Cudatext: Some HTML color value background coloring make numbers hard to read

[bogen85]

Where in the CudaText code is the foreground determined for a given background?
Some of the existing foreground choices make the numbers hard to read.

I did a quick search, but I'm not sure what I should be looking for.

Some algorithm like this:
for each background red/green/blue 0 to 255 level:
if background < 128, foreground = 255
if background > 127, foreground = 0

Would always result in a readable combination, but the result would be 8 different foreground colors being used.

There would likely be a way to reduce the number of foreground colors used (or just use the average of all 3 from the above algorithm, and if the result is < 128 use white, and greater than 128, use black).

This is a suggestion for an enhancement. I'm willing to do the work and submit a PR. but I'd like to discuss it here first and get an idea of where in the code to look.

[bogen85]

So, I decided to start here: https://wiki.freepascal.org/CudaText#HTML_color_codes_underlining

Searching for "underline_color_files" led me here: https://github.com/Alexey-T/CudaText/blob/master/app/formmain_loadsave.inc#L994
Which led to: https://github.com/Alexey-T/CudaText/blob/master/app/formframe.pas#L880

And I ended up in this area: https://github.com/Alexey-T/CudaText/blob/master/app/formframe.pas#L969

      if bColorizeBack then
      begin
        C.Font.Color:= _ContrastColor(NColor);
        C.Font.Style:= [];
        C.Brush.Color:= NColor;
        CanvasTextOutSimplest(C, X1, AY, Copy(AStr, NStartPos, NLen));
      end
      else
      begin
        C.Brush.Color:= NColor;
        C.FillRect(X1, Y-NLineWidth, X2, Y);
      end;

[bogen85]

And _ContrastColor is here: https://github.com/Alexey-T/CudaText/blob/master/app/formframe.pas#L858

function _ContrastColor(AColor: TColor): TColor;
var
  bLight: boolean;
begin
  bLight:= ((AColor and $FF) + (AColor shr 8 and $FF) + (AColor shr 16 and $FF)) >= $180;
  Result:= UiOps.HtmlBackgroundColorPair[bLight];
end;

[bogen85]

OK, so the existing code is based on an average, and determining black or white... ($180 is sum of $80+80+$80)

[bogen85]

For the green, the first hard to read hard in my example, 124+252+0 = $178, so white is chosen...
Likely hard to read, because green is perceived as a higher brightness than red or blue.
So maybe each of red/green/blue should be scaled differently...

[bogen85]

Yes, it seems like one does need to scale each one differently to get the perceived brightness...
Calculating the Perceived Brightness of a Color
HSP Color Model — Alternative to HSV (HSB) and HSL

The Naïve RGB Algorithm

If RGB 0,0,0 is black and RGB 255,255,255 is white then R+B+G
should give us a good approximation of the color brightness,
shouldn’t it?

The problem is that some colors are brighter than others,
red is brighter then blue and green is brighter then red

[bogen85]

Checkpoint 2.2 - Ensure that foreground and background color combinations provide sufficient contrast when viewed by someone having color deficits or when viewed on a black and white screen

From the W3 page, color brightness is determined by the following formula:
((Red value X 299) + (Green value X 587) + (Blue value X 114)) / 1000

So maybe we could say Red gets scaled by 3, Green by 6, and blue by 1.

So rather than $80+$80+$80 and using $180 (384), we use ($80 * 3) + ($80 * 6) + ($80 * 1) or $80 * 10 = $500 (1280)

I may give that a try, but would like some feedback first.

[bogen85]

Alright, I went ahead and tried it...

function _ContrastColor(AColor: TColor): TColor;
var
  bLight: boolean;
  red, green, blue: integer;
begin
  red := AColor and $FF;
  green := AColor shr 8 and $FF;
  blue := AColor shr 16 and $FF;
  bLight := (red*2) + (green*6) + blue > $500;
  // bLight:= ((AColor and $FF) + (AColor shr 8 and $FF) + (AColor shr 16 and $FF)) >= $180;
  Result:= UiOps.HtmlBackgroundColorPair[bLight];
end;

[bogen85]

All the greens that were hard to read before look much more readable now:

I plan to submit a PR.

[bogen85]

Alright, created the PR...

[bogen85]

Oh, I see I messed up (red *2, instead of red*3). I'll withdraw the PR and redo it tomorrow.
I'll also likely use the scaling given by W3: ((Red value X 299) + (Green value X 587) + (Blue value X 114)

[bogen85]

Alright, I submitted PR #3636

[Alexey-T]

Thanks for the finding; applied the PR.