Browse Prior Art Database

Simple Data De-Interleaver

IP.com Disclosure Number: IPCOM000111483D
Original Publication Date: 1994-Feb-01
Included in the Prior Art Database: 2005-Mar-26
Document File: 2 page(s) / 65K

Publishing Venue

IBM

Related People

Barraclough, KR: AUTHOR [+4]

Abstract

Disclosed is a simple method for grouping together data of two or more types which are interleaved with each other. The example given is for 8-bit digital video chrominance signals which are generated by a video decoder and stored in a 32-bit memory array for processing. The technique may be adapted to other data widths and interleave patterns. Benefits are a reduction in bus width or pins on a VLSI chip, a reduction in components and the formatting of data.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Simple Data De-Interleaver

      Disclosed is a simple method for grouping together data of two
or more types which are interleaved with each other.  The example
given is for 8-bit digital video chrominance signals which are
generated by a video decoder and stored in a 32-bit memory array for
processing.  The technique may be adapted to other data widths and
interleave patterns.  Benefits are a reduction in bus width or pins
on a VLSI chip, a reduction in components and the formatting of data.

      When the YUV encoding scheme for color video is used, there is
less information stored in the U and V components (chrominance
values) than in the Y component (luminence).  Typically this means
that either the U and V data can be stored in less significance,
fewer bits per sample, or the sample rate (bandwidth) can be reduced.
If this last method is chosen, it is possible to use the same data
bus for both the U and the V data (at separate times).  The V and U
samples could be written, as they become available, into two separate
areas of memory, one for V samples and one for U samples, but this
can be wasteful of components.

      Another method is to write samples away as they appear which
causes them to alternate throughout memory on a sample-by-sample
basis, e.g., VUVU VUVU.  A discrepancy then arises due to the fact
that V and U, the two chrominance values, are available from the
decoder interleaved with each other, but need to be processed as two
separate arrays, one of V and the other of U data.

      In a typical case for high bandwidth systems, the Random-...