Duration of Afterimage Disability After ViewingSimulated Sun Reflections

Roger L. Saur and Slavko M. Dobrash

A scene was simulated of an automobile driver's view through his windshield. It contained a roadway, anidentification target consisting of a Landolt C in various orientations on the road 96 m ahead, and a glarereflection of the sun arising from the direction of the driver's windshield wiper arm. The product ofluminance times area of each of the four glare sources employed was about 0.1 of the maximum implied inthe Federal Safety Standards. The target required a visual acuity of 20/40 on the Snellen scale, equal tothat imposed by driver license requirements. After being exposed for realistic intervals while lookingdirectly at the glare source, the duration of the interval required by thirty-five observers for identificationof the target orientation was measured. Averages showed that the increase of identification time overthat required after no glare exposure varied from 0.8 sec to 2.7 sec. Reduction of apparent area of theglare source reduced afterimage disability more than did reduction of intensity.

IntroductionThe federal government, in its Motor Vehicle Safety

Standard (MVSS) No. 107, limits glare presented toautomobile drivers from sun reflections by requiring theuse of matte surfaces on certain metallic trim items.'The gloss of such surfaces is limited to 40 gloss units*maximum. Because of instrumentation specified, it isinferred that image size is equal to that of the sun.Glare is thus reduced by decreasing the intensity of thesun's reflection to no more than 2% of that of the sun.

Visual inspection of sun reflections from metal sur-faces showed that comparable reduction of glare im-pression could be obtained by demagnification of theapparent image size by use of a curved trim surfacewhich was mirrorlike. A subsequent investigationshowed that glare impression was proportional to theproduct of the apparent area and the intensity of theglare source. 2 3

An additional program was initiated to study after-image disability imparted by simulated sun images ofvarious sizes and intensities. This program is describedherein.

The authors are with the Electrochemistry Department,Research Laboratories, General Motors Corporation, Warren,Michigan 48090.

Received 11 December 1968.* The origin of the selection of this value is obscure.

Apparatus and ProcedureEach test subject with chin in a chin cup observed a

scene outlined on a flat, vertical sheet (153 cm X 102 cm)as diagrammed in Fig. 1. The surface was paintedmatte, either white or gray. Observation distance was91.5 cm.

The scene depicted a driver's view of a straight levelhighway of white cement extending to the horizon, withgray areas representing the roadside scene, similar tothat used by Christie.4 The horizon line was locatedlevel with the eyes of the test observer. Scene lumi-nances, obtained with tungsten lamps were: for thesky, 400-700 mLt; for the roadside, 250 mL; for the roaditself, 300400 mL. These approximate the valuesfound by Allen5 in actual roadway scenes.

A visual target was located at an apparent distance of96 m ahead on the road. The target was a Landolt Cwhich subtended an angle of 10 min of arc at the ob-server, equal to that of a 20/40 Snellen letter. Theopening and letter thickness each subtended 2 min ofarc. The C was projected from the reverse side intothe plane of the simulated scene by lenses. Luminanceseen within the lens system by the test observer was460 mL, and that of the target C was 350 mL; contrastratio was therefore (460-350) 100%/460 = 24%.Orientation of each of 12 target C's was up, down, left,or right. Individual presentation order was random.

The glare source was located 5 deg below the visualtarget, thus simulating the direction of a sun reflectionarising from a windshield wiper arm. The glare source

t Footlamberts = 0.938 X millilamberts.

