Viewing conditions - Graphic technology and

photography

(Revision of ISO 3664 - 1975, Photography -

Illumination conditions for viewing colour
transparencies and their reproductions)

i
ISO 3664:2000(E)

Contents Page

Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv

1 Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

2 Normative references . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

3 Terms and definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

4 Viewing condition requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

5 Test methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Annexes

A Summary of ISO viewing conditions specified in this International

Standard . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

B Experimental data leading to selection of metameric indices and reference

illuminant for this International Standard . . . . . . . . . . . . . . . . . . . . . . . . 14

C Guidelines for judging and exhibiting photographs . . . . . . . . . . . . . . . . 19

ii © ISO 2000 - All rights reserved

 ISO 3664:2000(E)

Foreword
ISO (the International Organization for Standardization) is a worldwide federation
of national standards bodies (ISO member bodies). The work of preparing
International Standards is normally carried out through ISO technical committees.
Each member body interested in a subject for which a technical committee has been
established has the right to be represented on that committee. International
organizations, governmental and non-governmental, in liaison with ISO, also take
part in the work. ISO collaborates closely with the International Electrotechnical
Commission (IEC) on all matters of electrotechnical standardization.

International Standards are drafted in accordance with the rules given in the
ISO/IEC Directives, Part 3.

Draft International Standards adopted by the technical committees are circulated to

the member bodies for voting. Publication as an International Standard requires
approval by at least 75% of the member bodies casting a vote.

Attention is drawn to the fact that some of the elements of this International
Standard may be the subject of patent rights. ISO shall not be held responsible for
identifying any or all such patent rights.

International Standard ISO 3664 was prepared jointly by Technical Committees

ISO/TC 42, Photography and ISO/TC 130, Graphic technology, with input from
ISO/TC 6, Paper, board and pulps.

This second edition cancels and replaces the first edition (ISO3664:1975) which has
been technically revised. This revision of the 1974 version of the International
Standard meets the current needs of the Graphic Technology and Photographic
industries and minimizes differences between viewing equipment. It should be
noted that this revision contains multiple specifications, each of which is
appropriate to specific requirements. Users should ensure that they employ the
specification which is appropriate to their application.

Annexes A to C of this International Standard are for information only.

© ISO 2000 - All rights reserved iii

ISO 3664:2000(E)

Introduction
While colour and density measurements play important roles in the control of colour
reproduction, they cannot replace the human observer for final assessment of the
quality of complex images. Colour reflection artwork, photographic transparencies,
photographic prints, and photomechanical reproductions such as on-press and off-
press proofs, or press sheets, are commonly evaluated for their image and colour
quality, or compared critically with one another for fidelity of colour matching.
Paper and other substrates contribute to the colour appearance and controlling the
colour of these is equally critical. However, it should be noted that the paper
industry has its own set of International Standards for unprinted paper which differ
in illumination conditions from those recommended in this International Standard.

There is no doubt that the best viewing condition for the visual assessment of colour
is that in which the product will be finally seen. Where this is known, and it is
practical to do so, the various people in the production chain may sensibly agree to
use this viewing condition for all evaluation and comparison. However, it is
important that this be properly agreed upon in advance and that it be specified that
such a viewing condition is NOT ISO-defined.

Unfortunately, such agreement is often not practical. Even if a particular end-use

condition is known, it may be impractical to provide everybody in the production
chain with sufficiently consistent viewing apparatus. Since deficiencies in light
sources and viewing conditions, and inconsistencies between colour viewing
facilities, can distort the colour appearance of substrates, reproductions and artwork,
they are likely to cause miscommunication about colour reproduction and
processing. This International Standard provides specifications for illumination and
viewing conditions that, when properly implemented, will reduce errors and
misunderstandings caused by such deficiencies and inconsistencies.

The illumination used to view colour photographic prints, photomechanical

reproductions, and transparencies needs to provide adequate amounts of radiant
power from all parts of the ultraviolet and visible spectrum to avoid distorting their
appearance from that observed under commonly used sources of illumination such
as daylight. The ultraviolet content is important where fluorescent samples, which
are excited in this region, are encountered; a phenomenon associated with many of
the paper substrates on which images are reproduced as well as with some of the
dyes and pigments themselves.

To ensure consistency with the 1974 International Standard, as well as the majority
of equipment in current use, the reference spectral power distribution specified in
this International Standard is CIE Illuminant D50. Many of the reasons for the
selection of illuminant D50 in 1974, as opposed to any other CIE daylight illuminant,
are equally applicable today. Much consideration was given to changing the
reference illuminant to be CIE F8, a 5 000 Kelvin illuminant more typical of
fluorescent lamps. However, it was felt that this would provide only a minimal
conformance advantage (as shown in informative annex B) and the actual goal is for

iv © ISO 2000 - All rights reserved

 ISO 3664:2000(E)

the illumination to simulate natural daylight.

Because it is very difficult to produce artificial sources of illumination which

closely match the spectral power distribution of daylight, it is important that the
tolerances specified within this International Standard provide a compromise
between that required for lamp manufacturing purposes and that for consistent
viewing. In this International Standard three constraints which define the colour of
the light falling on the viewing plane apply, one directly and two indirectly, and all
three must be met simultaneously if a viewing apparatus is to be in compliance.

The chromaticity, which directly defines the colour of the illumination at the
viewing surface, is specified as that for illuminant D50 and the tolerance by a circle
in the CIE 1976 Uniform Chromaticity Scale (UCS) diagram having a specified
radius around that value. To establish the compliance of the spectral power
distribution of the illumination to that of illuminant D50 the methods defined in CIE
Publications No. 13.3 and No. 51 are both specified. One defines the colour
rendering quality of a lamp; the other its ability to correctly predict metamers. Both
requirements are important to the graphic technology and photographic industries.

Because CIE Publication No. 51 does not currently address illuminant D50,
additional virtual metamers for this illuminant, for both visible and ultraviolet
evaluation, were calculated and are defined in this International Standard. They
were derived from those published in CIE Publication No. 51 and are equivalent to
them. Also, based on experimental work described in annex B, a practical tolerance
of acceptability has been defined, alongside a Colour Rendering Index requirement.

The perceived tonal scale and colours of a print or transparency can be significantly
influenced by the chromaticity and luminance of other objects and surfaces in the
field of view. For this reason, ambient conditions, which may affect the state of
visual adaptation, need to be designed to avoid any significant effects on the
perception of colour and tone and immediate surround conditions need to be
specified also. Such specifications are provided in this International Standard.

Experience in the industries covered by this International Standard has revealed the
need for two levels of illumination; a high level for critical evaluation and
comparison, and a lower level for appraising the tone scale of an individual image
under illumination levels similar to those under which it will be finally viewed.
This International Standard provides these two levels of illumination.

The higher level is essential to graphic technology where comparison is being made;
such as between original artwork and proof, or to evaluate small colour differences
between proof and press sheet in order to control a printing operation. It is effective
in these situations because it enhances the visibility of any differences. The high
level of illumination is also appropriate in photography when comparing two, or
more, transparencies or when critically evaluating a single image to assess the
darkest tones that can be printed.

Since, despite adaptation, the level of illumination has quite a significant effect on
the appearance of an image, the lower level is required in order to appraise the
image at a level more similar to that in which it will be finally viewed. Although it
is recognized that quite a wide range of illumination levels may be encountered in
practical viewing situations, the lower level chosen is considered to be fairly
representative of the range encountered. For this reason it is applicable to aesthetic
appraisal, including the conditions for routine inspection of prints.

The viewing of transparencies is specified both for direct viewing and by projection.
Additional conditions are also specified for those conditions where transparencies
are to be compared to a print. The particular surround specified for transparencies

© ISO 2000 - All rights reserved v

ISO 3664:2000(E)

recognises the way that a transparency should be viewed for optimum visibility of
the dark tones, but acknowledges that practical viewing equipment is likely to have
ambient conditions that introduce some viewing flare. The combination of surround
and flare produce an appearance that is fairly representative of how the transparency
will look in a typically lighted room.

Small transparencies are commonly evaluated in graphic technology by direct

viewing. When it is necessary to view transparencies directly, they should be
viewed according to the conditions specified for that situation. However, for some
purposes, smaller transparencies are not viewed directly because the viewing
distance for correct perspective and perception of detail is too small for visual
comfort. Furthermore, when small transparencies are reproduced for publication or
other purposes, they are usually enlarged. To ease comparison, it is helpful to
enlarge the transparency image when comparing it to the print. For these reasons, a
viewing condition may be required which provides a magnified image when viewed
at an appropriate distance.

Colour monitors are increasingly being used to display and view digital images in
graphic technology and photography. In order to ensure consistency of assessment
in this situation it is important that the viewing conditions in which the monitors are
placed are reasonably well specified. However, it should be noted that adherence to
these specifications does not ensure that the monitor will match the hardcopy
without provision of a defined colour transformation to the displayed image, or use
of proper colour management. This aspect of matching is beyond the scope of this
International Standard. In practice, even with high quality colour management, an
accurate match is difficult to achieve because the luminance levels generally differ
significantly between hardcopy (print or transparency) and softcopy (monitor).

