What is the decimal RGB scale used in Matlab called?

Hi there,

I have been trying lots of websites to find colours for my plots. The colours i find on webistes that are RGB are ont on the decimal scale. Therefore is there another scale I should search for, or is there a way to convert them?

Example:

• RGB: (37, 147, 174)

Many thanks!

[Rik]

The colors in Matlab are generally 8bit, meaning that the colors are scale 0-255. So I don't know what you would mean by converting.

Answer 1

[Adam Danz]

As Rik Wisselink mentioned, RGB values are scaled to 0-255 (8 bit means 2^8 which equal 256). Matlab normalizes each RGB value to this 256 range. So a value of 0.5 means 50% of the 256 range.

For example, the color crimson is represented as [0.85938 0.078125 0.23438]. When you multiply this by 256,

 [0.85938     0.078125      0.23438] * 256
ans = 
    [  220      20       60]

Likewise, if you have an RBG value of [0, 250, 100], matlab will represent it as

    [0, 250, 100] /  256 
    ans = 
        [ 0      0.97656      0.39063]

[UPDATE]

Scaling by 255 makes more sense than by 256 for reasons explained in the comment section below.

[Image Analyst]

I often use the eyedropper tool in Photoshop to pick a color I like. Photoshop will give you the color in the range 0-255. Then I go into MATLAB and use Adam's last snippet to get the RGB value MATLAB will like for things like drawing lines/curves with the plot() or line() functions:

crimson = [0, 250, 100]  /  256; % Dividing by 256 is needed!
plot(x, y, 'Color', crimson);

[Adam Danz]

Here's an additional source I often use. If you click on a color name, you'll see the 8-bit RGB numbers on the right.

https://en.wikipedia.org/wiki/List_of_colors_by_shade

There are links to additional color palettes at the bottom.

[Steven Lord]

There is a tool, uisetcolor, that allows you to select a color (either from a fixed list of standard and recent colors or from a gradient) and returns the RGB value for that color or changes the color of the object whose handle you passed into uisetcolor.

[Stephen23]

@ Adam Danz & Image Analyst: errrr... why divide by 256?

Why not just divide by 255 (that being the maximum value per channel, and is exactly what MATLAB's inbuilt functions do, e.g. im2double)? Is there a secret reason why you want to scale the colors a smidgen darker, and totally ignore white?

>> im2double(uint8(255)) % correct scaling
ans =  1
>> 255/255 % correct scaling
ans =  1
>> 255/256 % your suggestions
ans =  0.99609

See also the MATLAB documentation:

https://www.mathworks.com/help/matlab/creating_plots/working-with-8-bit-and-16-bit-images.html

which explains:

"Conversely, divide by 255 after converting a uint8 RGB image to double:"

RGB64 = double(RGB8)/255

[Adam Danz]

That's a good point and the better scaler is 255 for that reason.

The rgb() FEX function scales all values < 240 using 256 and then essentially scales values 241:255 using 255. The author notes that scaling by 256 more often gives rounder numbers such as [0 128 0] / 256 = [0 .5 0] but that rarely matters.

Answer 2

[TADA]

some functions like plot use color intensity values (between 0 and 1) to represent RGB values, instead of what you are used to and is used in web browsers (values between 0 and 255).

you can just take your RGB vector and divide it by 255

also I found this rgb function very useful

[Stephen23]

+1 at least someone got the scaling right.

Answer 3

[Walter Roberson]

uint8() the decimal values to get something that MATLAB will understand.