There seem to be three different colour numbering systems for colors in PD:

• The positive integers, which sort colors in a seemingly random order;

• The negative integers, which sort colors in the RGB scheme (so that -1 to -256 is blue, 256 * (-1 to -256) is green, and 256^2 * (-1 to -256) is red;

• The numbers that are used to save colored objects in the .pd file, which confusingly are different than the RGB colors. (color an object with a positive or negative number, save the file and open it in a text editor, and you will not find the same numbers as you used to color the object).

Does anyone know the relationship between these three systems? Actually, I'm mostly looking for the relationship between 1 and 2, ie. I am looking to find the relationship between the positive color numbers and their RGB equivalents. The third one is just a curiosity.

Hi Liam,

You can use [ds-color-sel] from jmmmp's library to see how the colour numbers relate to the colours themselves. This object displays 100 colours per line in 9 lines in total (so 1st line has colours 0-99, second line has colours 100-199, etc.).

Take care,
Gilberto

Is [ds-color-sel] the same as [pd-colors]? I see the latter in the jmmmp library, but not the former. This seems to add a fourth color numbering system, the one used for data structures! Unfortunately, though, it still doesn't solve the problem for me...

Basically, I am trying to make an abstraction that will calculate the inverse of a particular color. Actually, it doesn't matter if the new color is the exact inverse of the old; what matters is that it is in high contrast to it. The objective is to color a slider with only one number, and for the label to be automatically colored in high contrast, so that it is readable against the background.

The attached patch isolates the Red, Green & Blue components of an RGB color and then addss 128 to all of them. This achieves satisfying contrast for negative RGB colors, but not for positive numbers, which don't seem to follow any pattern at all. (Input +139, for instance, gives two very similar colors which are hardly distinguishable from one another).

If there is a formula for converting from positive colors to RGB, I could probably generate a contrasting color for them. On the other hand, maybe there is an easier way to find the inverse of a color?

colortest.pd

Just did some research: to find the inverse color, you subtract the old RGB colors from 255, ie.

new_red   = 255 - old_red
new_green = 255 - old_green
new_blue  = 255 - old_blue

I updated the patch to reflect this, however, it still doesn't give an inverse color for positive color numbers.

inverse.pd

Is [ds-color-sel] the same as [pd-colors]? I see the latter in the jmmmp library, but not the former.

Sorry Liam, in fact [ds-color-sel] was created by the same author (Joao Paes) but it's not included in the jmmmp library. I don't recall where I downloaded it from, but you can download it from here: ds-color-sel.zip

It's basically the same system as [pd-colors], but in a much more handy object.

But I agree with you that these systems are very confusing and I also cannot possibly think of a way right now to systematize them, much less to find the inverse of a colour. Sorry

Take care,
Gilberto

I feel like there must be a system, though. Either there is a formula for converting from positive color numbers to RGB values, or someone typed them in individually, which seems very unlikely (there are at least 10^8 of them!)

I feel like there must be a system

Oh, most certainly there is! I can't imagine a programmer creating a random system just for the sake of it. Now how the system is organized is a completely different story...

I've come up with a solution based on the observation that there are actually only 30 positive color numbers. After 30, the series repeats. This patch just has those 30 numbers programmed in, to calculate the contrast. It's nothing fancy, but it works reasonably well. Maybe someone else will find it useful.

highcontrast.zip

I was still finding the contrast a little hard for reading, so I decided just to use black and white for the text color. Here is a version that outputs three values: one for the input number, one for a high contrast color, and one for black or white (whichever contrasts more).

highcontrast1.2.zip