Lawrence Lile <lilel@toastmaster.com> wrote:

> > Reserve 2 bytes, add the new sample to the 16 bit number, grab the
> > MS byte as the average. Is this what you're trying to do?
>
> Very close.  What you are describing is the same as (257)*(Old
> average) - (Old average) + (New data point) .  This would give a
> running average of 257 points.

Lawrence:

I think you (or I) may have misunderstood the paragraph you quoted.
It seems to me that what he's doing is simply accumulating samples
and ALWAYS dividing by 256, no matter how many samples have actually
been taken.

This method isn't equivalent to a running average of 257 points...
In fact, it's only accurate right after the 256th sample has been
taken, and its accuracy falls off dramatically both before and after
that point.

-Andy

=== Andrew Warren - fastfwd@ix.netcom.com
=== Fast Forward Engineering - Vista, California
=== http://www.geocities.com/SiliconValley/2499