Call Search
     

New to Ham Radio?
My Profile

Community
Articles
Forums
News
Reviews
Friends Remembered
Strays
Survey Question

Operating
Contesting
DX Cluster Spots
Propagation

Resources
Calendar
Classifieds
Ham Exams
Ham Links
List Archives
News Articles
Product Reviews
QSL Managers

Site Info
eHam Help (FAQ)
Support the site
The eHam Team
Advertising Info
Vision Statement
About eHam.net

donate to eham
   Home   Help Search  
Pages: Prev 1 [2] 3 Next   Go Down
  Print  
Author Topic: human detection of weak CW signals  (Read 4664 times)
W3TTT
Member

Posts: 266




Ignore
« Reply #15 on: July 14, 2017, 10:20:24 AM »

What is the bandwith of your IF filter Joe?
I used 300-2800 Hz, but by shifting you can make it (300+n) till (2800+n) with n a positive or negative real number. and the CW carrier still within that band. Any idea about n and the tone pitch of the received code you experienced as optimal?
This is using a 30 year old Icom 730, with no cw filter.  Oh, I do wish I had the IF CW filter for contests etc. 
On my Icom 718, by the way, I installed a used 250 hz CW filter.  When I switch it in, the noise level goes UP.  Is that supposed to happen?  I thought the noise level would go down.  Or is the filter (eBay) defective?  The filter does filter out any signals up or down the pass band.  But the noise goes up.  Please advise.
thanks
Joe W3TTT
Logged
N3QE
Member

Posts: 4880




Ignore
« Reply #16 on: July 14, 2017, 10:51:09 AM »

Lower pitch should have a better detection because the human mind has a built in filter about one octave wide.

It is not obvious that lower pitches are automatically easier to decode in broadband noise. In your example I felt that I could decode the high pitched tones just as easily as the low pitched tones.

While music is built around octaves it is not obvious to me that pulling a signal out of the noise works with that exact part of the musical brain.

In different noise environments - say intentional nearby QRM where I have to distinguish between two signals just 10Hz apart - I am willing to believe, there may be value in working at lower pitches, where 10Hz is a bigger part of an octave.
Logged
PA0WV
Member

Posts: 388




Ignore
« Reply #17 on: July 14, 2017, 12:19:45 PM »


This is using a 30 year old Icom 730, with no cw filter.  Oh, I do wish I had the IF CW filter for contests etc.  
On my Icom 718, by the way, I installed a used 250 hz CW filter.  When I switch it in, the noise level goes UP.  Is that supposed to happen?  I thought the noise level would go down.  Or is the filter (eBay) defective?  The filter does filter out any signals up or down the pass band.  But the noise goes up.  Please advise.
thanks
Joe W3TTT
Joe, the VOLUME changing of noise may be due to the design of the set, that tries to compensate for filter loss with additional gain when you switch the filter in. Compensate for that with your volume control.

However the PITCH of the noise has to change because it becomes band limited to the CW bandwidth of your filter.
Example of noise in an SSB band 300-2800 Hz on
http://pa0wv.home.xs4all.nl/eham/SSBRUIS.WAV
Example of noise with a switched in CW filter 712 Hz center frequency 150 Hz wide on
http://pa0wv.home.xs4all.nl/eham/CWRUIS.WAV


Lower pitch should have a better detection because the human mind has a built in filter about one octave wide.

It is not obvious that lower pitches are automatically easier to decode in broadband noise. In your example I felt that I could decode the high pitched tones just as easily as the low pitched tones.

While music is built around octaves it is not obvious to me that pulling a signal out of the noise works with that exact part of the musical brain.

In different noise environments - say intentional nearby QRM where I have to distinguish between two signals just 10Hz apart - I am willing to believe, there may be value in working at lower pitches, where 10Hz is a bigger part of an octave.
I think you are right. So I was amazed to be able to copy the different pitches all with hardly notable difference. Some hams however, reported in this thread a preference for lower pitches. I guess you are correct that the octave-ratio is important for distinguishing of nearby stations not for one stn present.
So I stand corrected and thanks for your remark.

