Novelty of the Present Work

The author of the present work believes that for large colour stimuli the viewing situation in computer controlled environment and the set of displayable colours offer better inspection. A new psycho-physical experimental design can be introduced, which is different from a static visual situation and more appropriate to be used resulting in a more precise description of the large-field domain of the colour size effect. Temporal control of the large-size colour stimuli seems to be able to e.g. exclude or reduce chromatic adaptation and throw new light upon the phenomenon; therefore a more ‘sterile’

experimental setup is able to be built, with more focus on the effect of large size. In this

chapter, this near-immersive self-luminous aspect is investigated by simulating an almost immersed scene by the aid of a large plasma display panel (PDP) extended by a viewing chamber of mirrors and a typical cathode-ray tube (CRT) monitor. The possible underlying visual functionalities are also reflected on. These functionalities might have been neglected when observing surface colours in real illuminated painted rooms due to the diverse colour appearance in such situations. Considering the experimental set-ups of previous and recent surface colour researches, the author’s opinion is that, though it better resembles real situations, the complex scene has a noisy influence on the appearance of the colour of the walls of the painted rooms. The mixed illumination situation, possible shadows, textures of the surface, structural inhomogeneity may all confuse the single effect of the stimulus size, and therefore distract its essence and the possibility to trace its origin. This is one possible reason why the studies reviewed above had largely deviating outcome and that the overall average error of the predictions of the models built upon the experimental results decreased only to approximately ∆Eab*=9 from 18.

It was already described in Section 1.3 that the CIE 1931 colorimetric observer contains the colour matching functions (CMFs) of an average trichromatic observer, collected from colour matching experiments where the subtended visual angle of the matching fields was 2°, owing to the fact that there is no influence of rods in a viewing field of 2° in the human retina. To apply this model correctly, the background should be neutral and of medium intensity. Colour appearance models usually assume a 2° colour sample size in a 10° background and a set of parameters defining the viewing situation.

Experimental data e.g. for CIECAM02 were collected with set-ups satisfying the above conditions. From the cited references^42-47 it is apparent that (though, they were not considered to) neither the CIE observers nor CIECAM02 was designed to predict the effect acquired by visual experiments using different experimental techniques to assess the appearance of colour stimuli observed at 20° of visual angle and above. In the rest of the work the term ‘immersive scene’ refers to a general photopic viewing situation where a homogenous colour stimulus fills the visual field of the observer centrally or even totally, here simulated by what subjects were exposed to when observing the PDP by leaning to the opening of the booth.

1.6.1. Questions to Be Answered

In the two forthcoming Sections (2 and 3) two separate experimental series are discussed, correction formulae are introduced for (almost completely) immersive self-luminous

scenes and several questions are intended to be answered in connection with the colour perception of exceptional size stimuli, such that:

1. How systematic is the size effect regarding self-luminous stimuli, in terms of the correlates of these perceptual attributes in a defined colour space (e.g. CIELAB hue and/or chroma)?

Can a mathematical correction formula be given to predict, which attributes of colours change as it has been previously given for the surface colours? Is it sufficient to extend the recently suggested size-dependent model, or a new one should be developed for extremely large scales? Are these shifts uniform (independent) or are they cross-related? Is it possible to separate the perceived lightness change of the sample from its hue? I.e. do the same shifts occur for red tones as for instance bluish colours?

2. Does the 2° or the 10° observer yield a closer match to the colour appearance of the large stimulus?

It is likely that, perceptually, the 10° sample is closer to the extremely large area, as the CIE recommends it above 4° of visual angle. Since the difference between the colour matching functions of the 1931 and the 1964 colorimetric observers is negligible compared to the expected magnitude of the size effect (literature data) simply using 10° CMFs

instead of the 2° ones may not yield sufficient results either, though it seems reasonable that it performs better. No literature data was found that a 10° matching ends up with a better result for large stimuli, in either size-effect related studies.

