隨著LED亮度愈來愈亮，製作LED投影機不再是夢想，使用LED取代傳統燈泡光源不僅減小投影機的尺寸，還可以降低系統複雜度，LED的主要優點是顏色純、壽命長、低功率消耗、體積精巧、高亮度。不過LED也有缺點，其中一個是低照度值，加上LED投影機的集光效率不夠好，這些原因限制了LED投影機只能在完全暗室裏使用，另外一缺點則是適合的投影距離小於一公尺，理由是照度隨著距離平方成反比，如此投影畫面將不會大。

此研究主要探討的是解決在暗階(dim levels)時量測儀器靈敏度不夠高導致量測的三刺激值不準的問題，解釋如下：
所有投影機皆會在輸入訊號為零時漏光，這稱為black level，主要的原因是因為內部的雜散光。Black level會導致每一channel的色座標值隨著亮度變化而變化，我們已經知道在色彩特性量測中把black level量測準確是非常重要的，而解決儀器對black level靈敏度不高的方法已經被提出，不幸的是LED投影機的輸出值特別地低，量測儀器甚至是在輸入訊號不高時也不靈敏，而不只是black level而已。

所以為了增加準確性，在此研究中引用了兩個方法，其中一個是black level預測最佳化，另一個是把black level切割成三個channel的個別貢獻。對於暗階的問題，於此文章中我們提出一個迴圈程序，迴圈包含已經提出的black level預測最佳化以及用非線性內插取代暗階三刺激值，也就是說我們修正了所量測的black level和暗階三刺激值。

除了儀器靈敏度的問題之外，還有測量三原色加法性、色域、對比、均勻度、時間穩定性、頻譜穩定性、空間不均勻性、channel獨立性、空間獨立性，這些特性皆與某篇LCD投影機色彩特性論文裏的數據做一比較。最後比較各種model的色彩預測準確性，以及探討減少大量的測量資料及時間的可能性。

As brightness of LEDs is brighter and brighter, making LED projector is possible. Taking the place of conventional UHP (ultra high pressure) lamp with LEDs reduces the projector size and system complexity. The major advantages of LEDs are natural pure color, long lifetime, low power consumption, small size, and high brightness. LEDs also has defects, one is low lux value, it limits LED projector to perform only in absolutely dark room, another is the suitable projection distance is below 1 meter. The reason is that illuminance varys with inversely proportional to the square of projection distance, and gathering efficiency of LED projector is not good enough. Additionally, power and brightness of LEDs are not high enough to be an illuminant source at present. If so, the projection image will be not large.

The major research of this article is solving the problem of low sensitivity of measuring device at dim levels, that will cause inaccurate tristimulus, and explained below：
When digital input values are zero, display still outputs light. The phenomenon is referred to as display’s black level. The basic reason is internal scattering. The black level will cause chromaticity variation. It is already known measuring black level precisely in colorimetric characterization is very important. The low instrument sensitivity for black level had been solved and proposed. Unfortunately, output tristimulus for dim levels of LED projector are especially low. Measuring device is not sensitive enough for low digital values, not only black level.

To improve accuracy, two proposed method are employed. One is black level estimation, another is cutting black level into respective contribution of each channel. For dim levels problem, a routine procedure including proposed black level estimation and nonlinear interpolation of low output tristimulus in place of raw data is proposed, namely tristimulus of every dim level output as well as black level are corrected.

In addition to sensitivity problem, primary additivity, color gamut, contrast, uniformity, temporal stability, spectral stability, spatial non-uniformity, channel independency, spatial independency are measured, All the characteristics are compared to data in a paper about colorimetric characteristics of a LCD projector. Color prediction accuracy of various model are compared eventually, and we probed into possibility of reducing large number of measuring data and time.

[1] http://www.fuji.com.tw/shownews.asp?RecordNo=529
[2] www.pida.org.tw/optolink/optolink_pdf/87111810.pdf
[3] Matthijs H. Keuper, Gerard Harbers, Steve Paolini, “RGB LED Illuminator for Pocket-Sized Projectors,” SID Symposium Digest of Technical Papers, Vol. 35, pp. 943-945 (2004)
[4] G. Harber, M. Keuper, and S. Paolini, "Performance of high power LED illuminators in color sequential projection displays," 10th Int. Display Workshop, IDW'03, pp. 1585–1588 (2003)
[5] Won Yong Lee, Young Chol Lee, Kirill Sokolov, Hee Joong Lee,“Led Projection Display,” Proc. of SPIE, Vol. 5529 (2004)
[6] 大田 登原著, 陳鴻興、陳君彥編譯,”基礎色彩再現工程”，全華科技圖書有限公司 pp. 4-2~4-7(1992)
[7] R.S. Berns, R.J. Motta, M.E. Gorzynski, “CRT colorimetry, part I:theory and practice,” Col. Res. Appl. Vol. 18, pp.299-314 (1993)
[8] R.S. Berns, “Methods for characterizing CRT displays,” Displays Vol. 16, pp. 173-182 (1995)
[9] IEC 61966-6, Multimedia systems and equipment – Colour measurement and management – part 6: Front projection displays, International Electrotechnical Commision, April (2004)
[10] Y. Kwak and L. MacDonald, “Characterisation of a desktop LCD Projector,” Displays vol. 21, no. 5, pp. 179-194 (2000)
[11] M.D. Fairchild, D.R. Wyble, “Colorimetric characterization of theApple Studio Display (Flat Panel LCD),” Munsell Color Science Laboratory Technical Report (1998)
[12] ANSI IT7.215-1992, Data Projection Equipment and Large Screen Displays─Test Methods and Performance Characteristics
[13] J E Gibson and M D Fairchild, “Colorimetric Characterization of Three Computer Displays,” Munsell Color Sci Lab Tech Report (2000)
[14] Madár G, Kránicz B, Schanda J, “Colour reproduction with different types of projectors,” Laboratory of Colour and Mulimedia, University of Veszprém, Hungary, CORM (2003)
[15] Roy S. Berns, Scot R. Fernandez, and Lawrence Taplin, “Estimating Black-Level Emissions of Computer-Controlled Displays,” Color research & Application, Vol. 28, no. 5, pp. 379-383 (2003)
[16] A. Cazes, G. Braudway, J. Christensen, M. Cordes, D. DeCain, A. Lien, F. Mintzer, and S. L. Wright, “On the color balancing of liquid crystal displays,” in Flat Panel Display Technology and Display Metrology, SPIE Proceedings, Vol 3636, pp. 154–161 (1999)
[17] J.Y. Hardeberg, L. Seimea, and T. Skogstadc, “Colorimetric characterization of projection displays using a digital colorimetric camera,” Proc. SPIE, Vol. 5002, pp. 51-61 (2003)