ASTM E1808-96(2003)
(Guide)Standard Guide for Designing and Conducting Visual Experiments
Standard Guide for Designing and Conducting Visual Experiments
ABSTRACT
This guide is intended to help the user decide on the type of viewing conditions, visual scaling methods, and analysis that should be used to obtain reliable visual experimental data. It is also intended to illustrate the techniques that lead to visual observations that can be correlated with objective instrumental measurements of appearance attributes of objects. This guide includes a review of issues regarding the choice and design of viewing environments, an overview of various classes of visual experiments, a review of experimental techniques for threshold, matching, and scaling experiments, a review for data reduction and analysis procedures. The three different threshold and matching techniques namely, the methods of adjustment, limits, and constant stimuli, are explained. Perceptual scaling techniques reviewed include ranking, graphical rating, category scaling, paired comparisons, triadic combinations, partitioning, and magnitude estimation or production. Brief descriptions and examples, along with references to more detailed literature, are given on the appropriate types of data analysis for each experimental technique.
SCOPE
1.1 This guide is intended to help the user decide on the type of viewing conditions, visual scaling methods, and analysis that should be used to obtain reliable visual data.
1.2 This guide is intended to illustrate the techniques that lead to visual observations that can be correlated with objective instrumental measurements of appearance attributes of objects. The establishment of both parts of such correlations is an objective of Committee E12.
1.3 Among ASTM standards making use of visual observations are Practices D 1535, D 1729, D 3134, D 4086, and E 1478; Test Methods D 2616, D 3928, and D 4449; and Guide E 1499.
1.4 This guide does not purport to address all of the safety concerns, if any, associated with its use. It is the responsibility of the user of this standard to establish appropriate safety and health practices and determine the applicability of regulatory limitations prior to use.
General Information
Relations
Standards Content (Sample)
NOTICE: This standard has either been superseded and replaced by a new version or withdrawn.
Contact ASTM International (www.astm.org) for the latest information.
Designation:E1808–96(Reapproved2003)
Standard Guide for
Designing and Conducting Visual Experiments
This standard is issued under the fixed designation E1808; the number immediately following the designation indicates the year of
original adoption or, in the case of revision, the year of last revision. A number in parentheses indicates the year of last reapproval. A
superscript epsilon (´) indicates an editorial change since the last revision or reapproval.
1. Scope E1478 Practice for Visual Color Evaluation of Transparent
Sheet Materials
1.1 Thisguideisintendedtohelptheuserdecideonthetype
E1499 Guide for Selection, Evaluation, and Training of
of viewing conditions, visual scaling methods, and analysis
Observers
that should be used to obtain reliable visual data.
1.2 This guide is intended to illustrate the techniques that
3. Terminology
leadtovisualobservationsthatcanbecorrelatedwithobjective
3.1 The terms and definitions in Terminology E284 are
instrumental measurements of appearance attributes of objects.
applicable to this guide.
The establishment of both parts of such correlations is an
3.2 Definitions:
objective of Committee E12.
3.2.1 appearance, n—in psychophysical studies, perception
1.3 AmongASTM standards making use of visual observa-
inwhichthespectralandgeometricaspectsofavisualstimulus
tions are Practices D1535, D1729, D3134, D4086, and E1478;
are integrated with its illuminating and viewing environment.
Test Methods D2616, D3928, and D4449; and Guide E1499.
3.2.2 observer, n—one who judges visually, qualitatively or
1.4 This guide does not purport to address all of the safety
quantitatively, the content of one or more appearance attributes
concerns, if any, associated with its use. It is the responsibility
in each member of a set of stimuli.
of the user of this standard to establish appropriate safety and
3.2.3 sample, n—a small part or portion of a material or
health practices and determine the applicability of regulatory
product intended to be representative of the whole.
limitations prior to use.
3.2.4 scale, v—to assess the content of one or more appear-
2. Referenced Documents ance attributes in the members of a set of stimuli.
3.2.4.1 Discussion—Alternatively, scales may be deter-
2.1 ASTM Standards:
mined by assessing the difference in content of an attribute
D1535 PracticeforSpecifyingColorbytheMunsellSystem
with respect to the differences in that attribute among the
D1729 Practice for Visual Appraisal of Colors and Color
members of the set.
Differences of Diffusely-Illuminated Opaque Materials
3.2.5 specimen, n—a piece or portion of a sample used to
D2616 Test Method for Evaluation of Visual Color Differ-
make a test.
ence With a Gray Scale
3.2.6 stimulus, n—any action or condition that has the
D3134 PracticeforEstablishingColorandGlossTolerances
potential for evoking a response.
D3928 Test Method for Evaluation of Gloss or Sheen
3.3 Definitions of Terms Specific to This Standard:
Uniformity
3.3.1 anchor, n—the stimulus from which a just-perceptible
D4086 Practice for Visual Evaluation of Metamerism
difference is measured.
D4449 Test Method for Visual Evaluation of Gloss Differ-
3.3.2 anchor pair, n—apairofstimulidifferingbyadefined
ences Between Surfaces of Similar Appearance
amount, to which the difference between two test stimuli is
E284 Terminology of Appearance
compared.
