SMPTE ST 2073-3 PDF

VC-5 Video Essence – Part 3: Image Formats

SMPTE ST 2073-3 – VC-5 Video Essence – Part3:Image Formats

This standard specifies extensions to the VC-5 Elementary Bitstream defined in SMPTE ST 2073-1 for representing R′G′B′(A) images, Y′C′₁C′₂(A) images without color difference component sub-sampling, and Color Filter Array (CFA) images including Bayer image formats. This standard specifies the default color space for R′G′B′ and Y′C′1C′2 color values.

This standard extends SMPTE ST 2073-1 by adding new parameters and chunk elements:

• The ImageFormat parameter specifies the type of image: R′G′B′(A), Y′C′₁C′₂(A), CFA, or Bayer;
• The PatternWidth, PatternHeight, and ComponentsPerSample parameters specify the structure of CFA images, including Bayer images;
• The MaxBitsPerComponent parameter specifies an upper bound on the number of bits needed to represent all values in the component arrays present in the bitstream;
• The VendorSpecificData chunk element allows private data to be embedded in a VC-5 bitstream;
• The InverseTransform, InverseTransform16, and InversePermutation chunk elements specify how to transform and reorder decoded component arrays into the same color space and order as originally presented to the encoder;
• The UniqueImageIdentifier chunk element specifies the image sequence and position within the sequence of an image represented by a VC-5 bitstream.

This standard extends SMPTE ST 2073-1 by defining three stages for transforming decoded component arrays into images: inverse component transform, inverse component permutation, and image repacking process.

This standard extends SMPTE ST 2073-1 to specify levels of performance for decoder implementations.