Motoe: Sony was particular about high picture quality from analog broadcast era. Since then we have been continuing with development and research. Sony first introduced our WEGA Engine for the CRT (Cathode-Ray Tube) televisions in 2003. Until then, the circuit consists of several blocks. To consolidate together in one small board, we named it the WEGA Engine.
CRT and LCD (Liquid Crystal Display) televisions have very different resolutions, especially when compared in terms of image noise. Improvements were made on the WEGA Engine and the birth of Sonyfs BRAVIA line introduced the BRAVIA Engine. The BRAVIA Engine helped evolve this noise-reduction technology.
Yamamoto : We developed algorithms (signal processing) for the picture quality circuit according to ever-changing trend and circumstances of TV technology and the market; such as the launch of digital broadcasting. We expanded BRAVIA Engine into BRAVIA Engine 2, and then BRAVIA Engine 3. We began to face challenges with noise-reduction technology in digital broadcast and gradation because of the panelfs limited specs.
From there, the picture quality circuit of BRAVIA Engine 3 evolved into gX-Reality PRO/X-Realityh. gThe X-Reality/X-Reality PROh engine incorporate Sonyfs unique super-resolution technology. Other picture quality improvements include high definition, noise reduction, and dynamic color reproduction technology. Considering the support needed for todayfs wide range of video content, improving picture quality is the strongest point; from native low-grade image quality to Full HD.
Motoe: To restore resolution analyzing one image is single-frame method, while to restore resolution analyzing motion from multiple serial images is multi-frame analysis, which brings big difference in motion picture resolution. Furthermore, classifying input video signals to several thousand image patterns into database and processing super-resolution by referencing each pattern data enables us to restore images as similar to original image as possible. Thanks to the super-resolution optimized by database and multi-frame analysis, resolution restoration even for motion picture can be done with low noises.
gX-Reality PROh is a combination of gX-Realityh and gXCA7h, a specialized picture engine. Noise reduction is improved by gX-Realityh, while gXCA7h improves super-resolution processing and gradation. XCA7 stands for gExtreme Creation Architectureh.
Motoe: gBeautiful video from low-grade content to High-definitionh; under this concept efforts began to develop. Driving force was to help our customers enjoy more beautiful picture on Sony televisions sold globally.
Yamamoto: I still remember vividly that the first simulation image was so fascinating that our development team members exclaimed gI want to buy this television definitelyh. All together we promised ourselves gI will develop what I can be content withh.
Yamamoto: It was a continuous struggle (laughs). Switching multiple databases and adopting various signals while maintaining the basic performance of super-resolution performance was quite difficult. We overcome difficulty through trial and error, utilizing the research result in high resolution technology including DRC. Through this effort, we were able to reach the outcome we dreamed from the beginning.
Yamamoto: First, let me explain super-resolution optimized by multi-frame analysis. The conventional single-frame analysis was a technology to restore resolution. This information is calculated by referencing each surrounding pixel of a still image in a single frame. Of course, analyzing each pixel in a single frame shows improvement, it is still more efficient and more precise when being able to analyze pixels not only in one frame but also in multiple frames; we refer to this information on the time axis. By doing this, we can better understand the characteristics of the image.
Yamamoto: We refer to the database when we restore the information obtained by analyzing multiple frames. It becomes possible to create more realistic and optimal super-resolution by process of analyzing wave form from video signals and increasing resolution of various patterns within picture characteristics, instead of simply interpolating pixels. This database has thousands of patterns that are available for High Definition (HD), Standard Definition (SD), SD up-conversion, IPTV and so forth. The database cannot be developed in a short period. It has evolved from DRC development and other technologies we accumulated from the past and can bring very high super-resolution quality.
Yamamoto: After improving the image quality based on the database, Super Bit Mapping for Video (SBM-V) applies a process to smoothen tone. Compensating for the tone of low bit rate video, SBM-V reduces banding noises that were visible in the tone of sky, for example, so that it looks as natural as possible. X-Reality PRO also has a process to maintain high gradation according to the panel resolution.
Motoe: X-Reality PRO can control motion by the sub-pixel. The motion picture cannot be well reproduced on a pixel-by-pixel basis, as in many cases the picture moves by less than a pixel from one frame to the next. X-Reality PRO uses the sub-pixel control allowing better enhancement of the video resolution. Sub-pixel control may cause side effects such as adding unintentional noise without exact and appropriate control. We can control these effectively with X- Reality PRO.
Motoe: It is possible to improve picture quality for digital satellite broadcasting and digital terrestrial broadcasting. For example, gX-Reality PROh can reduce flickering and compression noise due to limited transmission bandwidth of digital broadcasting. This can improve the overall picture quality including line flickering and jagged noise even from conventional SD broadcasting, DVD and other SD content.
Yamamoto: X-Reality PRO also supports low-resolution images, which have increased in recent years, such as Internet Protocol Television (IPTV); YouTubeTM for example. Such content was intended to be viewed on a much smaller screen and the feeling of low-quality was noticed when moved to big-screen. X-Reality PRO improves the quality to reduce jagged noise.
Motoe: For example, a scene where the video camera moves slowly such as aerial scenes is one of typical scenes. With actually broadcast program, we donft have absolute still image, but either the camera or subject is moving. It is possible to improve the total picture quality by removing outstanding noises especially in slowly moving scene. Experience the beautiful imagery of gX-Reality PROh in current BRAVIA line.