Can have and can use of technology

Leonardo Chiariglione

Humans are often driven by collective dreams of acquiring some goods that exist in the physical space. Illustrious examples are provided by the Promised Land of the Old Testament, the El Dorado in the New World, the search of a route to the spices alternative to the Ottoman Empire or California’s Gold Rush.

At about the same time of this last collective dream for immediate physical riches, there was probably the first major case of a collective technology-induced dream. The telegraph, invented a few years earlier, sparked the establishment of competing companies, particularly in USA, all intent on exploiting the magic new technology enabling the creation of corporate empires. Half a century later there was another similar collective dream – wireless telegraph when again companies and governments alike battled to create economic empires.

In the first half of the last century a more collectivistic view did not allow other dreams – radio and television – to play a similar role and more sober collective dreams prevailed. In most countries, the major exception being the USA, governments simply took over the exploitation of radio – and later television – either directly or through tightly controlled agencies.

The field of recording provided another opportunity to create technology-induced dreams. First the vinyl disc, then the compact cassette, then the V2000/Betamax/VHS video cassette recorders and finally the CD and DVD. The process that led to the eventual adoption of VHS at the expense of Betamax triggered another collective by consumer electronics companies to acquire media companies and thereby control future media gadgets.

In the second half of last century a new collective dream took shape with the name "media digitisation". People knew that the heaviest generators of information – graphics, pictures, speech, audio and video were bound to become digital and that carriers – storage, networks, broadcast channels etc. – would be converted to carry digital signals. But people also knew that straightforward digitisation would generate such an unreasonably high number of bits, that practical use of media in digital form would be delayed by decades.

At that time collectivistic views were no longer kept in high regard, so the collective dream to reduce the number of bits without affecting the message started. I was myself hired in the early 1970s to support the "media digitisation" dream of the company that eventually became Telecom Italia.

Those were different times than today. Bandwidth (bitrate) was a much more precious asset, with compression even more in need than today. Knowledge on Digital Signal Processing, human perception etc. was scarce, algorithms primitive, access to electronic computers to simulate algorithms was limited, the costs of converting an algorithm into electronics were staggering and technology constraints were so high that implementation was a nightmare.

The good side of the story was that in those early times companies had the virtue to wait. Unlike today, when people are stingy and want exploitable results tomorrow, at that time people had the stomach to pour money in the "compression well" for decades.

Eventually they got something that was exploitable. But to convert it into products they had essentially two approaches to choose from:

  1. The CE approach – make the right level of investment, make a good product, get it accepted by the market and conquer the world. This was tried for motion video for CD-I, the HDTV codecs by GI and Telettra, etc.
  2. The telco approach – get a standard out and let manufacturers make products based on it. This was the case of several CCITT speech codecs (G series), video codecs (H. series) etc.

The two approaches differed only in the means, not in the goal which remained the selection of technology. Indeed once that was achieved, the two industries would move along in the same way. The need to find an easy way to define this common starting point was the high cost of converting this radically new technology into products.

MPEG was a totally unexpected phenomenon that developed a hybrid CE-telco approach in technology selection, possibly leaning more towards the telco approach. The departure was in the decoder-only specification (which was also used in CCITT video coding standards) and the provision of conformance testing because of the need for interoperable devices in a competitive environment.

In the second half of the 1990s the technology landscape changed considerably. Compression, a technology that used to be abstruse and unwieldy, and extremely costly to make products from, became reasonably well known and easily implementable on PCs. In a rapid sequence we saw software decoding of full bandwidth stereo audio (MP3), moving video of increasing picture size (Real Network, Windows Media and Quick Time for streaming video – all proprietary and incompatible solutions), MPEG-2 Video developed by a number of vendors, MPEG-4 Video for movies (DivX), etc.

The success of the different solutions/players has been shaped by different factors:

The success has also been shaped by the way licensing of technology has been applied. Smart licensing terms for the technology is one reason of the continuing success of MP3, while free availability of decoders and some limited forms of encoders is one reason for the wide acceptance of Real Network, Windows Media, Quick Time and DivX.

The MPEG-4 Visual licensing terms retain the "per codec" and "per disc" royalty of MPEG-2, but the latter has been extended to important cases of content use on networks and broadcast channels, including cases in which a service is offered on content.

MPEG-4 Visual is a moderate success, the main dampening factor being its licensing terms. But a similar, if not worse, fate may well await the new MPEG-4 Advanced Video Coding (AVC) standard. Unlike ten years ago when the development of the MPEG-2 licensing started, today compression technology is plentiful, a large number of implementations are for programmable devices and there is a range of licensing terms on offer. An international standard is still considered a valuable asset, but the availability to use it on the part of industry players is contingent on its licensing terms being at least comparable with those of other proprietary solutions.

Not much is known yet about the future licensing terms of AVC, but what is known does not bode well for the future. Royalty on encoders and decoders are probably a necessity and the news about the intention to treat software decoders differently from hardware decoders are a good sign. Royalties on discs are now a constant in the industry and should pose no big acceptance problems, but encumbering network distribution of valuable content with royalties, no matter what are the logical justifications given, is a sure way of killing the standard in networked environments.

If the matters stay as they are presented today, my forecast is that AVC stands a chance of being used in package media and broadcast, while proprietary solutions will be adopted in networked environments (fixed and mobile). The problem with package media is that it is not at all clear that more (and different) compression is needed for next-generation DVD. Also – and I know this is a heretical opinion – the last entry in the broadcasting industry’s list of needs is more compression. Where compression is really needed is in networked environments, but here the chances of adoption of AVC vs. proprietary solutions are minimal because of the licensing issue.

So, is there a future in compression research? It depends. Of course more compression is always desirable but then digital audio and video compression are not the field it used to be ten years ago. Powerful formats – MP3 and MPEG-2, not to mention MPEG-4 Visual and proprietary formats – are entrenched. Displacing them requires substantially more compression or support of new features needed by applications or both. Above all – you guess it – it needs enticing new licensing terms.

An effort that would be clearly welcome by many in the industry is a standard compression algorithm that does not infringe on currently valid patents. This is not very likely to happen, though. The first impediment are the rules in ISO that would require the creation of a new organisation outside of it, the second is the lack of lustre of such a project from the academic viewpoint and the third is the difficulty of putting together the necessary expertise because of the vested interest of some industrial players.

In a sense, however, compression is no longer such a relevant issue. Looking back at the last few decades when industrial interest in digital technologies started to coalesce, one can see the interest (and the fightings) of industry moving from the low OSI layers upwards like a wave. Every time the battle on one layer was over, the interest shifted one layer up. In the first half of the 1990s the battle on the transport layer was fought and concluded with two winners. In the second half of the 1990s the battle was on compression and that is being concluded. In the first decade of the 21st century the battle is in the provision of a DRM solution on top of the media (compression) layer.

The Digital Media Project is the place where industry is called to develop an interoperable DRM platform collaboratively.