@NI0C and VK3EEE mni tks for the file that I received in the mail. I investigated it, it turns out not to be of use for the following reasons:

1. The sampling rate is 4000 samples/s so you can't reproduce recorded noise above 2000 Hz, that is insufficient for CW in an SSB receiver pass band of 300-2800 Hz. CW pitch is 600 Hz.
The noise that I generated within that passband you can find the spectrum of in
http://pa0wv.home.xs4all.nl/eham/ssbnoise.jpg

2. His noisespectrum is flat over the whole band 0-2 kHz, so the author  obviously made easy white noise with a random number generator. He did not cut away  noise below 300 Hz and limited the high side by reducing the sampling rate.

3. The noise is heavy in saturation, nearly a block wave, so it will be undersampled for reproduction and give aliasing.

My conclusion: You can't conclude anything objective  by listening to that file.

« Last Edit: July 14, 2017, 12:34:06 PM by PA0WV » Logged

PA0WV
Member

Posts: 388




Ignore
« Reply #18 on: July 14, 2017, 01:00:00 PM »

@VK5EEE
Lou, previous post is closed for modify.
Well thank you very much for the CW2DB report.
I must look for a level that makes it harder to find out what frequency is best as center frequency, according to the idea of IF-band shift and RIT control together to change the pitch of the noise.

Last insight N3QE presented is that the detection reliability should be equal, and in a 150 Hz filter you can probably easy get rid of neighbouring stations.

What you call. the sensitivity for  hard sound, is called hyperacusis. It often goes together with tinnitus. I have known a guy that had such a loud tinnitus that the hyperacusis made living a hell.

I can't actually  detect the frequency I hear, I estimate 15625 Hz ; the previous TV line frequency in Europe, but I wonder when I put it in a loudspeaker or a headphone whether or not I can introduce a beat.
Logged

VK5EEE
Member

Posts: 1053




Ignore
« Reply #19 on: July 14, 2017, 05:57:03 PM »

Thanks OM, I too have tinnitus, when I clench my teeth on one side, the volume of lower sounds (I estimate around 3kHz) goes up 1000dB :-)
So I did not know this, that this is what can produce "hyperacusis" and not just crankinessisis
Logged

Long Live Real Human CW and wishing you many happy CW QSO - 77 - CW Forever

Support CW and join CW clubs. QTT: FIST#1124, HSC#1437, UFT#728, RCWC#982, SKCC#15007, CWOPS#1714, 30CW#1,
PA0WV
Member

Posts: 388




Ignore
« Reply #20 on: July 16, 2017, 07:05:14 AM »

the samples of a signal can be devided in a number of amplitude categories.
I take 45 categories. The level can be between -32768 and +32767 for a 16 bit sampled signal.

A picture with horizontal axis one of the 45 amplitude categories and vertical axis the number of occurances, is called a probability density function; acronym pdf.

The pdf of the by NI0C  QSP  -10 dB file is found in
http://pa0wv.home.xs4all.nl/eham/pdfAG1LE.bmp

It tums out to be flat (random numbers for the samples) and heavy over excited (max and minimum values are heavily overpopulated) file.

When you listen to the file minus12dBf.wav on this website however, the noise is "white" between 300 and 2800 Hz, being the SSB band of a receiver,

The pdf of minus12dbf.wav is for inspection available on:
http://pa0wv.home.xs4all.nl/eham/pdfmin12dB.bmp



Logged

W3TTT
Member

Posts: 266




Ignore
« Reply #21 on: July 18, 2017, 05:46:44 PM »

Example of noise with a switched in CW filter 712 Hz center frequency 150 Hz wide on
http://pa0wv.home.xs4all.nl/eham/CWRUIS.WAV

Yes, that is about what I hear when I switch in my 250 hz filter.  But the background noise gets louder than before it is switched in.  The filter works to filter out signals outside of the passband.  But for weak signals, I can hear better with the filter off.  Is that the way it should be?
W3TTT
Logged
N3QE
Member