Thus, it should be noted that the specifications for images viewed on colour
monitors, provided in this International Standard, are for images viewed
independently of any form of hardcopy; conditions for direct comparisons between
hardcopy and softcopy (even where a suitable colour transformation has been
applied) are beyond the scope of this International Standard which can be seen as
being primarily relevant where successive viewing of hardcopy and softcopy takes
place. ISO 12646, Graphic Technology - Colour proofing using a colour display,
currently at Working Draft level in TC 130, is being prepared to provide more
detailed recommendations where direct comparison is required. In general it may
be stated that for such comparisons it is desirable to view the colour monitor under
the lower levels of ambient illumination specified in this International Standard and
with the maximum level of luminance achievable, and the hardcopy sample at the
lower levels of illumination specified for printed matter in this International
Standard (and their equivalent for transparencies). However, it should be noted that
this will, in turn, affect the perceived tone and colourfulness of the hardcopy.

vi © ISO 2000 - All rights reserved

 ISO 3664:2000(E)

Viewing conditions - Graphic technology and photography

1 Scope ISO 5-2:1991, Photography — Density measurements —

Part 2: Geometric conditions for transmission density.
This International Standard specifies viewing conditions
for images on both reflective and transmissive media, such ISO 5-3:1995, Photography — Density measurements —
as prints (both photographic and photomechanical) and Part 3: Spectral conditions.
transparencies, as well as images displayed in isolation on
colour monitors. Specifically, it shall be used for: ISO 5-4:1995, Photography — Density measurements —
Part 4: Geometric conditions for reflection density.
— critical comparison between transparencies, reflection
photographic or photomechanical prints and/or other ISO 12646:____1), Graphic technology — Displays for
objects or images, colour proofing — Characteristics and viewing conditions.

— appraisal of the tone reproduction and colourfulness of CIE Publication No. 13.3, 1995, Method of measuring and
prints and transparencies at illumination levels similar specifying the colour rendering properties of light sources,
to those for practical use, including routine inspection, 2nd edition.

— critical appraisal of transparencies which are viewed CIE Publication No. 15.2, 1986, Colorimetry.
by projection, for comparison with prints, objects, or
other reproductions, and CIE Publication No. 51, 1981, A method for assessing the
quality of daylight simulators for colorimetry.
— appraisal of images on colour monitors which are not
viewed in comparison to any form of hardcopy. CIE Publication No. 17.4, 1987, International lighting
vocabulary.
This International Standard is not applicable to unprinted
papers.
3 Terms and definitions

2 Normative references For the purposes of this International Standard, the

following terms and definitions apply.
The following normative documents contain provisions
which, through reference in this text, constitute provisions 3.1
of this International Standard. For dated references, chromaticity
subsequent amendments to, or revisions of, any of these property of a colour stimulus defined by its chromaticity
publications do not apply. However, parties to agreements co-ordinates, or by its dominant or complementary
based on the International Standard are encouraged to wavelength and purity taken together
investigate the possibility of applying the most recent [CIE Publication No. 17.4:1987, 845-03-34]
editions of the normative documents indicated below. For
undated references, the latest editions of the normative 3.2
document referred to applies. Members of IEC and ISO colour rendering index
maintain registers of currently valid International measure of the degree to which the psychophysical colour
Standards. of an object illuminated by a test illuminant conforms to

1) To be published.

© ISO 2000 - All rights reserved 1

ISO 3664:2000(E)

that of the same object illuminated by the reference

d
v
illuminant, suitable allowance having been made for the Lv

state of chromatic adaptation. dA × cos  × d
[CIE Publication No. 17.4:1987, 845-02-61]
where d
v is the luminous flux transmitted by an
3.3 elementary beam passing through the given point and
correlated colour temperature propagating in the solid angle d containing the given
temperature of the Planckian radiator whose perceived direction; dA is the area of a section of that beam
colour most closely resembles that of a given stimulus at containing the given point;  is the angle between the
the same brightness and under specified viewing conditions normal to that section and the direction of the beam
[CIE Publication No. 17.4:1987, 845-03-50] [CIE Publication No. 17.4:1987, 845-01-35]

3.4 3.10
flare off-press proof print
light falling on an image, in an imaging system, which does print produced by a method other than press printing whose
not emanate from the subject point purpose is to show the results of the colour separation
SEE image flare, veiling flare, and veiling glare. process in a way that closely simulates the results on a
production press
NOTE Veiling glare is also sometimes referred to as flare
3.11
3.5 on-press proof print
hardcopy print produced by press printing (production or proof press)
representation of an image on a substrate which is self whose purpose is to show the results of the colour
sustaining and reasonably permanent separation process in a way that closely simulates the
SEE softcopy, print, and transparency results on a production press

NOTE Examples include prints and transparencies. 3.12

original
3.6 illuminance the scene or hardcopy from which image information is
<at a point of a surface> quotient of the luminous flux obtained, using an image capture device, in a reproduction
incident on an element of the surface containing the point process
by the area of that element
[CIE Publication No. 17.4:1987, 845-01-38] NOTE As used in graphic technology, the original is typically a
print or transparency, and the capture device is usually an input
3.7 scanner or, occasionally, a process camera. In photography the
illuminant term original scene is sometimes used.
radiation with a relative spectral power distribution defined
over the wavelength range that influences object-colour 3.13
perception print
[CIE Publication No. 17.4:1987, 845-03-10] two-dimensional hardcopy form of an image intended for
viewing
3.8 SEE hardcopy, softcopy, transparency
image flare
light from a subject point that is scattered by the optical NOTE In still photography and graphic technology, the term print
is reserved for reflection hardcopy; a medium designed to be
system to areas of the image plane other than the
viewed by reflected light.
appropriate image point

NOTE The distribution of image-flare light resulting from any 3.14

subject point is specified by the image point spread function. relative spectral power distribution
Point spread functions tend to fall off rapidly as the distance from ratio of the spectral power distribution of a source or
the image point is increased, are variable for different image-point illuminant to a fixed reference value which can be an
locations and are typically not radially symmetric for image points average value, a maximum value, or an arbitrarily chosen
some distance from the optical system axis. value of this distribution.

3.9 3.15
luminance (in a given direction, at a given point of a softcopy
real or imaginary surface) representation of an image produced using a device capable
quantity defined by the formula: of directly representing different digital images in
succession and in a non-permanent form

2 © ISO 2000 - All rights reserved

 ISO 3664:2000(E)

EXAMPLE the most common example is a monitor samples, which provide metameric matches for specific
SEE hardcopy standard daylight illuminants.

3.16 NOTE Virtual metamers are used to test and classify illumination
source sources which simulate daylight according to the method provided
primary emitter of electromagnetic radiation in CIE Publication No. 51. This classification is accomplished by
calculating the average of the colour differences obtained for
these metamers between the illumination source in question and a
3.17
CIE standard illuminant. Although it may be possible to construct
surround physical realizations of some virtual metamers, the fact that they
area adjacent to the border of an image which, upon may not be real allows greater flexibility in their design.
viewing the image, may affect the local state of adaptation
of the eye
4 Viewing condition requirements
NOTE The surround, which can have a significant effect on the
perceived tone and colour reproduction of an image, should not be
confused with any border immediately surrounding the image, 4.1 General requirements
such as any unprinted white substrate for reflection copy or the
unexposed border present on many transparencies. For a colour In this clause, the requirements that apply to all of the
monitor, the border will normally be dark grey or black, and specified viewing conditions are stated. The requirements
hence the same as the surround. However, when simulating specific to each of these viewing conditions are defined in
hardcopy it will be similar to that hardcopy, both in terms of 4.2 (Critical comparison), 4.3 (Practical appraisal of prints)
lightness and width. and 4.4 (Projection viewing of small transparencies).

3.18 NOTE For ease of reference, each viewing condition described in

transparency the International Standard has been given an alpha-numeric
two-dimensional hardcopy form of an image designed to be designation. This may be useful in describing or specifying
viewed by transmitted light conditions: e.g. ISO viewing condition P2 as specified in this
SEE hardcopy, softcopy, print International Standard.

3.19 4.1.1 Viewing apparatus

transparency illuminator
apparatus used for back illumination of a transparency To comply with this International Standard, the values
specified shall be achieved at the surface of viewing. The
3.20 specified relative spectral power distribution applies to the
veiling flare illuminated surface rather than to the source (or lamp)
relatively uniform but unwanted irradiation in the image because the light from the source may be modified by
plane of an optical system, caused by the scattering and reflecting and transmitting components of the apparatus,
reflection of a proportion of the radiation which enters the and the required relative spectral power distribution may be
system through its normal entrance aperture where the obtained from a mixture of light from different sources.
radiation may be from inside or outside the field of view of
the system The source, image being viewed, and observer's eyes shall
be positioned to minimize the amount of light specularly
NOTE Light leaks in an optical system housing can cause reflected toward the eyes of an observer on or near the
additional unwanted irradiation of the image plane. This normal to the centre of the viewing surface.
irradiation may resemble veiling flare.
The surround of a print or transparency shall have a
3.21 diffusing surface and shall have a CIELAB chroma value
veiling glare: Light falling on a radiant image surface, such no greater than 2; i.e. shall appear neutral.
as a back illuminated transparency or monitor, which adds
to the luminance of the image. 4.1.2 Spectral conditions for the reference illuminant

NOTE Veiling glare lightens and reduces the apparent contrast of The relative spectral power distribution of the reference
the darker parts of an image. It differs from veiling flare in that it illuminant for both prints and transparencies shall be CIE
is used exclusively for the perception of images in which no illuminant D50 as defined in CIE 15.2 (see Table 1). This
entrance aperture is defined.
represents a phase of natural daylight having a correlated
colour temperature of approximately 5 000 K. The
3.22
chromaticity coordinates of illuminant D50 are x10 = 0,347 8
virtual metamer
and y10 = 0,359 5 in the CIE chromaticity diagram
set of spectral radiance factors, not based on physical

© ISO 2000 - All rights reserved 3

ISO 3664:2000(E)

Table 1 — Relative spectral power of reference illuminant D50.

