[time-nuts] Timestamping audio waveforms

[Magnus Danielson]

Hi Tim,

On 03/31/2018 04:41 PM, Tim Shoppa wrote:
> I would like sub-millisecond timestamps for a mono audio radio signal that
> I have in the shack.
> 
> The timestamps could be continuous (Every sample) or just every "frame"
> where maybe a frame is a second to a minute.
> 
> I would like to calibrate both the absolute time as well as the delta time
> between samples based on the timestamps. (I would expect that say a nominal
> 48kHz sample rate would be off by many tens of ppm because of crystal
> tolerances.).
> 
> I have a both Windows and Linux based PCs running Audacity with a local
> GPS-based LAN refclock and ntpd. I trust the ntpd time to be stable to the
> sub-millisecond.
> 
> Can Audacity do this kind of timestamping for me based on the system clock?
> 
> Or should I, say, take the PPS from a GPS, and feed it into channel 2, with
> the audio going into channel 1, and make a stereo recording? I suppose I
> could then manually label the filename with the second the recording was
> begun. There would likely be some delta (maybe half the sample interval?)
> between the two channels but I'm fine with that as long as it is at the
> sub-millisecond level.

Have a IRIG-B signal recorded on the second track.
IRIG-B has been decoded before, here is one I threw together:

https://github.com/sa0mad/irigb

You can readout the time for each sample.

Cheers,
Magnus