Optimization of Lighting Projects Including Photopic and Circadian Criteria: A Simpliﬁed Action Protocol

: Lighting projects that consider parameters related to circadian light remain rare. Using controlled lighting on both photopic and melanopic levels, this study aims to simplify the design of circadian lighting projects based on traditional photometric parameters and calculations. A real classroom is used to illustrate the behavior of horizontal (visual stimuli) and vertical (circadian contribution) illuminances under di ﬀ erent design parameters, such as the varied reﬂectance of walls, ceiling, and ﬂoor; varied spatial distribution curves, including the number and position of luminaires; and across the spectral power distribution of a variety of LEDs. In this work, we seek to clarify and simplify to the greatest possible extent the meaning and scope of various lighting standards while establishing simple protocols. Our results will enable designers to carry out optimized lighting projects from both the photometric and circadian perspectives. light levels, minimize disruption to the body’s circadian system, and support good sleep quality. Feature 54: Circadian lighting design recommends as an approximation a parameter known as equivalent melanopic lux (EML), which is one of the ﬁve α -opic illuminance outputs used as criteria for “circadian lighting design”, with reference to the equi-energy illuminant instead of D 65 [19].


Introduction
Most interior lighting projects must comply with a series of requirements imposed by current regulations for various spaces, including levels of illuminance, uniformity, glare, correlated color temperature (CCT), or color rendering index (CRI) [1,2], and nearly always including budgetary concerns and a desire for low energy consumption [3,4]. Recently, there has been an increase in lighting projects that include, as new criteria, temporal variations in the light, both spectrally and in terms of intensity, with which our body is familiar and that regulate the circadian cycles of our biological clocks. Such projects represent the response to the presence of intrinsic photosensitive retinal ganglion cells (ipRGCs) in humans and the nonvisual effects of light [5][6][7][8][9], and there are numerous studies that highlight the substantial influence that variations in the intensity and tone of light have on health, mood, and many other factors related to the tasks performed by individuals who work for long periods of time under artificial light [10,11]. These lighting projects that consider the possible effects of light on people, optimizing them to create the greatest possible well-being in the short-, medium-, or longterm, are termed human-centric lighting (HCL) projects. However, the increasing number of such projects does not correspond to the importance that circadian light should have according to relevant studies, perhaps for several reasons, including the absence of specific clear regulations; manufacturer and market inertia; unacceptable costs; a lack of appropriate and properly characterized products; and a lack of a sufficient number of trained technicians, product promotion managers, or lighting designers.
Biologically, there are two dependent pathways for light in the brain: visual and nonvisual. The well-known visual tasks of rods, S-cones, M-cones, and L-cones have been widely described, however the role of ipRGCs remains under study; their contribution to both functions has been described [12]. Since the discovery of ipRGCs, different action spectra and metrics have been proposed in which both photopic and melanopic contributions can be easily quantified and interrelated between current metrics, applying it to a group of commercial LEDs and a lighting project. In this initial approximation, it is assumed that the materials of the luminaires and walls of the rooms do not modify the SPD of the vertical illumination that reaches the corneal plane.

