File formats

Audio and video material goes through many steps as it moves from a camera to a TV receiver and archives. The following is an overview of parts of a broadcaster's material chain and the resulting file format implications.

Acquisition and contribution

When material is captured as part of the program production cycle, quality is key. Metadata support that identifies shot and take information, shooting location, timing and links to the script will improve production efficiency. In news acquisition, capturing the “who, what, when, where, why and how” information during shooting can improve the access to this source material in later phases of production.

Sports programs combine both on-site and recorded segments. These live events require substantial facilities, planning and crews. While sports events are planned, news events typically are not, except for news conferences. Common elements with these two types of programming include the need to manage latency, have reasonable quality and perform real-time effects. This usually requires an uncompressed routing and signal processing system.

On-air operations

The day-to-day operation of a broadcast facility presents a different set of requirements. What is important to one part of the process may be irrelevant to another. Most of the rich editing and material production metadata needed for production and post production are not needed in the finished material.

However, metadata about ownership, content usage and related information are useful. An on-air operation needs access to program content with latencies of perhaps a few seconds. Cuts-only video edits, audio fade in/out and voice-overs are used for last-minute content changes.

Post production

A post-production house can benefit from a file format that can save work in process. Re-edits are simplified so one can easily undo editing that did not gain a customer's approval. While stream formats can be used to import and store material in these systems, a richer format is needed to maintain the source material in a form that permits effective editing.

A typical post system stores the metadata and an Edit Decision List (EDL) in a database. This means that post-production systems need to support rich editing models. The objective is to have fast, simple and effective editing. A final program, commercial or other content is made by compiling or conforming to a streaming format. This step creates a finished version for distribution that does not contain the EDL, the unused audio/video material, and most of the production and editing metadata.


Each facility will impose different requirements on its archives. Let's review some of the important file format requirements.

When a project is in post production, it may be desirable to save a copy of the work in process. This simplifies later production of derived works.

An archive supporting the storage of a master program editing version needs rich metadata and the source material referenced by the EDL.

Short-term archives are desirable for the production of news and sports. These frequently have all of the material captured from inbound feeds and other material acquired by an organization. This material may be kept for several days to a few weeks. A fraction of the material is sent to a long-term archive.

A broadcaster may wish to save copies of the final version of the actual broadcast. The metadata needed with the material are limited. Most archival metadata will probably be kept in a database for fast access.

A single format

Unfortunately, fundamental conflicts make it difficult to build a single format that meets these diverse requirements. For example, if a file is stored as audio or video tracks, an NLE can work with it quickly and effectively. A compilation turns the tracks into low-latency time-multiplexed streams that are ready for playout. Part of the stream compilation is the rendering of dissolves and other transitions. Filtering the metadata is also accomplished during the compile process.

Minimizing the number of formats a broadcaster uses will greatly simplify what the operators have to do and improve their workflow efficiency. However, selecting a single format that does not work well over the complete material chain produces the opposite result. Stream and file formats should match the workflow.

Compressed formats such as DV DIF and MPEG elementary streams are well suited for the transport and emission of uncut works. MXF supports a wide variety of video compression systems and rich metadata. The focus of the MXF architecture is material interchange between devices. It can also be used on the disk as a storage file format in some applications.

For editing systems, fast and effective editing is the highest priority. The AAF Association has delivered a file format, an API specification and sample software. AAF is an evolution of the OMF project. It has been adopted by several vendors and is currently being standardized by SMPTE. The format supports a wide range of features needed for post production.

Audio and video are important parts of today's personal computers. Microsoft, Apple, RealNetworks and others have developed formats and tool sets to support audio and video on PCs.

The feature sets in these formats focus on desktop computing, not broadcasting. However, because the personal computer market is large, a significant number of applications are available. This means that in some broadcast applications, prosumer PC solutions may be the proper choice.


The number of features any stream or file format provides varies. As designers try to solve more problems with a single format, the complexity increases dramatically. Sometimes designers are able to invent simple solutions to complex problems. Unfortunately, this is not always the case.

Formats that support simple edits and have low latency remain the best solutions for news, sports and on-air operations. Such formats are not appropriate for post production or high-end editing applications.

End users should select a format that is supported by the devices they wish to install and matches their workflow requirements. Users may find that tomorrow's devices will support multiple formats, thereby giving operators the best of many options.

Bob Edge is manager, standards and technolgoy, Thomson Grass Valley.