Turner's Steve Fish: Wrapping the Future In MXF
(originally published in August 2007 issue Turner Broadcasting
has built its formidable media presence, in part, on its ability to handle and manage its assets. The purchasing of the MGM library in 1986 marked just one of the company's pioneering leaps in repurposing already existing material. It should be no surprise then that Turner should today still be at the forefront of asset management for the 21st century in the person of Steve Fish, VP of Engineering at Turner Broadcasting Europe.
At the Henry Stewart Events DAM conference
at the Portman Radisson SAS in London this past June, Mr. Fish delivered a presentation called "The MXF Driven End To End Tapeless Production Proof Of Concept At Turner Broadcasting", about the MXF (Material Exchange Format)
asset management standard. Looking, in his glasses, very much like a Steven Spielberg ca. 1975, but unspooling information like a university physics professor, even those of us still grappling with understanding the format, left substantially enlightened.
Turner is not the only company committed to MXF in its media management. Recently Warner Bros. studios adopted media storage solutions supplied by HP which will allow the movie studio to operate entirely within a 4K video environment. The HP system employs a primitive version of MXF.
In brief, MXF is a set of standards hammered out over the past decade by SMPTE
(Society of Motion Picture and Television Engineers) that describes the parameters of a specific variety of digital wrapper. This envelope can contain a broad spectrum of audio, video, and metadata. The core of the format is called SMPTE 377M, but over 25 additional SMPTE documents both describe and prescribe the various intricacies of the format, all of which are available for download on the SMPTE site (www.smpte.org).
Advantages to the MXF format include its keeping only unique items in multiple versions. If, for example, an additional soundtrack is required for a foreign version of a show, only that new soundtrack need be created for the foreign version. Other types of asset management might require duplicate files for every single different version, which quickly balloons into a massive – potentially absurd - storage problem. MXF also provides a method of putting description metadata into header tracks. One or more may be used and each can be labeled uniquely. A key upside to MXF is its platform-agnostic simplicity and the its forward-thinking design intended to make it a adaptable to future sound and image formats.
The MXF, like many industry standards, a work in progress, adapting to industry changes over time. Some of the original research documents were written as far back as 1997, but the work on honing and streamlining the standard is ongoing. Steve Fish is candid about describing the limitations of MXF. In fact, it is vital that the limitations are openly discussed if better solutions are to be found, particularly as the MXF enters the current period of transition from mere standard to practical deployment.
For one, Fish noted that MXF is still too complex with too many options. He pointed to MPEG2 as an example to learn from. MPEG2, in the early days of its usage, also offered many options and parameter, but ultimately only a few of them were useful and user-friendly for the industry. Such an over-flexible system also makes manufacturers wary of implementing it. Already they are threading the needle in trying to provide the exact technologies their customers want. Having so many options to choose from raises the stakes too high. Also at least 79 specification sets still need to be fixed. SMPTE, like any responsible standards institution, operates with great care and diligence – but not great speed. The fixes will not be solved in the next year, nor would any sane person expect them to be.
Steve Fish emphasizes a lesson learned and practiced at Turner: "Don't try to change the whole world at once. Don't try to solve the entire workflow." Throughout the DAM Conference different versions of this same statement were repeated. In the 1990's many people had dreams of digital asset management systems that would allow a single person to control an entire production from end to end. Now that we have come back to reality a bit, virtually everyone agrees that not only is it impractical to have a single, universal solution for all data problems in all spheres of the industry, but that such a blanket solution is unwise.
Fish has stated that the goal of MXF is "the creation of a simple system with the potential to be as ubiquitous as tape." At this year's NAB Convention
in Las Vegas, he oversaw a tech demonstration, of an entire MXF workflow, which surprised even him in how well it worked. True ubiquity of MXF may be some years away, but if the industry does adopt the standard, it will have done so at the end of a lengthy, exhaustive trial period.
Recommended for those wanting to know more are "The MXF Book" the standard text on the format as well as membership in the Advanced Media Workflow Association (www.amwa.tv)
Labels: media production