What is 32-bit color?

Digital photographic color is generally defined as 24-bit, with 8 bits each
in the red, green and blue components. Most recent monitors can be set to
32-bit color. What are the extra 8 bits, and what purpose do they serve?

Cooter wrote:

Digital photographic color is generally defined as 24-bit, with 8
bits each in the red, green and blue components. Most recent
monitors can be set to 32-bit color. What are the extra 8 bits,
and what purpose do they serve?

32bit color for the vast amount og graphic cards means exactly the same
number of colors as 24bit. It's just a matter of byte alignment.

Due to technical issues, it's often easier (read: faster) for computer
hardware to move bits in blocks of 32 instead of 24. My graphic card
doesn't even have a 24bit setting. The extra 8bits are just wasted
space.

One notable exception is the Matrox Parhelia cards, that allows for
30bit colors (divided equally in R, G & B color channels).


If we look at graphic formats instead, 32bit sometimes means that there
is a byte reserved for alpha-level transparency.
--
Toke Eskildsen - http://ekot.dk/

In article ,
says...

32bit color for the vast amount og graphic cards means exactly the same
number of colors as 24bit. It's just a matter of byte alignment.

Due to technical issues, it's often easier (read: faster) for computer
hardware to move bits in blocks of 32 instead of 24. My graphic card
doesn't even have a 24bit setting. The extra 8bits are just wasted
space.

Exactly right, computer cpu chips (including on the video card) only
understand 8, 16, or 32 bit word sizes. Therefore 24 bit color would
otherwise require three operations to move three 8 bit bytes, which is
slow. Instead, they package the three 8 bit bytes into a 32 bit word and
32 bit buss, which is only one move operation instead of three. Same
result, still 24 bit color, but the 32 bit video mode is simply faster.
It also requires 33% more video memory for the unused fourth byte, but
which is a small price to pay for the speed.

--
Wayne
http://www.scantips.com "A few scanning tips"

On 2007-07-05 09:44:31 -0700, Wayne said:

In article ,
says...

32bit color for the vast amount og graphic cards means exactly the same
number of colors as 24bit. It's just a matter of byte alignment.

Due to technical issues, it's often easier (read: faster) for computer
hardware to move bits in blocks of 32 instead of 24. My graphic card
doesn't even have a 24bit setting. The extra 8bits are just wasted
space.

Exactly right, computer cpu chips (including on the video card) only
understand 8, 16, or 32 bit word sizes. Therefore 24 bit color would
otherwise require three operations to move three 8 bit bytes, which is
slow. Instead, they package the three 8 bit bytes into a 32 bit word
and 32 bit buss, which is only one move operation instead of three.
Same result, still 24 bit color, but the 32 bit video mode is simply
faster. It also requires 33% more video memory for the unused fourth
byte, but which is a small price to pay for the speed.

Correct wrt byte alignment.. Wrong wrt the extra 8 bits being wasted
(sometimes). High end graphics workstations used the extra 8 bits for
the Alpha (transparency) channel. This is _usually_ only prevalent in
real time visual simulation (flight simulators etc) driven by High end
image generators e.g. Silicon Graphics ONYX line or the competing Evans
& Sutherland IG's. Today some of the high end Gaming GFX cards use HW
alpha as well.

On graphics chipsets without HW alpha, 32-bits is a waste. The extra 8
bits simply padding for the byte alignment.

--

thepixelfreak

In article 2007070509580175249%not@dotcom, says...

On graphics chipsets without HW alpha, 32-bits is a waste. The extra 8
bits simply padding for the byte alignment.

Yeah, not much concern if the source is 24 bit JPG images.

But only a waste of bit space. Not a waste as a price for speed.

--
Wayne
http://www.scantips.com "A few scanning tips"

thepixelfreak writes:

On 2007-07-05 09:44:31 -0700, Wayne said:

In article ,
says...

32bit color for the vast amount og graphic cards means exactly the
same number of colors as 24bit. It's just a matter of byte
alignment.
Due to technical issues, it's often easier (read: faster) for
computer hardware to move bits in blocks of 32 instead of 24. My
graphic card doesn't even have a 24bit setting. The extra 8bits are
just wasted space.
Exactly right, computer cpu chips (including on the video card) only
understand 8, 16, or 32 bit word sizes. Therefore 24 bit color
would otherwise require three operations to move three 8 bit bytes,
which is slow. Instead, they package the three 8 bit bytes into a 32
bit word and 32 bit buss, which is only one move operation instead
of three. Same result, still 24 bit color, but the 32 bit video
mode is simply faster. It also requires 33% more video memory for
the unused fourth byte, but which is a small price to pay for the
speed.

Correct wrt byte alignment.. Wrong wrt the extra 8 bits being wasted
(sometimes). High end graphics workstations used the extra 8 bits for
the Alpha (transparency) channel. This is _usually_ only prevalent in
real time visual simulation (flight simulators etc) driven by High end
image generators e.g. Silicon Graphics ONYX line or the competing
Evans & Sutherland IG's. Today some of the high end Gaming GFX cards
use HW alpha as well.

You're talking about graphics textures there. In the framebuffer
where it all gets rendered for display there is generally no concept
of transparency. Some special hardware has multiple layers of
framebuffers where transparency does matter, but such things are not
usually found in PCs.

On graphics chipsets without HW alpha, 32-bits is a waste. The extra 8
bits simply padding for the byte alignment.

Some old high-end graphics cards could use the extra 8 bits per pixel
as a palette-mode overlay with a key colour in the 24 RGB bits.

--
Måns Rullgård