Technology that minimises video quality degradation during video relays has been developed by Fujitsu Laboratories. The new technology has been adopted as a “Recommended Practice” of the Society of Motion Picture & Television Engineers (SMPTE), an international professional association establishing standards in the motion imaging industries. The adoption of video transmission systems that adhere to this recommended practice further improve the quality of video content.
Conventionally, to broadcast news reports, sports and other events, weather forecasts, and other on-site relays, video is transmitted across multiple relay points. Every time video signal passes through the video encoders installed at each of the relay points, slight degradation occurs in the colour component of the image. The accumulation of this degradation has been known to lead to visible colour blurring.
The new technology, which is compatible with existing equipment, minimises the occurrence of colour blurring.
During video relays of footage such as news reports, sporting events and other kinds of events, and weather forecasts, communications satellites and networks such as land-based wireless networks and the Internet are generally employed. Because of the high-frequency in which such video relays are repeatedly reused for different programs and media, there has a need for high-quality video to be transmitted at a relatively low rate of around 10MB/sec. On-site video transmission usually involves setting up around five different relay points, and when video is repeatedly transmitted between these relay points, the eventual occurrence of visible colour degradation has become a considerable challenge.
In general, images can be separated into brightness components and colour components. It is also widely known that the human visual system has a lower ability to discern the resolution of colour components compared to brightness components. Consequently, when transmitting video at a low rate, the video transmission process usually involves using H.264/AVC(2) and other video encoding technology to shrink by half the horizontal resolution of only the video’s colour components with a video encoder. Technology has been widely used to then expand the resolution of the colour components to their original resolution using a video decoder, and then re-render the video together with the brightness components before playing.
Between the video input and output stages, however, the colour components of the video after transmission become slightly degraded due to the process of shrinking and expansion. After a single video transmission, the degradation of the colour components is extremely minor and has relatively no effect on video quality. But after multiple stages of video relay to multiple relay points, the colour component degradation accumulates, ultimately resulting in visible colour degradation.
In order to lower any potential hurdles to adoption so that the industry can quickly enjoy the benefits of the newly developed technology, it includes the following two features:
Designing the new technology with an emphasis on compatibility with existing equipment makes it possible for companies to gradually adopt the new technology. Degradation resulting from the halving and doubling of video colour components falls roughly into two categories: 1. Colour blurring resulting from the degradation of colour details; and 2. Colour displacement resulting from the misalignment of an image’s brightness and colour components. Colour component shrinking and expansion used in existing video encoder devices are often plagued by colour blurring. Although research has been done on the “perfect reconstruction condition” in which images can be shrunk and expanded without degradation, it has been impossible to commercialise such technology because of the unavoidable colour displacement that occurs when connecting it to existing video encoders.
By mathematically formulating the necessary requirements and seeking an optimal solution, however, Fujitsu Laboratories has developed one of the world’s first technologies for shrinking and expanding colour components that can simultaneously minimise colour degradation and colour displacement. Just by using a video device that employs the new technology, it is possible to minimise colour blurring, even after repeated video relays, and colour displacement does not occur when connected to existing equipment.
With existing technology, colour blurring occurs on the character after five and ten rounds of processing. When using Fujitsu’s new technology, however, even after ten rounds of shrinking and expansion, no degradation can be seen.
Fujitsu
