Telluric "Sounds" Comments

I have received a large number of responses as a result of the Telluric "Sounds" files.   Instead of attempting to reply individually I thought it would be better to summarise the responses here for all to see.    In order of most frequent:-

4.5 Second Period Sinusoidal Signal

Many of you suggested that it sounded like the unstable birdies from the power mains.   This was my impression too.   A quick calculation had shown that the necessary error for a beat with the sampling frequency was much larger than the observed drift from many measurements.   Another suggestion was an error in the sampling itself (Stewart).  I had also thought of this but had dismissed it as I had calculated that the PC clock would have to be out by over 8 minutes in 30 hours to cause this effect - this is clearly not the case.   Nevertheless, I agree that this is the most likely cause.  In particular I suspect that the 1 second sampling rate from the provided Protek software is missing samples as 1 sample a second is close to the conversion time for the DVM.    Missing one sample in 200 to 300 samples would do the trick.

Rhythmic "Zulu" Beat

 Most responses were in line with what I suspect - the pattern is too regular to be anything other than man-made.   The problem is what is it ?   My first idea was that it was our 'fridge.   Unfortunately there was no correlation between the signal and the fridge switching off and on.   Also we live in the "sticks" or bush (farmland in non-Aussie lingo).    The other problem is that the pulses are not narrow in time.   See below and you will observe that the third pulse in the train (the one which is to the right of the 1.7 minute marker) has a decay time of over two minutes.    This is difficult to explain away as a switching transient.    The only thing I can think of is that this is the result of sudden increases in mains earth current which depolarises the contact potential associated with the earth rods.   This could be checked by monitoring the mains current noise level simultaneously.

Nevertheless, the regular nature of the pulses almost certainly point to a man-made source - I will continue to try and isolate the cause.   If others can carry out the same experiment at another location this would serve to throw some light on the puzzle.

WAV File Format

There were several queries about the actual WAV ( and Real Audio) files presented.   Several people who have analysed the files have correctly identified the processing I have done.    The original text file (floating point text readings from the DVM) was normalised to +-1.00 and then converted into a WAV file at a sample rate of 8000Hz.   The sample rate was chosen because the WAV to Real Audio conversion application I use will only accept "standard" sample rates (8000Hz being the lowest it will handle).    The problem is that that the resultant audio output from this sample rate is too fast to hear the rhythmic pulses.   Therefore before conversion to a Real Audio file  I use transpose processing to halve the frequencies to a slower rate (equivalent to a "playback" rate of 4000Hz).    As you would know this will produce image products which I filter out by low-pass filtering with a cut-off frequency of 2000Hz.     Saving as an 8-bit file re-introduces some quantisation noise above the cutoff frequency.    This slowed down WAV at  a sampling rate of 8000Hz is converted to Real Audio files with substantial savings in space.

Here is the original 8000Hz rate 16-bit file for your own analysis.

If you have any more ideas let me know by email.