Keys to Solving the Strategic Challenge Posed by MSOs’ Reliance on MPEG-2

SP1113_MI_Euclid

Euclid Discoveries’ Software Delivers Big Gains on Existing Encoders

By Richard Wingard, CEO, Euclid Discoveries

Introduction

As cable companies strive to maximize encoding efficiencies across an ever-more complex pay TV service matrix, there’s nothing more vital to strategic success than to find ways to drive better compression performance on legacy MPEG-2 channels.

In fact, it can be argued that any strategic game plan that aims to sustain competitive advantage without incurring untenable capital costs absolutely depends on a cable company’s ability to increase quality and efficiency in MPEG-2 encoding. Everyone understands that migration to MPEG-4 H.264 and, eventually, all-IP distribution of premium content is essential, but it’s also well understood that prudent execution of that migration will require reliance on the embedded base of MPEG-2 set-tops with allocation of 6 MHz QAM channels to MPEG-2 transport for a long time to come.

Consequently, not only must cable operators continue to find ways to reduce the number of QAM channels devoted to MPEG-2 in order to free up QAMs for multiscreen services and DOCSIS broadband capacity; they must also meet the quality challenges posed by ever larger HD screens and the growing prevalence of 1080p HD from other sources, including Blu-ray, fiber-based telcos, DBS, and even over-the-top (OTT) suppliers, if they want to ensure the quality of their customers’ viewing experience is second to none.

Over the past several years, ongoing gains in hardware density and processing techniques have allowed encoder manufacturers to produce significant improvements in MPEG-2 compression efficiency. This enabled cable operators using variable bit rate (VBR) statistical multiplexing techniques to achieve MPEG-2 stream densities of up to three or even four HD channels per QAM. But such advances come at great cost, requiring replacement of existing hardware and frequently imposing compromises in quality.

Moreover, the prospects for further hardware-based advances in MPEG-2 encoding are dim at best, given the amount of effort that’s already been expended to achieve maximum performance. Indeed, as shall be seen, techniques that consume extraordinary amounts of processing power to achieve incremental improvements in encoding performance are counterproductive, diverting valuable resources from the opportunity to exploit higher hardware densities to reduce the costs per encoded stream.

A more promising vein of development lies with Euclid Discoveries’ EuclidVision software platform, where innovations complementary to but extending beyond the techniques embodied in MPEG-2 specifications can deliver ongoing improvements in compression efficiency while freeing up processing power to increase output volume on high-density hardware systems. These gains are achieved without requiring any changes in the decoding process, which means they can be fully realized on existing MPEG-2 set-tops.

Euclid brought its innovations to light with the debut of EuclidVision at the 2013 CableLabs Summer Conference in Keystone, Colo. There, and at an ensuing encoding “bake-off” at CableLabs’ headquarters in Louisville, Colo., Euclid demonstrated under rigorous testing procedures that its platform can be integrated into an MPEG-2 compliant encoding system to produce significant gains in compression efficiency, measuring up to 30 percent on high-action video content.

Equally important, Euclid demonstrated that when its processes are applied to certain types of pictorial content and scene changes that are especially challenging to traditional MPEG-2 encoding processes, the resulting improvements significantly impact the viewer’s perception of picture quality. By addressing these weak spots in HD rendering on MPEG-2, operators can prevent them from becoming impediments to a good viewing experience on large-screen TVs.

Clearly, if there are ways to significantly enhance MPEG-2 encoding performance at far lower costs on existing hardware platforms, cable operators have much to gain by letting suppliers know that such enhancements will be taken into account as they assess investments in encoding technology. By elevating improvements in MPEG-2 encoding to the same priority level accorded developments with H.264 and its successor, H.265, also known as HEVC (High Efficiency Video Encoding), operators will give suppliers a much-needed incentive to focus on opportunities to improve MPEG-2 performance as they compete for new business.

In fact, the incentive to implement EuclidVision on manufacturers’ encoders becomes even greater when one considers that the technology advancements introduced by EuclidVision are also applicable to H.264 and HEVC. As with the EuclidVision MPEG-2 software suite, techniques applied in H.264 and HEVC encoding are compatible with the standard compression technologies and do not require any enhancements or changes to the decoding processes at the set-top or other receiving devices.

