Skip to the content of the web site.

Dirac Video Decoder

The Dirac Video Decoder project aims to implement the pro version of the Dirac video decoder. The decoder will be implemented in VHDL, targeted for an Altera FPGA.

Dirac is an open video codec being developed by the BBC. Unlike H.264, VC-1 and MPEG-2, Dirac uses an inverse discrete wavelet transform instead of an inverse discrete cosine transform.

We are implementing the Dirac Pro version of the codec. It is currently in the process to be reviewed and standardized as VC-2. The specifications will eventually be released Here, but in the mean time we have a copy in our repository.

Dirac Procedure

Video decoding can be broken down into four general steps:

  1. Bitstream decode
  2. Inverse Discrete Wavelet Transform (iDWT)
  3. Motion Compensation
  4. Post processing

Resources

The project has a Mailing List and stores it’s code in a Subversion Repository

Development Team

  1. Active
    • Nathan Buchanan (Lead)
    • David Cananzi (VGA)
    • Andrew Davidson (Arithmetic Coder)
    • Roxanne Guo (Arithmetic Coder)
    • Kevin Park (SRAM)
    • Thebe Parsons
    • Nigasan Ragunathan (VGA)
    • Jiewon Ryu (VGA)
    • Justin Wong (SRAM)
  2. Past
    • Richard Baverstock (Lead)
    • Terence Hampson (Lead)
    • Jason Lor (Lead)
    • Uzma Syed (SRAM)

This list is updated at the end of each term with a list of those actively involved with the project. If your names is missing, let us know.

Original Dirac Project Resources

See the kick off presentation for more details on the decoding scheme: Dirac Kick-Off Presentation.