[Top] [All Lists]

Re: [TenTec] Orion Filter Controls

To: Rick@dj0ip.de, tentec@contesting.com
Subject: Re: [TenTec] Orion Filter Controls
From: Sinisa Hristov <shristov@ptt.yu>
Reply-to: tentec@contesting.com
Date: Mon, 05 Jan 2004 18:53:29 -0500
List-post: <mailto:tentec@contesting.com>
Rick Westerman wrote:

> It's an established fact that the bulk of the power of audio
> frequency is in the low tones (say, below 500 Hz), while the bulk
> of the intelligence is in the area of 500 to 1800 Hz
> (approximately).  Indeed I've often filtered out the low tones so
> that I could turn the volume up to a higher level without my
> loudspeaker (or headphones) distorting from the excessive audio
> power.  This gave me greater intelligence from the received
> signal.

It depends on the transmitted signal.
If nothing is transmitted below 500 Hz,
then there is no point in receving below 500 Hz :-)

Let's forget QRM, because it may come from all "directions".
Let's also forget S9+++ signals.

Let's talk about a weak signal in white noise.
In such a condition, inteligibility cannot be
improved by cutting out those portions of spectrum
that carry significant energy. This is a basic
principle of communication.

Intelligibility depends upon our aural system being able
to detect spectral characterists of speech in time.
One can even compare speech to DTMF ("Touch Tone")
signalling:  aural system detects spectrum shapes
in the low band (~300-1100 Hz) and high band (~1400-2800 Hz),
computes a kind of ratio of the two and that's it.

When signal to noise ratio is very low, it takes very 
little frequency response distortion to "confuse"
the aural system. Would you expect "intelligibility"
given a couple of dBs signal to noise ratio,
with RX frequency response randomly distorted
by another couple od dBs? Of course not.
"Formants" (spectral peaks) will appear in false 
positions and aural system will get confused.

Spectrum between 300 and 500 Hz is quite important
for vowel recognition. Contrary to popular myth,
vowels ARE important for real-world communication.
All parts of speech are important. Nature has been
perfecting speech for milleniums, and there isn't much
that can be removed harmlessly (1100-1400 Hz range excepted).

* * * *

But before you try this on your RX, you must check
it's frequncy response. Long time ago receiver
manufacturers discovered a clever way to produce
"quiet" receivers. They simply suppress most of the
audio spectrum. So, RX sounds quiet, indeed :-)

Here are some measurements as I recall them at the moment:
  * TS570:      -15 dB at 2700 Hz
  * MkV:        -12 dB at 2700 Hz (slightly better with DSP demodulation)
  * IC706MkII:  -20 dB at 3 kHz in FM mode
  * FT290R:     -12 dB at 1 kHz, -20 dB at 2 kHz, -30 dB at 3 kHz in FM mode.

Values do NOT include crystal filter attenuation,
this is post-detection distortion only!

Crystal filter ripple is also important, but that's another story.

It may well be that when cutting lows,
one is actually trying to rectify RX frequency response,
making it as FLAT as possible.


Sinisa  YT1NT, VA3TTN

TenTec mailing list

<Prev in Thread] Current Thread [Next in Thread>