What is the strange 22k signal please?

[neil wilkes]

Hi Everyone.

I was sent these images as someone wanted me to tell them what the odd signal at about 22k is on these files. The first shot shows the initial view of the anomalous signal as a solid block but the mystery deepened once we zoomed in & saw the obvious modulation pattern.
I have never in all my years as an audio engineer seen anything like this, and was hoping someone could possibly shed some light on this for me please.

[videoh]

22K satellite switcher tone?

[SeeMoreDigital]

Good call. That's a very logical answer...

[videoh]

Thank you. But how is it leaking into his audio files?

(Hehe, don't forget I worked for STMicroelectronics on, among other things, satellite tuner front-ends. ST has since dissolved their set-top box division after being crushed by Broadcom and others. Not my fault!)

[Emulgator]

Neil, as long as the screenshots take for approval:
How were these files obtained ?
Another thought without being able to see your signal:
If 22.05kHz: this would be half the CD sampling rate...
If someone had his source resampled using unfit algos across different resolution/sample rates, well possible.

[videoh]

Good point, Emulgator.

[StainlessS]

BUMP !!!

As the screenshots are now visible, wadya say now.

[Sharc]

What is the original source?
Sampling rates of 22.050kHz (or even 11.025 kHz) have sometimes been used for lower quality audio on Macintosh and PC-compatible computers. Games, web-sites and other multimedia productions utilized 22.050kHz (or lower) 8-bit audio to save disc space or bandwidth in the past.

FM stereo broadcast transmits a pilottone at 19 kHz which is used to generate a subcarrier of 2*19kHz=38kHz for the demodulation at the receiving end. FM receivers suppress this 19kHz pilot with a filter. If not properly filtered the tone will leak through. Sometimes this carrier is AM-modulated for the transmission of low-speed data which would slightly broaden its spectrum.

Edit: Just saw the picture with the tone at about 22kHz, so the FM case is less likely.

[StainlessS]

I know pretty much zilch bout this stuff, but do remember that Mp3Pro (used by MusicMatch/Nero and others) was supposed to
encode 44.1KHz into 22.05KHz mp3 with little loss of preceived audio clarity on decode, and half size encodes.

Probably nothing whatever to do with it, but:- https://en.wikipedia.org/wiki/Mp3PRO

EDIT: IIRC, a version of MusicMatch somewhere about v7.0, had free encode to Mp3Pro (usually some limit on
free number of encodes [EDIT: unless you got the paid version of MusicMatch]), maybe they forgot to enable the free encode limited timebomb or something.

EDIT: MusicMatch (about v7.0) stopped working properley with the introduction of Internet Explorer 7.0 (dont know why).
MusicMatch still my all time favourite MP3 player, till ownership changed and they messed it up, wish it still worked.

[Ghitulescu]

What is the source of the audio?
Digital audio is immune to satellite noise, for instance, and analogue signal has the resamploing errors filtered.

[neil wilkes]

This is from a rip of the US Vinyl version of 'Smallcreeps Day' by Mike Rutherford.
It does not appear on the UK version of the same album, so had me absolutely baffled.

[Boulder]

Very strange indeed. Can you hear it in any way if you separate and amplify it?

[SeeMoreDigital]

This is from a rip of the US Vinyl version of 'Smallcreeps Day' by Mike Rutherford...Do you know what equipment and software was used to create this analogue to digital back-up?

And... What audio format is the digital back-up? And what is it's sample-rate and bit-depth?

[Sharc]

Very strange indeed. Can you hear it in any way if you separate and amplify it?
Maybe the OP should ask his dog or cat......

[Boulder]

Maybe the OP should ask his dog or cat......
True :D
I was thinking more on the lines if it spreads a bit over to the area where you may be able to hear something like a pattern. I'm sure it's something like the signal in Contact :eek::cool::D

[jpsdr]

I was curious of the spectrum of bats, but it's not on the sheet... :D

[SeeMoreDigital]

This reminds me...

I checked in on my father yesterday and could hear a really annoying high pitched bleeping coming from his kitchen. It turned out that the battery in his carbon monoxide detector was running out!

