I've noticed this effect myself. At super narrow bandwidths, the
signal and remaining noise are both essentially on the same frequency.
With nothing else around to provide a base-line reference, discrimination
becomes a function of real-time delta power measurement rather than tone
detection. My ears and brain are more sensitive to frequency information
than power information, and perform poorly under such conditions. I
suspect this is true for most other people as well. Open the bandwidth up
a bit, and the ear/brain system is back on familiar ground.
Some experiments are suggested, eg replacing the signal with a clean
generated tone at the dynamic power level as measured by the DSP,
injecting surrounding "clean" white noise around the filter, or both if we
want to make it sound like radio even if, at this point, it's a virtual
signal :).
On Wed, 9 Feb 2011, Scott Ellington wrote:
> I agree with both Don and Bill. There are really two issues:
>
> In the presences of just noise, the DSP with a wider crystal filter is
> adequate. The additional noise reduction of a narrow crystal filter is
> negligible. (QRM is another matter.)
>
> The brain/ear system, at least for many operators, is very good at copying
> weak signals in noise, even with a broad bandwidth. If there's no QRM
> nearby, I always find it easier to copy with a larger bandwidth, though
> anything above 500 Hz makes little difference. (Except for quickly tuning
> across a mostly empty band.) The 250 Hz bandwidth, in my experience, ALWAYS
> makes it harder to copy weak signals, unless there is adjacent QRM. For that
> reason, I prefer to operate with 400 Hz bandwidth most of the time, switching
> to a narrower bandwidth only when necessary.
>
> This seems to contradict signal theory, which says that a narrower bandwidth
> improves S/N ratio until the filter bandwidth is equal to that of the signal.
> What signal theory fails to take into account is the matched filter in a CW
> operator's head.
>
> 73,
>
> Scott K9MA
>
> On Feb 9, 2011, at 7:51 AM, Don Wilhelm wrote:
>
>> Ed,
>>
>> If there are no strong signals on the band or nearby your receive
>> frequency, then any filter, even the stock 2.7 kHz is sufficient. The
>> only purpose of the roofing filter is to keep strong signals out of the
>> receiver passband so they do not activate the AGC and de-sense the
>> receiver for the signal you are trying to hear.
>>
>> 73,
>> Don W3FPR
>>
>> On 2/9/2011 3:07 AM, Edward R. Cole wrote:
>>> Let me ask a filter question for a different situation. I am
>>> interested in receiving very-weak CW and currently have the 2.8 KHz
>>> and 400-Hz 8-pole filters. Will narrowing the DSP bw down to 200-Hz
>>> work as well as using a 200-Hz filter? There are no strong stations
>>> either nearby or on the band.
>>>
>>> BTW I am impressed with the 400-Hz filter on the K3 when compared to
>>> the 400-Hz DSP on my FT-847 (no surprise).
>>> I find trying to hear extremely weak-CW that narrowing down to 200,
>>> 100 or even narrower makes the difference.
>>>
>>> I left a blank filter space on the main receive for adding a narrow
>>> SSB filter at a later time. I also have the 13-KHz filter for
>>> FM/AM. Sub-Rx has 2.8 and 13-KHz.
>>>
>>>
>> ______________________________________________________________
>> Elecraft mailing list
>> Home: http://mailman.qth.net/mailman/listinfo/elecraft
>> Help: http://mailman.qth.net/mmfaq.htm
>> Post: mailto:[email protected]
>>
>> This list hosted by: http://www.qsl.net
>> Please help support this email list: http://www.qsl.net/donate.html
>
> Scott Ellington
> Madison, Wisconsin
> USA
>
>
>
> ______________________________________________________________
> Elecraft mailing list
> Home: http://mailman.qth.net/mailman/listinfo/elecraft
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:[email protected]
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html
>
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[email protected]
This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
