Computer Graphics with Photoshop - Photoshop color display per wavelength

Author

Photoshop color display per wavelength

Joseph Chamberlain, DDS

2005-05-24, 7:17 pm

Dear listers:

I am trying to find out if there is a way to have Photoshop display the
color that corresponds to a certain wavelength in the visible spectrum. For
instance, When entering the value of 480 nm (nanometers) it would display in
a graphic or pre-designed shape with the color corresponding to this exact
wavelength (in this case a shade of blue).

Thank you in advance for your help.

Best regards,

--
Dr. Joseph Chamberlain, D.D.S.
Oral and Maxillofacial Surgery

hoffmann@fho-emden.de

2005-05-24, 7:17 pm

Joseph,

spectral colors can be shown only by a monochromator. A monitor
doesn't allow the reproduction of spectrally pure colors.
This doc (650kBytes)
http://www.fho-emden.de/~hoffmann/ciexyz29082000.pdf

shows on p.2 the CIE chromaticity diagram. The RGB space sRGB is
near to CRT monitors. The D65 white point is marked by a black
circle.
An example: lambda=3D550nm. Draw a straight line from the white
point to the spectral locus. This line intersects the edge R-G.
The RGB values are B=3D0 and R,G according to the proportions of
the two pieces of the edge (barycentric coordinates).
The nearest visualization (approximation) is found by this method.

Another cheap advice: assign wavelengths 700nm to 465nm to the hue
angle 0 to 270=B0 in HSB. This will be entirely wrong, unless a very
nonlinear mapping is used.
P=2E3 shows some spectrum bars, which can be considered as reasonable
illustrations (averaged informations, using several sources).

A shorter answer would have been: not directly possible in PhS.

Best regards --Gernot Hoffmann

Ctein

2005-05-24, 7:17 pm

Dear Joseph,

It's physically impossible for a monitor to do that. The monitor only emits
red, green and blue, with a fixed spectral distribution for each. It creates
the illusion of true spectral color by combining various amounts of these
primaries, but it's only an illusion-- there's no way to vary the
wavelengths the monitor emits.

pax / Ctein
==========================================
-- Ctein's Online Gallery http://www.ctein.com
-- Digital Restorations http://photo-repair.com
==========================================

Bart van der Wolf

2005-05-24, 7:17 pm

"Joseph Chamberlain, DDS" <drjchamberlain@earthlink.net> wrote in
message news:BEB41C67.1F0A4%drjchamberlain@earthlink.net...
> Dear listers:
>
> I am trying to find out if there is a way to have Photoshop display
> the
> color that corresponds to a certain wavelength in the visible
> spectrum. For
> instance, When entering the value of 480 nm (nanometers) it would
> display in
> a graphic or pre-designed shape with the color corresponding to this
> exact
> wavelength (in this case a shade of blue).

As others have explained, it is physically impossible to exactly match
spectral wavelength colors.

However, this little application produces results that are 'close',
given the physical limitations:
<http://www.efg2.com/Lab/ScienceAndE...ing/Spectra.htm> .
The links on that page can help to further understand the issues
involved.

Bart

hoffmann@fho-emden.de

2005-05-24, 7:17 pm

Bart,

an interesting doc, but as long as the RGB space is not
defined the results will be somewhat meaningless.

The author shows the luminance formula for NTSC primaries
(which is applied to gamma encoded values RGB' ) and the
correct formula for ITU-R BT.709 primaries which are used
for sRGB. This doesn't clarify the situation.

Best regards --Gernot Hoffmann

hoffmann@fho-emden.de

2005-05-24, 7:17 pm

Bart,

thanks for the interesting discussion (which is perhaps not helpful
for the OP). A suggestion how to INTRODUCE spectral colors in PhS:

1. Set Y=1 .
2. Calculate from chromaticity coordinates xy on the spectral locus
the values XYZ.
3. Convert XYZ to Lab.
4. While Not[ (0<=L*<=100) And (-128 <=a*<+127) And (-128<=b*<=+127)]
Do (Y=Y-0.05; Repeat step 2,3 )
5. Input Lab in PhS.

Now we have at least Lab values which STILL describe the correct hue
and which are in the allowed code range.

These can be converted into any RGB space - of course with some even-
tually surprising gamut compression. Not tested, so far.

Best regards --Gernot Hoffmann