[argyllcms] Re: Using VRML to visualize RGB data
- From: "C." <ml-argyllcms712@xxxxxxxxxx>
- To: "C." <ml-argyllcms712@xxxxxxxxxx>
- Date: Sun, 23 Dec 2007 14:28:23 -0800
Forgive my somewhat naive suggestions; I'm sure it's quite probable
that there are already better tools within ArgyllCMS or other
color-centric applications for your purpose. I'm just not
sufficiently familiar with those options yet to suggest them
in any detail, so I'll share what did come to mind.
Perhaps some other kind of free software
3D visualization / viewing / plotting application could help.
I've used OpenDX to visualize various multi-dimensional datasets,
and I know it has fairly flexible support for user defined input
formats so you could tell it how to access the color samples in
your input data. I'm not sure if you intend to plot the RGB
space of the IT8 "chart" data or if you intend to also take scanned
IT8 images like the one you submitted and visualize those also
(possibly differentially wrt. the IT8 precise source values to see
gamut distortions). I'd guess that telling it to import "binary"
"image" data would be a little harder than an ASCII format, but
if you saved your scanned image data as something like a raw RGB
bitmap of 8R/8G/8B bits per pixel or 16R/16G/16B bits / pixel
I suspect it'd be trivial to have it import that. Perhaps
there are free TIFF / JPEG /PNG image import options for OpenDX too.
In a similar way you could conceive to to this kind of 3D plot
in things like MS Excel, OpenOffice Calc, perhaps the plotting
capabilities of GNU Octave, Inria's free SciLab,
commercial MatLAB (if you have it), et. al. could be used.
I'm sure I've seen "image processing toolboxes" for things
like SciLab / MatLAB / Octave that could sample image file
colors and then plot the 3D arrays.
I'd guess that the free "R" statistical / mathematical data
analysis system should be able to trivially do this as well
since it's built for number crunching and almost certainly
has flexible input parsing / import capacities (including image
I'd suspect). I'm not quite sure how good the 3D plotting is,
but I wouldn't be surprised if it was workable.
If you exported the data to NetCDF / HDF formats (for multi-dimensional
arbitrarily complex datasets) then many different free and commercial
visualization programs could easily analyze / plot it. Though just
a raw BMP is probably fine too for your case.
Overall it looks like OpenDX should be facile for your needs at first
glance, though it's more "heavyweight" than simply removing the CIE space
coordinate transformation from one of the plots available from ArgyllCMS if
there's
an option for that.
http://opendx.org/index2.php
http://www.scilab.org/
http://siptoolbox.sourceforge.net/
SIP stands for Scilab Image Processing toolbox.
http://www.octave.org/
http://www.r-project.org/
http://www.randomfactory.com/lfa/toc.html
http://www.cs.cmu.edu/~cil/v-source.html
Roger Breton wrote:
I'm looking for a way to have my students visualize RGB code values
"natively", not linked to a CIE space of some sort.
My goal is to be able to demonstrate the color space sampling of, say, an
IT8.7 target such as the one attached (hope it will come through).
I found a few neat java applet that interactively allows visualizing
movements inside RGB color space, in 3D. But nothing that would allow me to
load a text file, that would contains all the RGB data points of an IT8.7
(such as the one attached) and that would allow me to visualize it in 3D
space (without the usual CIELab distorsions). Ideally, I would like VRML to
show the vertices of the RGB cube along with the positions of my list of
data points inside the cube. I believe such an exercice for the students
would have a high pedagogical value.
A side effect of this application would be to generate what I would consider
a more "even" sampling of RGB space for the purpose of developping color
scanner targets.
Right now, I intend to have the attached IT8.7 file outputted at my local
photofinisher on photographic media. Then I should measure all the patches
and use that data file in the classroom for the students to build scanner
targets. But I know there are big "holes" in the IT8.7 dataset.
Best regards and happy holidays!
Roger Breton
------------------------------------------------------------------------
Other related posts:
