This article was downloaded by: [University of California, San Francisco]

On: 28 October 2014, At: 14:01

Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK

LEUKOS: The Journal of the Illuminating Engineering

Society of North America
Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/ulks20

Practical Use and Calculation of CCT and Duv

a
Yoshi Ohno
a
National Institute of Standards and Technology, Gaithersburg, Maryland, USA
Published online: 18 Oct 2013.

To cite this article: Yoshi Ohno (2014) Practical Use and Calculation of CCT and Duv, LEUKOS: The Journal of the Illuminating
Engineering Society of North America, 10:1, 47-55, DOI: 10.1080/15502724.2014.839020

To link to this article: http://dx.doi.org/10.1080/15502724.2014.839020

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions
 LEUKOS, 10:47–55, 2014
ISSN: 1550-2724 print / 1550-2716 online
DOI: 10.1080/15502724.2014.839020

Practical Use and Calculation of CCT and Duv

Yoshi Ohno
National Institute of Standards ABSTRACT Correlated color temperature (CCT) is often used to represent chro-
and Technology, Gaithersburg, maticity of white light sources, but chromaticity is two-dimensional, and another
Maryland, USA dimension, the distance from the Planckian locus, is often missing. Duv is defined in
ANSI C78.377 for this purpose but is not yet widely used. In this article, the use of a
Downloaded by [University of California, San Francisco] at 14:01 28 October 2014

combination of CCT and Duv is proposed as an intuitive expression of chromaticity

of white light sources for general lighting. In addition, this article presents practical
calculation methods to calculate CCT and Duv, having sufficient accuracy, within
an error of 1 K, in a wide range of chromaticity, from 1000 to 20,000 K in CCT and
−0.03 to 0.03 in Duv.
KEYWORDS chromaticity, correlated color temperature, Duv, light source, Planckian locus

1. INTRODUCTION
The chromaticity is one of the critical parameters for light sources for general light-
ing and is normally specified with chromaticity coordinates CIE (x, y) or (u , v ).
However, these two numbers do not provide the color information intuitively. For
practical purposes, correlated color temperature (CCT) is commonly used to pro-
vide the chromaticity information of general illumination sources. CCT, however,
provides only one dimension of the chromaticity, and there is another dimension,
which is the position of chromaticity with respect to Planckian locus. For this pur-
pose, “duv” or similar terms as the distance from the Planckian locus have been
used in some parts of the industry, but these had not been officially defined in any
standards. Recently, Duv was defined in an American National Standards Institute
standard [ANSI 2008].
For color quality information of lighting products (for example, in data sheets),
only CCT and the color rendering index [CIE 1995] are often used, and Duv is
not shown though it is also critical. The chromaticity specification for fluorescent
lamps [ANSI 2001] specifying (x, y) coordinates has long been used, and that for
solid-state lighting products, specifying the ranges of Duv, is also available [ANSI
This article not subject to US copyright
law. 2008]. However, not all products in the market meet these requirements. It is a
Received 6 April 2013; revised 22 August
problem that some products do not meet the requirements in these standards because
2013; accepted 23 August 2013. inappropriate Duv may cause rejection by customers. It has also been anecdotally
Address correspondence to Yoshi reported that chromaticity points slightly below the Planckian locus are preferred.
Ohno, National Institute of Standards Duv is important for the users’ acceptance and preference of light sources. Related
and Technology, 100 Bureau Dr. MS
8442, Gaithersburg, MD 20899-8442. research on white point for light sources has been reported [Rea and Freyssinier-Nova
E-mail: [email protected] 2013].

47
 A Duv value provides information on the distance
and direction of a color shift from the Planckian locus c1 λ−5  c 
2
−1
Le,λ (λ, T ) = exp −1 (1)
(yellowish/greenish or pinkish), whereas the “duv” or other π λT
term often showed only the distance and no informa-
where L is the spectral radiance, T is the temperature (K),
tion on the direction of the shift. With the use of CCT
and λ is the wavelength in the medium. Note that the
and Duv, the two numbers can provide full information
refractive index of the medium, n, which is normally used
on white light chromaticity of light in an intuitive man-
in the Planck’s equation for radiometric applications, is not
ner, and these two numbers can replace (x, y) or (u ,
used for the calculation of CCT. c 1 , c 2 are the radiation
v ) in test reports and specification data. In this arti-
constants; the value of c 1 (3.741 7749 × 10−16 W·m2 ) is
cle, the use of CCT and Duv is proposed to express
not relevant to the value of CCT, and only c 2 is relevant.
chromaticity of white light sources for lighting, whereas
The value of c 2 used in the calculation of CCT should fol-
color differences should use distances in (u , v ) coordi-
low that adopted in the current International Temperature
nates. Because Duv is not widely known, its definition
Scale (ITS-90) [Preston-Thomas 1990], namely, c 2 =
together with that of CCT is first introduced in sec-
1.4388 × 10−2 m·K. See Appendix E of CIE 15:2004
Downloaded by [University of California, San Francisco] at 14:01 28 October 2014

