On Mon, 14 Dec 2015, Adam Sampson wrote:
Will Godfrey <willgodfrey@xxxxxxxxxxxxxxx> writes:
Not for audio, but having JACK run at 96000 Hz is quite useful for
low-cost software-defined radio...
Explain.... please.
A common architecture for an SDR receiver uses a quadrature sampling
detector: you mix the incoming radio signal with a pair of clocks 90
degrees apart, and sample the two resulting signals (I/Q). Some DSP
then allows you to "tune in" any signal within the bandwidth that you've
sampled -- if my detector's clock is running at 7 MHz and I'm sampling
I/Q at 48 kHz, my SDR receiver can demodulate any signals between 6.976
MHz and 7.024 MHz (and can receive several signals simultaneously
provided they're within that range of frequencies).
Ahh.
So the advantage of sampling at a higher rate is that you can receive a
wider chunk of the radio spectrum without having to retune. High-end SDR
receivers have a dedicated ADC that can run at a high sampling rate, but
cheap radios like the SoftRock kits (http://wb5rvz.com/sdr/) use a
regular soundcard.
The "dedicated ADC" says a lot as the average Audio ADC that runs at even
192k or higher still has analog ciruitry BW limited to around 20K. I would
imagine for this use, 12 bits at 96000hz with analog circuitry that is
bandwidthed from around 10k to 50k might work better than an "Audio card".
I am not sure how much the bit depth affects this process but do know that
the noise floor of an off air signal is not anywhere near 96db below any
peak signal.
I'm using a SoftRock RXTX with dttsp (and sdr-shell/softrig/fldigi),
which uses JACK for audio IO. Running JACK at 96 kHz for this means that
I can see most of the 30m amateur band at the same time...
Sounds very useful.
--
Len Ovens
www.ovenwerks.net
_______________________________________________
Linux-audio-user mailing list
Linux-audio-user@xxxxxxxxxxxxxxxxxxxx
http://lists.linuxaudio.org/listinfo/linux-audio-user
