Digital audio and video file formats (Guideline 22)
See information on:
- Choosing a format/codec for creating digital audio or video
- Choosing a format/codec for digitising analogue audio or video
- Criteria for choosing formats for preservation purposes
- Formats and codecs for digital audio preservation
- Formats and codecs and formats for digital video preservation
The terminology used for digital audio and video can be extremely confusing, particularly as terms are often used interchangeably. The major terms are defined below.
Coding and encoding
The way binary numbers (1 and 0) are used to represent sound or images: the binary numbers can be ‘read’ by a computer. All video and audio on the internet and in computing systems are encoded in some way.
‘Encoding’ is often used to refer to both the format of video/audio streams and the file type of digital video or audio files. For example .mp3 and .dv are often referred to as ‘encodings’ but the former is an audio file format and the latter an encoding format for digital video streams. Some encoding formats are ‘wrapper’ formats (see Wrapper format below).
The structure of a discrete chunk of digital data is a known as a file. A file format is a way of capturing the encoded data stream as a logical unit and storing it. Codec software is then needed to interpret the file (see Codec below). Some file formats are known as wrappers.
Wrapper (or packaging) format
A wrapper is a file format for encapsulating different bitstreams (or files) and metadata into a single file. For example, in an audio ‘wrapper format’ like WAVE (.wav), audio streams or files and other chunks of metadata are encapsulated into a single file. AVI, the standard Windows Media wrapper, and MOV, the standard QuickTime container, encapsulate audio, video, metadata, and other control information together as a single package. MXF is also a wrapper format that wraps video, audio and other bitstreams.
Short hand for (en)coder/decoder or compression/decompression. A codec is the method by which an audio or video bitstream is encoded and decoded. For some uses it is important to reduce the size of the data streams e.g. for Internet delivery, so many codecs compress audio and video data. Most of the publicly documented audio file formats can be created with one or more codecs. Without the correct codec software to decode the encoded bitstream, an audio or video file may not play.
Codecs are usually:
- Lossy: A lossy codec discards certain portions of the signal to achieve a smaller file size. For example, mp3 codecs attempt to identify and remove portions of the signal that would not result in a perceived loss of quality of the sound to make the file smaller. The more compression, the more data is removed and more reduction in audio quality. Once removed, the data cannot be recovered.
- Lossless: A lossless codec achieves smaller file sizes through other methods rather than removing data. For example, a variable bit rate may use fewer bits to encode silences as compared to an active section of music. Lossless codecs do not compromise the audio or video quality even when they hold compressed data.  A lossless compressed format requires much more processing time than an uncompressed format but is more efficient in space usage.
Some codecs may be configured to be either lossy or lossless.
 All definitions are sourced from Arts and Humanities Data Service (AHDS), Digital moving images and sound archiving study, August 2006, available at: http://www.ahds.ac.uk/about/projects/archiving-studies/moving-images-sound-archiving-final.pdf
 National Archives and Records Administration (NARA), Frequently asked questions about digital audio and video records, available at: http://www.archives.gov/records-mgmt/initiatives/dav-faq.html