Topband: CW Speaker Mystery - My Experience

[donroden at hiwaay.net]

Slightly off-topic, but got me to thinking about DSP/ CW / MUSIC.

In my profession, broadcast engineering, we use audio processors that 
shape and limit multiple bands of audio.

"Processor Overshoot" was an issue .... trying to stay as close to 100% 
usually meant that the compressor was either too aggressive (90% when 
you wanted 100%)  or not aggressive enough ( 110% ).

Then someone created what was referred to as a "Look Ahead" processor.

If you could "look ahead" and see that the incoming audio was going to 
be too loud, you could manually or electrically turn down the volume 
milliseconds before the loud part even reached the point of needing to 
be turned down.   ( like knowing WHEN  to cover your ears just before 
the cannons hit during the 1812 Overture.

Now, you "could" do this in the analog world ( clunky as it might seem ) 
by using audio tape with the "control channel audio head"...  ( the 
compressor) just in front of the "totally unprocessed" audio head on the 
tape path.    I'm more of an rf guy than an audio engineer, so don't 
shoot me !!    Anyway, guys smarter than me said, "We can do this by 
using digital memory and charge $10,000 for every box we sell to radio 
stations that want to be the loudest in their area".        And they 
did.

If something similar could be applied to a narrow spectrum of an audio 
range, like what typical analog CW filters use, then it seems to me that 
costs would drop dramatically.

Once something like CW spectrum is converted to a block of digital 
"audio", it all becomes numbers .    CW+ adjacent interference  + 
inverted adjacent interference = CW without interference.    Or any 
percent of interference eliminated.     Techniques like FT8 that copy 
below the noise floor could possibly be applied to CW below the noise 
floor.    Predicted ( expected ) responses below the noise floor would 
be a possibility

On 2026-02-07 11:19 am, Tree wrote:

>> So is the takeaway that the older superhet top line rigs are then the
> optimum for 160-meter DX?
> 
> I guess this begs a question about FT8.  Many people are copying 
> signals
> way below the noise level with their DSP radios and FT8.  How is that
> possible if that signal just isn't there?
> 
> Tree N6TR