Wikipedia:Reference desk/Archives/Computing/2021 June 12

Computing desk
< June 11 << May | June | Jul >> Current desk >
Welcome to the Wikipedia Computing Reference Desk Archives
The page you are currently viewing is a transcluded archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.


June 12

edit

Color 50% 50% 0%

edit

We know that all these colors are:

  • Red = 100% 0% 0%
  • Yellow = 100% 100% 0%
  • Green = 0% 100% 0%
  • Cyan = 0% 100% 100%
  • Blue = 0% 0% 100%
  • Magenta = 100% 0% 100%
  • Black = 0% 0% 0%
  • White = 100% 100% 100%
  • Gray = 50% 50% 50%

However, what about 50% 50% 0%?? On the DOS computers of the early 90's it was kind of orange or brown; on Windows computers it is more olive green. Why this mismatch?? Georgia guy (talk) 23:51, 12 June 2021 (UTC)[reply]

I think it's to do with the type of graphics adapter. The 8-bit IBM CGA adapter with a 16-color palette did this sort of thing, which is 50-50-0 with 8 bits:
dark yellow
6 #AAAA00
but RGBI monitors starting with the IBM 5153 with extra circuitry added some intensity to give this alternate brown:
brown
6 #AA5500
The Enhanced Graphics Adapter#Color palette and later 16-bit adapters with a 64-color palette do this, which is 50-50-0 with 16 bits.
olive
48 #555500
MinorProphet (talk) 02:22, 13 June 2021 (UTC)[reply]
To make things (perhaps) a little clearer: as you note, Gray=50-50-50, but the intensity bit allows two shades of gray, 'dark white' and 'bright black' - see also List of monochrome and RGB color formats#4-bit RGBI. These correspond to the 50-50-0 of dark yellow and olive in the examples above.
Light gray
7 #AAAAAA
Dark gray
8/56 #555555
MinorProphet (talk) 12:19, 13 June 2021 (UTC)[reply]
On a roll, now: the most legible contrast is something like this:
Cmd prompt
#FFFF00 on Blue
#0000FF skies, MinorProphet (talk) 13:14, 13 June 2021 (UTC)[reply]

The situation is a lot weirder than you think. See [ https://www.youtube.com/watch?v=wh4aWZRtTwU ]. Related: [ https://www.youtube.com/watch?v=uYbdx4I7STg ]. --Guy Macon (talk) 17:30, 13 June 2021 (UTC)[reply]

Guy Macon, that channel is priceless! Thanks for thinking of it! Elizium23 (talk) 17:34, 13 June 2021 (UTC)[reply]
  • Good catch! After the sun goes down, I am going to try the experiment from the above video; viewing that greenish-brown or brownish green (to my eyes) #808000 box on a black background in a dark room. Will become orange as the video predicts?
50% 50% 0%
#808000
Some older monitors (and some image formats?) show a noticeable difference between     #7F7F00     and     #808000    . --Guy Macon (talk) 22:30, 13 June 2021 (UTC)[reply]
Although they have been quite rightly reverted, I would like to apologise for my previous intemperate and ill-judged comments, which were out of place. Sorry. MinorProphet (talk) 02:09, 14 June 2021 (UTC)[reply]
Despite the fun and games with the brown sticky stuff (which you apparently can't see anyway), is anyone going to actually respond to the OP's question with a concise and technically correct answer? I'm itching to find out. MinorProphet (talk) 07:28, 14 June 2021 (UTC)[reply]
Isn't the answer given in Color Graphics Adapter § With an RGBI monitor: "Color 6 is treated differently; when using the formula above, color 6 would become dark yellow, as seen to the left, but in order to achieve a more pleasing brown tone, special circuitry in most RGBI monitors, starting with the IBM 5153 color display, makes an exception for color 6 and changes its hue from dark yellow to brown by halving the analogue green signal's amplitude."? (Essentially the first part of your first reply, except that this is a 4-bit colour space, not 8-bit.)  --Lambiam 09:52, 14 June 2021 (UTC)[reply]
And within that colour space, there was no such thing as 50% 50% 0% or #808000. Expressed in hexadecimal, each of the RGB bytes was either #00, #55, #AA, or #FF. Moreover, except for the exceptional treatment of color 6, either all three bytes were even, or all three were odd. That leaves 16 possibilities. The closest to #808000 in that space was #AAAA00, which was changed to #AA5500.  --Lambiam 10:16, 14 June 2021 (UTC)[reply]
Thank you. I was talking about an '8-bit adapter', which as you say only uses 4 bits for the color space: I was aware that my answer was somewhat imprecise. I used try to program the Intel 8051 UART serial adapter with Intel 8086 assembler, making up my own cables etc., but I have forgotten 98% of it except how much fun it was. I used to get paid for this sort of thing, but after a stroke some years back the details have become a little hazy. Maybe I shouldn't attempt to answer this sort of question any more. Suggestions on the back of a DIP module, please. MinorProphet (talk) 13:53, 14 June 2021 (UTC)[reply]