Perseus, a new video compression technology unveiled at the NAB Show by V-Nova promises to allow more efficient use of spectrum to deliver SD, HD and even 4K ultra HD as part of a next-generation digital TV standard than previously thought possible.
Television broadcasters have a new technology to consider as they piece together the jigsaw puzzle that is their over-the-air transmission future.
Perseus, a new video compression technology unveiled at the NAB Show, which closes today in Las Vegas, promises to allow more efficient use of spectrum to deliver SD, HD and even 4K ultra HD as part of a next-generation digital TV standard than previously thought possible.
“Just by adding one or 2 Mb/s on top of an existing MPEG-2 signal, you can achieve the beauty of 4K ultra HD with Perseus,” said Guido Meardi, CEO and co-founder of V-Nova, the London-based company responsible for the new compression technology, during a phone interview April 15 from a party celebrating the favorable reception Perseus received at NAB.
While specifying MPEG-2 as part of a new TV standard is unlikely, the new Perseus compression technique also can work together with the latest MPEG compression standards or by itself to achieve higher video compression efficiency.
V-Nova, a five-year-old company with London headquarters, has kept a low profile while developing the compression technique. It unveiled Perseus to the press April 11 at a Mandarin Oriental Hotel ballroom in Las Vegas.
During the press conference, V-Nova executive chairman Eric Achtmann said Perseus achieves a greater than 50% compression improvement over existing compression techniques.
It tackles image compression differently than other algorithms by imitating some aspects of the human visual system, added Meardi.
“When you look at a person, you first see a head, and then a face and then features, then maybe a beard and then the white whiskers of the beard,” he said.
Perseus uses mathematical transforms that leverage correlations in an image to preserve detail efficiently during compression in a way that mimics vision.
“The important thing is that the mathematics to be able to do this need to be amenable to a hierarchical structure,” he said.
The fundamentals of MPEG compression, such as blocks, frequency transforms and motion vectors “do not scale hierarchically,” he added.
Perseus also employs a holistic treatment of image content, much as human vision does, rather than chopping images into blocks to be processed that later can become visible under certain conditions when decoded, as is the case with MPEG, he said.
The new compression technique is also “massively parallel,” taking advantage of the true processing power that has been available for years, Meardi said.
V-Nova waited until now to introduce Perseus to the television industry at large because the company wanted to have an “enterprise-level system” through which to demonstrate the performance of its algorithm to an industry known for skeptical engineers who can smell vaporware from a mile away.
However, while developing Perseus, V-Nova has had input from various TV industry sources, including Sky Italia, which is now implementing the compression technology for commercial distribution of content, Achtmann said.
On the NAB exhibit floor, Perseus is being shown in four different booths, including the Hitachi booth (SL3910), where it is being used in an ultra HD ecosystem that includes Hitachi’s 4K SK-UHD4000 camera and Hitachi Data Systems servers.
During the press conference, the company said Perseus also can compress HD to SD bitrates and SD to sub-audio-channel bitrates. Even greater compression efficiencies will possible as the technology matures, said Achtmann.
In the ballroom where the press conference was held, V-Nova demonstrated Perseus decoding 4K ultra HD video on an i5-based Intel NUC. That 4K content, encoded with the algorithm, was running at 4.5 Mb/s, he said.
The demo made the point that the Perseus decoder can be loaded onto certain legacy devices to deliver video compressed with the new algorithm.
“Broadcom supports what we are doing because of our ability to take a box that can accept an update and then provide a higher quality of service using an existing box,” said Achtmann.
During a brief interview April 15 with Rich Chernock, in the Triveni Digital booth, the company’s chief science officer and the person leading the ATSC 3.0 standardization effort, confirmed that the new standard will be “extensible,” in other words able to adapt to and adopt new, improved technologies as they become available.
While saying he was unfamiliar with Perseus, Chernock said such developments are precisely the reason why ATSC 3.0 will be extensible when it is finalized.
The television industry learned its lesson, being locked into MPEG-2 as part of the ATSC (A/53) standard, precluding it from taking advantage of at least two successive generations of compression algorithms, and the competitive advantages they might have offered.
It won’t do that again, Chernock said.