[Lowfer] west coast - request KFI 'exact' carrier frequency

[Bill de Carle]

At 06:26 PM 9/21/2009, you wrote:
>Looking for someone on the west coast to get an accurate reading on 
>KFI's carrier frequency. I'm
>comparing a couple directional antennas and would like to use the 
>KFI carrier as a signal source. A
>number of years back KFI's carrier developed quite a 'wiggle' and 
>was easily identified and seen
>(Argo) here on the east coast. For those interested see the story at 
>www.w3eee.com 'KFI'.

Hi Jay:

I see someone responded with the exact frequency of the KFI carrier 
and you mentioned it might be difficult
to resolve with all the other carriers so close by.

Some years back I wrote a DOS program called CRUNCH - the basic idea 
was to leave a computer running overnight to record QRSS CW, then 
play it back the next morning at a much faster speed (e.g. 320x) - 
*without changing the pitch of the CW note,
nominally 800 Hz*.  The program worked by first mixing the 
800-Hz-centered target spectrum down to some low audio frequency, 
typically in the 16 or 17 Hz area, time-compressing by reducing the 
number of samples stored (decimation), then bringing the audio back 
up to 800 Hz by frequency-multiplication.  In this case we did the 
frequency-multiplication by playing back the resulting .wav or .aud 
file at a normal (e.g. 8000 s/s) sampling rate even though we had an 
audio signal at a much lower frequency, sampled at a correspondingly 
lower rate.  The "target" frequency (center of the input spectrum we 
processed) is user-specifiable, as is the compression ratio.  So, for 
example, we could record 8 hours of QRSS30 CW, time-compress it by a 
factor of 320, then play it back next morning in 90 seconds, sounding 
like regular 13 wpm CW at a pitch of 800-Hz.  I thought it was a real 
neat program but not many people tried it, probably because it ran on 
DOS (not Windows) and had support only for my old Sigma-Delta 
interface board (7200 s/s) and for the particular sound chip in my 
laptops, based on the ESS688/1688/1888 LSI parts.  It isn't 
particularly user-friendly or well-documented.

However....

I noticed a rather interesting side-effect from doing the 
time-compression: there is an accompanying *spectral dilation* in the 
resulting playback file.  In other words, while the precise target 
frequency (e.g. 800 Hz) plays back at exactly 800 Hz as intended, 
interfering signals on either side of that original 800-Hz are pushed 
out spectrally so they appear further away from it in the output 
file.  That makes it much easier to separate out the various 
frequencies which were previously too close together, much easier to 
copy the resulting CW signal by ear.  For example, assuming we used a 
320x compression factor and were centered on an 800-hz carrier, an 
interfering signal originally at 799 Hz (1-Hz away from the signal we 
want) would appear (320-Hz away) in the output file at a frequency of 
1120 Hz!  The output spectrum is inverted IIRC; if you need to get 
around the inversion you can always do the compression in 2 steps, 
say at 32x and then at 10x.  Two inversions restore the spectral 
sense back to normal.

Even after all these years the CRUNCH program is still available for 
download from my website: www.nrtco.net/~ve2iq

If it might be of use you're welcome to try it.  Although intended to 
compress and record soundcard audio on-the-fly to require greatly 
reduced hard-disk space, it does have an option to take the audio 
from a previously-recorded file instead and so can also be used for 
DSP Crunching of regular audio files.

If you decide to play around with it, let me know if you have any 
questions or difficulties.  It has been a long time now and I am 
getting to be an old man :-) but I should be able to remember how to 
use it, and can still make any changes you might require in the 
program to accommodate some specific needs, etc.

73,
Bill VE2IQ