File Formats and Codecs

Each granular element of audiovisual content is saved digitally in its own audio or image file format determined by the chosen codec (short for coder/decoder). A codec may be compressed (either lossy or lossless) or uncompressed. For proper preservation and long-term archive, audio and image are typically saved in industry-standard uncompressed or losslessly-compressed formats in order to preserve all original information in the elements. With a lossy codec some information is lost through the compression (although a good lossy image codec may be effectively “visually lossless”); lossy codecs are typically used for distribution masters because they are much smaller in file size compared to lossless or uncompressed codecs. The file format wraps the coded data into a file, carries additional metadata about the element, and provides the file extension for identification.

  • Digital Picture eXchange (DPX, SMPTE 268) — A generalization of the Cineon file format originally developed by Kodak in the early 1990s for digitally encoding scanned film, this SMPTE standard was later adopted as the de facto standard for digital intermediates when film was scanned and then color-corrected for finalization in a digital environment. Each DPX file represents one frame. The format is still used in scanning and mastering of motion pictures.
  • Waveform Audio File Format (WAV) — The standard file format used for uncompressed audio. It is used to carry the audio in DCPs and ProTools sessions.
  • Tag Image File Format (TIFF) — Used instead of DPX for final picture elements in some instances (including, notably, the Digital Cinema Distribution Master [DCDM] which is then compressed for creating a DCP). Each TIFF file represents one frame.
  • Apple ProRes— A family of lossy video codecs and formats proprietary to Apple, in widespread use in the entertainment industry. ProRes is documented as a SMPTE Registered Disclosure Document (RDD 36).
  • OpenEXR — A high-dynamic-range file format and software toolkit originally developed for visual-effects file exchange, it has become useful to archivists because of its ability to store complex image information, as well as being used for digital intermediates as a way to store higher fidelity than DPX. A subset of EXR has been standardized (SMPTE 2065-4). Each EXR file represents one frame.
  • FFmpeg — An open-source video codec that has become useful to archivists who do not want to use proprietary video codecs.
  • JPEG 2000 — A family of image codecs that includes both lossless and lossy versions, this codec is used in DCPs and IMFs.
  • File-format recommendations from the National Archive.