Ok, math/color theory question.

[mainegreen]

The math of color theory ain't exactly a strong suit of mine.

I've got a problem here, and I'd figure I'd toss it out and hope the answer is obvious to someone here.

I've got a color pair of rgb(179,216,140) - rgb(236,245,226).
Looks like so:

I've got another color rgb(125,65,153), like so:

It logically is being treated as the pair to the rgb(179,216,140) color. What is the associated lighter color match to the purple one?


The best I'm coming up with is rgb(165,74,247) like so:

But thats not really right. The color I was expecting should be much lighter, but the same damned kind of purple, just less saturated. No idea how to calculate that. I've got some sort of ratio here, but I know I'm screwing up because I'm not taking into to consideration the fact that I'm working on a 0-255 scale. I think.

Ah fuck, I don't really know. Any thoughts?

Does this even make sense?

Please?

[redqueen]

1. mainly because half the people I was with believed it

until I explained the situation to them

[mainegreen]

2. Hmmm, multiple DU tabs open, methinks?

:rofl:

[redqueen]

3. Nah I just had no clue what you were talking about

but it sounded interesting so I just wanted to reply.

:P

[Rabrrrrrr]

4. I can't tell you now what they're working on, but it'll be hell on the day of rectification

[mainegreen]

6. It's a hell of a day now!

x(

[suninvited]

5. If it wasn't for my horse,

I wouldn't have spent that year in college

[mainegreen]

8. Evidently.

I guess my question only makes sense to me.

:crazy:

[suninvited]

11. Actually I do kind of understand the problem

rgb(179,216,140) - rgb(236,245,226)= rgb(125,65,153) - rgb(x,x,x) and you need to figure out the math to come up with the second set that would be mathematically equal to the first set.

You dont think your shade of purple looks right.

[Westegg]

7. You need to suspend the pigment in a synthetic resin...

...Yves Klein proved this over 40 years ago. It's not about numbers. It's your eyes. I mean on the canvas. But if you're talking about computer screens, forget it. They can't compute or compete. Not even close.

[Dr. Strange]

9. How about rgb(180,131,203)?

Does that seem like a better fit?

[mainegreen]

10. Better, but not what I'm looking for.

Somehow, ever color I try comes out too dark. I'm beginning to think rgb math like this is not really valid, because I either get the right lightness, but come out with something too pink, or come out with the right shade relatively, but the wrong brightness.

I may just eyeball it, but I would love to be able to base it on a mathematical principle. Seems more appealing to me.


How did you arrive at that particular shade?

[Dr. Strange]

17. I plugged the rgb values into microsoft paint...

and looked at the hue, sat, lum values. For the first pair, those values are
(59, 118, 168) and (59, 117, 222).
For the purple color, it gave
(187, 97, 103)
so I put in (187, 97, 157) and it gave an rgb of (180, 131, 203).

Maybe this would work for a formula:
http://en.wikipedia.org/wiki/HSL_color_space#Conversion_from_RGB_to_HSL_or_HSV

[Rabrrrrrr]

12. What about 182, 94, 239?

I know nothing of RGB coloring, though, so don't trust me.

But mathematically, it's the same color shift as in the first pair. Number-wise, anyway.

[meegbear]

13. Thats what I would suggest as well

[suninvited]

14. It looks good on the chart, too

how do you do the math?

[Rabrrrrrr]

16. I just did the math as though they were coordinates

(a, b, c) and (d, e, f) being the coordinates of the two colors in the first set, then taking the given (g, h, i) of the purple one, made the coordinate (g+d-a, h+e-b, i+f-c)

I have no idea if that's proper math for RGB, and clearly it won't work in all cases since there's a maximum of any value of 255 and a minimum of 0, but it seemed worth a try.

[mainegreen]

15. How did you arrive at that?

It's slightly closer to what I want, but I'm also beginning to suspect that what I'm actually trying to do is not what I initially thought. I'm beginning to suspect that I need to treat this more as a distance problem, and that I need to figure out where I'm measuring distance from. I think I need to be treating these more like a cylinder. Maybe.

