R help - Oct 2007 - Speeding up image plots

Dear R-users,

I wonder whether it is possible to speed up plots of 2-dimensional 
images (such as image() or levelplot()) in order to export lower quality 
versions of the resulting plots.

I intend to export a large number of images in order to construct 
animated gifs. For this purpose, I used the png()-function in 
conncection with image() and levelplot(). But the underlying arrays are 
so big (900x900 entries) that the export of every single image takes 
very long. Using caTools for exporting animated gifs from 3-dimensional 
arrays (dim(array)=c(900,900,No of plots) works extremely fast, but 
unfortunately, the level of flexibility for plot design is very low (no 
labelling, legends etc.).

I guess the problem is that levelplot() or image() first plot a full 
quality image. Only on export, the quality settings from png() or jpeg() 
(e.g. for resolution) become effective. How could I provide quality 
settings before the time consuming plotting process starts?

Thanks a lot in advance for any kind of help!

Best regards,
Maik

> I wonder whether it is possible to speed up plots of 2-dimensional
> images (such as image() or levelplot()) in order to export lower quality
> versions of the resulting plots.
> But the underlying arrays are
> so big (900x900 entries) that the export of every single image takes
> very long.
I think the only way is to upscale your arrays before calling image()
or levelplot(), maybe with a simple fortran or C routine who scales
the original to a 450x450 (mean of 2x2 pixel blocks, if your values
are summable).
Or you just do a sample() of your original array, provided it has a
certain degree of continuity.
I very often have to produce 512x512 images, and it doesn't seem to be
any particularly intensive procedure; I mean, it is done within
seconds.
ScionForbai