Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Dithered 16-Bit Direct Color for Natural Images

IP.com Disclosure Number: IPCOM000115806D
Original Publication Date: 1995-Jun-01
Included in the Prior Art Database: 2005-Mar-30
Document File: 6 page(s) / 273K

Publishing Venue

IBM

Related People

Rackley, DP: AUTHOR [+2]

Abstract

Disclosed is a simple and efficient method for improving 16-bit Direct Color data for Natural Images in the conversion to 24 bit Color for display by three 8-bit DACs. The disclosed method provides Natural Image display quality that is superior to conventional 16-bit Direct Color by developing the "Missing Bits" in a pseudo-random manner known as "Dithering".

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 24% of the total text.

Dithered 16-Bit Direct Color for Natural Images

      Disclosed is a simple and efficient method for improving 16-bit
Direct Color data for Natural Images in the conversion to 24 bit
Color for display by three 8-bit DACs.  The disclosed method provides
Natural Image display quality that is superior to conventional 16-bit
Direct Color by developing the "Missing Bits" in a pseudo-random
manner known as "Dithering".

      In the field of MultiMedia, the display of Natural Images
requires a large number of data bits per displayed pixel to achieve
excellent "front of screen" display quality.  Ideally, with a display
subsystem having three 8 bit DACs, the display data for Natural
Images should have 24 bits per pixel to fully utilize the display
capabilities of the subsystem.  However, when encoded as 24
bit-per-pixel data, a Natural Image requires large amounts of storage
and display memory.  As a compromise, Natural Images are often
encoded as 16 bit-per-pixel data, which is probably the minimum
necessary to produce acceptable "front of screen" display quality.

      The presently described method is applied to MultiMedia display
subsystems, and to computer graphic display subsystems, including
extensions and enhancements to the IBM* PS/2* (XGA) Extended Graphics
Adapter  display adapters.  This method is applicable to any display
adapter used in a MultiMedia system or subsystem, or to any display
subsystem used for the display of Natural Images.  For example, this
method may be used in the Serializer Palette DAC SPD portion of a
display subsystem.  This method can be added as a standard feature to
a future SPD chip design or retro-fitted to current SPD chip designs
to enhance their function.

      When 2 bytes (16 bits) are allocated to each pixel to represent
its color, a form of "Direct Color" representation is commonly used.
In general, the 16 bits are divided among the three primary colors in
a roughly equal manner.  In current graphics products, the bits for
Direct Color are divided among divided among the primary colors in
the following ways:
  565RGB - 5 Red, 6 Green, 5 Blue
  664RGB - 6 Red, 6 Green, 4 Blue
  555RGB - 5 Red, 5 Green, 5 Blue + 1 control bit

      For example, the 565RGB designation means that, of the 16 bits
used to represent each pixel, the first 5 bits are used as the Red
component, the next 6 bits as the Green component, and the last 5
bits are used as the Blue component.  For Red and Blue, there are
five bits of data allowing 32 different intensity values to be
defined for each color.  For Green, there are six bits of data
allowing 64 different color intensity values to be defined.  For
Direct Color, it is usual that the color intensity values are
linearly
ascending, with all bits zero yielding the lowest intensity and all
bits one yielding the highest intensity.  Thus, with 16-bit Direct
Color, it is possible to define 32x64x32 = 65,536 different colors.

      The 8-bit DA...