Alberto di Bene wrote:

[.....]

an audio frequency. This is not of course the case, as the DDS is used to 
generate
the carrier at 136kHz. The frequency values reported in that table are to be 
added
to the nominal value of the carrier generated by the DDS. Sorry for the 
confusion.

[.....]

Yes, of course. I have already computed the DDS phase increment offsets for the
32 tones. My present guess of the delta-f is:

5512.5(standard Fs/2) * 3(FFT bins per step) / 0x10000(something with FFT & 
power
of two)... ? Please correct me if I am wrong.

By the way, what happens if a signal slides from one FFT bin to another?
Will it be a smooth transition so that the power is approximately -3dB in the 
bins
that surrounds the signal when the signal is in the middle? I guess there is a 
way
to get the correct total signal power even if the signal is not exactly on an 
FFT bin
(frequency wise)? I've had no experience with this windowing stuff, I've only 
read
a little about it... I'm just curious...

73
Johan SM6LKM