Wavelength Relative power for Wavelength Relative power for

nm illuminant D50 nm illuminant D50
300 0,02 550 102,32
305 1,03 555 101,16
310 2,05 560 100,00
315 4,91 565 98,87
320 7,78 570 97,74
325 11,26 575 98,33
330 14,75 580 98,92
335 16,35 585 96,21
340 17,95 590 93,50
345 19,48 595 95,59
350 21,01 600 97,69
355 22,48 605 98,48
360 23,94 610 99,27
365 25,45 615 99,16
370 26,96 620 99,04
375 25,72 625 97,38
380 24,49 630 95,72
385 27,18 635 97,29
390 29,87 640 98,86
395 39,59 645 97,26
400 49,31 650 95,67
405 52,91 655 96,93
410 56,51 660 98,19
415 58,27 665 100,60
420 60,03 670 103,00
425 58,93 675 101,07
430 57,82 680 99,13
435 66,32 685 93,26
440 74,82 690 87,38
445 81,04 695 89,49
450 87,25 700 91,60
455 88,93 705 92,25
460 90,61 710 92,89
465 90,99 715 84,87
470 91,37 720 76,85
475 93,24 725 81,68
480 95,11 730 86,51
485 93,54 735 89,55
490 91,96 740 92,58
495 93,84 745 85,40
500 95,72 750 78,23
505 96,17 755 67,96
510 96,61 760 57,69
515 96,87 765 70,31
520 97,13 770 82,92
525 99,61 775 80,60
530 102,10 780 78,27
535 101,43
540 100,75
545 101,54
NOTE The wavelength specification has been extended beyond the normal visual range because of the need to consider brighteners or
dyes which may fluoresce.

4 © ISO 2000 - All rights reserved

 ISO 3664:2000(E)

and u'10 = 0,210 2 and v'10 = 0,488 9 in the CIE 1976 mechanism for indicating degradation.
Uniform Chromaticity Scale (UCS) diagram.
However, it is the responsibility of the user, both before
NOTE Chromaticity is specified for the CIE 1964 standard and beyond this time limit, to undertake measurements as
colorimetric observer to ensure compatibility with the method specified in clause 5 to ensure compliance, unless it can be
specified in CIE Publication No. 51 which is used to define the otherwise demonstrated that the equipment remains within
degree of compliance of the illumination to the reference tolerance.
illuminant in 4.2.2.

4.2 Conditions for critical comparison (ISO

4.1.3 Colour rendering index
viewing conditions P1 and T1)
The CIE general colour rendering index of the viewing
surface shall be measured as specified in CIE Publication 4.2.1 Applicability
No. 13.3 and shall have a value of 90 or higher. In
addition, the separate special colour rendering indices for This subclause specifies viewing conditions for the critical
samples 1 to 8 as specified in CIE Publication No. 13.3 comparison between two (or more) copies of an image.
shall each have a value of 80 or higher. The comparison is usually either between the original and
its reproduction or between different copies of a
4.1.4 Ambient conditions reproduction, such as samples from a press run or multiple
photographic prints. The images being compared may be
The visual environment shall be designed to minimize on the same media (reflective or transmissive), or on
interference with the viewing task. It is important to different reflective media (including photographic or
eliminate extraneous conditions that affect the appraisal of photomechanical prints and press proofs or off-press
prints or transparencies and an observer should avoid proofs), or even between transmissive and reflective media
making judgements immediately after entering a new such as that pertaining when a transparency is compared to
illumination environment because it takes a few minutes to a proof of its printed reproduction. The high illumination
visually adapt to that new environment. levels specified permit more critical evaluation of colour
and tone gradation in higher density areas, which may not
Extraneous light, whether from sources or reflected by be perceived under most practical viewing conditions.
objects and surfaces, shall be baffled from view and from
illuminating the print, transparency, or other image being The condition for viewing a print is specified as condition
evaluated. In addition, no strongly coloured surfaces P1; that for viewing a transparency directly on an
(including clothing) should be present in the immediate illuminator having a diffusing screen (compared to viewing
environment. by projection) is specified as condition T1. The latter will
normally be the case for transparencies larger than 10 cm
NOTE The presence of strongly coloured objects within the by 10 cm and in graphic technology is generally the case
viewing environment is a potential problem because they may for smaller transparencies also.
cause reflections which cannot easily be baffled and may
influence viewer adaptation. NOTE In the graphic arts industry the primary viewing
application involves comparison, which requires that level P1 be
Walls, ceiling, floors, and other surfaces which are in the used. However, when it is important that tone reproduction that
will be perceived under lower levels of illumination is assessed, it
field of view shall be baffled or coloured a neutral matt
is recommended that P1 be supplemented by level P2, or the
grey, with a reflectance of 60 % or less. It should be noted
expected actual viewing condition. It should be noted that the
that it may be easier to minimize these problems by using a same correlated colour (D50) is specified for both P1 and P2.
viewing booth, rather than designing an open area for
viewing within a room. Such apparatus can also make it 4.2.2 Illumination
easier to meet the specification for surround conditions
specified in 4.2.4 and avoid the excessive flare which may The illumination at the plane of viewing shall approximate
otherwise cause problems on transparency illuminators. that of CIE standard illuminant D50. It shall have u'10, v'10
However, even with such apparatus, adaptation and chromaticity coordinates within the radius of 0,00 5 from
avoidance of extraneous light still need to be carefully that specified in 4.1.2 and a colour rendering index as
considered. specified in 4.1.3. When assessed using the method
defined in CIE Publication No. 51, but using the virtual
4.1.5 Maintenance metamers defined for the visible range in Table 2, it shall
fall within category C and should fall within category B.
Manufacturers of viewing apparatus shall specify the For condition P1, when assessed using the method defined
average number of hours during which the apparatus is in CIE Publication No. 51, but using the virtual metamers
expected to remain within specification. The apparatus defined for the ultraviolet range in Table 3, it shall have a
should include a time-metering device or some other metamerism index (MIuv) of less than 4. (See informative

© ISO 2000 - All rights reserved 5

ISO 3664:2000(E)

Table 2 — Five sets of spectral reflection radiance factor data providing virtual metamers with the standard
data provided in CIE Publication No. 51. To be used for visible range evaluation for illuminant D50.

Wave Virtual metamers, visual range Wave Virtual metamers, visual range
length length
nm Set 1 Set 2 Set 3 Set 4 Set 5 nm Set 1 Set 2 Set 3 Set 4 Set 5
400 0,029 0,044 0,029 0,403 0,175 600 0,427 0,072 0,102 0,238 0,200
405 0,028 0,056 0,028 0,403 0,177 605 0,473 0,076 0,103 0,240 0,228
410 0,027 0,063 0,028 0,403 0,179 610 0,515 0,083 0,104 0,241 0,258
415 0,026 0,074 0,027 0,403 0,182 615 0,552 0,085 0,104 0,240 0,286
420 0,024 0,081 0,027 0,402 0,184 620 0,582 0,087 0,104 0,237 0,316
425 0,024 0,088 0,026 0,401 0,187 625 0,608 0,087 0,103 0,234 0,342
430 0,024 0,089 0,026 0,398 0,187 630 0,630 0,086 0,103 0,229 0,366
435 0,025 0,088 0,024 0,393 0,186 635 0,646 0,085 0,104 0,228 0,387
440 0,025 0,083 0,025 0,387 0,181 640 0,659 0,084 0,104 0,228 0,405
445 0,026 0,081 0,026 0,375 0,178 645 0,671 0,084 0,106 0,236 0,422

450 0,027 0,076 0,027 0,372 0,174 650 0,683 0,085 0,109 0,245 0,437
455 0,028 0,071 0,029 0,366 0,170 655 0,695 0,086 0,114 0,264 0,451
460 0,031 0,066 0,031 0,360 0,165 660 0,708 0,088 0,120 0,287 0,466
465 0,035 0,059 0,034 0,353 0,160 665 0,722 0,088 0,129 0,320 0,482
470 0,043 0,052 0,037 0,345 0,156 670 0,736 0,088 0,140 0,358 0,502
475 0,054 0,048 0,045 0,336 0,151 675 0,751 0,087 0,154 0,403 0,522
480 0,068 0,045 0,056 0,327 0,148 680 0,766 0,086 0,170 0,449 0,543
485 0,085 0,042 0,067 0,319 0,143 685 0,781 0,086 0,188 0,502 0,564
490 0,103 0,039 0,077 0,311 0,141 690 0,794 0,086 0,206 0,552 0,584
495 0,121 0,037 0,086 0,304 0,139 695 0,806 0,087 0,227 0,600 0,603

