Skip to content

Opposite color in R

6 messages · Atte Tenkanen, Peter Dalgaard, Ken Knoblauch +3 more

Original
Atte Tenkanen
# 
Hi,

I have tried to find a way to find opposite or complementary colors in R.

I would like to form a color circle with R like this one: 
http://nobetty.net/dandls/colorwheel/complementary_colors.jpg

If you just make a basic color wheel in R, the colors do not form 
complementary color circle:

palette(rainbow(24))
Colors=palette()
pie(rep(1, 24), col = Colors)

There is a package ?colortools? where you can find function opposite(), 
but it doesn?t work as is said. I tried

library(colortools)
opposite("violet") and got green instead of yellow and

opposite("blue") and got yellow instead of orange.

Do you know any solutions?

Atte Tenkanen
Peter Dalgaard
# 
Not directly, but a few hints: 

First read up on "complementary colors" in Wikipedia. In particular, note that the traditional color circle does not satisfy the modern definition of opposite-ness. E.g. red paint mixed with green paint is brown, not black or grey.

The construction of the color circle is simple in principle: red, blue, yellow go at 0, 120, 240 degrees, the other colors on the circle are formed by mixing two primaries in varying proportions: green (at 180 deg) is an equal mixture of blue and yellow, violet (at 60 deg) of blue and red, orange (at 300 deg) of red and yellow. Blue-green (at 150 deg) would be half blue, half green, alias three quarter blue, one quarter yellow. Etc.

The tricky bit is that the above mixtures are subtractive mixtures (mixing paint rather than light beams) and I don't know how to make a subtractive color mixture in the additive RGB space that we usually work in. Maybe there are tools in the colortools package?

-pd

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Ken Knoblauch
    

    

      
      #
    

  


  

      
peter dalgaard <pdalgd <at> gmail.com> writes:

            

      
      
      
    

        
<attenka <at> utu.fi> wrote:

            

      
      
      
    

        
or complementary colors in R.

            

      
      
      
    

        
like this one: http://nobetty.net/dandls/
colorwheel/complementary_colors.jpg

            

      
      
      
    

        
the colors do not form complementary color circle:

            

      
      
      
    

        
you can find function opposite(), but it doesn?t work as is

            

      
      
      
    

        
Wikipedia. In particular, note that the traditional color

            

      
      
      
    

        
of opposite-ness. E.g. red paint mixed with green paint is

            

      
      
      
    

        
in principle: red, blue, yellow go at 0, 120, 240 degrees, the

            

      
      
      
    

        
two primaries in varying proportions: green (at 180 deg) is

            

      
      
      
    

        
(at 60 deg) of blue and red, orange (at 300 deg) 
of red and yellow.

            

      
      
      
    

        
half green, alias three quarter blue, one quarter
 yellow. Etc.

            

      
      
      
    

        
subtractive mixtures (mixing paint rather than light beams)

            

      
      
      
    

        
color mixture in the additive RGB space 
that we usually work in.

            

      
      
      
    

        
To start with, you should be specifying your "colors"
or lights actually in an additive color space like
CIE 1931 xy,
https://en.wikipedia.org/wiki/CIE_1931_color_space
which you can do in the colorspace package.
But this is based on an average observer and
the results are unlikely to match a given 
individual's vision.  On top of that, decisions made
when this norm was specified are such that it
deviates from human vision for short wavelengths
so that you would be better off using a corrected
version like that proposed by Judd in the 1950's
or for the most recent suggestion see
ww.cvrl.org
under 
New CIE XYZ functions transformed 
from the CIE (2006) LMS functions

best, 

Ken

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Jim Lemon
    

    

      
      #
    

  


  

      
Hi Atte,
If you look at the colors produced by rainbow(12):

rainbow(12)
 [1] "#FF0000FF" "#FF8000FF" "#FFFF00FF" "#80FF00FF" "#00FF00FF" "#00FF80FF"
 [7] "#00FFFFFF" "#0080FFFF" "#0000FFFF" "#8000FFFF" "#FF00FFFF" "#FF0080FF"

they are complementary additive colors. That is, in the RGB color
space, the colors at the opposite sides of the wheel would add to
white (#FFFFFF) if mixed. The colors in the diagram you mentioned
don't look like additive colors. Perhaps that diagram represents a
subtractive (i.e. pigment) color space but based on the additive (red,
green blue) primaries. Also remember that WYSINNWOPG (what you see is
not necessarily what other people get)

Jim

        
On Sun, Jul 26, 2015 at 4:45 PM, ken knoblauch <ken.knoblauch at inserm.fr> wrote:

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Franklin Bretschneider
    

    

      
      #
    

  


  

      
Dear Atte Tenkanen,


Re:

            

      
      
      
    

        
Actually, yellow and blue are complementary colours, but red and green aren't. 

