Re: [Discuss-gnuradio] GNURadio and Sampling Theory: simple question

[Marcus Müller]

Hi Saulo,

that is really just an artifact of the filter design. One could argue that a signal in the pass band should go through unattenuated, but one could also argue that it doesn't matter very often; also note that the filters designed internally in the resampler block also trade of steepness of the transition and beginning of the roll-off in the passband for limited length of the filters. So, if the signal you're passing through the resampler is close to the nyquist band edge, then it might already get a bit attenuated.

There's a simple way to "test" the magnitude response of the resampler, ignoring aliasing due to non-perfect stopband attenuation: Connect a Noise Source to its input, and a Qt Frequency sink set to a large FFT length and high averaging to the output. You should be able to see the filter's "shape".

Best regards,

Marcus

On 14.10.2017 23:47, Saulo Queiroz wrote:

Hi all,

I played a bit with the example "E2.3. A Note on Resampling" provided in the GNURadio tutorial [1].

My question is about the changes in the signal's amplitude value after it is interpolated.  For a interp. factors of 16 and 32, for instance, I observed gains of about 1 dB and 5 dB, respectively, in comparison to the "normal" signal.

Could someone share some thoughts about this? Is this consistent with theory?

Thanks in advance.

[1] https://wiki.gnuradio.org/index.php/Guided_Tutorial_Extras_Sample_Rates

--

Saulo Jorge bq
- "Beware of bugs in the above code; I have only proved it correct, not tried it." 

Donald Knuth.

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