More study is needed, because I think I'm not really fully grasping what question I should be asking.

I'd still like to know how you arrived at those numbers. It's much closer, chroma wise.

[cyberswede]

18. How about 230, 223, 234

which I got just visually,

or 208, 191, 214

which I got using: http://www.easyrgb.com

[mainegreen]

19. oooh, the second one looks right.

Interesting site.

Must spend more time there....


Thank you.

[Jamastiene]

20. I understand what you are trying to do and

I tried a lot of different calculations using the hue saturation and luminosity. Nothing worked.

So, I loaded the 3 colors you have into PhotoImpact 4.0 and got the Hexadecimal values.

The way colors are set up with the hex values is that the rgb values are in triplets of 2

In other words the Hex #000000 color black is the same as 0,0,0 rgb value and Hex#FFFFFF is the same as all 255, 255, 255 rgb values.

The easiest way to look at it is:
the FF values in the Hex Triplet really means 15*16+15=255 which is true, because in Hexadecimal, you have a base 16 numbering system.

So, to get the r or g or b value, you take the first digit in the Hex number and multiply it by 16 and then add the second digit.

Hex Binary Decimal
0 = 0
1 = 1
2 = 2
3 = 3
4 = 4
5 = 5
6 = 6
7 = 7
8 = 8
9 = 9
A = 10
B = 11
c = 12
D = 13
E = 14
F = 15

So, each set of FF = 15 * 16 + 15 = 255, which gives you rgb(255, 255, 255)

Your numbers when converted end up as these Hex values:

rgb(179,216,140) = B3D88C
rgb(236, 245, 226) = ECF5E2
rgb(125, 65, 153) = 7D4199
rgb(165,74,247) = h1, h2, h3

You need to break up the Hexadecimal number into Hex Triplet and find h1, h2, and h3 in Hexadecimal.

So:
B3 D8 8C / EC F5 E2 = 7D 41 99 / h1 h2 h3

You can use the calculator that came with your computer from here.
Here is what I did:
Using the scientific view of the calculator that came with the computer, I set it on Hex and calculated each set of numbers of the Hex Triplet values as a ratio.

B3 D8 8C / EC F5 E2 = 7D 41 99 / h1 h2 h3

B3/EC = 7D/h1
B3h1 = EC*7D
B3h1 = 733C
h1= 733C/B3
h1= A4

D8/F5 = 41/h2
D8h2 = F5*41
D8h2 = 3E35
h2 = 3E35/D8
h2 = 49 (in Hexadecimal, not Decimal)

8C/E2 = 99/h3
8Ch3 = E2*99
8Ch3 = 8712
h3 = 8712/8C
h3 = F6

When you put them all together, you get A449F6

A4 = 10 * 16 + 4 = 164
49 = 4 * 16 + 9 = 73
F6 = 15 * 16 + 6 = 246

which gives you: rgb(164, 73, 246)
If you want it lighter from there, you can just adjust it.

[Jamastiene]

22. Also, and this does look much better...

You can simply load the colors into Paint and look at the hsl values of the first 2 colors and see what the offset is for them between those two colors.

rgb(179,216,140) = hsl(59, 118, 168)
rgb(236, 245, 226) = hsl(59, 117, 222)
rgb(125, 65, 153) = hsl(187, 97, 103)
rgb(165,74,247) = h, s, l

the h value between the first darker/lighter colors is the same, so the h value for your mystery color would still be 187.

the s value between the first darker/lighter colors is offset by -1, so the s value for your mystery color would be 96.

the l value between the first darker/lighter colors is offset by 54, so the l value for your mystery color would be 103+54=157

so rgb(165,74,247) = hsl(187,96,157)
I checked and that one looks much better than the Hex computations. Apparently the luminosity makes the difference.

The Hex Value of that color is #B483CB.

[Jamastiene]

23. Or...

You could just eyeball it.

























:evilgrin:


:hide:

[RandomThoughts]

21. Just do this

Open Microsoft paint
double click on one of the color boxes on the bottom
push 'define custom colors' button

Then find the color you want, it will give you the rgb.