September 1969 / Vol.- 8, No. 9 / APPLIED OPTICS 1799

153 cm

Fig. 1. Drawing of afterimage disability apparatus as seen by ob-server being tested.

was one of four backlighted apertures. Source char-acteristics are given in Table I. Sources 1 and 2 arosefrom a photoflood lamp; sources 3 and 4 arose from a500-W xenon arc lamp. * The filter employed to moder-ate the lamp intensities for sources 2 and 4 was a veryfine mesh screen with about 6% open area. Appropri-ate lenses were used to obtain uniformity of sourceluminance. Maximum intensity L of the glare sourcewas limited to 108 mL to reduce chance of retinal burns.6

Minimum angular area co was arbitrarily established at6.3 X 10-7 sr, equal to 1% of that of the sun. Thesecharacteristics represent source 1. The intensity andangular area for each of the remaining sources werevaried in steps of about 10 so as to keep the product Lxapproximately constant. The four glare sources thusformed offer about equal amounts of discomfort glare, orglare impression2 .3 ; their range of characteristics ap-proximates those of sun reflections, seen by drivers,from surfaces that conform to federal restrictions andalso from polished curved metal surfaces.

A minimum time of 0.2-sec exposure to the glaresources was selected as a representative duration offixation during involuntary eye movements.7 Theremaining exposures of 1.0 see and 5.0 see were repre-sentative of average and maximum time for voluntaryfixation.

A pair of +0.87 diopter antireflection coated specta-cles (clip-ons if the observer wore spectacles whiledriving) was used by each observer while being testedto relax accommodation, thus simulating distancefixation.

While being tested, each observer was instructed tolook at the glare source aperture while it was luminouswithout blinking or voluntary eye movement. Anelectrically operated program was then initiated whichfirst removed a shutter from the glare source. Aftera predetermined time (0.2 sec, 1.0 sec, or 5.0 see), theglare source shutter was closed, and simultaneously therecognition target shutter was opened and an interval

* Kneisley Electric Company, Toledo, Ohio.

timer with calibrations of 0.1 see was started. Whenthe test subject had recognized the orientation of theLandolt C, he depressed a switch which closed thetarget shutter and stopped the timer. He then verballystated the orientation of the target C. Correct identi-fication was required before recognition time (uncor-rected) was recorded. Recognition time was also deter-mined with the glare source not luminous. Presenta-tion order of sources with various duration, luminance,and size was randomized. One determination wasmade for each source. Thirty-five observers wereemployed.

A brief visual examination was administered to allobservers to insure that each had vision adequate forthe task. Visual acuity was also determined at 6.1 mwith a chart employing Sloan letters.t A Pritchardtelephotometerl was employed for all measurements ofluminance.

ResultsArithmetic averages of recognition time of all ob-

servers were obtained for each source at each exposureduration. The average values are shown in Table I,and are plotted in Figs. 2 and 3. Corrected values of

) 3.0

E

2.0

I1.0

0

A -5.0 sec exposureB -1.0 sec exposureC -0.2 sec exposure

A

1 2 3 4Glare Source

Fig. 2. Plots of average uncorrected recognition time vs glaresource following exposure of 0.2 sec, 1.0 sec, or 5.0 see to glare

source.

Z

_- 3.0

E

g 2.0

(y 1.0

A - 5.0 sec exposureB -1.0 sec exposureC -0.2 sec exposure

A

B

3 4Glare Source

Fig. 3. Plots of average corrected recognition time vs glare sourcefollowing exposure of 0.2 sec, 1.0 sec, or 5.0 see to glare source.

t Visual acuity chart manufactured by Goodlite Company,7426 Madison Street, Chicago, Ill.

$ PhotoResearch Corporation, Hollywood, Calif.

1800 APPLIED OPTICS / Vol. 8, No. 9 / September 1969

1

Table I. Average Recognition Time Required for Identification of Orientation of 20/40 Landolt C Following Glare Exposure

Glare Luminance Area Product Glare exposure Average recognition time (sec)source L (mL) X (sr) Lx time (see) Uncorrected Corrected

1 1.0 X 10, 6.3 X 10-7 63 0.2 1.1 0.81.0 1.5 1.25.0 1.7 1.2

2 1.4 X 107 6.3 X 10-6 88 0.2 1.2 1.11.0 1.6 1.35.0 2.2 1.7

3 0.6 X 106 6.3 X 10- 38 0.2 1.7 1.31.0 2.5 2.25.0 3.0 2.7

4 0.8 X 10 6.3 X 10-4 50 0.2 1.1 0.81.0 2.0 1.75.0 2.9 2.5

recognition time for each observer were obtained bysubtracting the recognition time required with the glaresource not luminous from the (uncorrected) time re-quired after exposure to the luminous glare source.