tion 2. The use of CCT and Duv is further discussed

[CIE 2004] for the recommended use of these values for
in section 3 by introducing a chart using these two
calculation of CCT.
numbers.
Duv (symbol: Duv ) is defined as the distance from the
Duv is not available in many commercial instruments,
chromaticity coordinate of the test light source to the clos-
and it is often asked how to calculate Duv. In Section 4,
est point on the Planckian locus on the CIE (u , 2/3v )
methods to calculate Duv from CCT and (x, y) or only
coordinates, with a plus sign for above and a minus sign
from (x, y) are presented, including an approximation for-
for below the Planckian locus [ANSI 2008]. Duv is nor-
mula developed by the author. CCT and Duv are generally
mally calculated in the process of calculating CCT, which
calculated together, and it is also often asked how to cal-
requires determining the closest point on the Planckian
culate CCT. In Section 5, practical methods to calculate
locus. By knowing the chromaticity coordinate (u , v )
CCT and Duv developed by the author are presented with
of the test source and that of the closest point on the
evaluation of their accuracies.
Planckian locus (u0 , v0 ), Duv is given by
There are several other methods available for simpler
formulae to calculate CCT directly using polynomials   2 12
(not minimum-search). Such methods tend to have lim- 2  2 
Duv = (u − u 0 ) + · sgn(v − v0 ),
2
v − v0
ited accuracies and limited chromaticity ranges, and such 3 3
small-size programs are becoming less important for the (2)
capacity of modern computers and thus are not cov-
ered in this article. The methods described in this article where sgn(z) = 1 for z ≥ 0 and sgn(z) = −1 for z < 0.
work accurately enough in a wide range of chromatic-
ity, from 1000 to 20,000 K in CCT and −0.03 to
+0.03 in Duv and can be easily extended further if 3. CCT-DUV CHART
necessary. As mentioned above, the two numbers of chromaticity
coordinates (x, y) or (u , v ) do not carry color infor-
mation intuitively. CCT and Duv can do so, and these
2. DEFINITIONS OF CCT AND DUV two numbers can be used to specify the chromaticity of
CCT (Symbol: T cp ) is defined as “temperature of the a white light source instead of (x, y) or (u , v ). It would
Planckian radiator having the chromaticity nearest the also be more convenient to judge the compliance to the
chromaticity associated with the given spectral distribution ANSI [2008] specification because it defines the ranges of
on the diagram where the (CIE 1931 standard observer chromaticity by CCT and Duv, see Fig. 1 (left).
based) u , 2/3 v coordinates1 of the Planckian locus It is also possible to create a chart using the coordinate
and the test stimulus are depicted” [CIE 2004, 2011, (T cp , Duv ) for graphic expression of chromaticity; as an
p. 36]. The Planck’s equation to be used for calculation of example, see Fig. 1 (right), where the quadrangles of the
CCT is ANSI [2008] specifications are plotted.

48 Y. Ohno
 Fig. 1 The ANSI specification [ANSI 2008] plotted on the (u, v ) diagram (left) and on a CCT-Duv chart (right) (color figure available
online).

This is an attempt to make such a chart for the whole (2) to calculate Duv. If the value of CCT is not available,
Downloaded by [University of California, San Francisco] at 14:01 28 October 2014

CCT range. In this case, a log scale is used for CCT to CCT needs to be calculated first from (x, y) or (u , v ) to
make the color space reasonably uniform, as shown in the use this method.
sizes of the seven-step MacAdam ellipses [DOE 2008] also
plotted. In this example, the CCT scale is reversed so
that the direction of the CCT scale is the same as that 4.2. Simple Approximation Formula
on the chromaticity diagrams. Charts with the CCT scale
This simple method can be used if the value of CCT is
not reversed can also be used. Such a chart can be used,
not available, and Duv needs to be calculated from (x,
for example, for quality control, by plotting (T cp , Duv ) of
y) or (u , v ) only and without using the full calculation
products measured.
procedures for CCT as described in section 5. First, the
A further consideration is that reporting Duv values
chromaticity coordinate of the light source (x, y) or
with many zeros after the decimal point is not convenient.
(u , v ) needs to be converted to the 1960
For more practical use of this quantity, a further idea would
(u, v)2 :
be a new quantity (for example, called “tint”) defined as
1000 times Duv, so that, for example, Duv −0.003 would
u = 4x/(−2x + 12y + 3) u = u
be “tint” −3.
or (3)

