View Full Version : Need help modeling extended vision and colorblindness with GIMP and/or PAINT.net

Snargash Moonclaw
11-02-2008, 08:40 PM
While this isn't a mapping problem, I suspect this is the best place to find those who can do this with the tools at hand. I'm exploring the ramifications of various races seeing different portions of the spectrum - affecting both visible and hyperspectral wavelengths and trying to find ways to show what this would look like. I suspect that the shifts in color perception (from partial colorblindness) can be readily modeled with graphic program color filtering - although I don't know enough about them to do so. More difficult would be to model ultraviolet and/or infrared vision but I hope to some degree possible as well.

If you're curious about tackling this puzzle please see the thread over on the Campaign Builders Guild on How Orkhs See http://www.thecbg.org/e107_plugins/forum/forum_viewtopic.php?56792.last - especially my later post in it with various graphics snagged from wikipedia.

Now playing: Nakai Eaton Clipman - Feather, Stone & Light (http://www.foxytunes.com/artist/nakai+eaton+clipman/track/feather%2c+stone+%26+light)
via FoxyTunes (http://www.foxytunes.com/signatunes/)

Now playing: Joan Baez - Suzanne (http://www.foxytunes.com/artist/joan+baez/track/suzanne)
via FoxyTunes (http://www.foxytunes.com/signatunes/)

11-02-2008, 09:43 PM
Wow, interesting topic. I'm no spectral analysis specialist, more an artist than scientist, though we've got members who are more scientist than I.

But, if someone can tackle this, the idea of different spectral views would be an interesting experiment in a VT app like MapTool. Creating maps perceived in the view of the eyes of an Orc or Elf, hmmm.

It might be worth attempting a map in the spectrum to an orc.

Have some REP for bringing up an intriguing topic. Besides, I know you are a well known member of the CBG, so you deserve some REP anyway! ;)


Snargash Moonclaw
11-02-2008, 10:10 PM
Thanks. This basically arose from switching to GURPS - which has a lot more mechanical possibilities for improved and impaired vision - including forms that are both. Unlike most systems, I find that different species can see better in certain ways and worse in other ways when compared to each other in a quantifiable (at least numerically,) fashion. Describing to players what those numbers really mean tho - to get e sense of the race's experience of the world in order to role play it is a little more challenging here.

11-02-2008, 10:47 PM
A race's sensitivity off the ends of the spectrum should effectively give them more colors to describe. The farther out you go, the more colors you get. If you're truly color-blind then all you get are shades, without even a concept of "color". Either of these has serious implications for cultural systems.

Here on earth we're very well adapted to the peak of sunlight. The sun's energy peaks at green and that's where our vision is most sensitive. If your star is red and your race has color vision then they will likely have a set of "colors" that humans would see as "red" or be totally insensitive to. Similarly, a race with a bluer star will probably tend to have their colors higher up in the spectrum, probably with an extra color or so in the UV and perhaps be missing a color or two in the red.

Color vision in humans follows a tri-stimulus model with a panchromatic base element. That is, there are three types of color sensors (called cones), each sensitive to different parts of the spectrum and one type of sensor (rods) that respond to the whole visual spectrum. Some things to consider:
(1) Color vision is slightly less acute than pure monochromatic vision. This result comes from the fact that cones are slightly larger than rods and from the need to mix cones of multiple types to get vision across the full spectrum range. If you need to add more types of "cones" to get farther out on the vision scale then visual acuity might decrease.
(2) Color vision needs more light to work. You can see this as light in the room gets lower at dusk. The world goes toward a grayer, grainier version as the world gets darker. Similarly, when you're outside looking at stars you're better off averting your gaze slightly to get the starlight to fall on the more-sensitive rods around the foveola.
(3) Light sensors are most concentrated at the primary focus of the lens get less concetrated the farther away from the center you get. In humans, the fovea is stuffed with cones, giving good color vision. There are fewer cones as you get farther away from the fovea and proportionally more rods. This means that your peripheral vision is lower acuity in both a spatial sense and in a color sense. The arrangement of light sensors works well at reducing the total amount of information that the brain has to process while still giving good results.
(4) Color blindness in humans tends to be more a result of defective cones of one or more types rather than the absence of those rods. People with color blindness may have somewhat lower visual acuity than someone with a full complement of vision.
(5) Visual predators tend to have a much greater visual acuity. In some species this acuity comes about by sacrifice of color vision. In others there are special structures to increase acuity (look at a raptor's eye).

