Do you recognize this sound?

Sounds like a 1 KHz calibrator on an HF receiver.

With poor selectivity. ;-)

Cheers! Rich

The signal originated from a voice communications radiotelephone system connected by telephone lines to a remote recorder.

I am particularly interested in the 900 Hz component. My signal processing tools are limited to a FFT written three decades ago and recently translated from FORTRAN. I interpret the transform as showing an amplitude and frequency modulation of a 900 Hz sinusoid.

If your sonar is an accurate representation of the 900 Hz component then I would appreciate further details on how you extracted it from the other signals.

Herbert

The sound i got lasted less than a second; way too short to characterize more definitively than "noise".

It is mostly noise, band-limited by what looks like a telephony system. It also contains another noise component, more heavily band-limited around 900Hz.

What it comes from is another matter. I have done a little processing on it, though, and it is now a submarine ASDIC. Just play it as a loop.

Pearce Consulting

Same here.

That is indeed a valid interpretation - you can make any complex signal as modulation on a carrier - doesn't mean it was created that way, though. It is much more likely that the 900Hz component and the broader band noise are unrelated.

That was just a bit of fun, created by amplitude shaping, copy and paste in the spectral domain and added reverb (Cathedral model). I did it with Adobe Audition and the Waves plugin bundle. It took no more than 30 seconds.

It is certainly your original file with no other signals added.

Pearce Consulting

If you want to extract the signal yourself using your FFT, then it is quite simple. Run the FFT. Set all the values up to the 800Hz point to zero. Do the same for points from 1kHz upwards. Now run the IFFT and you have a filtered version of the original, with just the 900Hz signal and a bit of noise - similar to what I posted, but without the extra joke effects.

d Pearce Consulting

to

Thanks for the advice, Don.

I have tried the technique and found that transient response of the filter to the rapid shifts of the signal dominate the output.

Herbert

Then instead of simply zeroing the values, write yourself a phase-linear filter and apply that to the signal.

Pearce Consulting

Hmmmm

sounds familiar to me...

does this have anything to do with JFK and Dallas?

i'ts not gun shots!

Mark

The file is from the Bowles tape of the Dictabelt and does not contain the alleged shots.

Herbert

When I look at the file in the frequency domain, it looks like a strong and nearly narrow band sinusoid surrounded by lesser noise. To me this spectrum shows simultaneous amplitude and frequency modulation of audio.

Herbert

Actually, when you look at the file in the frequency domain, it looks exactly like a grassy knoll.

Pearce Consulting

Like I said almost a week ago, a 1 KHz calibrator on a really crappy receiver.

looks

strong

this

I have seen 1 ms square wave sources on scopes and 100 kHz crystal oscillators on receivers. What in the world does a 1 kHz calibrator do on a receiver?

Herbert

No it doesn't. There is no information in that file that tells you

Pearce Consulting

Except it isn't even close to 1kHz.

Pearce Consulting

looks

strong

this

I seem to have missed your response.

The modulation index of the 900 Hz signal is relatively small. This enables my function generator to almost lock on the dominant signal from the file. Connecting the signal from the file to the X-axis and the output of the nearly locked generator to the Y-axis produces a nearly stationary scope patten. Examination of the sine wave output of the generator shows the frequency modulation.

Herbert

of

modulation

What is your interpretation of a FFT, which shows a prominent nearly narrow band sinusoid

and sounds like a loud noise?

The wave file contains 16384 points, which facilitates FFT analysis. Short term spectra show the sinusoid rapidly hopping around 900 Hz. Any insights you have would be appreciated.

Herbert