The implementation of these improvements across all generations of widely used MPEG video encoding modes would represent a win/win for operators and encoding suppliers alike. Because EuclidVision is available not as a proprietary enhancement specific to only one encoding platform supplier but as a benefit equally accessible to all, operators can stipulate their preference for such enhancements without undermining the advantages of a multi-supplier ecosystem.

Euclid is not an encoder manufacturer, nor is it positioned to compete with suppliers of encoding technology. Instead, its mission is to bring to the cable industry a suite of solutions derived from years of experience developing compression algorithms for advanced communications systems, algorithms it has refined to significantly enhance encoding performance in the pay TV sector. By virtue of this diverse experience and the specialized expertise of its developers, Euclid is introducing a broad range of patented techniques which until now have not been available to developers in pay TV compression technology.

Critically, suppliers of encoding platforms for the cable industry who elect to integrate the EuclidVision algorithms into their systems will be able to benefit from ongoing advances underway at Euclid, which will continue to generate performance enhancements on their platforms not only for MPEG-2 processing but for MPEG-4 and HEVC encoding applications well into the future. This is especially significant for suppliers of multi-protocol transcoding systems that are designed to process all the major formats used in pay TV distribution. By utilizing the EuclidVision SDK (software development kit) with support from Euclid experts to integrate the technology into their systems, these suppliers will be well positioned to augment their implementations of the standard codecs with significant performance improvements enabled by Euclid.

The MPEG-2 Encoding Challenge

While much of the cable industry’s recent development activity in compression technology has been focused on satisfying transcoding requirements for streaming premium content to IP-connected devices, operators and their suppliers cannot afford to let up in their search for ways to improve MPEG-2 encoding performance. Intensifying competition from telco, satellite, and OTT suppliers, combined with rising consumer expectations for a superior viewing experience, is raising the bar on requirements for MPEG-2 well beyond previous expectations.

Forces Extending the MPEG-2 Life Cycle

Despite the compelling efficiencies offered by MPEG-4 and HEVC, operators’ ability to compete while relying on MPEG-2 is a make-or-break issue for the industry. There are approximately 112 million legacy MPEG-2 digital set-tops in U.S. cable households, according to Frost & Sullivan principal analyst Dan Rayburn.1 Rayburn estimates replacing these with MPEG-4 capable set-tops could cost in excess of $11 billion.

The replacement conundrum confronts operators in other mature cable markets as well. With about 520 million cable TV subscribers worldwide,2 the majority of whom are connected via MPEG-2 digital set-tops, the replacement costs of migrating to MPEG-4 would total over $50 billion, based on Rayburn’s modest per-box cost of $100. That’s more than twice the $22.2 billion researcher IHS says was spent on set-tops of all types by cable, telco, and satellite providers worldwide in 2013.3

Of course, no such tidal wave of replacements is in the cards. MSOs have had great success keeping pace with market requirements for ever more HD linear and on-demand content without having to rely on MPEG-4 to reduce bandwidth consumption, even as they take steps to enable incremental transition to MPEG-4 in the years ahead. A major factor in this strategy has been switched digital video (SDV), which has allowed operators to continue using the core MPEG-2 service bundle to add dozens of channels without consuming more bandwidth. Analog bandwidth reclamation through use of DTAs (digital terminal adapters) to translate incoming digital signals for analog distribution in the home has been a major factor as well.

The deployment of hybrid MPEG-2/MPEG-4 set-tops and media home gateways (MHGs) has also helped conserve bandwidth, insofar as operators have begun to use MPEG-4 to support more bandwidth-efficient unicast distribution of VOD content to households equipped with these devices or to add linear content that’s available only on tiers that require installation of the hybrid devices. And, in cases where operators use MHGs with built-in transcoding and IP streaming capabilities, they’ve been able to support multiscreen distribution in the home without having to consume bandwidth to deliver all the content over IP from headends.

