Noise blanking method ?

[Externet]

Good day.
A receiver is tuned to a upper side band USB valid signal and the demodulated signal is very noisy from atmospherics, or interference.
If the same receiver has a second circuitry to receive lower side band LSB and is active at the same time; two output signals are obtained as double side band DSB.

These double side band demodulated signals are noisy from atmospherics or interference.
USB = demodulated valid signal + noise
LSB = noise

If the LSB demodulated noise is inverted and added to the USB demodulation; would reception be greatly noise-free by cancellation ?

Or, A receiver tunes a USB single side band valid signal with a double sideband demodulator obtaining two outputs :
USB = demodulated valid signal + noise
LSB = with no modulation + noise

Inverting the noise from the LSB demodulation and adding it to the USB demodulation; would it greatly decrease noise ?

How factible do you see it as a method of blanking noise on SSB communications ? My point is both USB and LSB have the same exact noise at the reception frequency. Invert one of the noises and sum to obtain just the valid signal demodulated.

Valid signal + noise - noise = clean valid signal ????

 

[betwixt]

I understand what you are trying to do. The problem is that for a given frequency, the USB is above it and the LSB is below it, they are not occupying the same band space although they are adjacent to each other. This means the noise will not be the same unless the noise itself is wider than the combined USB to LSB spectrum.

If you shifted USB and LSB to the same frequency by tuning to different frequencies a few KHz apart you would get the same noise but subtracting garbled audio from normal audio would probably result in something completely unreadable.

Brian.

 

[dick_freebird]

If you want to "blank" (as in suppress entirely) based on
adjacent "noise" that might work for situations where the
interference is a sliver of a single broadband source (so,
omnipresent).

When your noise is an aggregate of many random sources
this time-correlation falls away and you might well "blank
for nothing but random detects".

That's what old timey CB radios had the "squelch" knob
for - because noise is noisy and you never know how much,
when, but you'd like to do something -right now-.

Maybe an AI-enhanced squelch knob. Yeah. One that knows
how to distinguish noise from signal before it happens.

 

[Warpspeed]

You say atmospherics or interference, but a very great deal depends on the nature of the "noise" and its source.
Impulse interference at the lowest end of the HF spectrum, typically from remote lightning discharges, and nearby electrical sparking and arcing, have been a problem since the earliest days of radio.
Simplest very early solution, a pair of back to back diodes to clip amplitudes above what you are listening to.

A further refinement of that was the "noise blanker" which detected these high amplitude noise transient spikes, and instead of just clipping them, actually muted the audio for a few milliseconds (or whatever). Cutting small silent holes in the signal which gave a subjective improvement. Both these techniques reach back into the ancient valve and morse code era.

A very different noise problem at high and very high frequencies, is a continual background hiss. The cure for that is again to completely mute the audio, until a received signal exceeds some threshold just above the noise floor. That is the "squelch" well known to the CB guys. You turn it down until the background hiss just disappears.

Apart from that, the only noise reduction technique that works in an analog system, is severely limiting the bandwidth to no more than absolutely necessary. There is no such thing as "negative noise" as its completely random, adding one noise source to another just generates a greater noise.

There is no known way to separate random noise from a random analog signal (occupying the same bandwidth) once they are combined.