500 0,136 0,034 0,092 0,296 0,137 700 0,817 0,088 0,250 0,646 0,621
505 0,148 0,035 0,095 0,289 0,135
510 0,156 0,033 0,097 0,281 0,135
515 0,159 0,032 0,095 0,276 0,132
520 0,160 0,032 0,092 0,271 0,129
525 0,162 0,032 0,090 0,265 0,125
530 0,164 0,032 0,089 0,260 0,122
535 0,167 0,032 0,088 0,255 0,121
540 0,172 0,033 0,086 0,251 0,121
545 0,177 0,033 0,084 0,248 0,121

550 0,182 0,033 0,084 0,246 0,121

555 0,189 0,032 0,086 0,245 0,119
560 0,196 0,030 0,087 0,244 0,116
565 0,209 0,032 0,088 0,243 0,110
570 0,226 0,036 0,091 0,241 0,108
575 0,248 0,041 0,094 0,239 0,113
580 0,275 0,045 0,096 0,236 0,119
585 0,309 0,049 0,097 0,234 0,131
590 0,345 0,055 0,097 0,234 0,149
595 0,384 0,063 0,100 0,235 0,174
NOTE This data complements that in CIE Publication 51. The standard data, to which the above provides metamers, may be
obtained from that document. The above data are NOT available in CIE Publication 51.

6 © ISO 2000 - All rights reserved

 ISO 3664:2000(E)

Table 3 — Spectral characteristic of non-fluorescent samples providing virtual metamers with the
fluorescent samples provided in CIE publication 51 for illuminant D50.

Wave Virtual metamers, Wave Virtual metamers,

length ultraviolet range length ultraviolet range
nm 1 2 3 nm 1 2 3
400 0,662 0,505 0,212 600 0,838 0,838 0,838
405 0,687 0,589 0,293 605 0,839 0,839 0,839
410 0,711 0,668 0,401 610 0,840 0,840 0,840
415 0,742 0,723 0,507 615 0,842 0,842 0,842
420 0,767 0,764 0,613 620 0,844 0,844 0,844
425 0,797 0,805 0,737 625 0,846 0,846 0,846
430 0,822 0,838 0,842 630 0,848 0,848 0,848
435 0,824 0,845 0,868 635 0,850 0,850 0,850
440 0,820 0,843 0,866 640 0,851 0,852 0,852
445 0,816 0,836 0,857 645 0,851 0,854 0,854
450 0,810 0,826 0,845 650 0,852 0,856 0,856
455 0,808 0,822 0,845 655 0,852 0,856 0,857
460 0,807 0,820 0,843 660 0,852 0,856 0,857
465 0,807 0,816 0,837 665 0,853 0,856 0,857
470 0,804 0,813 0,830 670 0,853 0,856 0,858
475 0,806 0,813 0,828 675 0,853 0,858 0,859
480 0,810 0,816 0,827 680 0,853 0,858 0,860
485 0,812 0,817 0,826 685 0,853 0,858 0,861
490 0,814 0,818 0,826 690 0,853 0,858 0,862
495 0,816 0,819 0,825 695 0,853 0,858 0,863
500 0,818 0,821 0,825 700 0,853 0,859 0,864
505 0,822 0,825 0,828
510 0,826 0,829 0,831
515 0,830 0,831 0,834
520 0,831 0,833 0,836
525 0,832 0,833 0,836
530 0,832 0,833 0,835
535 0,832 0,833 0,834
540 0,833 0,834 0,835
545 0,833 0,834 0,835
550 0,834 0,834 0,834
555 0,835 0,834 0,834
560 0,835 0,834 0,835
565 0,835 0,834 0,835
570 0,836 0,835 0,836
575 0,836 0,835 0,835
580 0,836 0,836 0,836
585 0,837 0,837 0,837
590 0,837 0,837 0,837
595 0,837 0,837 0,837
NOTE This data complements that in CIE Publication 51. The standard data, to which the
above provides metamers, may be obtained from that document. The above data are NOT
available in CIE Publication 51.

© ISO 2000 - All rights reserved 7

ISO 3664:2000(E)

annex B for further explanations of these tolerances). 4.2.5 Luminance at the surface of the transparency
illuminator (T1)
NOTE No specification is provided for the ultraviolet emission of
the illumination for condition T1. In practice fluorescence is not The luminance at the centre of the illuminated surface of
an issue for photographic transparencies and the diffusing surface the transparency illuminator shall be 1 270 cd/m2 ± 320
of the illuminator normally absorbs the majority of any ultraviolet cd/m2 and should be 1 270 cd/m2 ± 160 cd/m2. Any
emission from the source.
departures from uniformity shall be gradually diminishing
from centre to edge such that the luminance (measured
The categories to which the equipment conforms at the
normal to the surface) at any point within the luminous area
time of manufacture shall be displayed on the equipment.
is not less than 75 % of the luminance measured at the
Where the ultraviolet metameric index is greater than 2, the
centre of the image plane.
manufacturer shall specify whether the contribution to
visible energy from ultraviolet excitation is greater than or
4.2.6 Transparency illuminator diffusion
less than the contribution of illuminant D50.
characteristics (T1)
4.2.3 Illuminance (P1)
The transparency illuminator surface shall provide diffuse
light such that the luminance of the surface measured at any
The illuminance shall be 2 000 lx ± 500 lx, and should be
angle between 0° and 45° from the normal shall not be less
2 000 lx ± 250 lx, at the centre of the illuminated viewing
than 90 % of the luminance at the same point measured
surface area. Any departures from complete uniformity
normal to the surface.
shall be gradually diminishing from centre to edge. For a
viewing area up to 1 metre square, the illuminance at any
4.2.7 Transparency surround (T1)
point within the illuminated area shall not be less than 75 %
of the illuminance measured at the centre of the illuminated
The surround shall be at least 50 mm wide on all sides. It
viewing surface area. For larger viewing areas, the lower
shall appear neutral compared to the source and shall have
limit shall be 60 %.
a luminance that is between 5 % and 10 % of that of the
surface of the image plane of the illuminator in the
4.2.4 Surround and backing for reflection viewing (P1)
direction of observation. A transparency mounted with an
opaque border may be viewed without removing the mount.
The surround and backing shall be neutral and matt. The
surround shall have a luminous reflectance between 10 % NOTE This condition is similar to that specified for direct
and 60 % with the specific value being selected to be viewing of transparencies in the previous version of this
consistent with practical viewing. For many applications, a International Standard. However, that version specified an
mid-grey of 20 % reflectance is very convenient and is "illuminated" surround when transparencies were compared to
recommended where no other condition is defined. prints. The purpose of this surround was to effect a reduction in
However, whatever value is selected, it is important transparency contrast to facilitate comparison to prints.
when images are being compared that the surrounds for Unfortunately, this method of contrast reduction significantly
each are similar; and as a result the ratio of the surround reduces tonal differentiation in the dark tones of the image. With
luminance shall be 1,0 (± 0,2):1. modern imaging systems, contrast reduction can be achieved
through a variety of means that maintain shadow contrast. The
illuminated surround approach could therefore result in a
NOTE 1 A wide range of surround reflectances is allowed in this
misleading interpretation of transparency shadow detail,
International Standard so that reflection hardcopy images can be
particularly for low-key subjects. The dark surround has therefore
evaluated in conditions which are similar to those used in practice.
been incorporated for all assessment conditions in this
However, extremely light or dark surrounds are not allowed
International Standard. In practice this condition may be met by
because of their large effect on appearance. Where no practical
using an opaque black mask; such a mask will appear to have a
condition can be specified, a mid-grey of 20 % reflectance should
luminance somewhat above absolute black because of viewing
be used.
flare and ambient illumination falling on the mask.