Theoretical Considerations
In this section, basic expressions required to develop illumination projects, considering both the photopic and melanopic pathways, are clearly and briefly introduced. The method that we propose is calculated from the traditional relationship between radiometric and photometric magnitudes and from the current standards as follows: Melanopic irradiance ( W m 2 ) E e,melanopic = 780 λ=380 SPD(λ) × S mel (λ)dλ, To establish a full analogy between photopic and melanopic components, K melanopic must be defined. The K melanopic factor is well defined by the normalizations imposed by the different standards.
Equivalent melanopic lux (EML) is defined by the WELL standard [19] with reference to the equi-energy illuminant (E) and can be calculated as follows: where R = 780 λ=380 SPD(λ)×S mel (λ)dλ 780 λ=380 SPD(λ)×V(λ)dλ For any source, in terms of melanopic parameters, the EML is defined as follows: With K melanopic,E = 831.8 lm/W melanopic,E . In addition, EML's relation to photometric illuminance is as follows: In accordance with EDI as defined by the CIE standard [16], normalization is proposed with the melanopic illuminance provided by the standard illuminant D 65 (daylight CCT = 6500 K). A light-source type D 65 furnishing photopic illuminance E photopic,D65 to provide the same melanopic illuminance E melanopic,D65 can be calculated.
If the photometric quantity is defined by E photopic,D65 = K m × E e,photopic,D65 , the same melanopic E melanopic,D65 = K melanopic,D65 × E e,melanopic,D65 for light source D 65 can be described in terms of equality as follows: For any source: In addition, EML's relation with photometric illuminance can be deduced as follows: Using these conversion factors and photopic values of illuminance, transformations from one melanopic metric to another are easily calculated: Similarly, Rea et al. [13,28] mathematically define the modelled spectral sensitivity of the human circadian system by the circadian light (CL A ) expression in terms of the spectrally weighted lux per unit area. In this model, all known photoreceptors contribute to the spectral sensitivity of the circadian system [29]. There is a sudden transition in the modelled spectral efficiency at 497 nm. The CL A efficiency at shorter wavelengths reflects both ipRGC-melanopsin and S-cone sensitivities, whereas efficiency at longer wavelengths is modelled by the ipRGC-melanopsin spectral sensitivity alone. By definition, 1000 lux of CIE illuminant A equals 1000 lux on the CL A scale and is identified by the subscript "A", as in Equation (12): CL A is the circadian light and CS is the circadian stimuli. The constant, 1548 lm/W, sets the normalization of CL A so that 2856 K blackbody radiation at 1000 lux has a CL A value of 1000 lux. SPD λ is the light source spectral irradiance distribution; MC λ is the melanopsin (corrected for crystalline lens transmittance) [30,31]; S λ is the S-cone fundamental; mp λ is the macular pigment transmittance; V λ is the photopic luminous efficiency function; V' λ is the scotopic luminous efficiency function; RodSat is the half-saturation constant for bleaching (rods = 6.5 W/m 2 , k = 0.2616, a b-y = 0.700, and a rod = 3.300).
The WELL standard, based on the model developed by Lucas et al. [15], and the CIE model assume equal-energetic and D 65 theoretical illuminants, respectively, as references for their calculations. This approach represents one of the main conceptual differences of these proposals from the Rea et al. model, which adopts the A illuminant as its standard. Another important difference is the discrepancy in the circadian spectral sensitivity of the retinal ganglion cells; while the CIE model is based on the spectral response of the photopigments in the ipRGC, cone, and rod photoreceptors, the CS model is based on the suppression of the hormone melatonin. Considering all the proposed models as valid, it is necessary to determine the congruencies among these theories to simplify matters for designers and architects.
