M
Mark McKay
I've been using Java for processing graphics for a few months, and
looking for some advice about effiency regarding the growing variety
of imaging options. So far I've stuck with the old producer/consumer
model, since it seems to me to be very speedy (my primary concern) and
it's possible to build the image data directly in an array. I've used
the Java2D functions a bit, and while I like the fancy tools like
spline drawing and gradient fills, I've stayed away because I've found
that BufferedImages take a great deal of time to modify and to render
to screen.
I've started reading the documents for JAI recently, and wondering if
it's worth while switching from the producer/consumer model to the
JAI. I'm worried that since (as far as I can tell) it focuses on
reproducing portions of an image on demand that there will be a lot of
excessive recalculations each time I render an image to screen.
Would anyone know:
Is the producer/consumer system accelerated with native code (for
things such as grabbing pixels or producing new images from arrays of
bytes)?
Given that all my image producers will be able to produce their data
immediately, will the JAI pull model give me any advantages?
Does the JAI work well with VolitileImages?
Thanks
Mark McKay
looking for some advice about effiency regarding the growing variety
of imaging options. So far I've stuck with the old producer/consumer
model, since it seems to me to be very speedy (my primary concern) and
it's possible to build the image data directly in an array. I've used
the Java2D functions a bit, and while I like the fancy tools like
spline drawing and gradient fills, I've stayed away because I've found
that BufferedImages take a great deal of time to modify and to render
to screen.
I've started reading the documents for JAI recently, and wondering if
it's worth while switching from the producer/consumer model to the
JAI. I'm worried that since (as far as I can tell) it focuses on
reproducing portions of an image on demand that there will be a lot of
excessive recalculations each time I render an image to screen.
Would anyone know:
Is the producer/consumer system accelerated with native code (for
things such as grabbing pixels or producing new images from arrays of
bytes)?
Given that all my image producers will be able to produce their data
immediately, will the JAI pull model give me any advantages?
Does the JAI work well with VolitileImages?
Thanks
Mark McKay