The surround shall extend beyond the materials being

4.2.8 Relationship between transparency luminance
viewed on all sides by at least 1/3 of their dimension.
and print illuminance (P1 and T1)
Where objects are being compared, they may be positioned
edge to edge. The backing should have a luminous
For critical comparison between transparencies and
reflectance of 2 % to 4 % to be consistent with the
reflecting materials, the illuminance at the reflecting
definition in ISO 5-4.
material surface shall be that specified in 4.2.3 [i.e. 2 000 ±
500 lx]. The transparency illuminator shall have a
NOTE 2 The above requirement may be met by appropriate
finishing of the viewing surface or by introduction of masking
luminance as specified in 4.2.5 [i.e. 1 270 ± 320 cd/m2].
devices. However, the combined tolerances must be such that the
ratio of the maximum luminance of the transparency
illuminator to the maximum luminance of a perfectly

8 © ISO 2000 - All rights reserved

 ISO 3664:2000(E)

reflecting and diffusing material, at the plane of the correspond to an office, library, or a relatively brightly
reflecting material, shall be 2 (± 0,2):1. The maximum illuminated area in a residence. By appraising images
luminance by reflection from the perfectly reflecting and under such conditions it is possible to ensure that they
diffusing material is equal to the incident illuminance provide a satisfactory tone reproduction; such a judgement
divided by
. cannot be made unambiguously at the higher level of
illumination specified for condition P1.
4.3 Conditions for practical appraisal of prints
(including routine inspection). (ISO viewing condition NOTE In the graphic arts industry the primary viewing
P2) application involves comparison, which requires that level P1 be
used. However, when it is important that tone reproduction that
4.3.1 Applicability will be perceived under lower levels of illumination is assessed, it
is recommended that P1 be supplemented by level P2, or the
expected actual viewing condition. It should be noted that both P1
The specifications in this subclause are applicable for the
and P2 have the same correlated colour temperature of D50.
appraisal of tone reproduction of individual images,
photographic image inspection or the judgement of prints.
4.3.2 Illumination
They are not appropriate for the simultaneous comparison
of media, where colour matching is the primary concern,
The illumination at the surface of viewing shall comply
such as any comparison between proof and photo-
with that described in 4.2.2.
mechanical print, transparency and proof (or print), or
between different photographic prints and transparencies.
4.3.3 Illuminance
The only exception is when comparing a print to a colour
monitor, because of the low luminance level exhibited by
The illuminance at the centre of the viewing surface shall
current monitors, but such comparisons are outside the
be 500 lx ± 125 lx. The illumination uniformity shall
scope of this International Standard which only deals with
comply with that described in 4.2.3.
appraisal of images on a monitor in isolation from
hardcopy (see 4.5).
4.3.4 Surround and backing
It should be noted that the relative spectral power
The surround and backing shall comply with that described
distribution characteristics specified for P2 are exactly the
in 4.2.4 except that a white backing may be used where
same as those specified for condition P1. Therefore,
appropriate, e.g. for display purposes. The surround and
images that match under the conditions of P1 will match
backing shall be assumed to be in compliance with 4.2.4
under the conditions of P2. However, the reverse is not
unless communicated otherwise to others involved in the
necessarily true, particularly if there are significant dark
production chain, in which case the reflectance of the
tonal areas involved.
surround and backing shall be specified.
Experience has shown that the high levels of illumination
4.4 Conditions for viewing small transparencies
specified for ISO viewing condition P1 can give a
misleading impression of the tone reproduction and by projection (ISO viewing condition T2)
colourfulness of an image which will ultimately be viewed
by the consumer in much lower levels of illumination. 4.4.1 Applicability
Images that appear quite acceptable when viewed at the
higher levels of illumination may not appear satisfactory The specifications for the equipment used for viewing a
when viewed at more typical levels of illumination. To projected image of a slide on a screen are given in 4.2 to
avoid this problem the illumination level for inspection of 4.8. These conditions are not to be confused with those
photographic prints is often set arbitrarily while many normally used for viewing slides in a commercial projector
graphic technology users take proofs into lower where the magnification is generally much greater and
illumination levels, of unknown conditions, to verify that there is no intent to compare such images with reflection
their tone reproduction will prove acceptable in use. prints.
Because neither the level or characteristics of the
illumination in these situations are controlled, this practice 4.4.2 Illumination
introduces uncertainties into the process and prevents
effective communication. The light emitted from the screen with an empty slide
mount in the gate shall comply with that described in 4.2.2.
The viewing conditions specified in this subclause are
intended to minimize those problems; the viewing 4.4.3 Luminance
conditions specified are for the appraisal of tone
reproduction, for photographic image inspection or The luminance at the screen in the direction of the observer
judgement of prints, under illumination levels that shall be 1 270 cd/m2 ± 320 cd/m2 when measured with an

© ISO 2000 - All rights reserved 9

ISO 3664:2000(E)

empty slide mount in the projector. for images viewed independently of any form of hardcopy;
conditions for direct comparisons between hardcopy and
4.4.4 Uniformity of screen luminance. softcopy (even where a colour transformation designed to
provide a colour match has been applied) are beyond the
Any departure from uniformity of screen luminance shall scope of this International Standard. Thus, these
be approximately radially symmetrical about the centre of specifications can be seen as being primarily relevant
the screen, the luminance gradually diminishing from the where successive viewing of hardcopy and softcopy takes
centre to the edges of the projected image of the open slide place. ISO 12646, is being prepared to provide more
mount. When the screen is viewed at any angle up to 25 detailed recommendations where direct comparison is
degrees from the perpendicular to its surface, and at the required.
normal viewing distance for the equipment, the luminance
of any point within the image of the open slide mount shall 4.5.2 Chromaticity
be not less than 75 % of that at the centre. The screen on
which the image is displayed shall exhibit no more speckle, The chromaticity of the white displayed on the colour
scintillation, or graininess than that exhibited by an monitor should approximate that of D65. It shall have u'10,
untextured flat, matt, front-projection screen. v'10 chromaticity coordinates within the radius of 0,025 of
u'10 = 0,197 9 and v'10 = 0,469 5 in the CIE 1976 Uniform
4.4.5 Surround Chromaticity Scale (UCS) diagram.

The surround shall comply with that described in 4.2.7. NOTE When viewed under the conditions specified in 4.5, the
monitor itself will provide the primary adapting stimulus to the
4.4.6 Stray light and flare eye. The chromaticity of the white of the monitor is not too
important in this situation although many users prefer that the
chromaticity of that white be close to that of D65. There is some
Provisions shall be made for shielding the screen from
evidence that, at the low luminance levels obtained with monitors,
stray light. The surfaces of the stray light shield(s) facing
a chromaticity close to that of D65 provides a better evocation of
the screen shall be matt black. white and, furthermore, such a chromaticity permits a higher level
of luminance to be achieved with current display technology.
Stray light and flare shall be such that, when evaluated with However, if the monitor is to be directly compared with prints or
a test transparency conforming to 5.2, the luminance at the transparencies then the chromaticity of the white of the monitor
centre of the spot image on the screen shall not exceed 1% should be close to that of the hardcopy to which it is being
of the maximum screen luminance for any point in the compared. This means that a colour monitor used for such a
surrounding field. purpose should have a chromaticity close to illuminant D50. Such
a chromaticity is within the tolerance specified in this
International Standard. The specification for comparing colour
4.4.7 Resolution
monitor to hardcopy is described in greater detail in ISO 12646.
The resolving power of the optical system shall be such
4.5.3 Monitor luminance
that, when evaluated with a test transparency conforming to
5.3, all patterns having a spatial frequency up to 40 line
The luminance level of the white displayed on the monitor
pairs per millimetre shall be resolved at any point in the
shall be greater than 75 cd/m2 and should be greater than
projected image.
100 cd/m2.
4.4.8 Distortion
NOTE With current display technology, the level of luminance
that may be achieved depends upon the chromaticity of the white
The projection system shall not exhibit noticeable spatial point of the monitor. As the correllated colour temperature is
distortion nor cause noticeable chromatic distortion of the increased, the level of screen luminance that may be achieved
projected image. becomes higher. A level of at least 100 cd/m2 is recommended for
this application but it is accepted that this may be difficult to
achieve on some monitors, particularly for older or lower cost
4.5 Conditions for appraisal of images displayed models, or where the monitor white point is set to the
chromaticity of illuminant D50.
on colour monitors.
4.5.4 Ambient illumination
4.5.1 Applicability
When measured at the face of the monitor, with a cosine
In order to ensure consistency of assessment of images
corrected photometer and with the monitor switched off,
viewed on colour monitors, it is important that the viewing
the level of ambient illumination shall be less than, or equal
conditions in which the monitors are placed are reasonably
to, 64 lx and should be less than, or equal to, 32 lx. These
well specified. However, it should be noted that the
limits must also be achieved when measured in any plane
specifications provided in this International Standard are

10 © ISO 2000 - All rights reserved

 ISO 3664:2000(E)

