SAN JOSE – The H.265 High Efficiency Video Coding (HEVC) standard is about to be ratified, promising a new generation of higher resolution and more compact digital video products. The bad news is chip makers are afraid to design products using it.
Vendors may have filed as many as 500 patents relating to the H.265 HEVC technology. But so far just who owns what and how much they expect in royalties is unclear.
The current video codec, H.264, has been the basis of digital video products such as cameraphones and digital video recorders for eight years. Thanks to a patent pool created by the MPEG Licensing Authority, the licensing rate is said to be a flat 25 cents per chip maximum, capped at about $12 million per vendor. An overview of the group's licensing terms is available online.
“That’s 40 million chips, but for cellphone chip vendors, that’s not much volume,” said an executive at one chip company who asked to remain anonymous.
The MPEG LA claims it is working on a follow on. “We plan to begin facilitation of an H.265 patent pool this summer,” said an MPEG LA spokesman.
But chip vendors say the group is having trouble getting support for the new pool. Some companies who have H.265-related patents including Mediatek and Qualcomm do not want to join the group but prefer collecting royalties on their own, said the semiconductor executive.
For its part, "Samsung has not made a final decision on its participation in the H.265 pool. However, we will actively take part in MPEG-LA's call for patents essential to the HEVC," said a company spokesperson.
The scenario is already playing out in other areas. Via Licensing recently said it will have this fall a 20-company patent pool for Long Term Evolution, the dominant 4G cellular technology. However, big patent holders including Ericsson, Nokia and Qualcomm are said to not be participating in the pool.
Chip makers expect the H.265 HEVC standard could be ratified within two months. That would open the door for them to start taping out their chip designs—if the royalty situation is clear.
“HEVC has so many patent holders and some of them say they will not be part of the pool but want to collect royalties themselves,” said the chip exec. “If say 20 people all want to collect royalties it will kill the standard—we need a fixed cost, it cannot be variable,” he added.
“People hate paying MPEG LA, but they were the only ones to collect royalties on H.264, so at least the business model was simple and clean,” said the source.
H.265 HEVC is expected to offer 30 to 50 percent boosts in compression efficiency over today’s codecs, lowering the bit rate needed to support quality video transmissions. The technology could be a big enabler for a new generation of compact, wearable cameras. It will also enable a new generation of 4K resolution TVs and monitors.
Engineers envision wearable cameras the size of postage stamps. The lower bit rates of H.265 HEVC could enable such cameras more easily to send video wirelessly so their output could be monitored on remote smartphones or PCs and tablets via cloud services
Thus the standard is “very important for us, we will enable it as soon as possible,” said the chip exec. “I think 4K TVs and monitors are coming on strong, and will be a reality in two years,” he added.
By definition shouldn't a standard should be non-proprietary and non-discriminatory? Sounds like the vendor costs should be capped at the agreed upon $12 million unless there is some additional benefit provided that makes it worthwhile to negotiate an additional payment for some value added features.
I find it hard to believe that there are 500 key patents related to this standard. Many of those are probably extensions to the standard. Critical for some applications but not all. We have all seen what happens in the marketplace when standards stumble... they die fairly quickly while another one succeeds. It is in the best interest of all the parties involved in this standard to protect their investments to date by not being greedy. Otherwise they are at risk of losing any advantage or reasonable compensation gained by their efforts. The real issue is if some party is using a key patent to throw a monkey wrench into the works in order to gain an advantage elsewhere. Business is a cutthroat environment and most everyone has cards not visible.
One way to take the edge off of this is for a consortium to hold the key IP and license it for a single fee per use. How each consortium member gets reimbursed from that fee is then by agreement of the consortium members. Take the max limits off of the royalties and it would probably work better for everyone concerned as it would not give an advantage to the big number players. Let them profit from their IP investments instead. This promotes business harmony.
While this may be a noble goal, if there are 500 patents involved in this standard, can you imagine the massive amount of "human capital" that has been expended to come up with all the technologies that are needed to achieve the results of this standard?
To that end, the people involved in creating the intellectual property involved in those standards need to be paid as do the companies that invested the money, time, effort and risk in creating said standards.
It may seem like a noble thought "benefit and future of mankind", but really who will benefit? The companies that make the chips that include the compression? The companies that develop software and online services that use this compression in order to enable their business model? What makes them more worthy of the profit and spoils than the people who developed the underlying technology?
On a purely practical note, the odds of developing a complex standard such as this that is open source and not in the process violating patents is fairly slim. To that end, perhaps the best approach is to only include technology from companies willing to put their patents into a patent pool and/or at least set a fixed cost and standard legal agreement for said patents on submission so that their value could be weighed not only from a technical standpoint but economical and legal standpoint.
IEEE or Some other Standardization agency should come up with a open source variant of compression standard that can be accommodated easily and at less cost. It seems strange that all the chips are working day and night compressing the videos and the money goes to a single pocket from thought the world. The volume is so large that standardization agencies should work on it for the benefit and future of mankind.
"The H.265 High Efficiency Video Coding (HEVC) standard is about to be ratified, promising a new generation of higher resolution and more compact digital video products. The bad news is chip makers are afraid to design products using it."
The problem is that ratification process is heavily influenced by the companies who own these patents; so that no one else can get into designing/selling these devices or get a bite of the royalties. The simplest solution is that IEEE/Standard body should kick all these stupid companies and their technologies out of the ratification process, and only include technologies from companies who agree to the FRAND licensing terms.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.