Re: [Discuss-gnuradio] FIR filters and set_history()

[Dan Halperin]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

George Nychis wrote:
> What general_work() would do is use the first ntaps() samples as
> "history" and start producing output at in+ntaps().  I would then
> consume ninput_items-ntaps() to then keep those last ntaps() samples in
> the input queue as the "history" for the next series of input.
> 
> I'm sorry if my logic is slightly or completely off.  But then I look at
> most of the FIR filters, and see the use of set_history(), great!  This
> sounds like it is exactly what I want.  Furthermore, looking at the
> documentation it states that it is what I want:
> http://gnuradio.org/trac/browser/gnuradio/trunk/gnuradio-core/src/lib/runtime/gr_block.h#L62

What set_history actually does is prepend that many zero samples. But it
still adds forecast(noutput_items) samples to the input buffer before
calling work... So when work() is called with noutput_items,
history+forecast(noutput_items) samples are ready in the input buffer,
starting from index 0.

If you look at the implementation of filterN, it reads at samples
0..ntaps-1. So when you call filterN with noutput_items as the length
parameter it will eventually read items 0..noutput_items+ntaps-1. Here
you've set ntaps = history.

In the normal code in the trunk, this works just as expected. But when
you make your custom vectors, you only pull noutput_items out of the
input buffer, not noutput_items+history.

Also, my understanding of a normal matched filter is that it does the
normal complex multiplication (by the conjugate, but I'm assuming you've
done that already before building the taps), not the real-wise and
complex-wise ... inner product, I guess? ... as you've defined it.

- -Dan
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQFHqXwCy9GYuuMoUJ4RApVoAKDC3Xrm7spiGzCZZNaS5tvYuPXzzwCfbQzN
pjjttuBVpMgY4vPo6l5/KLc=
=QVtD
-----END PGP SIGNATURE-----