Video consumes a tremendous amount of Internet bandwidth—some reports indicate that Netflix alone consumes 28% of all bandwidth in the U.S. (“Report: Netflix hogs 32 percent of peak Internet bandwidth”). A new video compression technology, HEVC (High Efficiency Video Coding) could reduce the burden that video places on bandwidth, however. This technology could reduce usage by half, if Ericsson’s claims are to be believed (“New video compression tech will cut bandwidth use in half, Ericsson says”). But will greater efficiency really translate into lower usage, and what does the new technology mean for consumers (business and individual)?
The New HEVC Compression Technology
Uncompressed video requires tremendous bandwidth to transmit; thus, coding standards take raw video, compress it to reduce the amount of information without harming visual quality and then (after the data is transmitted) decompress the video for display. Greater efficiency means more video using less bandwidth—and preferably delivering similar or higher quality. Computerworld notes that the new HEVC “promises to reduce the bandwidth requirements for video delivery by over 50 percent compared to the best H.264/MPEG-4 AVC implementations, according to Ericsson.” Ericsson also believes that the efficiency of the new standard (assuming it lives up to expectations) will ameliorate the strain on networks owing to the growing consumption of bandwidth for viewing video.
Does Greater Bandwidth Efficiency Cause a Rebound Effect?
On the surface, an increase in video coding efficiency would seem to promise a corresponding decrease in bandwidth usage: users will consume less bandwidth for the video content they watch. But the situation may be more complex.
Ian Bitterlin of DatacenterDynamics has made numerous references to a theoretical economic phenomenon known as Jevons paradox, or the Jevons effect, in particular with regard to energy efficiency in data centers. (For his latest reference, see “Jevons Paradox may strike in New Zealand!”) But the same phenomenon could easily be seen applying in any case where a new technology increases the efficiency of resource consumption. The Jevons paradox, simply stated, is that these new technologies will ultimately lead to greater consumption of a resource, not less. Debates abound on the Internet regarding whether this effect truly occurs.
No doubt, as with almost any economic system, a number of factors come into play—reductionism is tempting, but it seldom leads to accurate predictions. Although the rebound resulting from increased efficiency may not completely wipe out all gains in reduced resource consumption, some loss is likely. Consider that increasing the supply of a good leads to lower prices, but lower prices tends to lead to greater consumption. In other words, if a flood of some item reaches the market, prices will drop, and consumers will tend to be more amenable to purchasing that good—and possibly in greater amounts than they previously did. If there’s a glut of ice cream, for instance, supermarkets may decrease prices to move excess inventory. The result is not simply that consumers continue to buy ice cream at their normal rates, simply picking up a little savings along the way: they will tend to see the reduced price as an opportunity to enjoy more of something they like, so the rate of consumption will increase, clearing store inventory.
The same situation applies in other cases—such as bandwidth consumption. Of course, it partly depends on the tendency of consumers to use more when prices drop: the extent to which prices drop with greater availability of a resource and the extent to which demand increases with falling prices can vary from resource to resource. Nevertheless, given the amount of bandwidth consumed by video, and given users’ voracious appetite for these services, the notion that a HEVC’s increase in efficiency will lead to a corresponding decrease in bandwidth usage seems a little naive. The result may not be as extreme as the Jevons paradox would predict: assuming HEVC lives up to Ericsson’s promises, bandwidth consumption may not ultimately increase as a result of the greater efficiency. Nevertheless, the decrease in bandwidth usage is likely to be much less than a simplistic assessment would suggest. Many users will see the reduced bandwidth as a golden opportunity to increase the resolution of the video they watch, and mobile users may see video as being more accessible within the confines of their limited data plans.
Even were decreases in bandwidth usage wiped out by the Jevons effect, the increased efficiency of the new video codec is certainly not a bad thing, as other positive economic effects can result from the availability of more bandwidth, providing more access to more services (even if that includes just video).
Thus, if HEVC delivers on its promises and sees fairly rapid deployment, some easing of bandwidth consumption owing to video may well result—but the extent of this easing is open to speculation.
But at What Cost?
Matters of bandwidth aside, HEVC isn’t a straight-out replacement for other codecs, like H.264. In this case, greater efficiency means more performance requirements—which translates to more power consumption, all other things being equal. Computerworld notes that “for HEVC to work, decoders on the receiving side also have to be upgraded, and since the codec is more advanced it will have higher performance demands.” This is not necessarily surprising: more work is required to pack a lot into a much smaller “space.”
But another concern is fees associated with the new codec. CNet (“HEVC, a new weapon in codec wars, to appear in September”) states, “Anyone using H.264 in mobile phones, videocameras, operating systems, Blu-ray discs, or TV broadcasting must pay patent royalties to a group called MPEG LA that licenses the pool of H.264 patents on behalf of a large number of patent holders. HEVC is headed in the same direction, calling on patent holders to disclose if they have any patents essential to the video codec.”
With the increasing amount of legal wrangling over patents, and the explosion in number of software patents issued, the cost of using the new codec won’t be limited to just greater performance requirements. Users won’t necessarily see these costs, but they will be tacked on to applicable software. And with the current litigation happiness in the industry, other roadblocks to deployment of the codec could arise.
HEVC promises a more than considerable increase in compression efficiency compared with current major video codecs. Whether that efficiency increase will translate into a corresponding decrease in bandwidth consumption, however, is dubious. Jevons paradox may or may not manifest itself in this case (and even if it did, the existing growth in bandwidth consumption could mask it), but it is doubtful that users (business and individuals) have yet sated their collective appetite for more bandwidth, largely by way of more video.
Photo courtesy of go_nils