The human visual system has three types of cones: red-sensitive, green-sensitive and blue-sensitive.
(the labels are approximate, e.g. red-sensitive cones have their optimum sensitivity at a wavelength we might call orange, but for understanding colours, R-G-B is the useful standard designation).
A certain combination of these three together, such as in sunlight, is seen as white. In the digital domain, the three "colour channels" of an image are usually scaled to 8-bit numbers, i.e. from zero up to and including 255. So, all three channels 255 makes white.

Leaving one of the three colors out yields yellow (no blue), magenta (no green) and cyan (no red). The pairs yellow-blue, magenta-green and cyan-red  are truly complementary colours.

Colours are the result of the wavelength of the light, so one would expect colours to lie on a linear scale, from about 700 nm (red), through 550 (green) to about 440 nm (blue).

There is a complication, however: the photosensitive pigment of our red cones has a second action peak past that of the blue cones, so past pure blue we see a sort of reddish blue, in other words violet or purple. Therefore, the colours can be plotted in a circle, where violet and purple fill the gap between blue and red.

Using a combination of the three ground colors R, G and B, any desired colour shade can be composed. Orange, for example, consists of (approximately) all red and half green. 

- - - - - - - - -

R has ample possibilities to compose colours or colour palettes, with which one can create (almost continuous) gradients or stepwise colour patches.
Examples are col2rgb():

            

      
      
      
    

        
[,1]
red    255
green  165
blue     0

            

      
      
      
    

        
[,1]
red    238
green  130
blue   238

Cindy Brewer wrote a fine set of colour functions, adapted to R by Erich Neuwirth. See package "RColorBrewer".

And much can be done with the standard R distribution:
 
The following code plots a some colours in a circle, with the complementary colours at opposite sides (so crudely what you're after):


# define colour triplets
reds =   c( 255, 255,  255,    0,    0,    0,    0,  128)
greens = c(   0, 127,  255,  255,  255,  127,    0,    0)
blues=   c(   0,   0,    0,    0,  255,  255,  255,  255)
n = length(reds)
#  compute circle to plot in
stp = 2*pi/n
th = seq(0,2*pi-stp, length.out=n)
x = cos(th); y=sin(th)
#  plot (on a Mac, for other OSses call the appropriate grahics window
quartz(w=5, h=5)
par(xpd=NA)
plot(x,y,pch=15, cex=8, col=rgb(reds, greens, blues, maxColorValue = 255), asp=1, axes=FALSE, xlab='', ylab='')
points(x,y,pch=0, cex=8, col="black")
# arrows connect the complementary colours
arrows(0,0, 0.7*x, 0.7*y, length = 0.25, col = "grey")

Hope this helps;
Best wishes,


Frank
------

Franklin Bretschneider
Dept of Biology
Utrecht University
bretschr at xs4all.nl

  
  


  


      

    

    
1 day later

    

      
      

        


  

        

      Steve Taylor
    

    

      
      #
    

  


  

      
I wonder if the hcl colour space is useful?  Varying hue while keeping chroma and luminosity constant should give varying colours of perceptually the same "colourness" and brightness.

?hcl
pie(rep(1,12),col=hcl((1:12)*30,c=70),border=NA)


-----Original Message-----
From: R-help [mailto:r-help-bounces at r-project.org] On Behalf Of Atte Tenkanen
Sent: Sunday, 26 July 2015 7:50a
To: r-help at r-project.org
Subject: [R] Opposite color in R

Hi,

I have tried to find a way to find opposite or complementary colors in R.

I would like to form a color circle with R like this one: 
http://nobetty.net/dandls/colorwheel/complementary_colors.jpg

If you just make a basic color wheel in R, the colors do not form 
complementary color circle:

palette(rainbow(24))
Colors=palette()
pie(rep(1, 24), col = Colors)

There is a package ?colortools? where you can find function opposite(), 
but it doesn?t work as is said. I tried

library(colortools)
opposite("violet") and got green instead of yellow and

opposite("blue") and got yellow instead of orange.

Do you know any solutions?

Atte Tenkanen

______________________________________________
R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.