Einige Gespenster gehen um in der Welt – die Gespenster der Zauberlehrlinge


The title of this article is inspired by two masterpieces of German philosophy and literature. The first is Karl Marx’s “The Manifesto of the Communist Party” with the metaphor of the spectre (of communism) going around in Europe while the powers of conservation try to stop it. The second is Johann Wolfgang von Goethe’s “The Sorcerer’s Apprentice”, the story of an apprentice who thinks he can enchant a broom and get it to do some work for him, because he has seen his master doing it. The broom gets out of hand, the master comes back, the apprentice implores the master to help and the master sorts out the apprentice’s mess.

I agree that all the above is still rather cryptic and it is not at all clear what these two German works “combined” have to do with the topics that I usually deal with in this blog.

So let me explain: the broom is MPEG, the sorcerers are the people who run MPEG and the spectres are the multinational apprentices who think they can handle the MPEG broom because they have seen it done by those who know how to do it.

The apprentices are labelled spectres because the word indicates “something widely feared as a possible unpleasant or dangerous occurrence”.

Let’s talk about the MPEG broom

As I wrote in Who “owns” MPEG?, the word MPEG is used to indicate several related but often independent things. In one instance, MPEG stands for the “MPEG community”, i.e. the ensemble of people and entities who are affected by what the “MPEG group” does: end users, industries, companies who do business using MPEG standards, universities and research centres, and individuals with an MPEG technical background. Each element in the list is a microcosm, but here we are particularly interested in the last microcosm – individuals with an MPEG technical background. This is composed of active MPEG experts, non-attending registered MPEG experts, researchers working in companies on MPEG standards without being registered members, researchers at large who are doing research in areas that are, or are expected to become, MPEG standardisation areas, and consultants in MPEG matters.

All these people are MPEG stakeholders (the others, too, but here we concentrate on this particular microcosm). They rely on MPEG because MPEG is serving them. MPEG owes part of its existence because they exist and operate. To a significant extent, these MPEG stakeholders can operate because MPEG exists.

The “MPEG group” is another microcosm where ideas percolate through different channels. As explained in Looking inside an MPEG meeting and How does MPEG actually work?, to become standardisation projects, ideas are processed in different ways. Requirements are developed, communicated to different environments and agreements “stipulated” with different industry stakeholders based on those requirements. Standardisation projects require technologies whose existence and performance levels must be verified. Technologies come into MPEG through different channels and are processed in different ways by different groups. Standards are verified against the stipulated performance. Finally, standards are living beings: they evolve and need maintenance.

MPEG is not a broom that operates with a magic, at least not in Goethe’s sense. You do not require spirits (“Geister”) to use it. Still, it is a sophisticated broom that has taken the shape it has as a result of a Darwinian process that has involved and is involving incremental adaptations to match the MPEG group to the needs of the expanding industry coverage, the continuous shift of the way industry operates and the accelerating technology cycles.

The shape that MPEG has gotten today is not final. If it were so, that would mean that MPEG is dead. MPEG is evolving and keeping it adapted to changing conditions is a serious matter. It cannot be left in the hands of some sorcerers’ apprentices.


Next to MPEG there is a group called JPEG. Everybody knows the word JPEG because of the .jpg extension of image files. The JPEG standard (ISO/IEC 10918-1, first released in 1992) has had a far-reaching impact on consumers because all handsets and computers can handle .jpg files and many important services have those files at the core of their business. But let’s make a comparison between MPEG and JPEG.

ConstituenciesImageBroadcasting & AV streaming
Capability to evolveStill working on imagesExpanded field
Number of projectsA few in parallelSeveral tens in parallel
Business models“Royalty free”“IP-encumbered”
Competition of standardsNoVery lively
Approaches                Holistic, top downBottom up
Industry/ academia mix1:13:1
Work force60 members600 members (+1000s outside)
Standards impactHuge (2 standards)Huge (many standards)
Future-oriented standardsLight field imagePoint cloud & immersive video

We see that the two groups are different in many key respects: the industries they serve (image vs broadcasting and streaming), the capability to make the best out of the field to serve industry needs, the number of projects (limited vs several tens), the type of standards they provide (royalty free vs encumbered), the competition (little vs a lot of competition), the approach used to develop standards (principle-based vs experience-based), the percentage of academia in the membership (50% vs 25%), the organisation (handling a few vs handling tens of parallel projects), the impact (2 standards vs many), future oriented standards (coding of light field images vs coding of point clouds & immersive video).