between the monitor and the observer. The correlated 5 Test methods
colour temperature of the ambient illumination shall be less
than or equal to that of the monitor white point. 5.1 Spectral measurements
NOTE The level of ambient illumination needs to be significantly To determine the colour rendering index of the
lower than the luminance level of the monitor white point. This is
illumination, as defined in 4.1.3, and the chromaticity and
partly to ensure that the observer is reasonably adapted to the
metameric index as defined in 4.2.2, it is necessary to
monitor but primarily to ensure that the full contrast range of the
monitor is not significantly reduced by the effects of veiling glare. measure the spectral power distribution of the illumination.
It is for these reasons that the level of ambient illumination needs This requires measurement over the range 300 nm to 730
to be less than 64 lx and, preferably, much lower. This is nm for conditions P1 and P2 and over the range 380 nm to
particularly significant where low- luminance monitors are 730 nm for conditions T1 and T2. The bandpass of the
employed. If the level of ambient illumination approaches the measuring instrument (spectroradiometer) shall be 5 nm or
higher level specified in this subclause, the chromaticity of the narrower. The sampling interval shall not be greater than
illumination should be approximately the same as the white point the bandpass.
of the monitor, in order to minimize chromatic adaptation
complications.
As in any measurement process, the measuring equipment
must be regularly calibrated. In this case, the calibration of
4.5.5 Surround condition the spectroradiometer must include assessment of stray
light, linearity, wavelength accuracy and spectral power
The area immediately surrounding the displayed image and accuracy.
its border shall be neutral, preferably dark grey or black to
minimize flare, and of approximately the same chromaticity
5.2 Illuminance and luminance
as the white point of the monitor. The luminance of the
border should be 20 % of the white point luminance, or
All illuminance or luminance measurements shall be made
less, and preferably 3 % of the white point luminance, or
with a photometer having the spectral responsivity of the
less.
CIE standard photopic photometric observer, V(), and
measuring an area having a diameter no greater than 1/20
NOTE When the monitor is being used to visualise images which
will be reproduced as hardcopy the recommended lightness of any of the shortest linear dimension of the illuminated surface
border displayed around the image will depend upon the area. For illuminance measurements, the photometer shall
comparison. In general, for comparison to prints, which may well be cosine-corrected.
be reproduced with a white border consisting of unprinted
substrate, the border of the image should be light to simulate this For projection-viewing apparatus, stray light and flare shall
substrate; for comparison to transparencies it should be dark. be evaluated with the aid of a test transparency that is clear,
However, it is generally preferable that any such border be no neutral and transparent everywhere except for an opaque
more than 1 cm to 2 cm wide, even if it would normally be wider central circular spot producing an image having a diameter
on the hardcopy reproduction. 1/10 of the smallest linear dimension of the projected
opening in an open slide mount. Measurement shall be
4.5.6 Environmental conditions made perpendicular, within 7 degrees, to the spot and 35
cm from the screen.
The monitor shall be situated so there are no strongly
coloured areas (including clothing) directly in the field of NOTE The measurement should be made over an area not greater
view or which may cause reflections in the monitor screen. than 30 % of the opaque (black) spot diameter, approximately
Ideally all walls, floors and furniture in the field of view centred on the centre of the spot.
should be grey and free of any posters, notices, pictures,
wording or any other object which may affect the vision of All measurements shall be conducted in the presence of
the viewer. environmental illumination that would normally exist when
the apparatus is used.
4.5.7 Glare
5.3 Resolution assessment for projection viewing
All sources of glare should be avoided since they apparatus
significantly degrade the quality of the image. The monitor
shall be situated so that no illumination sources such as A test target shall be used, containing square wave
unshielded lamps or windows are directly in the field of resolution patterns varying in spatial frequency and with
view or are causing discernable reflections from the surface two mutually perpendicular orientations. The range of
of the monitor. frequencies included should be at least 20 to 60 line pairs
per millimetre and should include 40 line pairs per
millimetre. The dark bars of the test patterns shall have an
ISO visual diffuse transmission density, as defined in ISO

© ISO 2000 - All rights reserved 11

ISO 3664:2000(E)

5-2 and ISO 5-3, at least 2,0 higher than the density of the confidence. The concept of "reasonable confidence" is
transparent background. A pattern having a given spatial intended to indicate a level of confidence that is somewhere
frequency shall be considered resolved if the sets of lines of between complete confidence and no confidence at all. All
the projected image of that pattern, oriented both radially to resolution evaluations shall be made with the optical
the centre of the field and tangentially to a circle about the system set at the same focus. A low-power magnifying
centre, are clear enough to be counted with reasonable glass may be used.

12 © ISO 2000 - All rights reserved

 ISO 3664:2000(E)

Annex A
(informative)

Summary of ISO viewing conditions specified in this International Standard

Table A.1 is deemed to be informative since it is included for convenience and introduces no new specifications. It simply
summarizes the main normative requirements specified throughout this International Standard.

Table A.1 Summary of ISO viewing conditions

ISO viewing Reference Illuminance/ Colour rendering Metamerism Illumination Surround

condition illuminant and luminance index (according to index (according uniformity luminous
chromaticity CIE 13.3) to CIE 51) (min:max) reflectance/
tolerance a luminance/
illuminance
Critical comparison

Prints (P1) Illuminant D50 2 000 lx ± 500 lx General index:
90 Visual: C or For surfaces up to < 60 % (neutral
(0,005) (should be ± 250 Special indices for better (should be 1m x 1m
0.75 and matt)
lx) samples 1 to 8:
80 B or better) For surfaces
UV: < 4 greater than 1m x
1m
0.6

Transparencies Illuminant D50 1 270 cd/m2 General index:
90 Visual: C or
0,75 5 % - 10 % of

Direct viewing (0,005) ± 320 cd/m2 Special indices for better (should be the luminance
(T1) (should be samples 1 to 8:
80 B or better) level (neutral
± 160 cd/ m2)b and extend at
least 50mm on
all sides)
Practical appraisal Illuminant D50 500 lx ± 125 lx General index:
90 Visual: C or
0,75 < 60 % (neutral
of prints (P2) (0,005) Special indices for better (should be and matt)
samples 1 to 8:
80 B or better)
UV: < 4
Transparencies Illuminant D50 1 270 cd/ m2 General index:
90 Visual: C or
0,75 5 % - 10 % of
Projection viewing (0,005) ± 320 cd/ m2 Special indices for better (should be the luminance
(T2) samples 1 to 8:
80 B or better) level (neutral
and extend at
least 50 mm on
all sides)
Colour monitors Illuminant D65 > 75 cd/ m2 Not applicable Not applicable Not applicable Neutral, and dark
(0,025) (should be grey or black c
> 100 cd/ m2)
Notes
a This specifies the relative spectral power distribution of the reference illuminant, except for colour monitors in which case it specifies the
chromaticity of the white point of the monitor. Permitted tolerances in chromaticity, from that of the reference illuminant, are given in parentheses.
These are specified at the plane of viewing, according to the 1976 u'10, v'10 UCS system.
b When comparing a transparency to a print, the ratio of the luminance of the transparency illuminator to the equivalent illuminance of the print
viewing surface shall be 2 (± 0,2):1.
c The ambient illumination for colour monitors should be less than or equal to 32 lx and shall be less than or equal to 64 lx.

© ISO 2000 - All rights reserved 13

ISO 3664:2000(E)

Annex B
(informative)

Experimental data leading to selection of metamerism indices

and reference illuminant for this International Standard

B.1 Introduction comment on the wide variation in colour of the illumination

they perceived between different apparatus that were
In order to establish a suitable reference illuminant for this assumed to meet the 1975 specification; because of this it
International Standard, as well as determining tolerances was considered that the specification of the tolerance in
which are both acceptable and practical, a good deal of chromaticity was therefore in need of some reduction.
experimental work was carried out. The objective of the Various research workers had noted that the media
work was threefold: encountered in the industries for which this International
— to determine which reference illuminant would be most Standard was appropriate were far more metameric than
appropriate for specifying viewing conditions for the had previously been the case. The colour rendering index
evaluation of hardcopy images for both graphic (which simply measures the difference in colour for eight
technology and photography; Munsell samples as they would appear under both test and
— to define the most appropriate measures of reference illuminants) provided no indication that different
acceptability between this reference illuminant and real apparatus conforming to this specification would render
sources (together with any other components of the such practical metamers in a similar way. Furthermore, it
viewing environment); and had been suggested that, since most viewing apparatus used
— to establish suitable tolerances for these measures. fluorescent tubes for simulating the reference illuminant
they could not be spectrally similar to D50; supporters of
Since it was considered important that any tolerances this position concluded that a new reference illuminant was
specified were not unreasonably stringent, a small number therefore more appropriate because of the metamerism
of practical sources were studied rather extensively. described earlier.
Subsequently the work was extended to verify that a
reasonably large number of existing viewing apparatus, in At the time this revision was started, proposals had already
use in industry, were not significantly deviant from these emanated from the USA (though not finally agreed
tolerances, although it would not be a concern if many were nationally) to reduce the tolerance on chromaticity, replace
somewhat outside. Since much of this viewing apparatus the reference source with F8 (a relative spectral power
had been in use for some time, this would not be distribution defined by CIE as "typical" of fluorescent tubes
considered unacceptable. It is inevitable that in such a with a chromaticity the same as D50) and add a "band
situation many viewing environments would be due to be method" to reduce the incidence of metamerism.
updated, or need to have their sources replaced. It would
be hoped, and expected, that such a situation would result The ISO joint working group felt that there was substantial
in the availability of sources of viewing apparatus that are justification for considering these proposals. However, it
more consistent with one another than with older was agreed that it was desirable to undertake some
equipment. experimental work to verify their acceptability. It was also
noted that the "band method" was not supported by CIE
who had, in 1981, published a procedure (CIE publication
B.2 Reference illuminant and acceptability 51) for verifying the acceptability of daylight simulators.
This provides a set of theoretical metamers for each of the
tolerances
three reference illuminants D55, D65 and D75 and specifies a
procedure for calculating a metamerism index.
In the 1975 edition of this International Standard the
reference illuminant had been specified as D50 and the
acceptability tolerances had been specified by a limit on the
permitted chromaticity deviation from D50, together with B.3 Metamerism index
colour rendering indices as defined by CIE publication
13.2. However, in the intervening years, various workers The CIE method specifies eight theoretical metamers; that
had expressed reservations about this specification for quite is, the relative spectral power distribution for colours which
different reasons. Many practical users had been heard to match (thereby giving a colour difference of 0) for a
specified illuminant and observer. Five of these are

