Compression Technology In the Next Generation Broadcast Environment
The evolution toward more advanced video compression codecs is creating new opportunities and challenges for content providers. The transition from MPEG-2 to MPEG-4 has helped to shrink the bandwidth required for SD and HD programming, but the transition requires new capital expenditures and business challenges.
A new standard is being developed for High Efficiency Video Coding (HEVC), or H.265, which promises to deliver 30 percent to 50 percent more efficiency than the H.264 standard. Dave Kummer, CTO, EchoStar, says that the first HEVC equipment is expected to rollout next year.
“H.265 will allow EchoStar to efficiently deliver 4K and better 3-D services to customers, and perhaps use it in green-field operations to significantly reduce bandwidth requirements,” says Kummer.
The transition towards HEVC could take five to seven years if the economy is slow, but may come faster if terrestrial TV broadcasters adopt it. Richard Doherty, a research director at technical analysis firm Envisioneering Group, says, “The real key part of this is what the broadcasters may do with H.265, since they are currently stuck with MPEG-2 for broadcast TV. If they go to H.265 for the next generation of over-the-air broadcasts, that would change things more dramatically. If they make the decision in 2014, that will have a huge effect on satellite and cable adoption.”
The MPEG-2 compression standard has matured significantly since it was first established in 1996. Better encoders have helped to reduce the bandwidth required from 8 Mbps to as little as 2 Mbps. At the same time, the cost of the encoders has dropped significantly. The decoders are also being built into standard digital TVs, which has helped to further reduce the costs.
“99 percent of all Americans will get TV in MPEG-2 between now and Christmas,” says Doherty. In the United States, all digital TV is MPEG-2 based, except for a few ATSC mobile DTV channels being delivered in MPEG-4.”
The Dawn of MPEG-4
MPEG-4 (H.264) content distributors are seeing a 40 percent to 50 percent gain in bandwidth, allowing almost a doubling of channels. However, this gain decreases for some of the live sportscasts that have to be at a higher quality rate or for other content when there is a contract clause.
MPEG-4 makes the most sense for DTH operators like Dish and DirecTV as it allows an operator to transmit more content over limited satellite bandwidth. Doherty notes, “In North America, we see Dish embrace MPEG-4 more than DirecTV, which still has a lot of MPEG-2.”
DTH operators are upgrading consumers when a hard disk drive used for PVRs fails. Doherty explains, “When you have a PVR box that fails that is a good time to upgrade. EchoStar has been doing it for four years, and DirecTV for about three years.”
However, this can lead to bandwidth inefficiencies when the DTH operator has to maintain parallel MPEG-4 and MPEG-2 broadcasts. Doherty said some channels are duplicated. For example while some of the newer HBO channels are only sent in MPEG-4, some of the older channels like HBO East and HBO West are still simulcast in both MPEG-4 and MPEG-2. Doherty explains, “My understanding is they still have millions of MPEG-2 only receivers out there.”
Kummer says EchoStar has already moved to MPEG-4 for all its HD services, and has taken the pain of replacing customer HD MPEG-2 set-top boxes in the field to allow them to move to MPEG-4. “We were fortunate that we made the switch to MPEG-4 as soon as possible to reduce the cost of field replacement,” he says. “In fact, we are certainly the largest MPEG-4 service provider in the world (remember, we have Eastern Arc all MPEG-4, as well as Western Arc at MPEG-4 for HD). The bandwidth benefits have certainly been well worth it in the long run.”
For over-the-air broadcast, some U.S. operators are testing out MPEG-4 for distributing smaller format video for mobile devices using only 10 percent to 20 percent of the bandwidth of a conventional MPEG-2 signal. Only three cities in the United States currently have a substantial lineup of MPEG-4 over-the-air broadcast channels including Washington, D.C., Chicago and Seattle. A few others cities including New York and Dallas are testing the services on one channel. These signals can be decoded using a pocket receiver produced by RCA, Samsung or Hauppauge Computer Works.
Vendors are also working on various approaches to squeeze even more bandwidth. Benoît Fouchard, chief strategy officer at Ateme, says, “On the compression efficiency front, our latest generation of hardware encoders now implements all the tools of MPEG-4, even some that were perceived redundant in the earlier days of MPEG-4, such as PAFF versus MBAFF. These are two techniques to switch between interlaced mode and progressive encoding. And we found that by combining the two, we get even more efficiency than by using just MBAFF.”
Ateme is also working on better visual models for quality that are perceived by the human brain as opposed to machine measurements. For instance, they have developed algorithms to analyze the content and identify areas of interest where the human eye will naturally focus so that it can then allocate more bits to code that area perfectly, possibly at the expense of a meaningless background.
Other techniques, like Ateme’s Smartrate, focus on optimizing video quality over multiple channels. “To deliver good and constant quality on a single channel, the encoder is making decisions on how to allocate the available bandwidth across multiple pictures, and across macroblocks and line within a picture,” Fouchard explains. “What the Smartrate does is that it extends that optimization across multiple video channels. So when one channel needs less bandwidth (for example, because it is a fixed camera with little motion), that bandwidth is made available to enhance another channel on the same pool. The whole pool bandwidth (typically a full transponder) is constant, but individual channels consume more or less bandwidth over time depending on their complexity.”
Backhauls and Fronthauls
Progress has been slower in transitioning to MPEG-4 in backhauls from field crews, and fronthauls to cable TV headends. In both cases, content providers face financial and technical challenges in maintaining compatibility with existing MPEG-2 equipment. There are tradeoffs in balancing the reduced satellite bandwidth requirements with the cost of more expensive MPEG-4 equipment. The costs of MPEG-2 receivers are now down to about $500 per IRD, compared to about $2,000 per IRD for an MPEG-4 receiver.
