Rik Strobbe wrote:

But ... this would require a 'Jason signal' using 25.2Hz steps instead of
0.252Hz steps (producing an audio signal of 800Hz +/- 400Hz instead of
800Hz +/- 4Hz).
I don't know how hard this it to implement ?

Hi Rik,
  the audio tones in Jason are generated using a software implementations
of a NCO (Numerical Controlled Oscillator), where a long sine table is
periodically sampled according to a phase accumulator whose increment
is a function of the wanted output frequency (among other things).

So, as long as the requirement is not to have a frequency resolution finer
than 0.084 Hz, which is a limit dictated by other constraints in Jason,
it is a matter of just changing a couple of values in the program to do
what you want.

What will take the greatest part of the time needed to do this, is the user
interfacing. There must be a new Options Menu choice, where you will be
able to specify, let' say, a multiplying factor which will correspond to the
division you will perform on the generated RF signal. And then a check must
be performed to see whether the resulting range is compatible with the
center ftrequency specified for Rx... negative frequencies exist only
in the analytical domain, not in the real world !
But it is a work of not more than one evening, maybe less.
So I think I will do it after the mod to the encoding method, to use the
17-slot format.

73  Alberto  I2PHD