Streaming Media Europe
HEVC is barely out of the gate, but the race to achieve better quality and lower bandwidth might soon leave H.265 in the dust, according to speakers at a panel at this year's IBC.
The amount of video coming down the track for carriage over mobile threatens to break the net unless encoding vendors and compression standards can keep pace. The lesson from a recent IBC panel is that this can happen—provided the industry pays for it.
"[Codec development] is not a charity. It is a business," said Tony Jones, head of technology, TV Compression, Ericsson.
McCann's Law [named after engineer and strategist Ken McCann] holds that the bitrate required to achieve broadcast quality video halves every 7 years. According to SMPTE fellow David Wood, current chair of DVB groups in 3DTV and UHDTV, and past chair of the World Broadcasting Union Technical Committee, this pattern was borne out in the developments of MPEG-2 and H.264/AVC and is also coming true for HEVC.
"In two years time the work on the next-generation MPEG compression after HEVC will likely follow the same curve," he predicted.
But this advance may not come quickly enough—or be good enough quality—for the needs of the mobile industry.
"When you understand that video spectrum is the most expensive spectrum on the planet you can see that compression is an area of huge focus for us," said Jones.
Ericsson—which has a stake at the table, developing encoders/decoders—made the point that there will be 13 times more video over mobile networks in 2020 than there was in 2014. Already half the traffic on mobile networks is video. By 2020 it will be more like 75%.
More HD services, and progressively UHD services, will drain capacity, he said. "We have to deliver video efficiently using technologies like LTE broadcast and we also have to compress it well."
Eric Achtmann, executive chairman of V-Nova, said broadcasters and operators already had a huge problem dealing with ever-increasing capacity demands.
"We recognise that current codec improvements are delivering a two-fold capacity improvement every five to seven years, but demand is growing at twenty-fold," he said. "Incremental improvements will not suffice and there’s a huge disconnect which needs addressing."
He argued that the installed base of receivers has become too large to make any change practical or cost-effective. "There are 2.6 billion smartphones in the market while 200 million payTV STB's are being deployed annually." The installed base is growing and it would simply cost too much to upgrade all of them, he said.
In addition, said Achtmann, the industry's "unprecedented" investment in capacity cannot carry on. "The U.S. is investing $1 billion every 10 days in extra bandwidth capacity and upgrades," he said. No matter how much money is thrown at the problem, he argued, that capacity will at some point be limited.
Having barely got out the door with UHD-phase 1, the industry is demanding a better pixelled version, variously with Higher Frame Rates, Wider Color Gamut, Higher Dynamic Range, and next-gen audio for which there are multiple contenders in each category attempting to pass through DVB, ATSC, and SMPTE committees.
"There is too much technology and not enough convergence," said Thierry Fautier, VP video strategy at Harmonic and president of the Ultra HD Forum. "We're living in a world where we can do all the encoding in software and decoders are programmable but this not helpful [to the complexity]."
Plus, if all of the above attributes are desired along with backwards compatibility to legacy hardware, then the question is "how can you process the raw video without introducing anything you don't want?" he posed.
Fautier reckoned HDR UHD might be standardised by 2017, and the full range of attributes including HFR by 2020 earliest.
"We need to develop better compression, and licensing is required if we are not to stifle that," he said.
"It's great that MPEG is royalty-free but I doubt the quality will match that of [a technology like] V-Nova," he added, in reference to Perseus, a proprietary codec developed by V-Nova which has been championed by cellular operator EE.
Jones laid out the arguments that create the conditions to improve compression efficiency. These include continued competition among encoder vendors; "without that nothing happens," he said. "If you have a monopolistic business you don't get innovation." He pointed to the telco industry, where voice codecs are specified and there's been little change.
"Wide adoption for wide scale of licensing makes R&D investment worthwhile," he said. "And we need standards. Open standards are critical. It's important to know you can rely on it, that it's peer-reviewed and it will do what it says it does. Free licensing is nice to have, of course, but free also inhibits the research investment."
Wood pointed out that if a company takes part in bodies like MPEG, ISO, or ITU then "no-one expects that you shouldn't be able to charge licence fees," he said, "but the rule says you should apply fair, reasonable, and non-discriminatory terms. The problem is that no one has defined what the word 'reasonable' means in this case."
This could create a field for lawyers as IP-patent holders argue over usage.
"There's no such thing as a free lunch," said Achtmann, whose company has yet to reveal the cost of licensing the Perseus technology. "When the cost of something critical goes to zero the investment drops. So the notion of saying 'free codec' and 'innovation' is a contradiction."
Page 1
No comments:
Post a Comment