4. CALCULATION OF DUV v = 6y/(−2x + 12y + 3) v = 2v /3.

CCT is commonly shown in commercial instruments, but

Then, calculate
Duv is often not available. Duv can normally be calcu-
lated in the process of calculating CCT as described in

Section 4.2. However, if Duv needs to be calculated sepa- LFP = (u − 0.292)2 + (v − 0.240)2
rately from calculation of the CCT, the following methods  
can be used. u − 0.292
a = arccos
LFP

4.1. Exact Calculation Method LBB = k6 a6 + k5 a5 + k4 a4 + k3 a3 + k2 a2 + k1 a + k0

If the value of CCT as well as (x, y) or (u , v ) of the light Duv = LFP − LBB
source is known, Duv can be calculated simply as follows. (4)
First, obtain the spectral distribution of the Planckian
radiation at temperature T of the CCT using (1). Then, where k6 = −0.00616793, k5 = 0.0893944,
from the Planckian spectrum, calculate the chromaticity k4 = −0.5179722, k3 = 1.5317403, k2 = −2.4243787,
coordinate (u0 , v0 ) using any program available. Then, use k1 = 1.925865, and k0 = −0.471106.

Practical Use and Calculation of CCT and Duv 49

Downloaded by [University of California, San Francisco] at 14:01 28 October 2014

Fig. 2 Errors in Duv using the approximation polynomial (color figure available online).

The accuracy of this formula is shown in Fig. 2. The

errors in calculated Duv in the range from 2600 to
20,000 K and the Duv range from −0.03 to + 0.03 (well
covering the range in the ANSI [2008] standard) are within
0.00003, which is practically negligible. Below 2600 K, the
error sharply increases to up to 0.0005 at ∼2000 K and
within Duv = ±0.03. This formula is included as informa-
tion in the revised version of ANSI C78.377-2011 [ANSI
2008].

4.3. Conversion from (CCT, Duv)

Back to (x, y)
In some cases, the values of CCT and Duv, (T cp , Duv ),
need to be converted back to (x, y) or (u , v ). This conver- Fig. 3 Principles of the conversion from (CCT, Duv) to (u, v )
sion can be made with the following simple calculations. (color figure available online).

First, calculate the 1960 (u, v) coordinates, (u0 , v0 ) and

(u1 , v1 ), of the Planckian radiation at T cp (K) and at T cp
 
+ T (K), respectively. T is a small change in CCT; for Then, (u, v) can be converted to (u , v ) or to (x, y):
example, 0.01 (K). Then, calculate the slope angle θ of the
Planckian locus at T cp . (K) as illustrated in Fig. 3 with the
formula u = u, v = 1.5 v
(6)
x = 9 u /(6u −16v +12), y = 2 v /(3u −8v +6).
du = u0 − u1 , dv = v0 − v1 ,
(5)
u = u0 − Duv · sinθ , v = v0 + Duv · cosθ
There is no approximation in this conversion calculation.
This conversion method can create pairs of (CCT, Duv)
where
and (x, y) or (u , v ), which are used for evaluation of
  the accuracy of CCT calculation methods described in
sinθ = dv/ du2 + dv2 , cosθ = du/ du2 + dv2 . Section 5.
50 Y. Ohno
 5. CALCULATION OF CCT AND DUV is the distance from (ux , vx ) to (ui , vi ), with the light source
chromaticity ux = 0.2193, vx = 0.3271 as an example.
CCT can be calculated by a simple minimum-search com-
Note that values shown in Table 1 are truncated. Values
puter program that searches for that Planckian temperature
in full significant digits should be used in this calculation.
that provides the smallest chromaticity difference between
The spacing of color temperature Ti should be chosen so
the test chromaticity and the Planckian locus on the (u ,
that the spacing in (ui , vi ) can be reasonably uniform.
2/3v ) coordinate. This can be achieved in many different
Table 1 uses color temperature steps of a 1% increase
ways. Robertson’s [1968] method, which uses a small table
from that in the previous row, and a total of 303 rows
of values with more complex procedures and was very suit-
is used to cover from 1000 to 20,000 K, easy to handle
able for the limited capacity of computers at that time, is
in a normal spreadsheet application. A Planckian (u, v)
well known. To use Robertson’s method, the table values
table using the above conditions is hereinafter called a 1%
need to be revised for ITS-90, and the errors can be several
CT-step table.
Kelvin at higher CCTs. Described below are the algorithms
The second step is to find the point i = m where di
developed by the author and used at NIST. These algo-
is the smallest in the table, as shown in the last column
rithms takes a larger data table but use a simple algorithm
Downloaded by [University of California, San Francisco] at 14:01 28 October 2014