More importantly, by seeding the market with hybrid MPEG-2/MPEG-4 set-tops and media home gateways (HMGs) as new subscribers are signed up, existing subscribers opt for higher tiers, and aging legacy boxes expire, operators have set in motion the process by which they will eventually be able to switch over to all-MPEG-4 distribution within acceptable capital cost parameters. But these hybrids also allow them to prolong their reliance on MPEG-2 to deliver a competitive service. And, as the need to deliver pay TV content to connected devices outside the home intensifies, the MHG is becoming less a factor in mitigating bandwidth requirements for IP TV.

Adding to the potential lifespan of MPEG-2 is the fact that new cloud-based technologies designed to support advanced navigation and other next-generation features on legacy digital set-tops promise to reduce pressure to replace those devices even further. If such capabilities, now in trials and commercial rollouts in several cable markets in the U.S., Europe, and elsewhere, gain wide-scale traction, the industry will be able to prolong the transition to MPEG-4 and all-IP services even longer.

Basic Requirements for Sustaining Competitive Viability with MPEG-2

As operators take such steps to extend the MPEG-2 set-top replacement cycle, they must ensure two fundamental requirements are met:

  • Ongoing dependence on MPEG-2 should not deter them from satisfying the bandwidth requirements of broadband-based services, including multiscreen pay TV;
  • The quality of the MPEG-2 service must keep pace with consumer expectations as screen sizes increase and 1080p becomes the norm in competing HD services.

The QAM Allocation Agenda

Industry research projections underscore operators’ intentions to free up ever more QAM capacity for non-MPEG-2 transport. Most dramatically, the move to distributing linear as well as on-demand pay TV content to IP-connected devices, now in full sway among top-tier MSOs in the U.S. and abroad, requires new bandwidth allocations representing a simulcast of the entire cable lineup. While these transmissions employ MPEG-4, thereby cutting the per-stream bit rate in half compared to MPEG-2, they are delivered in unicast mode over the access network, which will consume ever more bandwidth as such multiscreen services draw more users.

According to statistics provided by Conrad Clemson, director of strategy and product management for Cisco System’s Service Provider Video Technology Group, the average number of QAMs devoted to multiscreen services per cable service group on MSOs’ HFC plant in the U.S. rose from eight in 2012 to 12 going into 2013.5 That was a year or more ahead of full transition to comprehensive multiscreen service portfolios by most MSOs, which will eat up substantially more QAMs. At the outset of 2010 the average number of QAMs devoted to IP pay TV per service group was 0.

At the same time, Clemson, writing in a January CED Magazine article, says QAMs devoted to SDV rose from 20 in 2012 to 24 in 2013, doubling the number used in 2010. Combined with QAM allocations for high-speed data and VOD, which held fairly steady over that period, the total QAM allocations per service group for services other than linear MPEG-2 channels rose from 18 in 2010 to 48 at the start of 2013. Not only will the move to multiscreen services continue to drive this upward spiral; the emergence of 100 Mbps broadband as an ever more pervasive top-tier broadband service is sure to have an impact as well.

The New Quality Imperative

Meanwhile, even as operators search for ways to pack more SD and HD MPEG-2 channels into as few QAMs as possible, they must ensure that the consolidation ratios take into account the impact of ever larger flat panel screens on quality requirements. This is a tall order.

Even though, according to Leichtman Research Group (LRG)6, 75 percent of U.S. households now have at least one HDTV set compared to 23 percent five years ago, and 51 percent of HDTV households have more than one HDTV, the appetite for larger, higher-quality displays is proving to be a hedge for CE manufacturers against the downsides of a saturated market, thanks in part to falling prices at the high end. Twenty-two percent of U.S. households purchased an HDTV set over the 12 months prior to the LRG study’s release in February 2013, at a mean purchase price of $680, which LRG says was 30 percent below the mean price five years earlier.

The ever-increasing number of programs available in HD is having an impact on set demand as well. According to LRG, the mean number of HD channels that respondents to the survey reported were available from cable, satellite, and telco providers stood at 77, up from 63 two years earlier and 29 five years earlier.

