Hi Mike,
you are mixing concepts that should not be mixed there. The ~90 Hz
bandwidth figure is real, it results from the tone spacing and the
symbol rate. Your should really try and think of it as a continuum
rather than individual tones when talking about bandwidth. The decoder
extracts signal energy from every part of the 90 Hz bandwidth. The
candidate received signal is sliced into bins each of width equal to the
tone spacing.
PCM is a digital concept, it is a series of equally time-spaced
instantaneous voltage measurements (discrete-time signal) representation
of a continuous-time signal, usually sampled across a capacitor that has
been charged by the audio source signal for a short period of time. PCM
streams are just a series of numbers that when converted to voltages at
the inherent sample rate, and filtered with a suitable low-pass filter,
will recreate faithfully the originally sampled audio signal. The only
constraint is that the maximum bandwidth that can be represented is half
the sample rate (see the Nyquist-Shannon Sampling Theorem for details).
PCM is just a convenient model for analysing or transforming signals in
the digital domain, i.e. Digital Signal Processing.
Whatever way you look at it the maximum audio amplitude, which is
essentially constant and the same for all WSJT-X modes (1), should not
exceed the saturation point of the rig's SSB modulator. Other than that
levels are pretty much irrelevant other than their effect on the output
power of the transmitter. Note this is not the case for modulations
methods like voice SSB or PSK31 where audio levels can be below the
saturation level of a rig's SSB modulator but still cause signal
distorting IMD products along with the intended RF output that increase
the transmitted bandwidth beyond the optimal minimum, colloquially
referred to as splatter. These are due to non-linearities in the
transmitter such as PA saturation and ALC action. Note these effects do
not apply to signals like FT8 or FT4 where linear amplification is not a
requirement(1).
(1) Except for DXpedition Fox mode when using multiple concurrent signal
slots. MSK144 is not constant amplitude either but only because it is
filtered by the rig's SSB transmit filter.
73
Bill
G4WJS.
On 26/04/2019 19:14, Deisher, Michael wrote:
Hi Bill,
I realized that just after pressing send. The 90Hz bandwidth (I call it
acoustic bandwidth since it is encoded as a PCM audio signal) is occupied by a
spectrally narrow tone at any given point in time so my concern is not valid.
The concern would be valid for other modulation techniques that fully utilize
the 90Hz at all times during transmission. Never mind...
Thanks and 73, Mike KK7ER
-----Original Message-----
From: Bill Somerville [mailto:g4...@classdesign.com]
Sent: Friday, April 26, 2019 11:06 AM
To:wsjt-devel@lists.sourceforge.net
Subject: Re: [wsjt-devel] FT4 gain adjustment
On 26/04/2019 18:51, Deisher, Michael wrote:
FT4 acoustic bandwidth is nearly twice that of FT8.
Hi Mike,
that is not correct. The FT4 signal is one-tone GFSK. At any point in time
there is only one tone with constant amplitude. In this respect the difference
between FT8 and FT4 is that FT8 uses 8 different frequencies to encode symbols
and FT4 uses just 4 different frequencies to encode symbols.
I am not sure how "acoustic bandwidth" is relevant, whatever that is.
The baseband signal is used to modulate an RF carrier so the resulting signal
is not acoustic.
73
Bill
G4WJS.
_______________________________________________
wsjt-devel mailing list
wsjt-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/wsjt-devel
