Mixing (adding) audio from power squelched sources

I’m working on a flow graph to monitor several nbfm channels in the 2m
band and stream them to individual wav file sinks. Each channel is
power squelched to only record when the channel is active.

I’d like to monitor the audio from all channels simultaneously, so I
figured I’d just add the streams together before sinking to audio.
Problem is gr_add causes the program to exit (no errors) if one of two
streams is halted (squelched); no samples on the squelched stream.
Funny thing is I can add a single stream with a constant and it works

It’s not a problem of samples to gr_add inputs arriving asynchronously
since the gr_add works fine with both channels active. I wouldn’t think
they would be asynchronous anyway since the channel selection is from
same USRP2 source. I expanded the audio sink to two channels and have
the same problem, however, it’s not the audio sink. If I run the adder
into a null sink the program still exits, so it’s as if the adder can’t
combine a halted stream with a running stream. I assumed it would treat
the halted stream as a constant but I suppose not.

Any tips for a work-around? This is using the current “next” branch.


Well I figured out for each channel I can just do a
simple_squelch->nbfm->adder->resample->audio_44.1_kHz for real time
audio, then in parallel with that do a
power_squelch->nbfm->resample->wav_8_kHz. It takes two nbfm demods for
each channel since the squelch must be applied before the nbfm. Not
efficient, but it works. I have 3 channels monitored out of the entire
4 MHz 2m band. We’ll see how it runs over 24 hours. It’s in GRC now,
but next comes the Python script to expand it to N channels. Wonder how
big N will be?