Perfect, Brian, you've answered my question. I had not understood the
expanding effect of 10 on the filter, nor the relationship of 10%2 to
restore to the correct size.
The biggest thing I missed was that R *is* used. It's used in the final
calculation. I can explain how the inexact sizing on the shadow is
compensated for: (($W){.R) crops R to fit W so that W can be simply
subtracted from it.
Thank you!
On Fri, Jan 25, 2013 at 1:39 PM, Brian Schott <[email protected]>wrote:
> Tracy,
>
> I am just guessing here.
>
> You mention 1r5 and 2 but you don't mention the 10 in the maximizing
> filter. But 10%2 = 5. I am puzzled also though because the moving
> average of length 8 remove 4 items on each edge, too. So maybe the
> resizing cannot and is not exact. But the sizing is pretty close
> according to my approximate manual measurement on my screen.
>
>
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
