spectra

Now consider two differently designed sources that each provide 100 lumens per input electrical watt. One may have comparatively strong radiant output but distribute more of it into blue or deep red wavelengths with lower photopic weighting. Another may convert less electricity into total optical radiation but position more of that radiation where V(λ) is high. Their final electrical efficacy figures can coincide through different physical combinations.

To investigate why the numbers match, collect the absolute spectra and electrical inputs, calculate or verify LER, then compare the resulting optical and electrical power balances. This is an inference problem, not a reason to invent missing laboratory data. An excellent explanatory paragraph can say that the same lm/W figure does not determine a unique SPD or wall-plug efficiency. That sentence helps buyers, educators, and technical translators distinguish observable measurements from interpretations that require additional evidence.

Reporting uncertainty without burying the meaning

No measured spectral integral is infinitely precise. Sources of uncertainty may include spectral calibration, wavelength registration, detector linearity, optical collection, temperature drift, repeatability, and data-processing choices. Some errors influence the entire spectrum proportionally, while others change its shape. Because LER is a ratio of integrals, a uniform multiplicative error may cancel in that ratio, but wavelength-dependent errors generally do not. Absolute lumen output remains sensitive to the overall calibration scale.

A readable results section should first state the measured quantity and conditions, then provide the calculated value, followed by limitations proportionate to the decision. Avoid an impressive string of decimal places that the apparatus cannot justify. If the purpose is to rank two nearly identical lamps, uncertainty may determine whether the ranking is credible. If the purpose is to explain why a blue-rich and green-rich spectrum have different photopic weights, the main qualitative conclusion may remain robust despite modest numerical uncertainty.

Precision in English and cross-language explanations

Technical translation fails when ordinary words replace quantities that are not equivalent. “Power” may mean electrical input, radiant flux, radiant intensity, or density of radiant power. “Efficiency” may refer to any ratio of useful output to a particular input. “Bright” may describe perceived appearance, a luminance reading, or simply marketing tone. Translate the surrounding unit, measurement condition, and denominator, not just the noun.

The same habit that supports scientific interpretation supports language learning. eduKateSG’s cross-linguistic transfer guide explores when familiar language patterns help or mislead. Its translation accuracy and verification guide develops the complementary discipline of checking meaning against context. When a multilingual team sends photometric results to another audience, a short shared glossary prevents an optical watt from turning into an electrical watt and prevents “lumen” from being silently replaced with “lux.”

A practical verification path for a spreadsheet

Begin with the original spectral export. Save a clean copy, identify the wavelength column, and record whether the y-values are absolute or relative. Align the spectral measurements with a recognized table of photopic V(λ) data, making explicit any interpolation or resampling. Multiply each spectral value by the appropriate weighting and bin width, sum the weighted contributions, and apply the photometric scaling factor when the units support it.

Then perform two checks. First, calculate the unweighted radiant-flux integral and, if appropriate, divide the luminous result by it to obtain LER. Second, compare the computed luminous flux with an independent calibrated photometric result when one is available. Significant disagreement prompts investigation of geometry, instrument response, wavelength limits, baseline handling, normalization, and units. Do not “fix” a discrepancy by secretly adjusting a scale factor to force agreement. The transparent discrepancy often teaches more than a cosmetically tidy answer.

How to read a published graph without raw data

Sometimes a journal figure offers only a spectrum image, not an exportable table. You may still compare peak position, relative bandwidth, dominant emission regions, and changes between samples. Estimating numerical LER from a digitized curve is possible in principle, but it introduces digitization error and remains dependent on whether the vertical axis represents spectral density. Treat such numbers as estimates and explain their origin. Do not label them direct measurements.

If the paper includes an independent reported luminous flux or efficacy value, quote that value accurately and keep it separate from your graph-based visual interpretation. Readers appreciate an explanation of what the available evidence does and does not warrant. Scientific literacy is often the ability to stop at the correct boundary of inference. A graph is not incomplete because it lacks every conceivable metric; it becomes misleading when readers demand from it a number it was never designed to supply.

Learning route: from spectral shape to useful decisions

A productive study sequence begins by reading the SPD without calculating anything. Identify the axis units, peaks, baseline, bandwidth, and normalization. Next, introduce V(λ) as the standard photopic weighting function and practice matching wavelengths. Then calculate a simple two-bin example by hand before moving to a full spreadsheet. Finally, ask learners to explain in ordinary language why two sources with equal optical power may have unequal lumen output.

The follow-up should use contrasting problems, not a pile of interchangeable questions. One task might expose a missing nanometre interval; another might compare electrical and optical watts; a third might show equal lumens but different beam geometry. eduKateSG’s vocabulary transfer guide is relevant because learners must carry definitions into unfamiliar graphs, reports, and real applications. Mastery means explaining the same relation with a diagram, equation, unit audit, and careful sentence without changing its meaning.

Frequently asked questions about photopic V(λ) and luminous efficacy

Can I calculate lumens from any spectral graph? Only if the graph or associated data contains an appropriate absolute radiant-flux scale and the measurement geometry supports total flux. A normalized graph alone cannot establish absolute lumens.

What is the difference between LER and lamp efficacy? LER divides lumens by optical radiant watts. Electrical lamp efficacy divides lumens by input electrical watts.

Does a taller green peak always mean more lumens? No. The result depends on integrated spectral power, its photopic weighting, and the measurement’s absolute scale.

Is V(λ) the same as colour rendering? No. V(λ) supports photometric weighting, while colour-rendering methods compare how sources reproduce object colours.

Why can two lamps have the same lumens but different lux? The spatial distribution and measurement geometry differ.

Is a spectrum measured with arbitrary units useless? Not at all. It can support peak, shape, and sometimes ratio comparisons, provided its normalization and limitations are stated.

What does the 683 factor signify? It is the conventional photometric scaling near the photopic peak, grounded in the SI definition.

Does photopic weighting describe night vision? Not generally. Dark-adapted and intermediate regimes require different treatment.

What should a translator preserve? The physical quantity, its unit, denominator, geometry, and operating conditions.

How to Translate Easily | How to Read Photopic V(λ) Weighting, Luminous Efficacy and Lumens from a Spectrum