3.3.3 interval scale, n—a scale having equal intervals be-
This guide is under the jurisdiction of ASTM Committee E12 on Color and
tween elements.
Appearance and is the direct responsibility of Subcommittee E12.11 on Visual
3.3.3.1 Discussion—Logical operations such as greater-
Methods.
than, less-than, equal-to, and addition and subtraction can be
Current edition approved Dec. 1, 2003. Published December 2003. Originally
performed with interval-scale data.
approved in 1996. Last previous edition approved in 1996 as E1808 – 96. DOI:
10.1520/E1808-96R03.
3.3.4 law of comparative judgments—an equation relating
For referenced ASTM standards, visit the ASTM website, www.astm.org, or
the proportion of times any stimulus is judged greater, accord-
contact ASTM Customer Service at service@astm.org. For Annual Book of ASTM
ing to some attribute, than any other stimulus in terms of
Standards volume information, refer to the standard’s Document Summary page on
the ASTM website. just-perceptible differences.
Copyright © ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959, United States.
E1808–96 (2003)
3.3.5 nominal scale, n—scale in which items are scaled undermine the integrity of the experiments.The spectral power
simply by name. distribution of the illumination should be known or, if this is
3.3.5.1 Discussion—Only naming can be performed with notpossible,thelightsourceshouldbeidentifiedastotypeand
nominal-scale data. manufacturer. Information such as daylight-corrected fluores-
3.3.6 ordinal scale, n—a scale in which elements are sorted cent light, warm-white fluorescent light, daylight-filtered in-
in order based on more or less of a particular attribute. candescent light, incandescent light, etc., together with param-
3.3.6.1 Discussion—Logical operations such as greater- eters such as correlated color temperature and color rendering
than,less-than,orequal-tocanbeperformedwithordinal-scale index, if available, should be noted in the report of the
data. experiment.
3.3.7 psychometric function, n—the function, typically sig- 5.2 Viewing Geometry—Almost all specimens exhibit some
moidal, relating the probability of detecting a stimulus to the degreeofgonioapparentorgoniochromaticvariation;therefore
stimulus intensity. the illuminating and viewing angles must be controlled and
3.3.8 psychophysics, n—the study of the functions relating specified. This is particularly important in the study of speci-
the physical measurements of stimuli and the sensations and mens exhibiting gloss variations, textiles showing directional-
perceptions the stimuli evoke. ity, or gonioapparent (containing metallic or pearlescent pig-
3.3.9 ratio scale, n—a scale which, in addition to the ments)orretroreflectivespecimens,amongothers.Thiscontrol
properties of other scales, has a meaningfully defined zero and specification can range from correct positioning of the
point. source and observer and the elimination of any secondary light
3.3.9.1 Discussion—In addition to the logical operations sources visible in the specimens, for the judgment of gloss
performable with other types of data, multiplication and specimens at and near the specular angle, to more elaborate
division can be performed with ratio-scale data. procedures specifying a range of angles and aperture angles of
3.3.10 scale, n—a defined arrangement of the elements of a illumination and viewing for gonioapparent and retroreflective
set of stimuli or responses. specimens. When fluorescent specimens are studied, the spec-
tral power distribution of the source must closely match that of
4. Summary of Guide
a designated standard source.
5.3 Surround andAmbient Field—For critical visual scaling
4.1 This guide provides an overview of experimental design
work, the surround, the portion of the visual field immediately
and data analysis techniques for visual experiments. Carefully
surrounding the specimens, should have a color similar to that
conducted visual experiments allow accurate quantitative
of the specimens.The ambient field, the field of view when the
evaluation of perceptual phenomena that are often thought of
observer glances away from the specimens, should have a
asbeingcompletelysubjective.Suchresultscanbeofimmense
value in a wide variety of fields, including the formulation of neutral color (Munsell Chroma less than 0.2) and a Munsell
ValueofN6toN7(luminousreflectance29to42);seePractice
coloredmaterialsandtheevaluationoftheperceivedqualityof
products. D1729).
5.4 Observers—Guide E1499 describes the selection,
4.2 This guide includes a review of issues regarding the
evaluation, and training of observers for visual scaling work.
choice and design of viewing environments, an overview of
Of particular importance is the testing of the observers’ color
various classes of visual experiments, and a review of experi-
vision and their color discrimination for normality. Color
mental techniques for threshold, matching, and scaling experi-
vision tests for this purpose are described in Guide E1499.
ments. It also reviews data reduction and analysis procedures.