Posts: 4880




Ignore
« Reply #22 on: July 19, 2017, 03:38:18 AM »

Yes, that is about what I hear when I switch in my 250 hz filter.  But the background noise gets louder than before it is switched in.  The filter works to filter out signals outside of the passband.  But for weak signals, I can hear better with the filter off.  Is that the way it should be?
W3TTT

If there's no QRM, a lot of operators prefer to listen (with their ears) for weak CW signals with a wide filter rather than a narrow filter. "Wide" varies by preference but it might be anywhere from 600Hz to 3kHz wide.

Narrow filters "ring" in the presence of noise (especially impulse noise), and they ring at their passband frequency which is just the same as the frequency you are trying t hear. At 250Hz, ringing becomes especially obvious/painful especially if there is any lightning-type noise. But ringing can be an issue with pure non-impulse noise too because the filter is still ringing.

There are different filters that ring in different ways. The obnoxiousness of ringing depends on the filter shape at the edges. Filters with gently sloping walls tend to ring less than brick-wall filters. From a filter-design standpoint, it's large variations in phase shift in the passband that cause the ringing - there are special filter designs (like "Gaussian to 20dB") that attempt to move the bulk of the phase shift out of the passband.

Psychoacuostically, listening at 250Hz all the time just drives me nuts. I don't mind narrowing down to 250Hz for a little while if it gets rid of nearby QRM but there's no way I'd operate that way all the time.

Other operators insist they have to have filters 50Hz or narrower all the time. So it obviously varies by taste!
« Last Edit: July 19, 2017, 03:47:55 AM by N3QE » Logged
VK5EEE
Member

Posts: 1053




Ignore
« Reply #23 on: July 19, 2017, 05:39:06 AM »

Well said N3QE, I'm learning much from you and PA0WV on this subject, you sure both know your stuff!

On the 712 recording, thanks W3TTT, I cannot make out a single character anywhere! Can anyone else? I did well as the others, but this one, total failure.

On IC751 I have a narrow filter, not sure how narrow seems to be about 250Hz but it does not have a very sharp cut off, and I DO find it helpful to reduce my local QRM by 2 S points, and can then hear CW that I cannot hear otherwise as the SNR greatly improves. Clearly this is not the case with all rigs. But like you N3QE, I'd not like to listen on narrow filter to weak signals forever, generally I prefer a wider filter, even 3 or more kHz and to let my brain do the filtering.

I find though, and it's not available on all rigs, that switching off AGC completely, and then reducing RF gain, can in many situations (weak signals with QSB) help a lot, and in the narrow filter using additional PBT and Notch (on the noise), in addition to Noise Blankers and RIT, are all very helpful!
Logged

Long Live Real Human CW and wishing you many happy CW QSO - 77 - CW Forever

Support CW and join CW clubs. QTT: FIST#1124, HSC#1437, UFT#728, RCWC#982, SKCC#15007, CWOPS#1714, 30CW#1,
N3QE
Member

Posts: 4880




Ignore
« Reply #24 on: July 19, 2017, 06:24:07 AM »

Well said N3QE, I'm learning much from you and PA0WV on this subject, you sure both know your stuff!

I have some formal training in physics and information theory and have been a ham since age 10. I enjoy building my own hardware, instrumentation both analog and digital, and when all of these combine to yield the quarter million QSO's of experience that I have, then I am very happy!

