Local Dimming of LCD monitors
The LCD monitor and especially the LED backlight is the major power consumer (up to 60%) of mobile devices like smartphones, tablets and notebooks. For higher visual quality like high luminance, high resolution etc. the LED power consumption needs to be even increased. Local dimming is an effective method to ease this problem.
The image below illustrates the principle of local dimming. The left half image has 100% backlight (white) and set the transmission of the TFT pixels according to the gray values of the image. The right half image has a locally dimmed backlight (gray) and the transmission of the TFT pixels is individually increased/brightened. The combination of the dimmed backlight and brightened image produces the desired image at desired luminance.
The goal of the local dimming technology is to achieve various physical and visual advantages on an LED LCD monitor for multimedia applications as well as significant power savings to allow longer battery operation. The specific merits of higher visual quality include higher contrast, better lack level and perception of color and specific content-dependent effects.
The photograph on the left shows an Edge LED TV and an example of local dimming. The image on the panel consists of two half photographs with (left) and without (right) local dimming. The static contrast is considerably higher (>10 X). The clouding effect, visible at the corners of the panel, is decimated. In addition the power consumption is drastically reduced through local dimming.
In this project the local dimming process has to further consider the contents to be displayed, since many kinds of images exist for computer application.The architecture will allow the upper system to control the local dimming system for various image contents and applications. The outcome will be between maximum power saving at a noticeable visual quality enhancement and maximum visual quality enhancements at a noticeable power saving. The crucial method applied is a specific optimization approach. For example, the scanning sequence of the image pixels is sorted according to greedy algorithm, as the figure on the right depicts.
The specific objective of this project is a very efficient algorithm implemented on a GPU. Only a small/negligible portion of the GPU computation power should be allocated to calculate a brightened image for the TFT control unit and individual brightness values for the LED backlight in real time. The GPU based solution will offer significant benefits over special purpose hardware like ASIC or FPGA. The implementation is SW based and thus flexible to quickly adapt an OEM model. No HW cost is required. Beside these advantages, we will combine rendering and local dimming methods to achieve new features. Local dimming of 3D image is a further new terrain.