Rick Lyons
12-12-2003, 02:58 PM
Hi Guys,
I have a simple (I think) filtering question.
I've programmed (years ago) microcontrollers,
but I've never programmed any of the modern DSP
chips with their "zero-overhead looping" and
single-cycle multiply-and-accumulate features.
(I know, I know. I should get a DSP starter kit!)
So the DSP chips are efficient for implementing
FIR filtering.
Am I correct in assuming that programmable DSP
chips *cannot* take advantage of the "folded FIR
structure" that's possible when a FIR filter's
coefficients are symmetrical. (You remember that
"folded FIR" idea, ... where an M-tap FIR filter
can be implemented with roughly M/2 multiplies
per output sample.)
I've searched around on the Internet, but haven't
found a solid answer to my question.
Thanks,
[-Rick-]
I have a simple (I think) filtering question.
I've programmed (years ago) microcontrollers,
but I've never programmed any of the modern DSP
chips with their "zero-overhead looping" and
single-cycle multiply-and-accumulate features.
(I know, I know. I should get a DSP starter kit!)
So the DSP chips are efficient for implementing
FIR filtering.
Am I correct in assuming that programmable DSP
chips *cannot* take advantage of the "folded FIR
structure" that's possible when a FIR filter's
coefficients are symmetrical. (You remember that
"folded FIR" idea, ... where an M-tap FIR filter
can be implemented with roughly M/2 multiplies
per output sample.)
I've searched around on the Internet, but haven't
found a solid answer to my question.
Thanks,
[-Rick-]