色序型顯示器色分離現象與閃爍現象之研究

Abstract

現代人的生活運轉與顯示器脫離不了關係，舉凡電視、電腦螢幕、手機螢幕都需要顯示器。然而現在的LED液晶顯示器，因為彩色濾光片的存在而使得能量的使用效率降為原本的33%，且彩色濾光片佔顯示器成本高達五分之一。因此從永續發展與節能的角度來看，如何發展一技術來提升顯示器的效能是十分重要的。

色序型顯示器的概念於1940年由Peter Carl Goldmark所提出，因其無彩色濾光片的特性，可望成為次世代節能顯示器。除此之外，無彩色濾光片色序型顯示器更有著許多優點，例如：影像色域廣、材料成本低、螢幕解析度高等等。然而，色序型顯示器卻存在著一個致命的缺陷——色分離。色分離現象會使得影像顯示時的品質大幅下降，造成觀看時的不適感。為了解決色分離現象，近年來有許多的想法被提出，種類大致可分為：動態顯示補償、提升顯示頻率、集中顏色於單一色場、降低三原色飽和度等。其中，Stencil色序法和LPD色序法對於抑制色分離現象的表現最為出色，然而這兩種方法各自有著一些問題需要被解決。

第一個問題在於產品開發的可行性低，此兩種方法都極度仰賴背光的分區調控，此外LPD色序法的演算法複雜度更是相當高，背光的分區調控會使得背光模組的驅動電路設計困難，演算法複雜度高使得我們對晶片運算速度必須有所要求。上述原因都會讓硬體或軟體的負擔增加，使得產品開發的可行性降低。另一個問題在於潛在的螢幕閃爍情形，Stencil色序法利用的概念是將顏色集中於單一色場，這類想法能夠有效地抑制色分離現象，因此受到較廣泛且深入的研究，但是過度的集中亮度於單一色場會使得色場間的亮度差異過大，衍生出潛在的螢幕閃爍情形。

如上所述，本論文將分為兩大部分，第一部分針對產品開發的可行性為考量提出第一個演算法，以新穎的觀點——影像邊緣，來看待色分離現象，發展出運算複雜度低、軟硬體負擔少，但抑制色分離表現佳的新一代演算法，可望使色序型顯示器成為普及產品的想法獲得實現。第二部分針對減少螢幕閃爍的可能性為考量提出第二個演算法，為「集中顏色於單一色場來抑制色分離」這類的想法提供另一個需要同時被思考的方向。此演算法利用背光與液晶的調控，重新分配色場的亮度分布，使色場間的亮度差異減少，降低螢幕閃爍的可能性。

Apparently, we cannot get away from “displays” in our daily life. Displays are indispensable components for TVs, computer monitors, and mobile phone screens. However, the transmittances of color filters are only 33%, which means 67% of the power is wasted. Besides, color filters take up to 19% of the overall material cost. Considering that the concept of sustainable development is highly encouraged, it is necessary for us to moving forward toward the goal of eco-friendly products.

Field-Sequential-Color liquid crystal displays (FSC-LCDs) has been proposed to promote the light-emission efficiency of LCDs. Since a FSC-LCD does not require color filters to render a full-color image, the optical throughput is three times higher than the conventional LCDs, which is promising for eco-friendly applications. Furthermore, FSC-LCDs are notable for lower material cost, wider color gamut, and possibly three times higher screen resolution. However, FSC-LCDs typically suffering from color breakup (CBU), which deteriorates the visual quality and makes it uncomfortable to watch.

To reduce the CBU phenomenon, many concepts have been proposed, including motion compensation, increased field rate, concentrating luminance into a single field, and local-primary-desaturation (LPD). Among these approaches, the Stencil-FSC and LPD significantly reduced CBU. However, there are some issues about using these methods. The first issue is the feasibility of implementation. It is highly demanded for these methods to apply with a local dimming backlight system. And the computational complexity is quite high for LPD. The segmented backlight or the complex algorithm cause high cost in both hardware and software, which hindering the possibility of real-time applications. The other issue is flicker phenomenon, which is neglected but cannot be ignored. For the concept that suppresses CBU by concentrating most of the image luminance in a single color field, such as Stencil-FSC, the luminance difference between the color-fields might be too large, leading to a potential flicker issue.

According to the aforementioned issues, we divided this thesis into two parts. In the first part, we proposed an algorithm considering the feasibility of implementation. We develop the algorithm with the brand-new perceptive — Edge information of the image content. The algorithm is talented with low computational complexity, low cost in both hardware and software, and the performance of suppressing CBU is favorable, makes it possible to become a widely used product. In the second part, we proposed another algorithm considering the flicker issue. The algorithm avoids the potential flicker and suppresses CBU at the same time, by redistributes the luminance of color fields with the modulation of backlight and liquid crystal signals, which provides a new perspective of balancing the CBU suppression and flicker reduction.