While I do a lot of CW and enjoy it most, I also do a lot of RTTY, and I am trying to learn how to do phone! Phone is the hardest of all modes and I am currently trying to climb the very steep learning curve. After 40 years as a ham I am starting to feel like I am making some baby steps of progress on phone.
« Last Edit: July 19, 2017, 06:30:04 AM by N3QE » Logged
VK5EEE
Member

Posts: 1053




Ignore
« Reply #25 on: July 19, 2017, 06:53:37 AM »

You and me both! I'm about the same age and ham experience time span (except a lot less as I was QRT for 20 years or so) but I have nowhere near this level of knowledge on technical matters, so it is great to learn from you, Wim and others here! My main interests are of course CW (really everything except "5NN" reports), and propagation, wire antennas, I was interested in some data communications in the past (especially Pactor), but I've missed out on the home brew of anything but antennas, and hope to catch up a little on that given the opportunity in future. I hope some day to build a decent CW rig of QRP sorts, max 5W-20W would be fine, and to engage in more outdoor CW activities. I'm missing half the hobby!
Logged

Long Live Real Human CW and wishing you many happy CW QSO - 77 - CW Forever

Support CW and join CW clubs. QTT: FIST#1124, HSC#1437, UFT#728, RCWC#982, SKCC#15007, CWOPS#1714, 30CW#1,
N3QE
Member

Posts: 4880




Ignore
« Reply #26 on: July 19, 2017, 07:12:29 AM »

I hope some day to build a decent CW rig of QRP sorts, max 5W-20W would be fine

If you are interested in "filters that sound good for CW" research designs like "Gaussian to 12dB".

I've built ladder crystal filters with this shape in the past, but today most folks would just do it in DSP.

The noise simulations folks have been using here, do not have the low-band thunderstorm noise that permeates my location every summer. I will see if I can record some wideband lightning noise for filter demonstrations like PA0WV has been sharing.
Logged
VK5EEE
Member

Posts: 1053




Ignore
« Reply #27 on: July 19, 2017, 07:22:49 AM »

Thanks for tip and that would be interesting for sure. I actually like the sound of lightning QRN though (with AGC on fast but not off preferably!), so I'd probably not want to filter it out but will be interesting to hear!
Logged

Long Live Real Human CW and wishing you many happy CW QSO - 77 - CW Forever

Support CW and join CW clubs. QTT: FIST#1124, HSC#1437, UFT#728, RCWC#982, SKCC#15007, CWOPS#1714, 30CW#1,
PA0WV
Member

Posts: 388




Ignore
« Reply #28 on: July 20, 2017, 07:42:46 AM »


On the 712 recording, thanks W3TTT, I cannot make out a single character anywhere! Can anyone else? I did well as the others, but this one, total failure.

That is the result of a misunderstanding. I gave a link to TTT with the purpose to demonstrate the pitch of 150 Hz wide noise. That noise and the SSBnoise file do not contain any signal.

By the way:
Gaussian shaped filter has worse slope. PA0LQ published a number of years ago a filter for CW, active with a few opamps, still can be found with google by typing in his call. He managed with experimentation with a simulator to get minimum ringing in a steep sloped filter.

The pass band of the 150 Hz design here is shown at

http://pa0wv.home.xs4all.nl/eham/passbdcw.bmp

The filter is designed for 150 Hz bandwidth and plotted over 150 Hz.
Vertical axis dB, horizontal axis to be ignored.

Let me put one dit in its input 60 ms wide (a 20 wpm character E) on the center frequency  without finite leading and trailing edge of the envelope and look what happens with the output pulse.

When it is succesful you may find it on
pa0wv.home.xs4all.nl/eham/ringing.bmp
Logged

PA0WV
Member

Posts: 388




Ignore
« Reply #29 on: July 20, 2017, 10:03:31 AM »


When it is succesful you may find it on
pa0wv.home.xs4all.nl/eham/ringing.bmp
It is on the clickable link
http://pa0wv.home.xs4all.nl/eham/ringing.bmp

It represets 0.175 s when on t=0 the filter is fed with 480 samples
of a suddenly started sinewave immediately with full amplitude (16000) , 712 Hz.
x-as is sample number of output, y=as is amplitude of the samples. samplerate 8000/s. After 480 samples (60 ms) the input is immediately switched to 0.
Logged

Pages: Prev 1 [2] 3 Next   Go Up
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.11 | SMF © 2006-2009, Simple Machines LLC Valid XHTML 1.0! Valid CSS!