Still no luck. If I use -R for raw mode a little more data gets through (as seen by file sizes or the Play progress meter) but it seems to be less intelligible if anything. When I write to a file, the size is always in chunks of 16384 bytes (0x4000 or 2^14). At a sample rate of 24000 that's 1/3 second, which is about right compared to what I hear. Without using -R, this doesn't repeat for about a minute. That 16384 must be some buffer size or something in the rtl_fm program, but I haven't found it yet. I think this is supposed to repeat often enough to be continuous. Could computer speed be an issue? I'm on a single core P4 at 3.2 GHz.

From rtl_fm -f 146970000 -f 146985000 -f 146910000 -f 146940000 scan2.dat I got this which looks like bursts of noise as rtl_fm is scanning, still all within about 1/3 second:

http://ab1jx.webs.com/toys/dongle/scan2.dat.gif

There's a curious ripple in the baseline which doesn't quite look like power supply ripple, more like something from some filter.

Alan

-----
Radio Astronomy - the ultimate DX

----- Original Message -----
> From: Adam Nielsen <a.nielsen@shikadi.net>
> To: Alan Corey <alancorey@yahoo.com>
> Cc: "osmocom-sdr@lists.osmocom.org" <osmocom-sdr@lists.osmocom.org>
> Sent: Monday, December 31, 2012 1:28 AM
> Subject: Re: Trying to use rtl_fm, etc
>
>> With rtl_fm, I get a tiny burst of audio about once a minute.
>>
>> A fresh run:
>> freebie# rtl_fm -N -f 162550000 - | play -t raw -r 32k -e signed-integer -b
> 16
>> -c 1 -V 4 -
>
> Just FYI, you can see here that:
>
>> Output at 24000 Hz.
>
> However you have told 'play' to play the audio at a sampling rate of
> 32kHz, even though the audio data is only arriving at 24kHz. So you will get
> stuttering as the audio buffer keeps running out and waiting for more data to
> arrive.
>
> For me (under Linux), I get best results using the -r option to rtl_fm to set
> the output audio sampling rate to 48kHz, then tell play to play at 48kHz too.
> This way my system doesn't have to resample it to 48kHz before it can mix
> the stream into the system-wide audio output.
>
> Note that the -s option sets the signal bandwidth and -r sets the output audio
> sampling rate. A lot of people misunderstand the purpose of the -s option,
> however you shouldn't need it unless you are trying to receive data signals.
> -W and -N set -s to the correct values for voice transmissions.
>
> I would also suggest playing a .wav file with the same 'play' options
> just to make sure your system can play mono audio at low sampling rates. I know
> my sound card drivers won't (possibly because I am using a SPDIF connection
> to an external amplifier) so I need the Linux audio system to upmix it to 48kHz
> 16-bit stereo or I won't hear anything.
>
> Cheers,
> Adam.
>