Supersized LCD TV sets larger than 50 inches in the diagonal dimension accounted for 27 percent of U.S. LCD TV unit shipments in the first quarter of 2013, up from 15 percent one year before, according to a TV Systems Intelligence update from researcher IHS.7 These large sets represented over half of all U.S. LCD TV revenue, at 53 percent, up sharply from 39 percent one year earlier, IHS says.

While demand has produced a slight uptick in prices for such sets in recent months, the longer term trend line has seen a sharp reduction in costs, with prices of 50-inch LCD TVs that once cost several thousand dollars now below $1,000. What it all adds up to is a cycling of HDTV sets through the home, with yesterday’s high-end 42-inch set moving to the bedroom as the 50-inch plus sets take over the living room.

This is not good news for cable operators when it comes to their reliance on MPEG-2, insofar as the larger the screen size, the lower the quality of reception for any given level of compression. A bit rate that may have produced an acceptable picture quality on a 42-inch set may not be sufficient to sustain that level of quality on a 55-inch set. Moreover, insofar as competitors such as telcos and OTT players have been able to exploit H.264 encoding to deliver higher quality pictures as the bandwidth available to them increases, consumer expectations as to what constitutes a good quality HD picture are much higher than they once were.

Adding to the challenges HD purchasing trends pose to cable operators, for higher income households who were shelling out $3,000-$4,000 for 42-inch sets a few years ago, a new level of quality and performance with big screens is coming into affordable price ranges with the aggressive competition underway among CE manufacturers to sell Ultra HD or 4K sets. In the third quarter of 2013, just five months after announcing prices on a new line of 55-inch and 65-inch Ultra HD sets, Sony said it was cutting the prices by $1,000 to $3,998 for the 55-inch set and by $1,500 to $5,498 for the 65-inch version.

This followed a similar move by Samsung, although its price levels were about $500 higher than Sony’s. And it helped spur LG Electronics to cut its prices to levels matching Sony’s.8

Not surprisingly, such pricing is contributing to faster-than-anticipated consumer adoption of 4K. Even before the latest price cuts, NPD DisplaySearch was predicting 4K set shipments worldwide would exceed 500,000 in 2013.9 By 2016, the researcher said, shipments will top seven million, with over two million sold in North America, second only to China with 2.6 million in projected shipments.

Given that the bandwidth requirements for 4K are four times those of HD for any given screen size, the delivery of channels in 4K resolution over MPEG-2 does not appear to be in the cards for cable operators. But the growing popularity of such sets further raises the pressure on operators to deliver superior HD quality over MPEG-2.

Overcoming the Limitations of Existing Encoding Platforms

While advances in hardware encoding systems have enabled operators to compress three HD and even four HD MPEG-2 channels per QAM, those ratios were set based on quality expectations two to three years ago. Clearly, as outlined in the discussion above, the bar on quality expectations has risen sharply and will continue to rise in the years ahead, putting in jeopardy aggressive QAM-packing ratios that were once deemed acceptable.

Moreover, operators can only implement the best capabilities offered by any given encoding manufacturer cost effectively as it becomes necessary to replace aging equipment or to add new capacity to handle higher volumes of encoding. Thus, in most instances, getting to those maximum performance ratios is an evolutionary process from one headend to the next.

Consequently, what’s needed is a means to improve on the installed encoding base while realizing even greater gains from the latest generation gear when installation of that gear can be cost justified. The enhancements for MPEG-2 embodied in Euclid Discoveries’ EuclidVision (EV) platform are designed to do just that. Whatever level of compression efficiency might be achieved on an old or new platform, EV delivers improvements that give operators the flexibility to get more out of MPEG-2 in whatever ways make the most sense, whether through compacting more channels into each QAM or by ensuring the quality levels on channels at a given ratio per QAM are sufficient to meet higher quality standards.

This is possible because Euclid has designed EV as a software enhancement to existing encoding platforms that operates on the video signal input in the temporal prediction phase of the encoding process. This positioning allows EV to overcome limitations intrinsic to MPEG-2 encoding without interfering with the transform, quantization and variable-length encoding functions of the encoder. As a result, the enhancements are completely transparent to installed set-top decoders, which treat the output of the encoder just as they always do in the deconstruction of the transform, quantization and variable-length encoding processes. (See Figure 1.)