14 © ISO 2000 - All rights reserved

 ISO 3664:2000(E)

metameric in the visible region of the spectrum, the other practice. The third was to determine the effect of the
three are fluorescent samples which are metameric with standard colorimetric observer selected. (The method
respect to the ultraviolet radiation emitted by the source. specified in CIE Publication No. 51 uses the 1964
The specified procedure requires measurement of the Supplementary Standard Colorimetric Observer, whereas it
relative spectral power distribution of the test source and is common practice in the graphic technology and
calculation of the tristimulus values for these metamers for photograph industries to use the 1931 Standard
both the test and reference source, normalized by area Colorimetric Observer).
according to equation B.1, so that the assessment is
independent of the absolute level of illumination. In order to make this evaluation, two additional sets of
metamers were calculated, derived from practical samples.
100 S(
)
Sn(
)
(B.1) For the first set of metamers, the reference consisted of the
700 spectral reflectance data listed in ANSI CGATS TR 001-

S(
) ȳ10 (
) (
) 1995 for the 928 colours defined in ISO 12642. (These are
400
effectively the colours produced when the data set is
where printed according to the US Specifications for Web Offset
Publications - SWOP). The metamers to the SWOP
y10 is the CIE 1964 Supplimentary Standard patches were calculated for the cyan, magenta, and yellow
Colorimetric Observer Y10 function, dyes of a continuous tone dye diffusion proofing printer,
 is wavelength, using a tristimulus matching algorithm. Three reference
S is the irradiance, and the subscript n denotes the illuminant and observer combinations were used to derive
normalized quantity. three sets of metameric matches (D50 /CIE 1964
Supplementary Standard Colorimetric Observer, F8 /CIE
If the test source has a different relative spectral power 1964 Supplementary Standard Colorimetric Observer, and
distribution from the reference illuminant, the metamers are D50 /CIE 1931 Standard Colorimetric Observer). The
likely to fail to match under the former; thereby giving rise metamers calculated were defined such that they had a
to a colour difference between each pair of metamers. The colour difference (E*a b) of less than 0,02.
magnitude of this difference will depend upon the
differences between the relative energies of the test and The reference for the second set of metamers consisted of
reference illuminants. The CIE procedure then requires spectral reflectances for eight colour patches printed with
that the differences be averaged to produce two indices, offset inks according to ISO 12647-2 (paper type 1). The
one for the visible region and one for the ultraviolet. A metamers of these printed patches were calculated using a
category scale is defined which ranks the test source, special matching algorithm with eight sets of cyan,
covering the range from A (best) to E (worst). magenta and yellow colorants belonging to the imaging
systems shown in Table B.1.
Unfortunately, CIE Publication 51 does not provide
metamers for D50. To enable this method to be used by the Six commercial fluorescent lamps (labelled A to F), all
Joint Task Force, the necessary metamers were provided by nominally D50 simulators, were measured and evaluated
McCamy. He used the same reference data as that in CIE using all three sets of metamers described above (the set
Publication 51 and calculated them from the metamers provided by McCamy, the set based on the SWOP printing,
provided in that publication by extrapolation. The resultant and the set based on ISO 12647-2). Tables B.2 and B.3
metamers, defined for illuminant D50 and the CIE 1964 summarize the results of these calculations. The deviation
Supplementary Standard Colorimetric Observer, are given in chromaticity of the lamps from D50, calculated using the
in Tables 2 and 3. It is anticipated that they will be CIE 1976 Uniform Chromaticity Scale, is included in the
accepted by CIE as an additional data set for CIE first column of each table It was proposed, when this
Publication 51 in due course. The full details of the revision was started, that the tolerance be defined as a
derivation can be found in Color Research and. circle of radius 0,005 centred on the chromaticity of D50. It
Application, Volume 21, Number 3, pp 236-237, June should be noted that lamps C and F do not meet the
1996. chromaticity recommendation.

B.4 Experimental work Each table contains the CIE visible range metamerism
index (MIVIS) based on the McCamy metamers (the average
The experimental work undertaken initially was divided colour difference for the five metamers). As described in
into three stages. The first was to evaluate whether there CIE Publication No. 51, the MIVIS is converted to a
was any advantage in using F8 as the reference illuminant. category scale, in which a "C" has a value ranging from
The second was to determine whether the metamers 0,5 to 1,0, and a "D" ranging from 1,0 to 2,0. These results
proposed by extrapolation from those published in CIE show that most lamps fell within the category "C"
Publication No. 51 provided reasonable correlation with the (proposed as the minimum in this specification), but that
differences anticipated for practical metamers found in two of them were outside this criteria.

© ISO 2000 - All rights reserved 15

ISO 3664:2000(E)

Also shown in Tables B.2 and B.3 are the values for the because of its metamerism index and F on both counts.
average and maximum E*a b encountered with each However, it is probable that the first of these lamps could
practical data set. Comparing the metameric index MIVIS easily be made acceptable by suitable design of the
with the metameric differences obtained with different apparatus. This "pass" rate was considered acceptable in
imaging systems, it may be seen that the index values fall the interest of minimizing the difference between products,
between the average and the maximum colour differences. as requested, by many users and led us to confirm this
The “artificial” metameric pairs used for the index specification. Support was given to this when we were
calculation may therefore be regarded as valid indicators of supplied with data showing that a number of Japanese
metameric differences based on the comparison with lamps, not tested in this study, all conformed to the
differences occurring with real imaging systems. specifications proposed.

The testing with the metamers based on the 928 SWOP As a final validation of the proposals made in this
samples was also extended to derive two additional sets of specification, 61 viewing apparatuses were measured in
metameric matches for F8 /CIE 1964 Supplementary situ. It was found that only 11 % fully met the
Standard Colorimetric Observer, and D50 /CIE 1931 chromaticity, colour rendering and metamerism index
Standard Colorimetric Observer. Again the metamers criteria. Another 17 % met the latter two but failed slightly
calculated were defined such that they had a colour on chromaticity. 13% met the chromaticity and colour
difference (E*a b) of less than 0,02. Tables B.4 and B.5 rendering specification, but failed slightly on the
show these data. (It should be noted that although the metamerism index. 7 % met the metamerism index and
MIVIS and category data are shown in each table, these chromaticity criteria but failed slightly on colour rendering.
values are only calculated for D50 and the CIE 1964 The remaining 52 % failed on at least two criteria but in
Supplementary Standard Colorimetric Observer as defined many cases the deviations in both cases were relatively
in CIE Publication 51.) small.

The average E*a b statistics were similar whether D50 or F8 Given that the maintenance schedule on these units was
were taken as the reference (with the CIE 1964 unknown, it was this data that gave us the final confidence
Supplementary Standard Colorimetric Observer) and there that the specification proposed was about right. Given the
was no consistent change in deviation between the two sets requirement by the industry that deviations between the
of results. The magnitude of the values was generally viewing apparatuses be reduced, simply producing a
larger for the 1931 Standard Colorimetric Observer, but specification that includes all current apparatus that meets
this was not unexpected; it has been reported before. the 1974 specification would be rather pointless. By
reducing the chromaticity tolerance from that required in
However, based on the 928 SWOP samples, the relative 1974, and adding the metamerism index, the industrial
ranking between the lamps was very similar regardless of requirements should, in part at least, be met. Whilst it is a
which illuminant or observer was taken as the reference guess, it seems likely that the number of these existing
(see Tables B.3, B.4, and B,5). The worst three were lamps units fully meeting the specification could be significantly
F > A > C, while D, B, and E’s ranking varied slightly. increased by cleaning and replacement of old tubes. Others
This ranking was the same when the proposed CIE D50 could be brought into line by suitable modification of the
metamers were used. For this reason, it was decided that apparatus itself (reflectors, diffusers, etc). But most
we should continue to specify D50 as the reference importantly, knowing that lamps exist with satisfactory
illuminant in this International Standard, to enable properties, it is clear that equipment vendors can be
consistency with existing products and use the CIE method expected to modify their designs to meet the new
with the metamers proposed by McCamy. It was accepted specification. The only question must be whether the lamp
that the experimental data (despite the fact that it is fairly samples measured are taken from batches with large or
extensive in itself) was fairly minimal in terms of the small distributions of the various criteria but we have no
possible variations which may be encountered in practice reason to believe that those selected are in any way
and more data would always be welcome. When the data untypical.
for the 8 metamers based on the eight colour patches were
subsequently provided, it confirmed the view, as described Thus, users who wish to improve communication by
earlier, that the CIE method (based on illuminant D50) was reducing the deviation between their units and others can
probably robust enough for application in this International choose to try and modify existing apparatus or purchase
Standard. new apparatus. This must be the compromise between user
requirements and practical production so important in an
On the basis of the proposals made in this International International Standard. Of course, at the next revision we
Standard only three of the lamps tested would be may find that production has improved such that we can
acceptable. C would fail because of its chromaticity, A reduce the specification yet further; only time will tell!