As an aside, modeling color differences with a graphics program doesn't really "mean much" in absolute terms. For example, I can pull the IR filter off my webcam to get it to see farther into the IR spectrum. What I see on my screen is a strange pink image, not new colors as you would expect for a race that is actually sensitive to those parts fo the spectrum.
Mapping "new" parts of the spectrum into our "regular" part of the spectrum doesn't do much for me. Consider http://apod.nasa.gov/apod/ap070505.html as an example. It's nice that they can assign X-ray parts of the spectrum to "blue", visible parts to "green", and IR parts to "red" to get a new image that tries to take advantage of your color vision. The result is somewhat pretty but seems to just be a curiosity.

Snargash Moonclaw
11-03-2008, 05:41 AM
Some interesting points here. A few of them I've covered over on CBG (but I'm crossposting your post into that discussion as well). I'm actually seeking to model, or at least describe well, a few different visual experiences. Elves, like bees, see in ultraviolet, but not red, while orkhs see in infrared but not violet - their crossbreeding produced humans (seeing "normally"). For all of them the number of colors is essentially the same, but none of them will have the same visual experience. I suspect that artists trying to create a realistic scene under bright afternoon light will all reproduce roughly the same pigment mixtures - and it would require a lot of comparison to recognize that they were seeing things differently. That is, regardless of how it appears to the individual - a mixture of pigments which matches their perception of the color of grass will result in agreement by artists of all 3 races that it's a reasonable match - even though one is seeing a much more bluish green and another a much more yellowish green. But, overtones from ultraviolet or infrared may skew them - I'm imagining a difference in luminosity, but not real sure here. I haven't started playing yet with vision extending into both hyperspectral bands - adding new colors at each end of the visual spectrum. (and I've got no idea what "Dark Vision" would actually look like. . .)

Since you mention it, solar color is critical, esp. in cultural development. It's a binary system with one larger red-orange star and a smaller, brighter yellow-white star. The smaller currently is nearly dead center upon the face of the larger, but their orbit around a shared gravitational axis has a slight offset, so over a 40,000 year period the smaller appears to cross the face of the large and then disappears behind it, leaving only the large one visible for about 10,000 (dim, cold) years of the cycle. I'm wondering how much UV the red-orange star would emit - without it the elves are in twilight that whole time - even with it I suspect it's still pretty dim.

Now playing: Track 1 (http://www.foxytunes.com/artist/-/track/track+1)
via FoxyTunes (http://www.foxytunes.com/signatunes/)

11-03-2008, 06:09 AM
The spectrum from stars follows the black-body curves which wikipedia has an entry. The graph within is here tho:

and this shows curves where there is a steep left side and a slow falling tail on the right. The point here is that if your star is a little cooler than our sun then you would a similar amount of infra red but the UV would be significantly reduced. Being coarse about it then the top end UV frequency is determined by the star temperature.

I would think that these new colours would be seen as monochrome or possibly new colours that we cant conceive of. Its like trying to describe red to a colour blind person. How do you do it.

As an aside, I thought elves had infravision as well.

11-03-2008, 08:50 AM
I would think that these new colours would be seen as monochrome or possibly new colours that we cant conceive of. Its like trying to describe red to a colour blind person. How do you do it.

Yea... thats why, even though this is a cool topic, I have not replied since thinking about such things just makes my brain hurt...

11-04-2008, 12:06 AM
As an example of what it means to be blind to a color band, consider the humble garden flower. Many flowers have marking that are only visible in UV, which bees and the like can see. http://elfefe.blogspot.com/2008/05/uv-flowers.html shows a variety of petunia that's white in regular light but has very distinct markings in the UV.