He couldn't hear it. And neither could the person who delivered his shopping a few minutes before I arrived :eek:

[StainlessS]

I was curious of the spectrum of bats, but it's not on the sheet... :D
Looks like they missed them off the image which can be found at below link.
https://en.wikipedia.org/wiki/Hearing_range
Bats

Bats have evolved very sensitive hearing to cope with their nocturnal activity. Their hearing range varies by species; at the lowest it can be 1 kHz for some species and for other species the highest reaches up to 200 kHz. Bats that can detect 200 kHz cannot hear very well below 10 kHz.[28] In any case, the most sensitive range of bat hearing is narrower: about 15 kHz to 90 kHz.[28]

Bats navigate around objects and locate their prey using echolocation. A bat will produce a very loud, short sound and assess the echo when it bounces back. Bats hunt flying insects; these insects return a faint echo of the bat's call. The type of insect, how big it is and distance can be determined by the quality of the echo and time it takes for the echo to rebound. There are two types of call constant frequency (CF), and frequency modulated (FM) that descend in pitch.[29] Each type reveals different information; CF is used to detect an object, and FM is used to assess its distance. The pulses of sound produced by the bat last only a few thousandths of a second; silences between the calls give time to listen for the information coming back in the form of an echo. Evidence suggests that bats use the change in pitch of sound produced via the Doppler effect to assess their flight speed in relation to objects around them.[30] The information regarding size, shape and texture is built up to form a picture of their surroundings and the location of their prey. Using these factors a bat can successfully track change in movements and therefore hunt down their prey.

"Some birds, most notably oilbirds, also use echolocation, just as bats do. These birds live in caves and use their rapid chirps and clicks to navigate through dark caves where even sensitive vision may not be useful enough."[33]



From a PDF somewhere on https://batslive.pwnet.org/
When bats use echolocation to hear and “see” things, they vary their sound signals so they don't waste energy “shouting” when the insect they want to eat is only a few feet away. We suspect bats can hear things from at least 40 feet away, and perhaps further.

This reminds me...

I checked in on my father yesterday and could hear a really annoying high pitched bleeping coming from his kitchen.
I have to apologise, I felt sure that your comment was gonna be a quip about your mother. :)

[scharfis_brain]

I've read somewhere that - if a record hasn't been cut carefully - the cutting head of the cutting machine might go into resonance, causing this high pitched noise.

Is the modulation in frequency by any means locked to one rotation of the record?

[Emulgator]

To me it looks like FM +-500Hz, for a ~10ms period, so ~100Hz sawtooth modulation frequency.
Could be an oscillating stylus driver amp within the cutting lathe, but both channels ? Rather not.
Later: But these are mechanically coupled, so well: Rather possible. Torsion ?

If intentional, IIRC, DMM mastering might have an intentional HF part overlaid, to ease cutting into metal,
just like the 80..160 kHz BIAS helps modulate the otherwise nonlinear ferrite particles on magnetic tape.
But I expected that to be rather around 50kHz, where the abilities for quadro cutting would end anyway.
Guesswork by me, I should ask somebody who has experience in cutting master discs.
I think I saw the service manual of the Neumann VMR70 once, mentioning poles at 24kHz ? 50kHz ?
Have to look that up once I have got time.

P.S. Just got up and looked at Neil's graph again: 29 cycles within 10,06-9,82s = 0,24s makes ~ 8,33ms cycle.

Beautiful: Then U.S. mains frequency is my closest suspect here, after rectifying you get a 120Hz voltage ripple,
which is formed by charging main caps in sinusoidal shape, then discharging in linear shape.
Here modulation seems to be inversed: Higher voltage guides lower frequency, now it looks like that rounded sawtooth.

My best guess is self-oscillation of cutting amp with the cutting head,
Amplitude is quite high, so the driver amp must have gone fully up until limiting had set in.
Now it becomes modulated by ripple of the fully loaded PSU while drawing full power because of that self oscillation.

There is a motional feedback loop involved for the cutting head, always a source for instability if not compensated correctly.
Any exaggerated loop gain or mass variation (after stylus replacement ?) might contribute to that.
Such cutting head might have been sitting there silently squeaking at suprasonic thoughout several productions,
if no pilot lamp cries out "murder".
A fault, not a feature, noone would want 22kHz residue at their tweeters. Please forget my DMM musings...

~25 years ago I got an Ortofon 2x 250W MOSFET amp in for service, derived from their cutting amp.
Had been used as PA amp for a small discotheque here, was instable too.
Had shot its MOSFETs before I repaired it and tamed it into stability IIRC.

Goes beautifully together with what scharfis_brain said.
And that 22kHz residue on that particular cut may very well have gone unnoticed if noone was using an analyser against that master.
Reading Smallcreep's Day up, it looks as if that vinyl, album and single exist in versions with swapped sides (US) and other production faults.