Based on the spectral distribution data in Figure 1, EDI (according to CIE S026:2018) and EML (according to WELL) for the different lamps at different CCTs were calculated for equal photopic illuminance (100 lux). These results were estimated at eye level. Based on the same spectral distributions, the circadian light (CL A ) and circadian stimulus (CS) as described by Rea et al. [14] were also calculated ( Table 2). To these calculations, we added calculations for the theoretical A, D65, and equal-energy illuminant, as described by the CIE. Table 2 shows the MAF calculated using Equation (4) from SPD, which is employed to find the melanopic illuminance with the different metrics. Differences among the outcomes occur, first, because of the normalization that is performed in the WELL metric with respect to the equal-energy illuminant, in the CIE metric with respect to the D 65 illuminant, and in the study by Rea et al. with respect to the A illuminant. The WELL values are always higher than those described by the CIE, with a factor 1.104 according to Equations (7) and (10). Rea et al. [14] adopted the function described in Equation (12); its discontinuity is found in Table 2 due to the differences in the definition of luminous efficiency in warm or cool lamps (part of the luminous efficiency has a negative contribution), where the circadian values that correspond to LEDs #16 to #21 (CCT from approximately 4100 K to 3500 K) have a lower melanopic efficiency when the calculations are performed with this metric rather than the other two ( Figure 2). Based on the spectral distribution data in Figure 1, EDI (according to CIE S026:2018) and EML (according to WELL) for the different lamps at different CCTs were calculated for equal photopic illuminance (100 lux). These results were estimated at eye level. Based on the same spectral distributions, the circadian light (CLA) and circadian stimulus (CS) as described by Rea et al. [14]    Regarding CCT, it has been determined that for the same CCT and equal photopic illuminance (100 lux), in general terms EDI and EML linearly increase with an increase in CRI ( Figure 3). The same behavior can be observed in the Rea metric, particularly at higher CCTs. At lower CCTs, it is more difficult to determine a fixed rule, perhaps because of the low number of LEDs evaluated with CRI 90 and CRI 95 . These characteristics are often observed. However, each lamp must be analyzed in detail, not only by CCT or CRI but also by its SPD. The MAF calculated with Equation (4) has a linear relationship with CCT, and it seems that the two well-correlated parameters can be used as references for the properties of the LEDs. Regarding CCT, it has been determined that for the same CCT and equal photopic illuminance (100 lux), in general terms EDI and EML linearly increase with an increase in CRI ( Figure 3). The same behavior can be observed in the Rea metric, particularly at higher CCTs. At lower CCTs, it is more difficult to determine a fixed rule, perhaps because of the low number of LEDs evaluated with CRI90 and CRI95. These characteristics are often observed. However, each lamp must be analyzed in detail, not only by CCT or CRI but also by its SPD. The MAF calculated with Equation (4) has a linear relationship with CCT, and it seems that the two well-correlated parameters can be used as references for the properties of the LEDs.

