Re: [Discuss-gnuradio] matlab/octave sample stream, 8 bit samples FEEDBACK

[Lee Patton]

On Wed, 2006-08-09 at 15:26 +0200, Vincenzo Pellegrini wrote:
> thanks Eric, thanks Lee.
> 
> octave import function worked straightforward, and was actually very
> useful.
> 
> I still have some doubts about the 8bit samples.
> 
> I have tried the usrp_fft.py, but when I do require 8bit samples through
> the provided command line option, the fft sink displays nothing at all,
> even if the applications seems to be running all right.
> the same happens when I import the python code for 8bit conversion into
> my scripts. any hint?

Turn the gain up.  The 8-bit samples are formed by sending only the 8
most significant bits of the 16-bit samples.  So, if your signal is too
small, it will be dropped.  With a 50 Ohm load on the Tx/Rx port of my
RFX-2400, I get no signal when the gain is at or below 53 dB.  The noise
floor appears when the gain is set to a value greater than 53 dB.

-Lee