Metadata For Multichannel

This site offers surround professionals an opportunity to exchange information, voice their opinions, and get their production questions answered by their peers. This was created for you to discuss issues pertaining to the surround sound revolution.

Music may be an in-the-moment experience for the musician, but these days that moment is headed for as many reincarnations as �junk� DNA. New formats continue to generate the need for additional metadata, or, by definition, the data about data. Whether your material is headed for high-definition video, an information appliance, or a mass storage server for a music library, future playback will depend on your attention to metadata.

While it is quite humbling to realize that we have no idea of what 51 percent of the human genome is for, it is far simpler to preserve the instructions for recreating multichannel audio as it migrates from generation to generation of performance platforms. For those of you in the process of restoring the spatial dimension of music, or creating new ones, generating metadata should be added to your �to do� list.

The hierarchy of metadata categories depends a bit on perspective. A number of groups, including the Audio Engineering Society(1), SMPTE, EBU, MPEG, the National Archives, National Institute of Standards, and the OASIS(2) project, are all at various stages of developing standards for audio metadata with multichannel sound in mind.(3,4) The AES/EBU digital audio interface is an early example of the use of audio metadata.(5) It includes both audio and user data. MPEG-4 was the first multimedia representation standard; coded in the form of �objects� it models audiovisual data as a set of objects. MPEG-7 is a Multimedia Content Description Interface that is being developed to facilitate the searching and processing of audiovisual content. MPEG-21 envisions the ability to, �enable transparent and augmented use of multimedia resources across a wide range of networks and devices.�

No matter the exact protocol followed, the multichannel recordist must envision the user of the future. What will it take to reconstruct the space you are creating in some as-yet-unspecified environment? Today�s multichannel performance is a digital item of the future to be accessed by humming a few bars or tapping out the rhythm of a phrase. Then it will be identified as one to be given freely to the community or to generate wealth by a specified payment algorithm.

The lessons we have learned from our experiences in archival re-recording of tape can serve us well as we consider the design of metadata systems in the future. The re-recording archivist must monitor such parameters as phase, azimuth, equalization, and tape speed, but the tasks don�t end there. After the basic preservation transfer is completed, �cleaning-up� the track for re-issue or generating a �user� copy begins.

What records of the processes should follow the �new� work? In this day, when we can retrieve discarded data, who is to say that yesterday�s aesthetic judgments may be the future�s noise or that an edited nuance may become a flash of unfolding ecstasy? What records will matter? So we not only have to carry information about the audio data or �essence,� but also a tremendous amount of other information must be carried along to uniquely specify what a cognitive psychologist(6) would call the audio object.

The essence is transported by some physical medium. All of the fields associated with the medium such as format, dimensions of medium, material makeup, physical condition, and chemical composition should be specified. For those in multichannel, it is obvious that content formatting information should include number of channels, track assignment information, and spatial or surround encoding, but there may be other parameters that enable more exacting spatial image recreation that are candidates for inclusion in a metadata set ranging from the director�s speaker placement recommendations, room acoustic conditions, and other experiential parameters or representations. What is the image heard by the artist at the moment of mixing?

Other categories that those working on audio metadata use to define an audio object include administrative data such as date/time of creation, file size, UNID (unique identifier number), audio transfer data information, device descriptor of what audio equipment was used to play, record, or manipulate audio, and profiles of the devices. As much of the work in the metadata arena is ongoing, it is a good time to think about what you wished you had as you repurpose old masters and how you imagine your work be heard in the future.

Surround Professional Magazine