b10

In the previous AES convention (October 2009, in NYC), I assisted to the panel titled “What will Perceptual Audio Coding stand for 20 years from Now?”.

On the panel there were some known figures in PAC: (copying the info from the AES website)

Moderator:

Anibal Ferreira, University of Porto – Porto, Portugal, ATC Labs, Chatham, NJ, USA

Panelists:

Jürgen Herre, Fraunhofer Institute for Integrated Circuits IIS – Erlangen, Germany

Youngmoo E. Kim, Drexel University – Philadelphia, PA, USA

Bastiaan Kleijn, KTH-Royal Institute of Technology – Stockholm, Sweden

Mark Sandler, Queen Mary University of London – London, UK

Gerald Schuller, Ilmenau University of Technology – Ilmenau, Germany

In general, the panelists showed a very gloomy and sad picture on the future of PAC, and the need or use of it in the future. Some of the arguments exposed were:

- The cost of bandwidth is going down, while the easy of access is going up: The trends show that the cost of data bandwidth has reduced dramatically in the past 10 years, and the available speeds of connection (specially wired) have increased in the same period of time. The assumption is that these trends will continue (but their rate of change might not be the same).

- Also, data storage costs are going down, and capacity of storage is going up.

- Distribution costs (moving data from/to places) are going down.

- Coverage is increasing (wired and wireless)

All of these suggest than from the point of view of the consumers, the cost of access per byte of data will be close to zero (if not free altogether). Therefore, the users have no real incentive to choose a lossy compressed media over the original file.

But, I have to say that I don’t agree with this point of view.

Although, from the point of view of consumers it is true that the costs (and time) for delivery and storage are coming down, close to free, the costs will still be there for the producers and distributors of the media files.

Online (or on the cloud) file sharing systems (like Youtube, or whoever will be hosting this material in the future) has an inherent cost per byte stored or transmitted (not mentioning the money paid to the content creators…. If any (pun intended)).

Each byte stored and delivered has costs related to their storage, electricity, computation and delivery. Even saving a fraction of a penny per transaction is important for these big companies.

In the distribution part, it is possible that wired connections to the home and office will be as ubiquitous as water and electricity networks are nowadays; but in the wireless real there will be a never ending competition for bandwidth. More devices will be connected wirelessly and have more needs for data. Streaming music from your phone to your clothes or to your beer mug anyone?

Not to speak of the increase in data in the content itself. More audio channels (5.1, 7.1, 10.1.. etc), higher resolution/aspect ratio for video, multi angle cameras, holographic displays and who knows what more. At the end, PAC (or perceptual media coding) compression will be an indispensable tool to keep facilitating the access to the ever growing library of media.

I see some problems (read: opportunities) for the PAC community to solve:

- New/improved algorithms to handle the growing resolution/quality demands of the new devices.

- What about the vast collection of media already compressed in the hands of users already? Think about all those MP3 files that you (or someone you know) already have in their hard drive. Probably people will not want to buy them again (…buy? ) for use in the new technology, so transcoders from MP3 to the new formats are necessary.

- Also, what about using the new knowledge and computer power to “clean” the artifacts introduced by old PAC? Wouldn’t it be great to make those old files at low bitrates sound even better?

So, in my eyes, the future of PAC is still a bright one. The motivations for lossy compression of media files might change, but the types of problems to solve and the solutions will still be the same.

-Rago