What said a famous female film editor in that film's extras? "If in doubt, cut it out !"
(and left many scenes of violence in, don't remember that movie for good now)
Here it is only a quick notch on the good old parametric EQ or the FFT EQ

Late edit: Reading up on
https://pure-analogue.com/de/die-schneidemaschine/
https://www.speakerscorner.de/
and other lathe sites showed that Ortofon cutting heads have proved instable, prone to repair and are avoided these days in favour of Neumann or Scully.
Im Einsatz sind hier heutzutage der Neumann SX-74 und Ortofon DSS-732. Letzterer gilt als sehr anfällig und wird daher nur selten und von keinem unserer Studios eingesetzt
Maybe it's just that oscillating driver amp loop (could be 2x500W from an Ortofon GO741) that burnt their precious coils ?

[Ghitulescu]

This is from a rip of the US Vinyl version of 'Smallcreeps Day' by Mike Rutherford.
It does not appear on the UK version of the same album, so had me absolutely baffled.
In vinyl reproduction there are lots of parasitic resonances.
If both (US+UK) albums have been played and digitised using the same hardware, then it can only be the master and a faulty filter/renderer (in the duplication chain), or, an overmodulated vinyl (but this would usually cause many harmonics which are missing). A lot of "analogue" vinyls originate nowadays from digital masters.

[neil wilkes]

In vinyl reproduction there are lots of parasitic resonances.
If both (US+UK) albums have been played and digitised using the same hardware, then it can only be the master and a faulty filter/renderer (in the duplication chain), or, an overmodulated vinyl (but this would usually cause many harmonics which are missing). A lot of "analogue" vinyls originate nowadays from digital masters.

I know this only too well (the last point) sad to say.
I can get the details of the chain used, but it's not the chain otherwise the UK version would look the same so the cutting head theory looks to me like the best solution so far.
It's absolutely looking like a modulated signal.....

[neil wilkes]

Do you know what equipment and software was used to create this analogue to digital back-up?

And... What audio format is the digital back-up? And what is it's sample-rate and bit-depth?

Will get this information (it's not my rip - I got involved when I was asked what the heck it was, as the ripper had never seen anything like it before.
Sample rate & bit depth is 24-bit 96kHz....

[neil wilkes]

BTW:
Thanks to everyone who has chipped in here - it really is appreciated.

[Ghitulescu]

It's absolutely looking like a modulated signal.....

It was a time when the companies amused themselves with all kind of copy-protection (prevention) methods. One of these used such signals amongst which 22kHz was a favourite. Such a parasitic signal would affect the Dolby system (like the pilot tone of FM Stereo).

[foxyshadis]

In vinyl reproduction there are lots of parasitic resonances.
If both (US+UK) albums have been played and digitised using the same hardware, then it can only be the master and a faulty filter/renderer (in the duplication chain), or, an overmodulated vinyl (but this would usually cause many harmonics which are missing). A lot of "analogue" vinyls originate nowadays from digital masters.

I was thinking EMF during capture could be another reason; even if the filters are working properly internally, the AC input is still the wild wild west, along with any other cables laying about. It might be worth making a second capture just to see if it's baked into the record at the same timestamps. If not, then problem solved. If so, well, too bad the recording engineer didn't catch it.

[neil wilkes]

It seems that we have found the cause of this odd signal - it turns out to be a Black Light lamp (https://www.adj.com/black-24blb) near to the setup!
Turn this off & the signal goes away, turn it back on & it returns.
So the above from foxyshadis seems to be on the mark - EMF during capture, probably caused by cheap, nasty screening of the PSU on the Blacklight or the bulb itself (probably not the bulb as the wavelength of the UV is 365nm, and that is far, far too high a frequency when converted so to my mind this must leave the PSU (AC is 100-120 @ 50/60Hz with a 13watt output).

So mystery solved - and many thanks for all your input in this thread.

[SeeMoreDigital]

Interesting...

It's just reminded me about this: https://www.bbc.co.uk/news/uk-wales-54239180

[Emulgator]

Ah, so it wasn't cut into the groove.
And I had assumed perfect capture conditions, not a SMPS close by.
22kHz fits the bill, and the US mains ripple modulation pattern too.

Just today a Neve preamp copy (NV73) came in for repair and what do I see:
A bloody SMPS to get +24V out of +-16V, sitting close to the precious µV-sensitive Microphone preamp, input transformer, physically touching the output pot. Photo anyone ?