Color Mixing Simulator For Display Surfaces Based On Human Color Vision

Abstract

This paper proposes a simulator that analyzes the color mixing process in large display units based on human color vision. The authors have previously proposed calculation model for the color mixing process. With this model, we can calculate the viewing distance at which humans distinguish the pixel structure, but the model does not give us the appearance of any particular image. Therefore, the proposed simulator enables us to understand how mixed colors are perceived without actually manufacturing a large display unit. It also supplies the output of the resulting image. First, the simulator makes test samples that reproduce the pixel structure. Next, the spatial frequency characteristics of each of the R, G and B pixel dots on the display surface are calculated. Thirdly, the visual characteristics in the spatial-frequency domain for each color are realized as a transfer function of a low pass filter, where the viewing distance is used as a parameter. Fourthly, in the spatial-frequency domain, the display surface characteristics are multiplied with the characteristics of human vision for each color. Finally, the simulator converts the spatial-frequency characteristics to the spatial characteristics, and composes the appearance of each color to show the resulting color mixing image. Using this simulator, we tried to produce the appearance of images at the different viewing distances. As a result, we found that the longer the viewing distance becomes, the less recognizable are the pixel dots and they look to have a more uniform, mixed color. These results agree qualitatively with the appearance of images on the screens of actual large display units and prove the validity of our simulator. INTRODUCTION A large display unit clearly shows the rows and lines of pixels and their structure when viewed from nearby, which prevents the displayed images from appearing with their colors evenly mixed. In order to discuss this inconvenience in visual perception, the authors have previously reported a calculation model that considers the characteristics of the display surface and human vision. This model treats the color mixing in images shown on display units as a phenomenon in which repeated patterns of pixel arrays are influenced by the visual characteristics in proportion to the viewing distance from the display surface. With this model, we can calculate the viewing distance at which humans do not recognize the pixel structure. However the model does not give us the appearance of any particular image and does not support easy understanding of the color mixing results. Therefore, this paper proposes a simulator that can output how the image appears based on this calculation model. First, a color mixture model based on the visual characteristics and how to realize the model in simulation programs are discussed in the second section. Next, to verify the validity of the proposed simulator, some simulation results are shown and discussed. COLOR MIXTURE MODEL A calculation model of color mixing The authors have previously proposed a modulation transfer function (MTF) model of the assimilation phenomenon based on visual characteristics (Nonaka et al. 2005). This is a model based on the spatial frequency response concerning luminance, and corresponds to Area II shown in Figure 1 (Hase et al. 1993). In this area, patterns are identified which react to the contrast of luminance in the presented samples. When discussing the assimilation phenomena of brightness and interference impressions caused by the pixel structures of display units, this area is one of contention. On the other hand, the color mixing phenomenon is based on the human characteristics of color vision. The phenomenon is influenced by the spatial frequency response concerning chrominance signals, and corresponds to Area I in Figure 1. This section describes the proposal of a color mixing simulator on the basis of these relationships. First, an outlines of the calculation model and the simulator are discribed. Next, the spatial-frequency characteristics of human color vision are expressed. Thirdly, the color mixing process of R-G stripes is interpreted by using the color mixing model. Proceedings 21st European Conference on Modelling and Simulation Ivan Zelinka, Zuzana Oplatkova, Alessandra Orsoni ©ECMS 2007 ISBN 978-0-9553018-2-7 / ISBN 978-0-9553018-3-4 (CD)