Are you or have you been an MPEG member?
Yes, we’re currently MPEG members
What do you think makes MPEG special?
The ability to gather leading edge technologists from potentially competitive organizations and drive them in a collaborative effort despite different interests and political positions. This makes MPEG comparable (even though on a smaller scale) to high profile organizations such as CERN (probably the whole ISO, ITU, 3GPP, DVB… community should be considered here). As per “big science”, results are world class.
What do you think is the most important MPEG impact?
The adoption of technologies such as MPEG2, AVC, HEVC, MP2/3, etc… speaks for itself: services and applications unconceivable 30 years ago are now commonplace on a global scale also thanks to MPEG (again, together with ITU, 3GPP, DVB, …)
Do you think MPEG is a good conduit for research?
Yes, Indeed. The most interesting aspect is the mixed participation from industry, public bodies and academia: the sharing of efforts make “reality checking” and deployment in real products possible of many pieces of technology that, under other conditions (proprietary developments with limited investment capacity), would remain a theoretical exercise or a proof of concepts. In this sense, it can be considered a good accelerator of innovation.
Can you comment on your MPEG experience?
Limited, from a contribution point of view but, still, very relevant at least as a means to understand the near future (developments in MPEG anticipate developments in industry of some 5 years)
Are you happy with MPEG standards?
Partly: some are a huge success, some other reveal to be dead ends. This is probably inevitable; innovation as such is a bet on future with a good level of uncertainty. At a higher level, there might be a problem of “path dependency” that constrains activities under different aspects:
– Standardization inertia (we’ve started working on issue X and will continue apparently despite the industrial relevance of the results without any assessment on real impacts)
– Technological inertia: as an example, hybrid codecs are the MPEG’s trademark in the field of video compression: it’s true that they have been extremely successful so far and no industrial grade competitor is present in the market but it is also true that it miles easier to incrementally innovate starting from the existing building blocks than from scratch. This for sure creates a bias (e.g.: VVC software is being developed as an evolution of HEVC HM). Maybe there’s no other possibility, industry being industry and standards being standards, maybe opening the discussion might be useful in the light of point 4 above.
Market inertia in form of difficulty in involving new actors/tackling new fields
– Limited to some specific cases, we have seen specifications technologies developed somewhere else becoming a standard just by rubberstamping. A rare event that hasn’t led to any meaningful commercial success; probably stronger antibodies should be developed.
Being MPEG an ISO WG, it has to be said that the standardization workflow might/should be streamlined, especially for an improvement in the time-to-market efficiency, especially when sister bodies are much faster in issuing twin specifications.
Do you think MPEG standards are the right choice?
We understand this question as meaning: do you think MPEG makes the right choices when deciding what to standardize? The answer in this case is yes and no. Where the track record has already been established as successful (video and audio coding), the choice is obviously right. In other cases, the “path-dependency” is more evident and brings to disputable results: choosing to standardize what is already at hand, without a wider attention to what is being done elsewhere, may lead to brilliant results without practical application (a first hand example, where we contributed directly: CDVS, finalized when Alexnet descendants where storming the field of image recognition and classification). Sometimes path-dependency translates in the impossibility to involve all the relevant actors
What do you expect from MPEG in the future?
MPEG maintaining the lead in the development of industrial grade technologies in the field of digital multimedia despite the arrival of competing entities trying to leverage on political issues more than technical value. MPEG has been great so far in creating shared value (always point 4 above); we should also probably ask ourselves how to preserve this value.
Another important point is how to rewards innovators for their technical/economical contribution in developing and standardizing new technologies. Today, the only available reward for them is obtained through the patent system, but the patent owners of technologies that became part of the standards are not enough protected against the “free riders” that use the patented technologies avoiding to take a license for their use.
