On Thu, Aug 12, 2010 at 9:05 PM, RussellMc <apptechnz@gmail.com> wrote:
>> > What I am saying is that right shifting by 1/(2**N) loses N bits of
>> > precision,
>
>> Actually we were talking about right shifting by N, which is the same as
>> dividing by 2**N or multiplying by 1/(2**N). =A0Right shifting by 1/(2**=
N)
>> makes no sense at all.
>
> I think it's probably a matter of nomenclature - I've little doubt
> that both commentators have a good grasp of what is involved.
>
> I'd read "right shifting by 1/(2**N) ... " as a short hand way of
> saying that 'you achieve division by a factor of 2^N by right shifting
> N bits, thereby losing N bits of precision, unless you use include an
> underflow register in your algorithm'. Among consenting experts in PIC
> who both know what the other guy knows I'd expect such dense compound
> meanings to be well enough understood. But, I may be wrong :-).
>
>
> =A0 =A0Russell

Olin is technically correct (the best kind of correct).  Right
shifting by N is what I meant.

My note about precision was not for experts like Olin, but to point
out a trap for young players.

Regards,
Mark
markrages@gmail
--=20
Mark Rages, Engineer
Midwest Telecine LLC
markrages@midwesttelecine.com

--=20
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.