Figure 1
fig1
The gains in compression efficiency enabled by EV make it possible to add another MPEG-2 channel to the QAM within the quality parameters previously set by the operator or to increase the quality of high-motion video at a given ratio of channels per QAM. Moreover, by tackling additional issues that are problematic to existing encoding platforms, EV can effect small improvements that can make a significant difference in the viewing experience, even though they may not register as having a big impact on quality analysis metrics.

Key Mechanisms Employed by EuclidVision

The Drawbacks of Traditional Temporal Prediction

By focusing on temporal prediction, EV is addressing one of the steps in MPEG-2 encoding, known as block-based motion estimation and compensation (BBMEC), that is most crucial to compression efficiency. BBMEC divides each video frame into a set of data blocks with standardized block sizes. It performs a local search for the best-matching region within previously coded video reference frames, thereby determining what degree of processing must be applied to capture the changes occurring in any given block of the target frame.

The limitations on compression efficiency resulting from how BBMEC is implemented in MPEG-2 have to do with the fact that finding useful data in block matches within a limited search range from previously decoded frames is a hit-or-miss proposition that can produce serious flaws in the encoded frames. For example, motion associated with a particular picture element is not always confined to the same data region from frame to frame. This is especially the case when the motion is fast, such as the feet of players in a professional soccer game.  Additionally, the moving element may not always remain oriented the same way even if it is confined to the same data region, as in instances of rotations, zooming or smoke, turbulent water flow, and other non-rigid motion. In any of these cases, the result is likely to be an inaccurate temporal prediction that produces a blurred or choppy rendering of the motion in the video.

There can also be problems with low-motion frame sequences, such as scene cuts, illumination changes, and occlusions (the blocking of one object by another). And because MPEG-2 does not perform spatial prediction, there can be problems in instances of high ambiguity, known as the aperture problem, where it becomes difficult to isolate the correct motion in a data block without reference to the surrounding pictorial elements in other blocks.

The EV Intelligent Modeling Process

Figure 2 provides a high-level view of the contrasting approaches of traditional MPEG-2 BBMEC and EuclidVision (EV). Essentially, EV neutralizes limitations in the MPEG-2 temporal prediction process though a much smarter approach to motion estimation that supports an intelligent search for additionaltemporal information across all blocks over a longer sequence of frames. This results in a significant improvement in temporal prediction that focuses the transform and quantization processes on the motion changes and other dynamics in the frame sequences that are consequential to the ultimate viewing experience.


Figure 2
fig2

The selective tracking enabled by EV’s modeling process serves to identify sequential changes resulting from fast motion wherever that motion occurs over multiple reference frames, from the target frame being encoded to the farthest reference frame. This focus on salient elements that are significant to the viewing experience eliminates wasteful processing on changes that are largely unnoticed.

At the same time EV uses advanced rate distortion optimization (RDO) to select the best temporal prediction among several candidates that may emerge in the search process. In other words, by virtue of tracking multiple macroblocks across multiple reference frames, the EV modeling process might generate multiple predictive solutions, which must be parsed to identify the one closest to the original source video. RDO then compares this result to the baseline MPEG-2 prediction, and if it’s an improvement, this becomes the information that is passed on for transform and quantization processing.

It should be noted that the expanded search employed with EV, by virtue of the selectivity of what is searched as determined by the intelligent modeling process, is very different from implementations of multi-macrobock, multi-frame searches that don’t employ such selectivity. Not only do such methods burn up a tremendous amount of processing power; they are not as accurate.  Thus, while RDO is used with all encoders, its application with EV’s intelligent modeling process produces a more accurate prediction that reduces the magnitude of the residual signal, greatly reducing the amount of processing required in all the downstream steps.

