On 5/9/07, Mark Probst <mark.probst@xxxxxxxxx> wrote:
> On 5/9/07, Øyvind Kolås <pippin@xxxxxxxx> wrote:
> > There is no real equivalent to a Channel Mixer operation at the moment.
>
> Does GEGL support single-channel (plus maybe alpha) output/processing?
Not sure what you mean by this, but it should be possible to rig up a
channel mixer using the arithmetic operations + component
extraction/muxing nodes.
> > When it comes to how an array of values are best expressed, there
> > probably are some examples in GIMP plug-ins.
>
> To be honest, I'm not sure if I want to do this with an array of
> values, for the simple reason that if you're using floats for image
> processing, your array either has to be pretty big or you're losing
> precision.
>
> I'd prefer either for GEGL to have some curve data structure that can
> do linear and spline interpolation, or for the operation to just take
> a user-implemented function which can compute any curve value. The
One, or two arrays, are that datastructures that would be needed to
express a curve. Passing in a function means having an expression
parser in-place, in general I feel that doing so is adding unwarranted
complexity. I think the curves tool in GIMP is using an array of
values + interpolation.
> Another thing that I'll need sooner or later is brush painting (for
> dodge/burn functionality). I saw in your FOSDEM talk that that's
> planned, or at least considered, for the future. Could you give an
> outline on how that would or could be implemented?
A brush system is outside the scope of GEGL, the only thing that would
be part of GEGL is actual procedural rendering of brushes. The rest of
the infrastructure needed to implement a paint-core is already present
in GEGL; and there is many (at least two, probably more) ways of
implementing such a system. Personally I do not want to start
experimenting with such things until after internal refactoring of
GEGL is done, since any profiling done before that will provide bogus
information.
/Øyvind K.
--
«The future is already here. It's just not very evenly distributed»
-- William Gibson
http://pippin.gimp.org/ http://ffii.org/
_______________________________________________
Gegl-developer mailing list
Gegl-developer@xxxxxxxxxxxxxxxxxxxxxx
https://lists.XCF.Berkeley.EDU/mailman/listinfo/gegl-developer
