Velocity Reviews - Computer Hardware Reviews

Velocity Reviews > Newsgroups > Computing > Digital Photography > the nature of color, reproduction and perception

Reply
Thread Tools

the nature of color, reproduction and perception

 
 
hbarta@comcast.net
Guest
Posts: n/a
 
      08-15-2003
Color is a result of various wavelengths of light from near infra-red
through ultra-violet. What we see as white light is an equal amount
of all frequencies. Or is it? Is there any fundamental reason for
the energy in the spectrum from the sun, for example, to be evenly
distributed across visible frequencies? It seems like our eyes are
capable of 'self white balance' without any conscious effort. How
many people complain about the green tint from fluorescent lighting,
for example.

A specific color is some mix of various frequencies. Is this
typically a combination of various frequencies or is the energy
distributed across a range of frequencies? e.g. what is red? I
believe that something like a red LED emits light at a specific
wavelength. Something like a LER (light emitting resistor emits
light over a range of frequencies that are commonly known as 'red.'

What does a digital color sensor measure? Does it measure a specific
wave length, or is an element sensitive to a range of wavelengths
that are characteristic of the given color? And regardless, it
seems like an imaging element provides a single measurement within
the given color range. There seems to be no way to differentiate
the actual spectral distribution. Given the previous example of an
LED or something hot enough to glow red, there is no way to tell
the difference for the red value. It seems to me that the only way
to fully characterize a given color is to measure the energy at all
visible wavelengths from near infra-red to ultra-violet. And yet all
digital color systems with which I am familiar do this with three
values. (OK, I'm thinking about RGB encoding. there is also CMYK
encoding that uses four values. Still not a lot of information for
what seems to be a potentially complex value.) Increases in color
fidelity seem to be oriented toward providing finer gradations for
the three or four channels rather than including more points along
the distribution.

Finally, what about reproduction? It seems to me that three or four
values for a given point can only approximate the spectral energy
distribution that represents a color at any given point. A CRT works
with three (RGB) inputs and so does a an LCD screen. I suspect that
the CRT has broader spectra for each of the three colors than the
LCD. Could that be why CRTs seem to be favored for work where color
fidelity is important? Printers use the same three or four color
values to produce their output. While some printers have more than
three colors of ink, that is to work around the problem that pigments
and dyes don't mix as easily as colored light.

If anything I've said is not correct or skirts the facts, please
follow up with corrections or clarifications.

I guess my basic question is the approximation of color by three
values. How come that seems to work so well?

thanks,
hank
 
Reply With Quote
 
 
 
 
Rafe B.
Guest
Posts: n/a
 
      08-15-2003
On Fri, 15 Aug 2003 02:57:49 GMT, http://www.velocityreviews.com/forums/(E-Mail Removed) wrote:


> I guess my basic question is the approximation of color by three
> values. How come that seems to work so well?



Because that's how human eyes work?

Actually, the 4th axis is intensity, which really provides
most of the detail. Color is just frosting on the cake.


rafe b.
http://www.terrapinphoto.com

 
Reply With Quote
 
 
 
 
Robert E. Williams
Guest
Posts: n/a
 
      08-15-2003


(E-Mail Removed) wrote:

> Color is a result of various wavelengths of light from near infra-red
> through ultra-violet. What we see as white light is an equal amount
> of all frequencies. Or is it? Is there any fundamental reason for
> the energy in the spectrum from the sun, for example, to be evenly
> distributed across visible frequencies? It seems like our eyes are
> capable of 'self white balance' without any conscious effort. How
> many people complain about the green tint from fluorescent lighting,
> for example.


White Light is not EQUAL amounts of all frequencies.
Technically, it is the distribution of wavelengths radiated from a "black
body" at some specified temperature.
Usually the temperature specified is 5500K. This corresponds to the
apparent surface temperature of the sun at noon. The spectral distribution at
this temperature peaks at about 480-490 nanometers , a kind of blue/green
and drops off at higher and lower frequencies.
Some folks specify 6000K. to allow for the effect of blue skylight.
The Eye/Brain Complex does an amazing job of "White Balancing".
I suspect that the eye records the actual color distribution on say an
overcast day, but the brain interprets the data like it thinks the subject
should look like from memory of similar scenes. It gets really hairy.
Bob Williams

 
Reply With Quote
 
Samuel Paik
Guest
Posts: n/a
 
      08-15-2003
I started writing a long response, but when it grew to around
a thousand words I figured it makes more sense to refer you
to books.

The short answer: light is a power spectra, color is human
perception of light through the sensors in the eye, there
are three kinds of sensors in the eye, (...lots of stuff
elided here...), you can hence parameterize any color
with three numbers, which is why RGB is so popular.

To learn more, you should hit some books, e.g.

Wyszecki & Stiles; Color Science: Concepts and Methods,
Quantitative Data and Formulae.

Some books on computer graphics have sections on color and the
human visual system, e.g.

Hall; Illumination and Color in Computer Generated Imagery

Lots of books on human perception (psychology) have sections
on the human visual system.

[my background is in computer graphics, not psychology, so I
can name the title of a book in that category off the top of
my head]

> I guess my basic question is the approximation of color by three
> values. How come that seems to work so well?


Because it is rooted in the human visual system. Dogs and aliens
may need a different number of parameters.

Sam
 
Reply With Quote
 
Don Stauffer
Guest
Posts: n/a
 
      08-15-2003
Light is composed of a stream of photons. Each single photon is of a
very specific frequency/wavelength. However, a wide band emitter like a
blackbody emits an absolutely staggering number of photons per second
(billions of billions of billions), and these photons span a wide range
of frequencies. An LED emits a narrow band of frequencies, but still a
band. A laser GENERALLY spans an even narrower band of frequencies,
depending on type. Some produce an extremely narrow band (certain gas
lasers). Most solid state lasers are relatively broad in spectrum
compared with gas lasers, but still narrow compared to LED sources. So
called 'single mode' lasers are more narrow in bandwidth than multi-mode
ones.

(E-Mail Removed) wrote:
>
>
> A specific color is some mix of various frequencies. Is this
> typically a combination of various frequencies or is the energy
> distributed across a range of frequencies? e.g. what is red? I
> believe that something like a red LED emits light at a specific
> wavelength. Something like a LER (light emitting resistor emits
> light over a range of frequencies that are commonly known as 'red.'
>
> snip


> thanks,
> hank


--
Don Stauffer in Minnesota
(E-Mail Removed)
webpage- http://www.usfamily.net/web/stauffer
 
Reply With Quote
 
 
 
Reply

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Visual perception and gamma encoding Johannes Digital Photography 3 10-08-2006 03:59 PM
Saving the db state for bug reproduction and tests George Homorozeanu ASP .Net 0 09-20-2005 04:48 PM
The Nature Of Life As Seen From Earth - Life Energy Particles - Perception At A Distance {HRI 20010829-pi9-V2.0} - (part issue 9 Version 2.0 on 21 Aug 2005) Koos Nolst Trenite DVD Video 1 08-28-2005 09:23 AM
University of Plymouth Computing Industry Perception Study Andy Microsoft Certification 0 06-26-2004 03:35 PM
Digital print reproduction on Epson 1270 and ink questions Dean Tran Digital Photography 1 04-19-2004 02:42 AM



Advertisments