Two other important components in EV are essential to superior performance as well. One involves scene motion analysis, which identifies scene changes so as to trigger resetting of the EV trackers. Here again the distinction from other utilizations of scene motion analysis processes has to do with the fact that the reset tracking processes are employing EV modeling to produce better temporal prediction.

The other component currently included in EV processing for MPEG-2 addresses the aforementioned aperture or ambiguity problem, where data searched in the traditional BBMEC temporal prediction process all looks the same despite the presence of motion, making it impossible to distinguish one prediction from another. To overcome this problem EV employs multi-resolution tracking, which first performs the motion search at a very coarse level of resolution. This creates a contextual means by which the motion can be pinpointed so as to enable a fine search that produces a meaningful prediction.

Near-Term Directions

It’s important to note that in addition to the EV processes discussed so far, which underlie the 10-30 percent efficiency gains as validated in recent industry demonstrations, Euclid is refining and testing additional processes for temporal prediction that will soon be added to the platform to produce additional gains. These include frame-level look-ahead processing; sub-frame, model-based adaptive rate allocation, and software and systems level optimizations to enable embedded, real-time processing.

These enhancements will result in improved compression performance not only for high-motion sequences but for complex videos with multiple scene changes and low motion/high-ambiguity videos as well. Looking further ahead, Euclid is also applying all the processes discussed so far along with additional processes for application in H.264 and HEVC encoding. These new processes include the use of EV with feature-based or object-based tracking.

All of this adds up to the ability to model data at multiple fidelities, according to what the source data supports. In so doing, EV is able to exploit whatever underlying structure exists in the data to improve compression across all generations of MPEG codes through higher-level modeling where it is needed, while reverting back to standard prediction modes when that is sufficient.

Testing Methodology

The performance gains cited in this document are based on analysis performed across a select set of representative HD test videos, including sports, action movies, TV shows, and animated movies, utilizing the widely accepted metric known as BD (Bjontegaard Delta) -rate, which is now the standard way for assessing encoding performance in the HEVC community. The BD-rate metric is a single number comparing the performances of two separately encoded versions of the same video. These performances are reflected in the average bandwidth reductions or integrated gains registered on their respective rate-distortion curves, which comprise a range of PSNRs (peak signal-to-noise ratios) plotted against a range of bitrates (Figure 3).


Figure 3
fig3

To assess EV performance, Euclid testing obtains the BD-rates for each video sequence in its test set for a reference implementation of MPEG-2, together with the BD-rates for the same reference platform with EV integrated. Euclid testing also involves human visual evaluations where the respective outputs are checked for qualitative performance, looking at side-by-side, toggled, mirrored and composite displays.

For Euclid testing, the reference encoder is based on a high-performance MPEG-2 system developed from files and algorithms made available through FFmpeg. Were the same tests to be run using a proprietary MPEG-2 encoding platform, the improvements reflected in the BD-rates would be roughly the same, insofar as EV is additive to whatever basic and augmented processes might be brought to bear by any given encoding platform.

As shown in Figure 3, the BD-rates, using rate-distortion curves plotted over four different target bitrates – 6.6, 9.9, 11.1 and 13.3 Mbps – are calculated by dividing the difference between the areas to the left of the test (EV) and the reference (FFmpeg) encoder rate-distortion curves by the area to the left of the reference encoder rate-distortion curve. If the quotient is a positive number, the result represents the percentage gain in compression efficiency achieved by the EV encoder. For the EV test set, the BD-rates were all positive, registering 10 to 30 percent on the sports and other high-motion videos in the test group.

As shown in Figure 4, frame captures from one of the compared videos provide dramatic visual evidence of the improvements registered by the EV-enhanced encoding process. The top frame is from the reference encoder, showing a PSNR of 32.7 dB on a 25 frame-per-second interlaced sequence running at a bit rate of 6.6 Mbps. At the bottom of Figure 4, the same frame measured at the same bit rate from the EV-encoded stream registered a PSNR of 34.9 dB.


Figure 4
fig4

Several differences between the two frames are immediately apparent. Notice for example the greater clarity of horses and riders in the lower frame, which represent the high-motion portion of the video singled out in the EV modeling process. There are improvements in static elements as well, such as in the clarity of the cloud formations and the woman’s hair in the right foreground.