On the visual acuity front, I was heading toward the difference between predator and prey when I sort of ran out of brainpower. A prey animal will tend to have a very wide field of vision with minimal overlap for 3D views while a predator will tend to have large fields of 3D vision. http://www.nhm.org/mammals/page010.html shows the general idea.

Sacrificing visual acuity in the central region to get enhanced peripheral vision is something I'd expect in prey animals.

11-04-2008, 09:40 AM
That is the problem I had.

I could hold up a piece of paper that says DON'T PANIC with ink that reflects all radiation in the visible light spectrum (looks white) but absorbs in the UV spectrum.

There is no real way to "show" this except using false colour, which really doesn't show anything :( It is just made up...

-Rob A>

11-04-2008, 11:08 PM
In the fun realm, races that don't have fully overlapping vision have an interesting possible result. Colored glass is simply a bandpass filter for light of a specific wavelength range. Glass that's opaque in the visible and IR regions but relatively transparent in UV would give an interior that's dark to everyone but the UV-seeing race. Same with IR-transparent glass. Race-specific light sources have a similar effect (the burning of typical items gives an IR-heavy result; UV bioluminescence might not generate any significant amounts of visible light).

Similarly, race UV can keep messages from race IR by simply using UV inks and paints. Totally invisible to the IR-capable race.

If the races have non-overlapping vision ranges then there might be huge differences in architecture and understanding. A well-light chapel full of art to one race mnight be a dark and foreboding den to the other.

Describing a color is a consensus activity. There is variation even among humans as to the peak of sensitivity for various colors. Everyone sees slightly different colors, but we learn to call them "red", "green", "blue", "aqua", and "cerulean".
I appear to have an "off-brand" of red sensor because I don't seem to see things in the red end quite the same as other folks I know. My opinion of the "most red" is definitely different than theirs. As I tell them, I'm not color blind, just color insensitive. Admittedly, that's more in regard to my choice of clothing colors, but you get the idea.

11-04-2008, 11:45 PM
So... vision like the TSA's millimeter wave scanners (http://www.tsa.gov/blog/2008/05/you-asked-for-ityou-got-it-millimeter.html)?

-Rob A>

11-05-2008, 12:28 AM
Terahertz sensing in a biological system isn't terribly likely because the structures need to be on the scale of the energy to be sensed. Plus, there isn't a good source of energy at those wavelengths out in the universe.

Seeing far into the IR range isn't terribly productive to a warm-blooded creature. Aside from the fogging issues from self-radiation, there's the issue of acuity. Longer wavelengths need larger sensing structures, which means either very large eyes or very poor acuity.

UV causes a problem for protein/DNA-based systems because those high energy photons tend to disrupt molecular bonds (hooray sunburns!) This absorption by molecules tends to limit the amount available for visbility. Not to say that the races wouldn't develop suitable bandpass items, but there is an upper limit to frequency sensing with water/carbon-based life if only because the radiation eventually becomes too penetrating to be captured (think X-rays).

Snargash Moonclaw
11-05-2008, 07:36 PM
Lots of cool stuff here guys, thanks!

Re: racial vision - depends on the system. AD&D (1-2e) IIRC gave elves ultravision & dwarves infravision. 3e did away with the terms (and the attempt at scientific rationalization) and referred to them as Low-light Vision and Dark Vision respectively. For my setting, using GURPS, I'm assigning Ultravision (literal UV sensitivity) to elves and infravision (also literal) to orkhs - interbreeding producing humans with normal vision. I haven't pinned down dwarves and halflings or gnomes as being a hybrid of the two like humans. GURPS Darkvision relies on "some means other than light, radar or sonar" with no attempt at explaining what. Hyperspectral vision ranges from UV to IR. Night Vision is simply a scalable reduction in the penalties imposed by poor lighting.