Three different threshold and matching techniques are ex-
6. Categories of Visual Experiments
plained, the methods of adjustment, limits, and constant
6.1 Visual experiments tend to fall into two broad classes:
stimuli. Perceptual scaling techniques reviewed include rank-
(1) threshold and matching experiments designed to measure
ing, graphical rating, category scaling, paired comparisons,
visual sensitivity to small changes in stimuli (or perceptual
triadiccombinations,partitioning,andmagnitudeestimationor
equality), and (2) scaling experiments intended to generate a
production. Brief descriptions and examples, along with refer-
psychophysical relationship between the perceptual and physi-
ences to more detailed literature, are given on the appropriate
cal magnitudes of a stimulus. It is critical to determine first
types of data analysis for each experimental technique.
which class of experiment is appropriate for a given applica-
4.3 Forreviewsoftopicsinotherthanvisualsensorytesting
tion.
within ASTM, see Refs (1, 2).
6.1.1 Threshold and Matching Experiments— Threshold
5. Viewing Conditions experiments are designed to determine the just-perceptible
difference in a stimulus, or JPD.Threshold techniques are used
5.1 Light Source—The illumination of the specimens in
to measure the observers’ sensitivity to a given stimulus.
scaling experiments must be reproducible over the course of
Absolute thresholds are defined as the JPD for a change from
the experiments. To achieve this, it is essential to control both
no stimulus, while difference thresholds represent the JPD
the spectral character and the amount of illumination closely in
from a particular stimulus level greater than zero.The stimulus
both space and time. Failure to accomplish this can seriously
from which a difference threshold is measured is known as an
anchorstimulus.Often,thresholdsaremeasuredwithrespectto
the difference between two stimuli. In such cases, the differ-
The boldface numbers in parentheses refer to a list of references at the end of
this guide. ence of a pair of stimuli is compared to the difference in an
E1808–96 (2003)
anchor pair. Absolute thresholds are reported in terms of the 6.1.2.4 Ratio Scales—Ratio scales have all the properties of
physical units used to measure the stimulus, for example, a the above scales plus a meaningfully defined zero point. Thus
brightness threshold might be measured in luminance units of it is possible to equate ratios of numbers meaningfully with a
candelas per square metre. Sensitivity is measured as the ratio scale. Ratio scales are often impossible to obtain in visual
work. An example of a ratio scale is the absolute, or Kelvin,
inverse of the threshold, since a low threshold implies high
sensitivity. Threshold techniques are useful for defining visual temperature scale. All of the mathematical operations that can
be performed on interval-scale data can also be performed on
tolerances, such as color-difference tolerances. Matching tech-
niques are similar, except that the goal is to determine when ratio-scale data, and in addition, multiplication and division
can be performed.
two stimuli are not perceptibly different. Measures of the
variability in matching can be used to estimate thresholds.
7. Threshold and Matching Methods
Matching experiments provided the basis for CIE colorimetry
7.1 Several basic types of threshold experiments are pre-
through the metameric matches used to derive the color-
sented in this section in order of increasing complexity of
matching functions of the CIE standard observers.
design and utility of the data generated. Many modifications of
6.1.2 Scaling Experiments—Scaling experiments are in-
thesetechniqueshavebeendevelopedforspecificapplications.
tended to derive relationships between perceptual magnitudes
Experimenters should strive to design an experiment that
and physical magnitudes of stimuli. Several decisions must be
removes as much control of the results from the observers as
made, depending on the type and dimensionality of the scale
possible, thus minimizing the influence of variable observer
required. It is important to identify the type of scale required
judgment criteria. Generally, this comes at the cost of imple-
and decide on the scaling method to be used before any scaling
menting a more complicated experimental procedure.
dataarecollected.Thisseemstobeanobviouspoint,butinthe
7.1.1 Method of Adjustment—The method of adjustment is
rush to acquire data it is often overlooked, and later it may be
the simplest and most straightforward technique for deriving
foundthatthedataobtaineddonotyieldtheanswerrequiredor
threshold data. In it, the observer controls the stimulus mag-
cannot be used to perform desired mathematical operations.
nitude and adjusts it to a point that is just perceptible (absolute
See Refs (3, 4) for further details. Scales are classified into the
threshold) or just perceptibly different (difference threshold).
following four classes:
The threshold is taken to be the mean setting across a number
6.1.2.1 Nominal Scales—Nominal scales are relatively
of trials by one or more observers. The method of adjustment
trivial in that they scale items simply by name. For color, a
has the advantage that it is quick and easy to implement.
nominalscalemightconsistofreds,yellows,greens,blues,and
However, it has a major disadvantage in that the observer is in
neutrals. Scaling in this case would simply require deciding
control of the stimulus. This can bias the results due to
which color belonged in which category. Only naming can be
variability of observers’ criteria and adaptation effects. If an
performed with nominal data.
observer approaches the threshold from above, adaptation
6.1.2.2 Ordinal Scales—Ordinal scales are scales in which
might result in a higher threshold than if it were approached
elements are sorted in ascending or descending order based on
from below. Often the method of adjustment is used to obtain
more or less of a particular attribute. A box of multicolored
a first estimate of the threshold, to be used in the design of
crayons could be sorted by hue, and then in each hue family,
more sophisticated experiments. The method of adjustment is
say red, the crayons could be
...
Questions, Comments and Discussion
Ask us and Technical Secretary will try to provide an answer. You can facilitate discussion about the standard in here.