Pgroup

Share This

A pgroup is also known as  pixel group. ‘Pixel’ here means a single colour element of the image, with each pixel being represented by a number of colour component samples. Samples may represent primary colour components or luma and chroma componenents. Some video standards have colour (chroma) components that are shared between a number of pixels. This is called colour subsampling, typically performed in YCbCr component format to reduce the amount of data required. Sample values are quantized, typically to 8, 10 or 16 bits per sample.

SMPTE ST 2110-20 and IETF RFC 4175 require that pixels which share sample values are transported together as a “pixel group”. If 10 or 12 bit samples are used such pixels would require a non-integer number of octets. In this case the samples representing several pixels must be combined into an octet-aligned “pixel group” (pgroup) for transmission. These pgroups must not be fragmented between packets.

A pixel group must fulfill both conditions below:

  1. An integer number of pixels that are entirely described by a group of sample values.
  2. The sample values describing the pixel group must fit into an integer number of octets without any stuffing or unused bits.

In short a pgroup is the  number of octets containing an integer number of  sample values which entirely describe one or more output pixels. Use of stuffing bits is prohibited as this would increase transmission overheads.

Using pgroups in this way simplifies receivers by ensuring that the packet payload is octet aligned and that samples relating to a single pixel are not fragmented across multiple packets.

See also: IETF RFC 4175, SMPTE ST 2110-20