of Table 1. This assures that the CCT to be obtained is

that is accurate to within 1 K for a CCT range of 1000 K
between Tm-1 and T m+1 . Then a triangle is formed by
to 20,000 K and a Duv range of −0.03 to + 0.03.
the chromaticity point (ux , vx ) of the light source, the two
chromaticity points on Planckian locus at Tm+1 and at
5.1. Triangular Solution T m-1 as depicted in Fig. 4. The Tx in the figure is the clos-
est point to the light source (ux , vx ) and is the CCT value
The CCT can be simply calculated by searching the clos- to be obtained. Assuming a linear line between the two
est point on the Planckian locus on the CIE 1960 (u, v) points Tm+1 and T m-1 on the Planckian locus, the black-
diagram by computation without a sophisticated approach body temperature Tx for the closest point is obtained by
like Robertson’s [1968] method, though the data table solving the triangle:
needs to be larger, which is not much an issue for mod-
ern computers. One of the approaches is the triangular x
Tx = Tm−1 + (Tm+1 − Tm−1 ) ·
solution. The first step is to prepare a table of (ui , vi ) l
coordinates of Planckian locus (called the Planckian (u, v) with
table) at certain intervals in the range of CCT needed and
(7)
then calculate the distance di from the (ux , vx ) of a test 2
dm−1 − dm+1
2
+ l2
light source. Table 1 shows an example for such a table for x= .
2l
the CCT range from 1000 to 20,000 K. Ti is the color

temperature of Planckian radiation, and di and l = (um+1 − um−1 )2 + (vm+1 − vm−1 )2

TABLE 1 Example of Planckian (u, v ) table

Ti (K) ui vi di (example) i

1000.00 0.448011 0.354625 0.230361 1

1010.00 0.445681 0.354826 0.228073 2
1020.10 0.443353 0.355026 0.225787 3
1030.30 0.441028 0.355223 0.223504 4
: : :
4404.38 0.218633 0.329547 0.002536 m−1
4448.42 0.218016 0.329042 0.002328 m
4492.91 0.217407 0.328537 0.002377 m+1
: : :
19, 592.54 0.183990 0.277410 0.060959 300
19, 788.47 0.183939 0.277254 0.061115 301
19, 986.35 0.183888 0.277100 0.061270 302
20, 186.21 0.183838 0.276948 0.061423 303

Practical Use and Calculation of CCT and Duv 51

 SIGN (z) = 1 for z ≥ 0 and SIGN (z) = −1 for z < 0.

The accuracy of this method using the example in Table 1

has been evaluated and the results are shown in Fig. 5. The
error in CCT increases significantly with the Duv deviating
further from zero in either direction. The accuracy of this
method is good only in the region close to the Planckian
locus. Note that, in Fig. 5, the errors do not minimize
exactly at Duv = 0.00; in particular, for 20,000 K, it
minimizes at around Duv = −0.02. This is because there
is another error not related to Duv, which is about 2 K
at Duv = 0.00 at 20,000 K, and the error related to Duv
at Duv = −0.02 cancels this error at 20,000 K. Beyond
Downloaded by [University of California, San Francisco] at 14:01 28 October 2014

this point, the actual CCT error is negative, but absolute

values are plotted in this figure. In addition, the errors
are generally higher at higher CCT but not always so,
because errors also depend on where the calculated CCT
value is positioned between the closest CCT points in
Fig. 4 Principles of the triangular solution (color figure avail-
the table.
able online). To further examine the errors around the Planckian
locus, Fig. 6a shows the errors at Duv = 0, Duv = 0.002,
Once Tx is determined, Duv is given as the length from the and Duv = −0.002 as a function of CCT. It is consid-
given point (ux , vx ) to the point on the Planckian locus at ered that these errors are caused due to nonlinearity of
Tx , thus obtained from the triangle: the CCT scale on (u, v) coordinates. These errors are
 2 1/2 much reduced by applying a correction, Tx,cor = Tx ×
Duv = dm−1 − x2 · sgn(vx − vTx ) 0.99991, as shown in Fig. 6b. This correction factor
value is optimized for this example condition (a 1%
with (8)
increment in CCT in Table 1) and it will change for
vTx = vm−1 + {vm+1 − vm−1 } · x/l different CT steps of the Planckian (u, v) table. For
CT steps of 0.25% or smaller this correction is not
and needed.

