New vga.getPixel function.

[Fernando Perez]

Although logic advises me to wait for you to publish instructions for its correct use, impatience leads me to try the new instruction now.
I want to use it in programs that use the 256 colors of VGA Annex, but for testing I have limited myself to the 24 named colors.
I have programmed and executed this:

Code: [Local Link Removed for Guests]

' Annex32-S3 CAN DMT VGA 1.51.2 qio opi LFS 
vga.delete
vga.init 2
vga.fill white
vga.show

i = 1 : fill = 0

for y = 140 to 300 step 80
  for x = 8 to 565 step 78
    READ c$
    command "fill = vga.color(" + c$ + ")"
    vga.rect x, y, 76, 76, black
    vga.rect x+1, y+1, 74, 74, fill, 1
    px = vga.getPixel(x+10, y+10)
    wlog c$, fill, px, fill-px
    incr i
  next x
next y  

vga.show

END

DATA "black","navy","blue","darkgreen","darkcyan","green","cyan","maroon"
DATA "purple","olive","darkgrey","skyblue","violet","brown","greenyellow","silver"
DATA "lightgrey","red","magenta","orange","pink","gold","yellow","white"

With these results:

image.png

I don't see a clear relationship between them. And, furthermore, in the colors Silver and lightGrey it returns the same result.

How is it used?

image.png

[cicciocb]

The logic is quite simple, there are only 8 bits for the colors.

The 16 bits 565RGB is converted to 8 bits 332RBG; this means that 8 bits are lost in the conversion

So, converting back from 8 bits to 16 bits, only 8 bits will be significative and the others will always be at 0.

The exact notation is

16 bits : RRRRRGGGGGGBBBBB
8 bits : RRRGGGBB

image.png

The conversion is done as shown in the image below :

image.png

As you can see, only 8 bits are transferred and the other 8 bits are "lost" in the conversion.

Converting back to 16 bits, the 8 "missing bits" are simply replaced by 0 :

image.png

From a practical point of view, the significative bits represents this number

image.png

So, the color read back, will be the = (original_color) AND &hE718

If you look at your table, you'll see that the column PX corresponds to column Fill and &hE718.


Example
violet -> 37212 -> 37212 and &hE718 = 33048

It is normal that different 16bits colors can have the same result but is simply because, at the end, they are converted to the same 8 bit color

[Fernando Perez]

I finally understood it.
I couldn't find a clear explanation on the Internet.
Thank you, Francesco.

[Electroguard]

The logic is quite simple...

Masterful - how to start an inferiority complex then stop it with a big and colourful explanation !