CIELAB、CMC、BFD、CIE94色差公式之績效評估——以ABS塑膠材料為例

The Evaluation on Performance of Four Color Difference Formulae-A Case Study of Abs Plastic Material

色差研究在工業界應用是非常重要的，諸如色彩容許誤差控制、色彩恒常及條件等色評價、不同色彩複製系統之色彩逼真性評估及提供工業設計界一個均勻色彩空間。一個理想的色差公式允許觀看參數改變，應該包含色樣本間隙、背景色彩、照度、樣本之明度、衡量技術、色差大小及樣本大小等因素。這些觀看條件之改變在色差知覺上有相當大的影響，因此對於色差知覺之參數效果進行一系統性觀察是有必要的。 很多觀測變數會影響到色樣對的知覺性色差。世界照明委員會（CIE）屬下之TC 1-28技術委員會所提出影響色差評估之參數，即用以調查此一問題。本研究進行一系列實驗之目的，在研究不同觀測參數之效果。以ABS材質之色樣對，探討不同色差大小、背景顏色、光源、樣本間隙等觀測參數對知覺性色差之影響。 色樣共分佈在紅、橙、黃、黃綠、綠、青、藍、紫、洋紅及灰色等10個色彩中心。每一個色彩中心包含11組色樣對，其中有5組僅含彩度差色樣對，另2組僅含明度差色對及4組同時含有彩度差及明度差色樣對。最後這些實驗結果也將用於測試CIELAB、CIE94、CMC及BFD等四個色差公式。 在測試四種色差公式績效上，以CIELAB、CMC色差公式績效較優，CIE94、BFD色差公式績效較差；但整體而言，四種色差公式在本研究中仍然有不錯的績效。

Color difference research is important for industrial applications such as tolerance control, assessment of metamerism and color constancy, color fidelity evaluation of different color reproduction systems and a uniform color space for industrial design. An ideal color difference formula should include factors to allow for the changes of viewing parameters, such as sample separation, background color, luminance, lightness of sample, scaling techniques, magnitude of color difference and physical size of sample. These changes in viewing conditions could significantly influence the perception of color difference. Thus a systematic investigation of parametric effects on color difference perception is necessary. There are many viewing parameters might affect the perceived color differences of sample pairs, and the CIE Technical Committee 1-28 on Parameters Affecting color-difference Evaluation was therefore formed to investigate this problem. The objective of this research is to study different effects caused by various viewing parameters. We used sample pairs made of ABS to investigate how different viewing parameters, such as intensity of color differences, color of backgrounds, source of light, interval of samples, could affect perceived color differences. All samples were prepared corresponding to ten color centers. They were named as red, orange, yellow, yellow-green, green, cyan, blue, purple, magenta and gray. For each color center have eleven color sample pairs, five sample pairs that exhibited chromaticity differences with negligible lightness differences were selected, i.e. ΔL（superscript *） was kept as small as possible, while the other pairs included two pair having only differences in lightness and four pairs that exhibited a mixture of chromaticity and lightness differences. At last, the final experimental results were also used to test four color-difference formula: CIELAB, CIE94, CMC, and BFD. As to the tests of effectiveness for each formula, CIELAB and CMC are better than CIE94 and BFD. However, they were all quite effective in general.