In article <3fae83c9.94003185@news-central.giganews.com>,
stevezinn@comcast.net (STZ) wrote:

Does anyone know the the format of the data field?

Berkeley and the author of WU2WAV.

Like:  How many bits, how are they packed and encoded in the WU.

The header tells you the number of samples.   You know the time they
cover.  Realising that the data looks as though it is uuencoded tells
you how many bits of usable data there are.  There are a lot of things
which tell you the frequency range covered by a work unit.  That really
tells you everything you need to know about the nature of the recorded
samples except that the centre frequency isn't 0, because of artifacts
of the splitting process.  At least, it should, if you know enough to
do any useful processing.

I don't believe the printable encoding is anything but the obvious.
There are very few sensible choices of bit order; to be sure you will need
to try to process a work unit with a test signal.  My guess would be that
all the bits for a sample are together, but you will have to experiment
to confirm whether bytes are little or big-endian.  The prototype was
on a Sun Sparc.  For most purposes, the order of the bits in a sample
probably doesn't matter.

I can't remember exactly what I was told about the splitting artifact,
but try looking at the splitting process.  I believe the WU2WAV author
experimented to find the answer.

You should consider adding sci.astro.seti to the newsgroups, as alt.sci.seti
only posting tend to be bragging and overclocking articles.