Lighting Project
This section describes the detailed parameters used in several lighting simulations visualized using the 3D modelling software program Dialux-EVO. Three virtual rooms measuring 10 m wide × 10 m deep × 2.80 m high were designed, and luminaires with regulation and different spectral distribution curves were chosen for each room: Lambertian (beam spread 120°), intensive (beam spread 54°), and extensive (beam spread 90°), (Figure 4).

Lighting Project
This section describes the detailed parameters used in several lighting simulations visualized using the 3D modelling software program Dialux-EVO. Three virtual rooms measuring 10 m wide × 10 m deep × 2.80 m high were designed, and luminaires with regulation and different spectral distribution curves were chosen for each room: Lambertian (beam spread 120 • ), intensive (beam spread 54 • ), and extensive (beam spread 90 • ), (Figure 4).

Lighting Project
This section describes the detailed parameters used in several lighting simulations visualized using the 3D modelling software program Dialux-EVO. Three virtual rooms measuring 10 m wide × 10 m deep × 2.80 m high were designed, and luminaires with regulation and different spectral distribution curves were chosen for each room: Lambertian (beam spread 120°), intensive (beam spread 54°), and extensive (beam spread 90°), (Figure 4).  In each room, 16 luminaires were regularly distributed in the ceiling. Their flux was 3750 lm. These luminaires can be equipped with LEDs with different SPDs and UGRs < 19, using an identical number of LEDs in the same Printed Circuit Board (PCB) format. The inner surface of the rooms was originally assumed to be ideally diffusive, with a reflectance of 80%. These parameters enable us to address the experimental design and analyze the importance of the deviations due to the spatial and spectral distribution curves of the LEDs with the materials that were employed. The effect on the luminous environment due to spectral reflectance of both furniture and surfaces (walls, ceiling, and floor) was analyzed using several specific cases. Illuminances in different planes (76 cm horizontal, 120 cm vertical) were calculated. The results exhibit varying reflectance factors (0%, 30%, 50%, and 80%), and different proportions of walls painted in black or white were analyzed in the same room (80% up and 0% down, 0% up and 80% down). These results reveal the maximum and minimum limits between which the variations in the illuminance are found.
A horizontal measurement grid was positioned 76 cm above the floor to represent a working plane, and a second grid was positioned vertically at the same 120 cm height to represent the corneal plane of the eye. The nine measuring points of the calculation area were chosen to have equidistant spacing, and we established an exclusion zone of 0.5 m wide from the walls to avoid untraveled areas. The adopted calculation parameters and the simulation results are provided in Table 3. All the results are for photopic illuminance. The intensive mode shows higher variations between the maximum and minimum values (horizontal and vertical), followed by the Lambertian and, in the third position, the extensive luminaire. The melanopic illuminance can be easily calculated from these values based on Equations (7), (10) and (12). Table 3. Relationship between the horizontal photopic illuminance (lux) in the work plane (76 cm) and the vertical photopic illuminance (lux) in the pupillary plane of the eye (120 cm). The variation in these values with the spatial distribution curves of the luminaires and with the reflectance of the environment is shown. The preceding results were calculated under the assumption that the observer is static. The same analysis can be performed for a dynamic observer with the ability to rotate his or her corneal plane around a vertical axis (horizontal plane located 120 cm from the floor) ( Figure 5). The results are shown in Table 4, where, to simplify the analysis, a wall reflectance of 80% was exclusively selected. These values clearly reflect a real situation in which the subject rotates his or her neck while, for example, performing an activity. Depending on the spatial location of the luminaire and the angle of vision, diverse illumination reaches the corneal plane, and as a consequence the melanopic contribution changes.
Appl. Sci. 2020, 10, x FOR PEER REVIEW  12 of 19 while, for example, performing an activity. Depending on the spatial location of the luminaire and the angle of vision, diverse illumination reaches the corneal plane, and as a consequence the melanopic contribution changes.  In another step, a more realistic scenario has been added. That is, the dynamic situation has been completed with a rotation with respect to a horizontal axis passing through the corneal vertex (120 cm from the floor), simulating a reading position ( Figure 6). To avoid excessive and repetitive simulations, two limit values of reflectance in the work plane (76 cm from the floor) are considered ( Table 5). The reflectance of the surfaces that is observed by a subject influences the illuminance at the eye level. Not only the reflectance of the wall or ceiling can elevate or diminish the amount of the light that reaches the corneal surface. To perform a more accurate simulation, the time that a subject spends looking in each position could be analyzed with an eye tracker. This approach could be useful to analyze realistic exposition times during working time.  Table 4. Relationship between the vertical illuminance at the eye-level plane (120 cm), 80% reflectance, and angular variation in the vertical axis. In another step, a more realistic scenario has been added. That is, the dynamic situation has been completed with a rotation with respect to a horizontal axis passing through the corneal vertex (120 cm from the floor), simulating a reading position ( Figure 6). To avoid excessive and repetitive simulations, two limit values of reflectance in the work plane (76 cm from the floor) are considered ( Table 5). The reflectance of the surfaces that is observed by a subject influences the illuminance at the eye level. Not only the reflectance of the wall or ceiling can elevate or diminish the amount of the light that reaches the corneal surface. To perform a more accurate simulation, the time that a subject spends looking in each position could be analyzed with an eye tracker. This approach could be useful to analyze realistic exposition times during working time.

Wall Reflectance 80%, Rotation around a Vertical Axis Photopic Illuminance (lux) E H-MED E H-MAX E H-MIN
( Table 5). The reflectance of the surfaces that is observed by a subject influences the illuminance at the eye level. Not only the reflectance of the wall or ceiling can elevate or diminish the amount of the light that reaches the corneal surface. To perform a more accurate simulation, the time that a subject spends looking in each position could be analyzed with an eye tracker. This approach could be useful to analyze realistic exposition times during working time. Figure 6. Graphical representation of the angular variation of the vision axis with respect to the horizontal axis; vertical illuminance at the eye-level plane (120 cm); wall reflectance of 80%. Table 5. Relationship between vertical illuminance at the eye-level plane (120 cm) with a wall reflectance of 80% and angular variation with respect to the horizontal axis. The design of the virtual room was carried out with dimensions and characteristics similar to a workspace or an office; the same procedure can be performed with other sizes and luminaires with different SPD and reflectance factors without a loss of generality.