Simply, it is a law of nature. If the size scales by an order of magnitude, everything ends up being different.

Enter the sorcerers’ apprentices

Now, the apprentices think that, because a part of MPEG is handling some technologies that JPEG, too, is handling, we should interpenetrate JPEG (60 people working on images as 30 years ago) and MPEG (600 people working on video, audio, systems and other data, who have designed the strategy that has made the media industry digital and fomented its continuous development) and create news groups creatively organised to manage the huge MPEG work program, the vast array of technologies, the network of liaisons and the large swathes of client industries. A similar fate may also be suffered by JPEG who is at risk to dissolve in the flood caused by the apprentices, as in Goethe’s ballade.

A disclaimer is needed at this point: this sort of idea has nothing to do with the proposal to elevate the MPEG Working Group (WG) to Subcommittee (SC) status presented in Which future for MPEG. The elevation to SC status seeks to change the WG envelope to an SC envelope, keeping the inside – the work and its organisation – exactly the same. The other seeks to upset a working machine thinking that the changes will work, much like Goethe’s sorcerer’s apprentice thought he could handle the magic broom.

A couple of expressions come to mind. The first is “elephant in a china shop”. The effects of the apprentices’ proposal will be exactly this: you will need to merge proud and accomplished people serving different constituencies; operating in environments of largely different complexity in terms of projects, number of people and industry; with different business models and approaches; operating in differently competitive environments; with different 30 years of history and experiences… After the elephant has entered the shop, forget finding any piece of chinaware intact.

The second expression is “A camel is a horse designed by a committee”. Unfortunately, this is not a joke but the harsh reality of some environments where people with a lot of self-importance operate in areas where they have little or no competence or experience. MPEG is mostly free from such people. Indeed, the sorcerers’ apprentices’ proposal comes mostly from people who left MPEG long time ago.

Some effects of the apprentices’ proposal

I could write a long list of negative effects, but let’s limit it to four.

  1. The MPEG brand. The proposed interpenetration will kill the MPEG brand affecting thousands of companies and researchers. Today researchers use their “I belong to MPEG” as a status symbol supporting their research. Tomorrow they will lose both their status and funding. The same applies to the JPEG brand.
  2. The MPEG credibility. The proposed interpenetration will mix two groups who share only a little part of one thing: technology. Technology is important, however, designing the structure of an industrial standards group like MPEG on the basis of technology, instead of constituencies’ needs, wipes off the credibility built by thousands of MPEG experts in 30 years of well-considered efforts.
  3. The MPEG standards. The proposed interpenetration will alter the process by which MPEG standards are defined and developed. Industry will shy away from this new generation of self-styled “MPEG standards” because they will not fit their needs and will look elsewhere. The only sensible thing MPEG will be left with is the maintenance of the 180 standards that were produced by the real MPEG.
  4. The MPEG productivity. The proposed interpenetration will dramatically affect the number and quality of standards produced. One value of MPEG standards is the breadth and depth of their scope. More important, however, is the fact that MPEG standards are not independent specifications but are designed to work together thanks to the painstaking efforts of hundreds of MPEG experts from different areas.

The sorcerers have their hands tied

Decades after decade generations of MPEG sorcerers have learnt the magic, but they are not free. If the MPEG broom is wrongly used, there will be no sorcerers coming back to help the apprentices to undo their misdeeds. The apprentices may well moan die ich rief, die Geister, werd’ ich nun nicht los (I cannot get rid of the spirits I called), but no one will be capable to stop the MPEG broom gone crazy.

Those who care about MPEG have better make themselves heard. At stake there are trillions of USD year on year, billions of users, millions of workers and thousands of highly skilled researchers.

Posts in this thread