3. Does the luminance of the background of the samples affect the phenomenon?

Can the prediction be made more accurate if the standard size patches are presented on an achromatic background of the same luminance as the large sample? The increasing size of the stimuli may trigger some contrast effects, too. Samples becoming larger tend to behave as unrelated colours regarding that their latitude inhibits the visual system to compare them to any other stimuli in the visual field, while for small samples it is effortless. Comparing the large colours with colorimetrically the same small sample on a same luminance neutral background suggest that their perception will be closer than on a different luminance background.

4. Related or unrelated colours are the case?

Though, the colour of the immersive stimulus is in itself an unrelated one, if it is compared to a related colour, what are the major attributes observers judge, match and relate? Will either physical measure of the immersive colour correlate with either perceptual attribute of the visually matching related colour stimulus?

5. Is the examination in terms of the perceptual colour attributes the best approach to understand what changes exactly if a colour stimulus is large?

Any systematic shifts at other levels of the visual process should be searched for, by referring directly to the origin of the phenomenon. If the effect creates systematic shifts in the perceived attributes of colours (e.g. lightness, chroma) then it can be suspected to take place in the later process (third stage) of the HVS, while if it creates regular modifications in the single LMS or the LMS-combined opponent signals then an earlier, near-retinal process could be responsible for it.

For peripheral colour perception phenomena, the critical region starts at 20°-25° off-foveal vision. Accordingly, for a 50° centrally viewed stimulus peripheral colour

perception influences are not intense (since the stimulus laterally ends at 25°). It was also reported that, for mediums size samples, peripheral colour perception phenomena can be suppressed and the same appearance can be obtained at 20° eccentricity than foveally if the size of the off-axis stimulus is physically increased.³² For a large (50°) homogenous field compared to a smaller off-axis patch can be considered an increased size peripheral stimulus, which further minimizes the possible peripheral influences. It is suspicious, however, that above this critical lateral visual angle (20°-25°), peripheral phenomenon, i.e.

opponent signal structure changes may affect the colour perception of centrally viewed stimuli that cannot be predicted any longer merely by the extension of the size dependent model⁴⁴ for broader visual angles. (50° was the maximum visual angle during the

observations, of which results the size-dependent models introduced in Refs. 43 and 44 were built on.)

6. Is this effect related to chromatic adaptation?

The size phenomenon for exceptional size surfaces was thought to be an adaptation effect.

In former experiments, this factor was discarded and there was no control of the viewing time (the time in which observers were exposed to the large stimuli). It is straightforward that the increased amount of adaptation might result in a fading of the chromatic content of

the stimuli if viewed continuously due to the bleaching of the cone pigments (practically the HVS’s effort to energy minimum). But the opposite was reported, namely, the

perceptual increase of chromatic content was observed for colour stimuli with the advance of their size.

The practical applications that would require modelling the colour size effect are numerous. There is an industrial demand to be able to predict the colour appearance of the same colour if applied on an indoor surface or on a façade surface of large size. Though, alluding to the size phenomenon through this architectural example was previously claimed to be inaccurate due to the mixed presence of several other effects, it is one of the major causes that are accountable for the visual mismatch of the dissimilar size stimuli.

Other applications for self-luminous stimuli include the more deliberate design of the display colours of large computer controlled colour devices that provide a large field. The newly set up Technical Committee No. 1-68 of the International Commission on Illumination is currently studying this phenomenon.

In the rest of this thesis, the term ‘large’ stimulus will be used for the near-immersive colour (extending almost the whole visual field) and ‘standard size’ or ‘small size’ colour for the 2° or the 10° matching field.

Data analysis and statistical calculations in the thesis as well as graphs and diagrams were prepared by the aid of Microsoft Office Excel spreadsheet program and SPSS for Windows application.

Figure I/3. Illustration of the experimental setup: the PDP with the mirrored booth, the separator and the CRT.