Lighting for Elderly Individuals
The age-dependent spectral correction factor for lens transmission k(λ,Y) defined for an observer of age Y is a parameter that should be considered to address nonvisual effects of light in lighting designs in which the age of the user is a key factor [32]. This value is defined as k(λ,32 years) = 1, and the higher that the age is, the lower the value. This correction factor enables us to estimate the real transmission factor of the eye, the lens in particular, and to determine how much light passes through the pupil and reaches the retina. Table 6 shows the calculations for 4 LEDs whose characteristics are provided in Table 1. The higher the CCT is, the lower the k mel,trans . This phenomenon combined with the age of the subject creates a wide spectrum of possibilities that must be considered in lighting projects. As described below, Equations (7), (10), and (12) were defined at the corneal plane or entrance of the pupil (EP). Now, they must be modified to determine how much melanopic illuminance reaches the retina: where EML Retina , EDI Retina , and CL A−Retina are the effective values required by their respective standards-for example, illumination levels in different hospital areas, schools, or residential homes-and that should be considered in lighting projects.

Example of Lighting Project
The transition from the photopic metric based on horizontal illuminance on the work plane to the melanopic metric based on vertical illuminance at the corneal level is illustrated with a particular lighting project, which was selected to clarify the previously described procedure. To optimize the photopic and circadian contributions, three main considerations must be considered.

•
Luminaires should have the most extensive spatial distribution curve possible, satisfying the UGR requirements characterized by the current normative for each area.

•
Diffusive walls and ceiling with the highest reflectance values.

•
LEDs with suitable CCT and CRI, correct price requisites and with the highest MAF possible.
A sample lighting project that accords with the normative with reference to the horizontal illuminance requirements was analyzed, and the vertical illuminance (120 cm from the floor) was calculated at several points. If a value of, for example, 200 EML is required to qualify for WELL certification, the MAF of the lamp could be calculated as follows: Analogously, to satisfy the CIE requirements and if a value of 200 EDI lux is required, the verification is as follows: When values are expressed in terms of CL A and CS, the recommended values could be CL A = 220 lux and/or CS = 0.3 and must be divided by the CL A or CS of the source. To obtain the E V value, the result must be multiplied by 100: In all cases, to obtain the best combination of parameters three approaches are feasible: • Substitute the simulated E V value obtained with the illumination program to obtain the minimum allowed MAF that can include LEDs in the luminaires. If the LEDs do not reach this value, another LED with a higher MAF must be selected.

•
Modify the design of the luminaire; another luminaire with a more extensive profile should be used because it contributes higher E V .

