|Howard Lukk explaining why Disney needs IMF|
Photo by David Cardinal
Very roughly, an IMF starts with a packing list (manifest) of source material, and a composition play list (CPL). Each is encoded in XML, as are Output Profile Lists (OPLs) which specifies how the CPL is used to create a set of track files from the source information. Pierre explained that XML was the clear choice for encoding, but that the actual packing and distribution was left up to specific implementations. To assist in extensibility, the format also allows for supplemental asset lists.
John Hurst, CEO of Cinecert, took the discussion further into the future, explaining how IMF files might eventually be delivered directly to consuming clients, instead of having to be transcoded for specific output formats in advance. For example, Internet vendors like Google, Apple, and Microsoft might eventually get a single IMF version of a movie that they in turn could render and play as needed, rather than needing to get final output versions.
The key to this is yet another file container format, Material eXchange Format (MXF), which is the structure used for organizing the source data for a movie or other IMF product. By implementing systems that can understand MXF, companies can in principle do their own version creation as needed directly from the IMF-version. John took attendees through the basic process of creating the needed code snippets to parse MXF and begin to build an IMF reader.
In response to audience members, the panelists explained that IMF itself wouldn't typically be used for Internet-based content delivery, but that by using it upstream it could improve the workflow of distributors and ultimately the viewing experience for users.
-- David Cardinal