Fig. 5 Errors in CCT [K] using the triangular solution with 1% CT-step table. Errors are in absolute value (color figure available online).

52 Y. Ohno
 Fig. 6 Errors of the triangular solution at Duv = 0, −0.002, and +0.002 using a 1% CT-step Planckian (u, v ) table (a) before and (b) after
the correction is made (color figure available online).

5.2. Parabolic Solution

Another approach for a direct search algorithm is to use a
Downloaded by [University of California, San Francisco] at 14:01 28 October 2014

parabolic function. In the method described in Section 5.1,

after determining Tm+1 and T m−1 (see Table 1 for an
example), the three values of d m−1 , d m , d m−1 are fitted
to a parabolic function as shown in Fig. 7. The polynomial
is derived from d m−1 , d m , d m−1 and T m−1 , T m , T m−1 as

d (T ) = a T 2 + b T + c, (9)

where
Fig. 7 Example of parabolic fit for d m-1 , d m , d m-1 (color figure
X = (Tm+1 − Tm ) (Tm−1 − Tm+1 ) (Tm − Tm−1 ) available online).

a = [Tm−1 (dm+1 − dm ) + Tm (dm−1 − dm+1 )+

Note that the value of this correction factor depends on
−1 the data table used, and the value above is optimized for
Tm+1 (dm − dm−1 )] · X
 2 the condition of Table 1. Once Tx,cor is determined, Duv
b = − Tm−1 (dm+1 − dm ) + Tm2 (dm−1 − dm+1 )+ is obtained by

2
Tm+1 (dm − dm−1 ) · X −1
Duv = SIGN (vx − vTx ) · (a Tx,cor
2
+ b Tx,cor + c) (11)
c = − [dm−1 (Tm+1 − Tm ) · Tm · Tm+1 +
with
dm (Tm−1 − Tm+1 ) · Tm−1 · Tm+1 +

dm+1 (Tm − Tm−1 ) · Tm−1 · Tm ] · X −1 . SIGN (z) = 1 for z ≥ 0 and SIGN (x) = −1 for z < 0.

The CCT of the input light is determined as the color tem- vTx is the v coordinate of the Planck’s radiation at temper-
perature T = Tx where the parabolic function reaches the ature Tx and may be calculated from the Planck’s equation
minimum and is obtained as or using the formula for vTx in (8) if the triangular solution
is also calculated.
b The accuracy of this method has been evaluated for the
Tx = − ∵ d  (T ) = 2aTx + b = 0 (10) example in Table 1 and the results are shown in Fig. 8.
2a
It is clearly shown that this method works very well for a
The same correction factor for the nonlinearity effect is wide range of Duv, but errors are very large around Duv
applied as = 0 (right on the Planckian locus). The reason for this
is explained in Fig. 9, which shows the case for CCT =
Tx, cor = Tx × 0.99991. 4000 K and Duv = 0. The parabolic minimum (solution)
Practical Use and Calculation of CCT and Duv 53
 Fig. 8
Downloaded by [University of California, San Francisco] at 14:01 28 October 2014

Errors in CCT [K] using the parabolic solution with the 1% CT-step Planckian (u, v ) table. Errors are in absolute values (color
figure available online).

Fig. 9 The case of parabolic fit for 4000 K and Duv = 0. The dashed line shows plots of actual distances. In this case the distances do
not follow a parabolic curve (color figure available online).