DiscussionThe graphs of Figs. 2 and 3 depict the average time

intervals required for identification. If all plots werehorizontal, then afterimage glare disability for a givenexposure time would be proportional to the product ofluminance L and apparent area co, or

Lw = k,

where k is constant for a given plot. However, therecognition time in any plot is the longest followingexposure to sun sized source 3, despite the slightly re-duced value for its Lx. product (38) below that of theother sources. Therefore, it can be stated that thesources which were smaller and more intense causedless afterimage disability than did the reduced intensitysource of sun size.

Apparent size of source 3 is no smaller than that of asun reflection from a surface which conforms to re-strictions in MIVSS 107; intensity is about 0.1 of theallowed maximum. Apparent size and luminance ofsource are comparable to those of sun's image seen in apolished metal surface with curvature such that thesize is 1% of that of the sun. Therefore, the data in-dicate that afterimage disability of sun reflections frommetal trim items can be reduced by use of curved brightsurfaces at least as effectively as by the matte surfacesnow specified. Curvatures necessary can be calculatedin terms of viewing distance.8' 9

The data indicated that neither age (from nineteento sixty-four years) of observer nor visual acuity (from20/15 to 20/40) controlled recognition time.

A frequency distribution study of the identificationtime required for each observer showed an approxi-mately gaussian (bell-shaped) distribution.

Summary(1) Proper curvature of mirrorlike surfaces reduces

both afterimage disability and discomfort glare ofsimulated sun images, at least as effectively as do the

matte surfaces now specified in Federal Standards onAutomotive Safety.

(2) Curvatures necessary for reduction of afterimagedisability are the same as were determined for dis-comfort glare.

(3) For drivers possessing at least 20/40 visualacuity, duration of disability following exposure tosimulated sun images with a glare potential (millilam-berts times steradians) between 38 and 88 was in arange of 1.1-3.0 sec for a visual target with criticaldetail subtending 2 min of arc and having 24%contrast ratio.

(4) Following exposure to simulated sun images,neither age nor visual acuity of observers licensed todrive an automobile affected the recognition time ofan identification target of size equal to a 20/40 Snellenletter.

The authors are grateful for the many fruitful dis-cussions with Glenn A. Fry of Ohio State University andPaul L. Connolly, Bloomfield Hills, Michigan. Thevisual examinations were also performed by Dr. Con-nolly. Randomization and data analyses were per-formed by H. W. Gugel of General Motors ResearchLaboratories.

References1. Motor Vehicle Safety Std. No. 107, Federal Register 32,

02411 (U.S. Department of Commerce, February 3, 1967).2. R. L. Saur, J. Opt. Soc. Amer. 58, 847 (1968).3. R. L. Saur, J. Opt. Soc. Amer. 58, 1309 (1968).4. A. W. Christie and A. J. Fisher, Trans. Illum. Soc. (London)

31,93, 114 (1966).5. M. J. Allen, Amer. J. of Optometry and Arch. Amer. J.

Optometry 40, 61 (1963).6. Private conversations with R. A. Weale (Institute of Ophtha-

mology, London) and N. D. Miller (Technology, Inc., SanAntonio, Texas).

7. H. Davson, Muscular Mechanism (Academic Press Inc.,New York, 1962), Vol. 3, p. 89.

8. R. L. Saur and S. M. Dobrash, presented at Automotive En-gineering Congress, sponsored by the Soc. Automotive Eng.,8-12 January 1968, at Detroit, Mich.; available as SAEPaper 680044.

9. R. L. Saur, private communication.

September 1969 / Vol. 8, No. 9 / APPLIED OPTICS 1801