On Sep 23, 9:33�pm, dbd <d...@ieee.org> wrote:
> On Sep 23, 10:18�am, "mahsad" <mahdi_kbm@n_o_s_p_a_m.yahoo.com> wrote:
>
> > hi
> > suppose you have a interpolation filter (with interpolating factor=3).
> > now,how can i implement this filter in assembly language with only 3
> > cycles?
>
.> You haven't said whether you mean 3 cycles per input sample or 3
.> cycles per output sample.
.> r b-j has suggested an answer for the former.
.> I'll suggest a box car filter for the latter.
On further thought, the first two observations seem reasonable. The
third seems to be a delusion, (perhaps brought on by overmuch
contemplation of Rick Lyon's samples on a ruler analogy in another
thread) unless you can scale an input and (+or-) accumulate in a
single cycle. And this takes advantage of Fred's point that the
interpolation filter has not been specified in your (homework?)
problem.
Dale B. Dalrymple
Reply by dbd●September 24, 20102010-09-24
On Sep 23, 10:18�am, "mahsad" <mahdi_kbm@n_o_s_p_a_m.yahoo.com> wrote:
> hi
> suppose you have a interpolation filter (with interpolating factor=3).
> now,how can i implement this filter in assembly language with only 3
> cycles?
You haven't said whether you mean 3 cycles per input sample or 3
cycles per output sample. r b-j has suggested an answer for the
former. I'll suggest a box car filter for the latter.
Dale B. Dalrymple
Reply by Fred Marshall●September 23, 20102010-09-23
On 9/23/2010 10:18 AM, mahsad wrote:
> hi
> suppose you have a interpolation filter (with interpolating factor=3).
> now,how can i implement this filter in assembly language with only 3
> cycles?
>
Would you care to say what that filter *is*? Or, is it up to us?
Fred
Reply by Dirk Bell●September 23, 20102010-09-23
On Sep 23, 2:40�pm, robert bristow-johnson <r...@audioimagination.com>
wrote:
> On Sep 23, 1:18�pm, "mahsad" <mahdi_kbm@n_o_s_p_a_m.yahoo.com> wrote:
>
>
>
> > suppose you have a interpolation filter (with interpolating factor=3).
> > now,how can i implement this filter in assembly language with only 3
> > cycles?
>
> perhaps, in 3 instruction cycles, you can perform 0th-order
> interpolation and copy each input sample into 3 adjacent output
> samples.
>
> dunno how many would call a zero-order hold much of an interpolator,
> but it is. �sorta.
>
> r b-j
For a generic signal, it has real bad imaging.
Dirk
Reply by robert bristow-johnson●September 23, 20102010-09-23
On Sep 23, 1:18=A0pm, "mahsad" <mahdi_kbm@n_o_s_p_a_m.yahoo.com> wrote:
>
> suppose you have a interpolation filter (with interpolating factor=3D3).
> now,how can i implement this filter in assembly language with only 3
> cycles?
perhaps, in 3 instruction cycles, you can perform 0th-order
interpolation and copy each input sample into 3 adjacent output
samples.
dunno how many would call a zero-order hold much of an interpolator,
but it is. sorta.
r b-j
Reply by mahsad●September 23, 20102010-09-23
hi
suppose you have a interpolation filter (with interpolating factor=3).
now,how can i implement this filter in assembly language with only 3
cycles?