> I replaced both functions as you suggested. No it looks a little
> better but not much.
> From my perspective the problem is that imlib2 stores color ranges
> internally in rgb.
In part.. As I mentioned, the use of the wrong scale was only
part of the problem. The other part is that it's doing an initial
interpolation in rgb color-space.. and that'll give incorrect results
since hsv<->rgb are not linear conversions. Furthermore, it also does
an initial hsv->rgb conversion to store the color range in rgb, as you
mention, and that causes some further (small) inaccuracies (since the
conversion hsv->rgb->hsv doesn't always give an identity).
> Maybe a better solution is to calculate the hsv color by myself and
> draw each line by myself.
> Although it may be less efficient however in this case I have one
> advantage to the imlib2 hsv color range function:
> I will not loose information from back and forth converting between
> colorspaces.
That would be one way.. and actually you might get something
which is even faster.
You can also try and do something like in evas, where though
the initial colors are input as argb, the actual interpolation is done
completely in ahsv space. You can do even better in imlib2 by, as you
mention, store the color range as ahsv values, and then proceed as
mentioned, thus avoiding any conversions to begin with.
So, if you have some time you could try and work out a patch
for imlib2 that would do this.
_____________________________________________________________
Click for free info on online masters degrees and make up to $150K/ year
http://thirdpartyoffers.juno.com/TGL2121/fc/Ioyw6i3nNjdFYdDR3fGPlZCD0wMLqBR5Prep3GzTlswkYigkTWnJcE/
-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
enlightenment-devel mailing list
enlightenment-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