•
Increase the value of the reflectance in the walls and ceiling, which increases the E V .
One should be careful when increasing the flux of the luminaires or their number to obtain a higher E V . This procedure could be flawed because E H will also increase, even at higher values than those required (and not recommended by the normative), and the installation could become energy-inefficient.
A numerical example with a classroom with exclusively artificial indoor lighting and insignificant natural daylight, providing the established minimum flux according to the normative EN 12464-1 (E H = 500 lux, UGR ≤ 19, and CRI ≥ 80), has been simulated [1]. The contribution of light necessary to stimulate the circadian system during daytime hours was stipulated at a minimum of 200 EML, 200 EDI, and 220 CL A , with 0.3 CS.
The values provided in Table 3 are required. The E H values that achieve the normative EN 12464-1 [1] are 500 lux ≤ E H ≤ 600 lux (upper limit calculated with the condition of avoiding exceeding the established medium levels by 20%). Lambertian illumination with a reflectance of 80% was selected to create a photopic lighting design according to the three described metrics. The circadian illuminance at eye level can be calculated as follows.
In the WELL metric, MAF = 164/288 = 0.569. Any LED from #1 to #19 can be selected for installation. Based on our experience, in a classroom a medium CCT (from #14 to #19) and CRI ≥ 80 are most appropriate.
According to the CIE metric, MAF = 181/288 = 0.628. Based on the previous reasoning, any LED from #1 to #18 can be selected.
To summarize this discussion, LED #18 can be selected according to the WELL and CIE models but cannot be selected according to Rea, while #23 is adequate according to Rea but does not satisfy the other models. A generally satisfactory solution would be to select LED #15 (CCT 4654K, CRI 97), which would be the best option for photopic norms while also meeting requirements based on the three circadian metrics. Another approach would be to modify the reflectance of the surroundings. A third would be to change the spatial distribution of the luminaires while searching for the best combination of the parameters.
To provide options, we can increase the number of luminaires of the project. We started with an equidistant distribution of 16 luminaires. This number can be changed to 25 adjustable luminaires. The normative limit of E H = 600 lux has been maintained to recalculate the circadian contribution of the installation. In this case, E V = 351 lux, MAF CIE = 0.515, MAF WELL = 0.467; CL A = 62.67 and CS = 0.085, indicating that the best circadian solution for the three metrics is LED #22 (CCT 3460K, CRI 99).

Discussion
The health of the visual and non-visual system is critical and has become one of the main concerns of modern society. Lighting designers require a precise and simple tool or guide to estimate eye-level photopic and melanopic illuminance and to determine how such illuminance can be improved in each situation according to the current normative. Illumination is a key factor in the environment but has not been accurately evaluated despite the recommendations that are available according to various measurement methods and equipment [25,33]. Improving the environment and appropriately regulating the amount of light that reaches eye level could have a positive impact on quality of life, well-being, and aging-related concerns.
This objective could be achieved by analyzing illuminance characteristics and regardless of the metric used. Having confirmed that the MAF represents a satisfactory parameter for calculating the melanopic contribution to a fixed photopic illuminance level (or vice versa), we have described several illumination designs to elucidate an appropriate method for quantifying these light types. We further show that measuring light in these terms facilitates predicting light levels, both photopic and melanopic, with different spectral distributions much more reliably than other methods that quantify metric-dependent contributions. To make the parameters more readily understandable, our approach is described based on three common metrics which seem entrenched in the lighting sector and whose interconnection and relationship with photopic illuminance is discussed in the first part of the paper. This approach seems to face more difficulty in extrapolating the CL A and CS parameters because of the intrinsic definition of these values, in particular for LEDs with CCTs ranging from 3500 K to 4100 K, where the differences related to melanopic performance reach their maximum values depending on the metric used.
Illuminations, both horizontal and vertical, critically depend on the dimensions of the room and the reflectance of its walls and furniture [23,34,35]. Room surface reflectance is an important factor in achieving a high illuminance and has even been described as being much more important than the light that arrives through the windows, particularly in the winter months and depending on the orientation of and the distance from the windows to the workstation [34]. Our results confirm those of the cited studies and add the reflectance of the observed surface (in our case, a table) as a parameter to be considered in calculations, whether for photopic or melanopic illuminance. The level of illuminance that arrives at the corneal plane, depending on the spatial distribution curve of the luminaires and the flow emitted by them, was investigated. The simulations using three selected profiles performed in this paper, which include the reflectance of the surfaces, facilitate verifying the importance of such reflectance in circadian lighting applications as a key tool with which to modify and control eye-level illumination [36].
When the melanopic component must be controlled, the SPD of the luminaires should be the parameter to be evaluated [25]. In a first approach, it would be sufficient to know the spectral parameters of the light sources to evaluate the contribution of the circadian component in the lighting project related to circadian rhythms (either WELL, CIE, Rea, or any other requirement imposed according to the objectives). However, in a second step knowledge of the characteristics of the luminaire and the surroundings will be necessary in order to design a complete and accurate lighting project. Comparisons among metrics have been proposed by various organizations [37]. A solution for a minimum circadian effect with optimal CRI requirements has been described using white LED solutions based on RGB LEDs [38][39][40], the combination and optimization of circadian effect and visual lit appearance (brightness or dimness) can be obtained [41], and spaces have been designed that accommodate seasonal changes in illumination at the eye level [34,42]. However, actual lighting projects are designed using the photometric characteristics of the luminaires, with specific SPDs as well as CCT or CRI. An easy procedure must be provided that includes a circadian technical solution. The described analysis enables us to recommend an approach for optimizing well-being lighting projects by considering together the parameters that influence the visual and nonvisual pathways.