The Opportunity for Encoding Manufacturers

Such results can be achieved on virtually any MPEG-2 encoder through integration of the EV software suite utilizing the EV SDK, which consists of an API (header files), libraries, and accompanying documentation. As a starting point, the SDK provides a mechanism to evaluate EV technology for MPEG-2 encoding prior to full integration, which will show compression gains achievable with EV but not necessarily the optimized computational performance.

The next step starts with completion of a technology access fee and agreement.  During this phase Euclid creates an SDK tailored specifically to the manufacturer’s environment and provides an SDK engineer to support the integration process, allowing the vendor to evaluate the EuclidVision performance in-house.  Euclid and the vendor then will work together to identify additional requirements.

The final step entails consummation of a licensing agreement, including the usual components of licensing fees, technology transfer, maintenance and support and upgrades. Once the SDK is fully integrated, a customer will be able to deploy EuclidVision in-house, within the customer’s reference codec implementation.

Conclusion

The cable industry is at a crossroads where failure to address the need for improved encoding performance in MPEG-2 could derail efforts to sustain a competitive pay TV service on legacy set-tops over the course of long-term migration to IP pay TV. That migration strategy requires allocation of ever more bandwidth to broadband and multiscreen services, which means consumption of bandwidth devoted to MPEG-2 must be held in check or even reduced, depending on how far any given operator has already gone in raising the ratio of MPEG-2 channels allocated per QAM.

But this can’t be done without significant improvements in encoding efficiency that will allow operators to keep pace with viewers’ quality expectations as households acquire ever larger flat panel TVs. Now that three quarters of U.S. households own at least one HD set, the buying trend has already swung toward acquiring larger sets with expectations that the purchases will lead to a better viewing experience in the living room. Cable operators cannot afford to disappoint those buyers by delivering a quality of experience that isn’t at least as good as if not better than it was on a smaller screen.

Efforts to wring greater performance out of MPEG-2 encoding through advancements in hardware have run their course. But, fortunately, there are great improvements to be found through software-based enhancements to those hardware platforms, if operators and manufacturers choose to incorporate the capabilities offered by Euclid Discoveries.

With 31 issued patents and 28 pending, Euclid is bringing techniques to the cable industry that are unique and additive to any other advancements that might be in play on existing encoding platforms. By virtue of the compatibility of the company’s EuclidVision platform with the MPEG-2 standard, implementation of the temporal prediction enhancements achieved through Euclid’s intelligent modeling and other techniques delivers significant gains in compression efficiency without requiring any changes in the decoding processes on set-top boxes.

The path opened by Euclid to ever greater performance in MPEG-2 encoding as it continues to add new techniques to its platform applies to H.264 and HEVC codecs as well. As a result, wide adoption of the capabilities introduced by Euclid will allow cable operators to hold the lead in quality performance across all encoding modes for years to come.


 

1“Why MSOs Should not Consider Switching Directly from MPEG-2 to HEVC,” Dan Rayburn, principal analyst,

Frost & Sullivan, September 14, 2013

2“Worldwide Pay TV Subscribers,” May 5, 2013, Multimedia Research Group

3IHS Set-Top Monitor Report, July 12, 2013

4See, for example, “2nd Life for Legacy Set-tops Looms as New Force in Migration to Cloud,” ScreenPlays Magazine, July 3, 2013, and “Liberty Global’s Cloud Plan Could Be a Game Changer,” ScreenPlays Magazine, September 30, 2013

5“The Future of MPEG-2 Video Transport over QAM,” CED Magazine, January 4, 2013

6“HDTV and 3D TV X,” Leichtman Research Group, February 8, 2013

7“TV Systems Intelligence,” IHS, May 31, 2013

8“Samsung Follows Sony in Cutting Ultra HD Prices,” Twice Magazine, August 26, 2013, and “LG Confirms Sept. 1 Ultra HDTV Price Cuts,” Twice Magazine, August 29,2013

9“Quarterly Design and Features Report,,” NPD DisplaySearch, January 29, 2013