Re: [Discuss-gnuradio] truncating bits

[Kunal Kandekar]

I doubt bit truncation will work too well as a method of filtering... if anything, it will probably end up introducing other artifacts into the signal. 

But to truncate bits in GNU Radio, you could try converting your signal to shorts and using an and-by-constant block (gr.and_const_ss) where the value of the constant is, say, 0xFFF0 (or whatever mask you wish to apply to zero out LSBs). There is no GR block to right-shift bits in a sample, but maybe you can do without it. This will effectively add a DC component to your signal proportional to the number of bits you're zeroing out, but I think that will be the least problematic side-effect.

Kunal

On Mon, Oct 3, 2011 at 10:58 PM, sirjanselot <address@hidden> wrote:

Hello,

I have a question in bit truncation.

Say I have a 14-bit FM waveform, digitized.  Underneath that signal
assuming they are both in the same frequency, there is another signal say
an AM signal.  The difference between the two is about 40 dB.  Is there way
such that I can remove the AM signal that is below the FM signal by bit
truncation?

I am curious as how the operation of truncating bits work.  If I do this,
since the FM signal is still about 56 dB higher than the AM signal, if I
were to take the MSBs and and truncate the LSBs that contain AM signal,
will I still be able to demodulate the FM signal?

I am having a hard demonstrating this in matlab to myself since I do know
how to use fixed-point toolbox (yet) so if somebody could please explain I
would be very greatful.

How can I do truncate bits in gnuradio?
--
View this message in context: http://old.nabble.com/truncating-bits-tp32586703p32586703.html
Sent from the GnuRadio mailing list archive at Nabble.com.


_______________________________________________
Discuss-gnuradio mailing list
address@hidden
https://lists.gnu.org/mailman/listinfo/discuss-gnuradio