Conclusions
We propose that illumination designers study real environments with commercial lights and evaluate the light from visual and nonvisual perspectives to determine a method to improve design feedback based on photometric parameters and in accordance with the current normative. Quantitative models have been proposed to evaluate light's nonvisual effects, resulting in confusing calculations, interpretations, inter-comparisons, and applications. It is important to develop a lighting design method that takes into account visual considerations according to current international standards and that can be easily interpreted and achieved independently of the nonvisual metric used. In this study, we define the parameter MAF, which is SPD-dependent and which facilitates the quantification and comparison of different metrics of circadian effects and visual appearance. These different circadian effects are discussed and contrasted using case studies to establish a pattern that lighting designers can follow in their illumination projects with commercial lights. Our paper introduces a simple protocol that should enable even novice designers to develop lighting projects using basic photopic and melanopic parameters with commercial luminaires and lamps.
The proper method for light measurement is preferably based on the SPD of the light source. The described conditions and control systems are critical to estimate the effects and to construct robust optic models. To study the effect of light on health, a multidisciplinary approach is required. Light and illumination are based on illuminance and color appearance. The spectral action of the photoreceptors with their three cone types and rods and that of the ipRGCs are completely different. The melanopsin function and illuminance equivalent to D 65 considered by the CIE, the equal-energetic illuminant considered by WELL, and the CL A or CS parameters of Rea are widely accepted as methods for light measurements at the corneal level in laboratory studies, although Chartered Institution of Building Services Engineers (CIBSE) and the BRE Group concluded in a literature review that the existing recommendations in the WELL Building Standard and DIN SPEC 67600 should be treated with caution. However, a standard for the visual field and environmental illumination does not exist in terms of spectral resolution. The most important contributions of this study are summarized as follows: be considered as it has previously been described. These equations can provide quick feedback for lighting manufacturers and designers regarding healthy circadian effects, and the correspondence between both metrics is easily determined.

•
An example of our method's application is provided using numerical simulation data obtained with DIAlux. Various lighting conditions are described, including different types of luminaire, various combinations of room surface reflectance, various orientations of the visual axis, and the observer's age. • Three ways are described to obtain the best photopic and melanopic illuminance levels in lighting projects: Substitute the simulated E V value obtained with the illumination program to obtain the minimum allowed MAF, CL A , and CS that can include LEDs in the luminaires. If the LEDs do not reach this value, other LEDs with a higher proportion must be selected. Modify the design of the luminaire; another luminaire with a more extensive profile should be used if it contributes a higher E V . Increase the value of the reflectance in the walls and ceiling, which will increase the E V .
• In this initial approximation, to perform a lighting project MAF, the CL A and CS parameters should be provided in the datasheet of the lamps. However, we have several limitations; the results reported in this study are based on a sample of SPDs, and our results could not be generalized to all LED spectra. In particular, these results should not be expected to have any predictive power for color-mixed LED solutions that employ arrays of narrow-emitting LEDs to generate nominally white light. It is assumed that the materials of the luminaires and walls of the rooms do not modify the SPD of the vertical illumination that reaches the corneal plane. In future research, these limitations should be considered.

Funding:
This research was partially funded by Ministerio de Ciencia e Innovación, grant number PID2019-107058RB-I00.