@Waldronate - I take it your comment about visual acuity vs field of vision among predators and prey implies you've read the CBG thread. I'm thinking of orkhs more along raptor lines as attached. Eyes are forward and nose bridge is minimal, providing much more binocular overlap, with the same peripheral scanning and movement detection. (I think their eyes would have to "bug out" slightly with the outer portion of the skull somewhat receded around the eyes - angling them back slightly.) Visual acuity in terms of detail is reduced - they make poor archers. Basically Paleolithic orkhs developed traits of both predator and prey - falling in a midrange on the food chain. (Of interesting note is Terrence McKenna's hypothesis that paleo/mesolithic humans may owe some success as a species to increased peripheral vision - resulting from consuming psilocybin - permitting them to hunt more effectively.)

Reduced UV would definitely result when only the orange-red sun is visible. Changes in ozone levels as a result of this would permit more of the UVA and B to reach the surface, but mitigating only slightly. Coupled with a protion of the visible light spectrum not being visible to elves could cause the entire race to suffer something along the lines of a 10,000 year collective case of Seasonal Affective Disorder. For any who care to read their ancient (meso-neolithic) history (http://www.thecbg.org/wiki/index.php?title=Fehladurh), this would actually explain a lot. It does warrant some slight revision of that document - incorporating the effects of Despair along with Pride, and probably having some influence on them during the 2nd (early Iron Age) Glacial Period. After the recurrence they would have identified the problem and now be considering plans to deal with/mitigate this when the next one comes around in another 15,000 years. (This is the equivalent of humans starting to plan for a major threat to the species in 1500 years, except that elves are psychologically more inclined to actually do so. . .)

I don't think taht UV or IR visibility would produce "colors" as such. While color (pigment) mixing and light mixing are different, I note that the artist's color wheel is complete. That is, the system is closed and colors blend consistently throughout the spectrum and back to the starting point on the wheel with no gaps, leaving no place where any additional hues might reside should a UV and/or IR sensitive artist seek to mix the additional color(s) on hir palette. Hence I'm inclined to expect these two perceptions to display some other characteristic - most likely producing radical alterations in shade/tint value and/or saturation of the visible colors as well as a slight spectral shift - the latter especially where the opposite end of the visible spectrum is opaque. Playing with the Hilbert Space below would be accomplished by extending one of the bottom corners while raising it's opposite and playing with the effect on the portion of polychromatic visible light below the central white locus where red and violet light mingle. I'm not sure what the extended portions would look lie - esp. for someone with hyperspectral vision extending both sides and adding a large new area across the entire bottom from the magenta section.

11-05-2008, 09:55 PM
By definition, a "color" is a light wavelength band. If you cannot detect a wavelength band then you cannot see that color. If your race can detect colors beyond human perception (IR and UV) then those wavebands would be perceived as additional colors and the race is likely to give them unique names.
The CIE chromaticity diagram you showed above is "full" only for a wavelength band between 380 and 700 nm (blue through red). The shape of the diagram was determined by asking volunteers to specify their perceptions of color and then determining the "typical" values. Thus, the diagram has three "corners" because humans have three waveband sensors. If your race has additional sensor types then the shape of the diagram will change.
As a practical application, consider that the color gamut (the displayable range of colors) for a typical CRT or print process is much smaller than the human visual response. TO see that this might mean, consider three races: humans, race RGB (has a visible range equivalent to the color gamut of a typical CRT), and race CMYK (has a visible range equivalent to the color gamut of a typical 4-color print process, or somewhat smaller than RGB). Humans can see many color differences that are invisible to both races CMYK and RGB. CMYK and RGB would not notice whole swaths of colors that humans could easily distinguish. RGB would be able to see some things that CMYK wouldn't be able to distinguish. Humans would find RGB art less satisfying than their own art because it has a much more limited range of colors and CMYK art less interesting still. Similarly, race RGB might see an uninteresting monochrome canvas when looking at a human painting of delightful shade of variations of saturated green.

I didn't read the CBG thread except at the first post, sorry.

CIE color space definition: http://en.wikipedia.org/wiki/CIE_1931_color_space
color gamut definition: http://en.wikipedia.org/wiki/Gamut
differences in color gamuts: http://dx.sheridan.com/advisor/cmyk_color.html