Border locking and the Café Wall illusion - Richard …

1 Border locking and the Caf Wall illusion1 Border locking andthe Caf Wall illusionRichard L Gregory, Priscilla HeardFrom: Perception, 1979, volume 8, pages 365-380 Brain and Perception Laboratory, University of Bristol,Department of Anatomy. The Medical School. University Walk,Bristol BS8 1TD. EnglandAbstractThe Cafe Wall illusion (seen on the tiles of a local caf ) is aM nsterberg chequerboard figure, but with horizontal parallellines which may have any luminance separating the rows ofdisplaced squares. Thcsc (these 'mortar lines) display markedwedge distortion which is especially affected by: contrast of thesquares ('tiles'; width of the 'mortar lines, and their luminancewhich must not be significantly higher than that of the lightsquares or lower than that of the dark squares for distortion tooccur. An experiment is described from which quantitative datahave been obtained by varying these parameters. It is suggestedthat contiguous regions of different luminance (and contiguouscolour regions) are normally held in spatial register by lockingfrom common luminance boundaries.)

2 The Caf Wall illusion isattributed to this Border locking producing inappropriate contourshifts from neighbouring regions of contrasting luminance whenseparated by narrow gaps of neutral luminance. Furtherimplications on the Border - locking notion are Introduction1 .1 BackgroundFigure 1. The original of the Caf Wall, St Michael s Hill, was noted some time ago (Gregory 1973) by a then memberof our laboratory, Steve Simpson, that the mortar lines of thechessboard-like design of tiles of a caf wall in St Michael's Hill,near our laboratory in Bristol, appear not parallel as they are, butto converge markedly in alternate-direction wedges (figure 1). Wemade models of similar patterns in which several parameters couldbe controlled, especially luminance contrast of the tiles , colourof the tiles , and their rectilinear proportions; and the width, thecolour and the luminance of the parallel mortar lines. Hundredsof subjects were shown these models, mainly in the informalsetting of lecture demonstrations, while we got a feel for thephenomena which were striking and remarkably consistent.

3 It wasat once noted that most of the effects persisted, or changed inrepeatable, consistent ways, over a very wide range of viewingconditions. So there were clearly hard data to be extracted. Thebasic figure of a chessboard with alternate rows of squares shiftedby half a cycle is the M nsterberg figure (figure 2).Figure 2. The M nsterberg our term Caf Wall illusion we refer to the much moregeneral case, especially where the mortar lines may have aluminance different from either the light or dark tiles . It tunis outthat the parallel black lines of the M nsterberg figure, in which thedark squares ( tiles ) are also black, is a special and limiting case;it does not reveal several features of interest in this unusualdistortion illusion , in which all lines are parallel or at right angles,and the figure is essentially symmetrical though the evokeddistortion is markedly these informal though rather extensive observations bymany observers several laws Laws of the Caf Wall illusion (i) The alternate wedge distortions occur in the samedirections, at all times, for all observers under the same viewingconditions; and for all observers the wedge distortions reversewhen alternate rows of tiles are pushed across half a cycle.

4 (ii) The distortion remains in the same direction for a widerange of shift of the alternate rows: there is no sudden switch ofwedge direction when alternate rows are shifted across thebisection, as might have been expected.(iii) Neither the amount of the distortion nor the direction ofthe distortion wedges depend on how the sides of the display aremasked. So it is not due to any kind of end effect of the sequenceof dark and light tiles at the sides. The wedge distortions aretherefore given by the repeated small scale asymmetrical featuresof the figure. (iv) The distortion is much the same for anyorientation of the figure. Moulden and Renshaw (1979) find,however, small changes with the M nsterberg figure.(v) The rectangles can have a wide range of vertical-to-horizontal length ratios. [For experiments described in this paperwe have used approximate squares.](vi) The distortion is highly dependent on the luminancecontrast of the tiles: it increases with increased luminance contrast.

5 (vii) The distortion occurs only when the luminance of themortar lies between the luminances of the dark and light tiles, or atleast, the mortar is not much darker than the dark or lighter thanthe light tiles (figure 3). The standard M nsterberg figure is alimiting and nonoptimal case where the luminances of the mortarand dark tiles are the same. The importance of what we call mortarluminance was realised by Fraser (1908).(viii) The illusion is retained with coloured tiles (say red andgreen) and coloured mortar; but not when the contrasting colouredtiles have the same luminance. This dependence on luminancecontrast, and loss of distortion at isoluminance was appreciated byFraser (1908); and confirmed by Yvonne Lammerich in ourlaboratory, as retailed by Gregory (1977) where it is also reportedthat the classical converging-lines illusions show no distortionwhen their lines and background are locking and the Caf Wall illusion2 Figure 3.

6 The Caf Wall display, showing the basic effect of change of luminance of the mortar lines. The illusion is only present when themortar luminance lies between, or at least is not far outside, the luminances of the dark and light tiles. The mortar width was controlled withspacers. (This model is made with squares of grey paper of two albedos, and not the white paper and retroreflecting material used in theexperiment.)(ix) The wedge distortion never reverses for any changes ofluminosities. (The wedges do, however, reverse with half-cycleshifts of alternate rows of the tiles as stated in the first law ,above.)(x) The distortion is clearly greater in somewhat peripheralvision (cf Moulden and Renshaw 1979), or for foveal vision whenthe display is blurred by a weakly de-accommodating is an indication that each tile is distorted into a separatesmall wedge: a problem is why these are seen as a continuous longwedge for each row, in alternate directions, though the figure hasonly repeated small asymmetry of the displaced tiles.

7 (This isenantiomorphic symmetry, cf Shuhnikov and Koptsik 1974.)Similar large scale distortions from repeated small-scaleasymmetries are found in many other illusions, such as the Fraserfigure (Fraser 1908).(xi) The wedge distortion occurs over a very wide range ofvisual angles for the display as a whole.(xii) The distortion occurs over perhaps the entire workingluminance range of the eye.(xiii) Tile distortion occurs only for narrow mortar lines; theymust not subtend more than about 10 min of arc (at high tilecontrast, less at low contrast) or the illusion is 4. The apparent spiral is in fact concentric circles. This isusually regarded as evidence of visual spatial integration, in thiscase from misleading line Further observationsDynamic effects occur while either the mortar-line luminanceor the luminance of the tiles is varied. The bounding borders of thetiles are seen to move. They creep across the mortar duringluminance changes.

8 Though a difficult observation, it seems thatthe movement is greatest for the borders having the lowerboundary contrast with the mortar, as the mortar luminance isvaried between the tile luminances. This slight asymmetry of theshifts of the borders is more easily seen when the tiles aredisplaced a quarter of a cycle, to give a chessboard pattern. It isclear by using colour contrast for the mortar, that this is notmerely loss of the mortar when it becomes isoluminant with thedark or light seems to us very important to distinguish between thedynamic shifts with luminance changes and the staticdisplacements observed at constant luminance. They could well beeffects or symptoms of different physiological processes, as, forexample, static wedge distortion does not increase with mortarluminances not much darker than the dark or much lighter than thelight tiles; but this is not so for the dynamic shifts, which aredramatic with extreme changes of mortar contrast.

9 We shallattempt a functional explanation in terms of processes that seemnecessary for maintaining registration of borders. it is hoped thatunderlying physiological mechanisms may soon be identified,explaining how the functions are Suggested explanation the Border - locking theoryFor visual displays such as printing or television, it istechnically exceedingly difficult to obtain precise spatialregistration at borders, and where contrasting luminances orcolours should meet without gaps or overlaps due tomisregistration. These Border discrepancies are annoying andconfusing. Registration may be achieved by high stability of themechanical or electronic components. but neural components arerelatively labile. This problem is exacerbated by the recent finding(Zeki 1976) that visual characteristics such as luminance, colour,and movement are mapped in separate cortical regions. Byanalogy with the display registration problem it is remarkable thatvision is normally free of spurious lines, gaps.

10 Or coloured edgesat borders where regions of different luminance or colour registration does, however, seem to be lost in someconditions: (a) with extremely high luminance contrasts,especially at low luminance levels, and (b) for contrasting colourspresented with no or very small luminance differences(isoluminance) (Gregory 1977). The former producesdiscrepancies during image retina movement, which is hardlysurprising, as under conditions of extreme luminosity contrastretinal receptors have very different response times, which image retinal movement, produce spatial discrepancies ofretinally signalled positions. Under the conditions of isoluminantcolour contrast. borders appear markedly jazzy (an effect used, ifunwittingly, in Op Art) and at isoluminance there is instability andthere are relative shifts with movement. So we find similarBorder locking and the Caf Wall illusion3phenomena for both extreme and zero luminance contrast.