On 17 May 2002 17:38:30 +0000, Mikhael Goikhman wrote:
>
> On 17 May 2002 08:28:52 +0200, Olivier Chapuis wrote:
> >
> > - The result can be good (and even really cool) if the shadow color
> > (fgsh) is well chosen (the font can come into the game too).
> > At present time fgsh is computed as sh is computed from bg. This is
> > not a definitive solution at all. Any idea is really welcome :o)
>
> In my humble opinion, there is a good effect when fgsh is:
>
> fgsh = (fg + 3 * bg) / 4
>
> All 3 rgb components are evaluated independently unsing this formula.
>
> I.e. fgsh should be something between fg and bg, 3 times closer to bg.
> At least it is much better for a default than "sh".
This formula produces soft shadows, it is ok for a default, because this
does not disturb a user too much, but usually a user wants a more hard
(contrast) text shadow. I think, this revised formula produces better
results in some cases:
fgsh = (5 * bg - fg) / 4
If bg is lighter than fg, the soft formula may be applied. And if fg and
bg are too close one to other, set fgsh to 0, or maybe to (fg + bg) / 4.
I used the following test (runs 2 minutes) for random fg and bg colors:
*FvwmIdent: Font "shadowsize=2:xft:Luxi Mono:Roman:pixelsize=13"
*FvwmIdent: Colorset 21
Next (XTerm) FvwmIdent
# the leading minus is important
-SendToModule FvwmPerl eval detach(); \
for (1 .. 20) { \
$f = 0; \
for $d (0 .. 2) { \
$a[$d] = int(rand(256)); \
$b[$d] = int(rand(256)); \
$c[$d] = int((5 * $b[$d] - $a[$d]) / 4); \
$c[$d] = int((3 * $b[$d] + $a[$d]) / 4) \
if $a[$d] < $b[$d] || $c[$d] < 0; \
$c[$d] = 0 if $a[$d] - $b[$d] < 32 && \
$a[$d] - $b[$d] > -32; \
} \
$e{fg} = sprintf("rgb:%02x/%02x/%02x", @a[0 .. 2]); \
$e{bg} = sprintf("rgb:%02x/%02x/%02x", @b[0 .. 2]); \
$e{fs} = sprintf("rgb:%02x/%02x/%02x", @c[0 .. 2]); \
command("Colorset 21 fg $e{fg}, bg $e{bg}"); \
sleep(3); \
command("Colorset 21 fg $e{fg}, bg $e{bg}, fgsh $e{fs}"); \
sleep(3); \
}
You may use system("xmessage 'fg $e{fg}, bg $e{bg}, fgsh $e{fs}'")
instead of sleep(3) if you want to see numbers too.
This algorithm is not perfect at all, but it produces good results.
Regards,
Mikhael.
--
Visit the official FVWM web page at <URL:http://www.fvwm.org/>.
To unsubscribe from the list, send "unsubscribe fvwm-workers" in the
body of a message to [EMAIL PROTECTED]
To report problems, send mail to [EMAIL PROTECTED]
