This should avoid the awkward handling.  The map makes sure that only the
central frequencies are passed around which have been used in the observation.

Also refactored some functions out of the TBB_Dipole clase since they are
not essential fot the object itself.

Before I accumulated the grand total over all sub-bands which stopped the
receiving of more sub-band data prematurely.
Also check for the band index of the incoming frame and log if it is suspicious.

The executable expects now a new parameter "subbandSize" which is the
expected number of complex values in each sub-band that the TBB will send.
This allows to stop receiving after we the right number of samples has been
received.  And it allows to report a problem if not.  Also knowing the number
of values per sub-band allows to receive bands out of order but store the data
in the right position in the file.

This variable represents exclusively either a sample or a slice number.
Therefore it makes sense to have a union for a variable itsSampleNr0 for the
transient mode and itsSliceNr0 for the spectral mode.

SW-509: TBB_DataWriter now writes meta-data using the new DAL as defined in ICD-001 v3.3 for spectral mode. All parameters are written, but not all are filled in yet because the names for the keys in the parset are still to be defined in SW-539. Adapted the tests and made them use new reference output. We probably need a new datadump as reference input from the boards with new TBB firmware later this/next week.