16 © ISO 2000 - All rights reserved

 ISO 3664:2000(E)

Table B.1 — Imaging systems used to assess the metamerism of light sources

Process Device

Electrophotography (dry toner) Canon CLC 800

Xerox DocuColor 40
Electrophotography (liquid toner) Indigo E-Print 1000

Ink jet (continuous) Scitex Iris

Ink jet (solid) HP Design Jet 750

Thermal transfer (dye sublimation) Tektronix Phaser 440

Silver halide Fuji Pictrography 3000

Photomechanical Fuji Color-Art

Table B.2 — Matched for D50, CIE 1931 Standard Colorimetric Observer,
8 imaging systems, 8 patches.

Lamp Metameric sample difference b

a
(chromaticity CIE 51 MIVIS CIE 51 category
difference) Average E*a b Maximum E*a b
A (0,003) 1,15 D 0,50 1,45
B (0,005) 0,88 C 0,42 0,91
C (0,008) 0,97 C 0,59 1,33
D (0,001) 0,73 C 0,55 1,34
E (0,002) 0,89 C 0,48 1,20
F (0,007) 2,42 E 1,60 3,87
F8 0,63 C 0,29 0,87
a Values refer to the CIE 1964 Supplementary Standard Colorimetric Observer.
b Values refer to the CIE 1931 Standard Colorimetric Observer.

© ISO 2000 - All rights reserved 17

ISO 3664:2000(E)

Table B.3 — Matched for D50, CIE 1964 Supplementary Standard Colorimetric Observer, 928 patches

Lamp Metameric sample differencea

a
(chromaticity CIE 51 MIVIS CIE 51 category
difference) Average E*a b Maximum E*a b
A (0,003) 1,15 D 0,94 1,75
B (0,005) 0,88 C 0,72 1,66
C (0,008) 0,97 C 0,74 1,69
D (0,001) 0,73 C 0,65 1,81
E (0,002) 0,89 C 0,59 1,67
F (0,007) 2,42 E 1,20 3,06
F8 0,63 C 0,34 1,67
a Values refer to the CIE 1964 Supplementary Standard Colorimetric Observer.

Table B.4 — Matched for F8, CIE 1964 Supplementary Standard Colorimetric Observer, 928 patches

Lamp Metameric sample difference a

a
(chromaticity CIE 51 MIVIS CIE 51 category
difference) Average E*a b Maximum E*a b
A 1,15 D 0,97 1,73
B 0,88 C 0,79 1,66
C 0,97 C 0,87 2,10
D 0,73 C 0,60 1,81
E 0,89 C 0,58 1,67
F 2,42 E 1,21 2,83
D50 0,33 1,78
a Values refer to the CIE 1964 Supplementary Standard Colorimetric Observer.

Table B.5 — Matched for D50, CIE 1931 Standard Colorimetric Observer, 928 patches

Lamp Metameric sample difference b

a
(chromaticity CIE 51 MIVIS CIE 51 Category
difference) Average E*a b Maximum E*a b
A 1,15 D 1,19 2,91
B 0,88 C 0,84 1,88
C 0,97 C 0,94 2,02
D 0,73 C 0,76 2,26
E 0,89 C 0,81 2,58
F 2,42 E 1,38 3,03
F8 0,63 C 0,43 1,88

a Values refer to the CIE 1994 Supplementary Standard Colorimetric Observer.

b Values refer to the CIE 1931 Standard Colorimetric Observer.

18 © ISO 2000 - All rights reserved

 ISO 3664:2000(E)

Annex C
(informative)

Guidelines for judging and exhibiting photographs

C.1 Introduction different source spectral power distributions. If the ISO
viewing condition recommendations are adhered to,
The subjective impression produced by photographs that metamerism should be minimized and colour rendering
are judged for contests or juried exhibitions is of particular should be appropriate. Unfortunately, many small
importance. It follows that photographs intended for such photographic associations do not have access to lighting
purposes should be optimized for a particular, known equipment that meets these strict requirements. The
viewing condition. The viewing conditions specified in question then becomes: if departure from this International
this International Standard are appropriate for this Standard is required by practical considerations, in which
application, but other considerations have led to the use of directions can tolerances be extended with minimal impact?
somewhat wider tolerances in certain directions, and
illumination levels that at first glance might seem In the photographic community, experience has indicated
inappropriate. that the smallest effect on the perception of photographs
results from allowing the correlated colour temperature to
The preferred illumination levels for judging and exhibition decrease. Tungsten sources can be filtered to produce
are those specified for conditions P1, T1 and T2. The excellent 5 000 K illumination, but there is a financial cost
higher illumination level is chosen to allow maximum due to the filters required, and the fact that more powerful
utilization of the density range and colourfulness sources are needed to produce the same illumination level.
capabilities of output media, particularly for judging. Fluorescent sources designed to meet the metamerism and
Prints and transparencies optimized for these illumination colour rendering index requirements of this International
levels will appear a bit dark when viewed under typical Standard are also expensive, and fluorescent sources tend
indoor illumination, but photographers produce the darker to be diffuse. Because tungsten sources produce more
images specifically for contests. The feeling is that the directional illumination, they can be positioned to minimize
increase in "impact" resulting from the larger dynamic surface reflections, thus affording the best view of prints
range is more important than having the prints look perfect that have a high dynamic range.
under typical indoor illumination. It is highly desirable,
therefore, that illumination levels similar to those specified The result of all these considerations is that the sources
for ISO viewing conditions P1, T1 and T2 be adhered to for preferred for judging photographic prints are moderately
judging. For exhibition, it may be impractical to achieve directional tungsten or filtered tungsten sources with
the high levels of illumination required, and lower levels correlated colour temperatures between 2800 K and
are acceptable. 5000 K. The use of filtered tungsten minimizes
metamerism and colour rendering problems, and
Another practical consideration concerns the relative differences in correlated colour temperature are dealt with
spectral power distribution of the illumination The human by the adaptation of the human visual system. Tungsten
visual system is excellent at adapting to the correlated and filtered tungsten sources are also used for projected
colour temperature of illumination sources which transparencies. Fluorescent sources which meet the
approximate a Planckian radiator between 2 800 K and requirements of this International Standard are preferred
6 500 K. Along this locus, typical photographic materials for the direct viewing of transparencies, where diffuse
exhibit larger perceptual shifts due to sources failing to illumination is desirable.
approximate a Planckian radiator than from changes in
correlated colour temperature. This is particularly true for C.2 Recommendations
sources which emit sharp spectral lines. A fluorescent
source may have exactly the same chromaticity as a Table C-1 summarizes the recommendations for the
particular Planckian source, but a photographic material judging and exhibition of photographs. These
may appear substantially different under the two sources. recommendations assume that comparisons will be made
only of similar media under the same illumination, and that
The purpose of the metamerism and colour rendering index observers are adapted to the illumination. If a non-standard
requirements in this International Standard is to minimize correlated colour temperature or illumination level is used,
the likelihood of substantial changes in appearance due to it should remain constant for all evaluations conducted. If

© ISO 2000 - All rights reserved 19

ISO 3664:2000(E)

prints and transparencies are to be compared, non-standard view the photograph. Also, the correlated colour
illumination should be avoided to the greatest extent temperature of any ambient illumination present when
possible. Any comparisons that involve different effective transparencies are viewed should be less than or equal to
correlated colour temperatures or illumination levels should the correlated colour temperature of the viewing
be successive, and observers should be given sufficient illumination.
time to adapt to the illumination before being allowed to

Table C-1 — Recommendations for viewing conditions for judging and exhibiting photographs.

Material Recommended viewing conditions

Judging

Photographic prints ISO condition P1 with 0°/45° viewing geometry, but the correlated colour temperature can be
as low as 2 800 K if tungsten sources are used a.

Transparencies ISO condition T1 (diffuse illumination).

(direct viewing)

Transparencies ISO condition T2, but the correlated colour temperature can be as low as 2 800 K if tungsten
(projection viewing) sources are used a.

Exhibition

Photographic prints ISO condition P1 with 0°/45° viewing geometry, but the correlated colour temperature can be
as low as 2800 K if tungsten sources are used1, and the illuminance level can be as low as
375 lx.

Transparencies ISO condition T1 (diffuse illumination), but the correlated colour temperature can be as low as
(direct viewing) 2 800 K if tungsten sources are useda, and the luminance level can be as low as 240 cd/ m2, or 1
000 times the veiling glare luminance, whichever is greaterb.

Transparencies ISO condition T2, but the correlated colour temperature can be as low as 2 800 K if tungsten
(projection viewing) sources are used1, and the luminance level can be as low as 40 cd/ m2, or 1 000 times the
veiling glare luminance, whichever is greater c.
a With sources at correlated colour temperatures other than 5 000 K, the metamerism and colour rendering index qualifications
described in this International Standard are not applicable. In such cases, the user must rely on the similarity of the tungsten source
relative spectral power distribution to that of a theoretical Planckian radiator. If the relative spectral power distributions are
sufficiently similar, the chromaticities, metameric differences, and colour rendering will also be similar.
b The veiling glare luminance in a typical direct viewing situation will be approximately 0,001 6 times the stray light illuminance
incident on the illuminator surface.
c The veiling glare luminance in a typical projection viewing situation will be approximately 0,3 times the stray light illuminance
incident on the projection screen.

20 © ISO 2000 - All rights reserved