Peter Ostapiuk, Intelsat’s vice president for media product management, notes that progress has been slowest in areas like Latin America with an established base of older IRDs. Improvements in MPEG-2 encoding are also helping to improve the attractiveness of staying with MPEG-2. “MPEG-2 has been finely tuned, and new features are being added in pre-filtering and preprocessing,” he says. “You can comfortably do SD in MPEG-2 at bitrates of 2 Mbps to 3 Mbps. When it came to market in the mid-1990s, the gold standard was 8 Mbps. With MPEG-4 it is down to 1.5 Mbps.”
There are greater financial incentives for transitioning HD content to MPEG-4, as it allows content providers to shrink 90 Mbps video streams down to 12 Mbps for transmission to the cable TV headend. There was also a much smaller base of MPEG-2 HD equipment. Ostapiuk says, “It made sense to do the MPEG upgrades for the HD signals first because the IRD population was smaller.”
The business case for moving to H.264 has improved as the equipment and algorithms have matured. Tom Lattie, vice president of product management at Harmonic, notes that it took several generations before H.264 compression equipment could fully deliver on its promises.
“Early adopters had to make concessions with regards to video quality and bandwidth gains in addition to contending with high capital expenditures,” says Lattie. “Based on how it has matured, there is little reason not to migrate to H.264 today. The codec fully delivers on its promise of exceptional video quality and optimized bandwidth allocation, while the mainstream nature of the codec technologies keeps capital costs at a reasonable level. Often broadcasters have to support MPEG-2 at the edge of their networks, but recent technology advancements have made it cost-effective to deploy high-quality transcoding technologies at the edge.”
But according to Ostapiuk, the promise of MPEG-4 has been held back to some extent by new complexities of managing multiple protocols. “Legacy cable systems around the world and particularly in the United States are MPEG-2 plants. Tens of millions of set-top boxes in consumer homes are only capable of MPEG-2. It was fine if you could take the signal off the satellite, which helped programmers control the cost of transmission, but then it had to be converted into MPEG-2 to go to consumer houses,” says Ostapiuk.
Equipment vendors have introduced a new category of equipment called a transcoder in order to provide the benefits of better compression with H.264 and maintain compatibility with existing MPEG-2 consumer equipment. Ostapiuk says transcoder receivers currently cost about $3,000 to $4,000 per channel.
In the backhaul, operators would like to find a way to move towards MPEG-4 without having to replace functioning MPEG-2 equipment with years of useful service life. This is creating a challenge in the business department in terms of calculating the value of transmitting more programming with the capital costs of replacing equipment. With H.265 just around the corner, there are also concerns that new MPEG-4 equipment purchased today will need to be replaced in a few years.
“Operators now investing in MPEG-4 are trying to amortize that investment over time. One of the difficulties in the video business is that the product lifecycle used to be much longer,” Ostapiuk says. “MPEG-2 equipment first came in 1995, and there is still a lot of it today in 2012. Meanwhile, MPEG-4 got started five years ago and H.265 will be coming out in two to three years when MPEG-4 is only seven years old. These new investments can only make sense when the equipment is being used for a certain number of years.”
“Broadcasters that have invested in MPEG-2 are slowly moving to MPEG-4 if it makes economic sense,” says Dick Tauber, vice president, transmission systems and new technology at CNN. “You cannot just throw out the encoders you bought a couple of years ago. If you are at a point of replacing the equipment, then moving from MPEG-2 to MPEG-4 makes sense. The distributors saw more immediate value and the potential for opening up new channels in going to MPEG-4. In order to work efficiently with the contribution feeds for our own equipment, we need to change everything over in a routine way that will get things done quickly, otherwise we would need to maintain two sets of gear and two sets of paths. We want to make the change in one fell swoop.”
Ericsson is offering multi-format encoders as a baseline feature to encourage customers to upgrade their old equipment today, while keeping their options open for the future. This allows an operator using MPEG-2 today to replace the receiver and continue to transmit in MPEG-2 until all of the receiver sites have been upgraded.
Matthew Goldman, head of TV compression technology at Ericsson, says, “If a content or service provider has been using MPEG-2 for years and would like to switch to MPEG-4 AVC, this may involve the need to convert tens to thousands of receivers. The operator would need to roll new equipment out slowly, and then can switch to MPEG-4 AVC when the rollout is complete. With a multifunction device, they don’t have to make a one-time choice that could impact them negatively in the future.”
Goldman expects the first release of the HEVC standard to be technically stable in the near future and to be approved in January. He expects it to be adopted initially by DTH broadcasters, which have historically been the first to use a new encoding scheme. The first HEVC decoding silicon will come next year, and the first receivers and set-top boxes likely would begin shipping by the first half of 2014.
“The next big step for HEVC will be the development of a profile to support fronthaul and backhauls,” says Goldman. “These feeds are normally transmitted in a higher quality 4:2:2 format, compared to the 4:2:0 format used in consumer video.”
HEVC could help pave the way for emerging standards for 4K and Ultra-HD video, which will deliver four times and 16 times as many pixels as HD. Ostapiuk adds, “Even with aggressive compression, Ultra-HD will still require over 1,000 Mbps using MPEG-4.” Once the standard has been ratified, Ostapiuk expects that all of Intelsat’s platforms will be upgradeable to pass the HEVC signal.
The transition towards HEVC could take five years to seven years if the economy is slow, but may come faster if terrestrial TV broadcasters adopt it. Doherty says, “The real key part of this is what the broadcasters may do with H.265, since they are currently stuck with MPEG-2 for broadcast TV. If they go to H.265 for the next generation of over-the-air broadcasts, that would change things more dramatically. If they make the decision in 2014, that will have a huge effect on satellite and cable adoption.”