Since autumn of 2008, the CREAS HD video chip has been implemented in all Blu-ray Disc (BD) recorder models, promoted as a technology that will enhance all HDTV images. Sony leveraged its rich storehouse of advanced digital video processing technology in creating a new chip capable of reproducing video images with their original rich color gradation and detail to reverse the degenerating effects of digitization and compression. The result is CREAS, the world’s first technology with the capacity to recreate incomparably realistic video images with 14-bit processing and 16,384 levels of gradation. Measuring just 1cm x 1cm, this tiny chip is packed with the technology needed to upscale the image quality of BD recorders and players to a new level. In this article courtesy of Sony technology highlights, we finally learn what CREAS truly means.
Digital broadcasting and BD have already become mainstream technologies, and high-definition (HD) content is, by definition, content offering highly-advanced quality and clarity, but will it still seem so attractive once HD is the everyday standard? Because of limitations imposed by the compression format, BD-ROM is recorded in 8-bit.While terrestrial digital broadcasts are classed as HD, they have only 1,440 horizontal pixels, compared with 1,920 pixels for full HD. Even if the original video content has been recorded at a higher bit density, it needs to be compressed for broadcasting. This means the original video content cannot be reproduced perfectly.
CREAS is based on the concept that image quality can be enhanced by recovering and reconstructing HD video data lost during compression and other processes. Another concept that is built into the CREAS system is the ability to reproduce video with color gradation equivalent to 14 bits even on a TV with an 8-bit or 10-bit panel. Sony has used two technologies to realize these two concepts in the CREAS system: HD Reality Enhancer, and Super Bit Mapping for Video. The HD Reality Enhancer analyzes the feature points of individual pixels in video input and automatically optimizes each pixel according to the feature points detected. Data lost through compression and other processes is recovered, allowing the image to be reconstructed with 14-bit precision.
The video analysis block has multiple feature point analysis modules and can detect flat areas, detailed areas, edges, noise, subtitles and other aspects at the pixel level. This data is used to select the optimal process, and to control the intensity of processing. The video analysis block uses video analysis to control the following two factors – enhance–reconstruction of lost details, and smoothing–restoration of natural color gradation.
The system detects detailed areas in the image and restores the sense of detail existing in the original video data. It enhances the image by subtly modifying detail at the individual pixel level to restore data lost during data compression. Earlier enhancement systems, which uniformly increase sharpness across the entire image, are not practical, since they cause outlines to become unnaturally conspicuous and also enhance noise. By analyzing images at the individual pixel level, HD Reality Enhancer is able to restore natural depth and clarity through a process that enhances only those elements needed to restore the original sense of detail. Even if the original video has excellent color gradation, subtle variations are lost during 8-bit quantization. This can result in color banding in areas that should have smooth gradation. The smoothing system accurately detects color banding and restores the original smooth gradation at 14-bit quality. Smoothing is not applied to areas where there is no banding, so there is no unintentional loss of precision in textured areas.
HDMI, which transmits digital signals unmodified, is becoming the standard interface for linking video equipment to TVs. Deep Color is a new HDMI standard that supports 12-bit, high-gradation output. CREAS was designed to extract the full potential of the Deep Color specification by adding 14-bit precision. Sony BD systems equipped with HD Reality Enhancer have the power needed to support Deep Color output.
This gradation conversion technology reproduces images generated by HD Reality Enhancer with the equivalent of 14-bit gradation, even with HDTVs featuring 8-bit and10-bit panels. Low sensitivity at high frequencies is a characteristic of human vision with high-frequency signals appearing, to the human eye, visually averaged. This means that it is possible to achieve visually smooth gradation, even with a limited number of bits, by superposing minute quantization errors in high-frequency areas where sensitivity is low (noise shaping). Also, since this system always uses high-frequency areas, perceived noise is far lower than with conventional gradation conversion processes. Compressed 8-bit output without this technology causes the smooth gradation to be lost, allowing color banding to reappear. Sony’s exclusive SBM-V gradation conversion technology brings out the full potential of superb image quality produced by HD Reality Enhancer. Now it is possible to enjoy this superior image quality on any HDTV set.