does not reach distance 0, though it should. In this case, within 0.5 K in the 2000 to 20,000 K range and −0.03
the actual distances follow V-shaped linear lines (dashed < Duv < 0.03 and the maximum error is 0.8 K (which
line) rather than a parabolic curve, causing the error in occurs at conditions other than shown in Fig. 10).
the obtained CCT. This occurs only at or very near the
Planckian locus.
5.5. Further Improvements
The error analyses presented above are for the cases using
5.4. Combined Solution
the 1% CT-step Planckian (u, v) table shown in Table 1,
The parabolic solution seems to work accurately except on consisting of 303 rows of data. The accuracy obtained in
or near the Planckian locus. To solve this problem, the tri- this condition is considered to be sufficient for most prac-
angular solution and the parabolic solution are combined tical applications. However, when necessary, the accuracy
so that the results of the triangular solution are used for can be easily improved by using a larger table. For exam-
|Duv | < 0.002 and the parabolic solution is used for other ple, by using a 0.25% CT-step table (1210 rows), the errors
regions. The threshold was determined where the errors will be reduced to one quarter of the results shown in
of both methods are at similar level. Figure 10 shows the Fig. 10, less than 0.2 K in the same ranges of CCT and
results of this combined solution. The errors are mostly Duv. By using a table with much smaller CCT intervals,
54 Y. Ohno
 Fig. 10 Errors in CCT [K] using the combined solution (1% CT-step table). Errors are in absolute value (color figure available online).
Downloaded by [University of California, San Francisco] at 14:01 28 October 2014

even higher accuracy can be achieved, though the size of NOTES

program will be very large. 1. The (u , 2/3v  ) coordinate is equal to that in the CIE 1960 (u, v) dia-
Another idea would be to start with a small Planckian gram, which is now obsolete. CCT has historically been calculated
using CIE 1960 (u, v). Changing to the 1976 (u , v  ) was investigated
(u, v) table (for example, 15% CT-step, only 20 rows)
when CIE 15:2004 [CIE 2004] was published; however, the CIE com-
for the above combined method and then automatically mittee then concluded to keep the CIE 1960 (u, v) as the basis for
expand the (u, v) table to smaller intervals (for example, calculation of CCT.
2. The 1960 (u, v) coordinate, now obsolete, is used in this article for
one tenth of the original interval) based on the first simplicity in expression and due to historical reasons for calculating
solution and then repeat expansion and calculations. CCT.
By expanding to three or four levels, this method would
achieve extremely high accuracy without requiring a very
large table. REFERENCES
[ANSI] American National Standards Institute. 2001. Specifications for the
Chromaticity of Fluorescent Lamps. Rosslyn (VA): ANSI. Publication
6. CONCLUSIONS No. C78.376. 16 p.
[ANSI] American National Standards Institute. 2008. American
Duv is an important parameter for color quality of light national standard for electric lamps—specifications for the
chromaticity of solid state lighting products. Publication
sources for lighting. The two numbers CCT and Duv can ANSI_NEMA_ANSLG C78.377-2008. Rosslyn (VA): American
specify the chromaticity coordinates of white light sources National Standard Lighting Group. 17 p. And its revision C78.377-
intuitively. The use of these two numbers, rather than 2011. 21p.
[CIE] Commission International de l’Éclairage. 1995. Method of measur-
(x, y) or (u , v ), is proposed to specify or report chromatic- ing and specifying colour rendering of light sources 3rd ed. Vienna
ity of light sources for general lighting. Several practical (Austria): CIE. Publication No. 13.3:1995. 16 p.
[CIE] Commission International de l’Éclairage. 2004. Colorimetry. 3rd ed.
calculation formulae to calculate Duv only or both CCT Vienna (Austria): CIE. Publication No. 15:2004. 72 p.
and Duv are presented. Two methods for calculating CCT [CIE] Commission International de l’Éclairage. 2011. ILV: interna-
and Duv using a minimum-search computer program, tional lighting vocabulary. Vienna (Austria): CIE. Publication No. S
017/E:2011.196 p.
triangular solution, and parabolic solution are presented. [DOE] Department of Energy. ENERGY STAR® program requirements for
Problems were identified for each of these methods. The CFLs ENERGY STAR eligibility criteria energy-efficiency criteria. Version
4.1. 2008. 21 p.
combination of the two solutions achieves high accuracy, Preston-Thomas H. 1990. The international temperature scale of 1990
less than 0.8 K, in a wide range of chromaticity from (ITS-90). Metrologia. 27: 3–10.
2000 to 20,000 K and −0.03 to 0.03 in Duv. If neces- Rea MS, Freyssinier JP. 2013. White lighting. Color Res Appl. 38 (2): 82–
92.
sary, the accuracy of calculation can be easily improved by Robertson AR. 1968. Computation of correlated color temperature and
using a table at smaller CCT intervals. distribution temperature. J Opt Soc Am. 58:1528–1535.

Practical Use and Calculation of CCT and Duv 55