AudioMasters

Today I bought a CD as a Christmas gift for my uncle. It is Charlie Haden's "Land of the Sun".  I am not very familiar with this type of music, but I heard that jazz-like thing are recorded and produced with great care. Just of curiosity, I listened to some tracks and heard something strange at high frequencies. When I looked at it in AA, it turned out that there is a constant 15840 Hz tone throughout the recording. There are also some weaker components at 17k and above. Please see the attachment.

I was really surprised by this. Is this normal? Is this a production flaw? Or something deliberate (but what for)?

I quickly checked several other CDs. The signal is not caused by the extraction process (something I wouldn't believe, but it also had to be checked) because it does appear on some disc and does not on others. When it does, the frequency and level differs across and within discs, and it is sometimes lacking before the music starts. Sometimes there are few tonal signal, or the signal is apparently modulated.

I extracted from one disc both with AA and EAC, and the result was the same.

Here are results of my tests:

CD, track (usually I looked at the beginning of each track)
frequency : level of most prominent signals (in AA's frequency analysis)

Charlie Haden 'Land of Sun', track 8
15840 : -67

George Michael 'Older', track 1
15590 : -75,  15660 : -79

track 8
15560 : -75, 15630 : -79, 15780 : -73, 15850 : -82

Slayer 'Diabolus in Musica', track 11
nothing found

Mozart Great Mass in C minor, Gardiner, track 1 and 11
nothing found

Brahms 4th Symphony, Kleiber, track 1
15680 : -91

track 2
15610 : -96, I noticed also a few smaller peaks both below and above

Beethoven 5th Symphony, Bernstein, track 1
15660 : -76

track 2
15620 : -70

Again: what is this  :?:  :?:  :?:

Possibly some interference from a TV related circuit somewhere in the recording process. In this country the TV line frequency is 15625Hz. Was any sort of video equipment used at any stage that you're aware of?

Cheers,
Alan

Yes, that's what I was thinking - are these tracks recorded at TV shows? Mind you, that line looks pretty solid - the only way that you'd tend to get it like that is if they recorded the audio from a master VT playback channel with no filtering on the output.

...If I haven't beaten you to it.  :twisted:

'Morning Steve!

Oops - got trumped by an edit!

Good evening, Alan!
Actually, I think it was me that got trumped...  :lol:

But for PQ's reference, there's a table here (http://www.sptv.demon.co.uk/tvstandards.html) with some relevant information in it.

TV was something that came to my mind, too.  I recalled that this annoying high-frequency noise emitted by TV-sets has a similar frequency (I had measured it once).
I know nothing about TV equipment during the recordings. All the tracks come from normal, commercialy available audio CD.  Are all these musicians, from classical orchestras to pop artists, watching TV during recordings? ;-)

I find that sort of thing on more than a few LPs I record for CDR. It is not at one particular frequency, however, it varies from LP to LP and sometimes it not consistent from track to track on an LP. It is generally not loud enough for me to notice, although perhaps in the days when I had to sit as far as possible from a TV, in order to escape its howl, the LPs might have been annoying too.

Quite a few studios use TV monitoring between control room and studio - often in both directions, but I don't think that this is the mechanism, because that line wouldn't be anything like as constant as it is. You'd have to have a mic pointing pretty directly at the source to pick this up, and I can tell you from first-hand experience that this really doesn't happen.

But if a performance gets recorded in a TV studio, and it's done the way that it often is these days, where the band or singer in the band sings to a backing track, and the entire result is recorded on VT, then it's not that difficult at all for line sync signals to pollute the audio lines. If you looked at the patching arrangements in most studios, you would begin to realise why!

It's not that unusual for 'cheap' CDs to be released from performances like this, and if no attempt is made to filter it out, there it stays...

I'm not saying that this is the definitive answer, but I'd be very surprised if TV line sync doesn't come into it somewhere...

I can't check the first (Charlie Haden) disk right now, but none of the other CDs with the signal was a cheap one... And they do not seem to be recorded in TV studios.
Brahms and Beethoven was recorded in Vienna, in Großer Saal of the Musikverein, in 1980 and 1977, respectively. The CDs are from Deutsche Grammophon.
George Michael's 'Older' was recorded at Sarm West Studio (Steve, don't you know the place?) in London probably around 1995. This CD is from Virgin.

If these are cheap and carelessly prepared CDs, which are expensive and properly recorded then?  :shock:

I have seen similar fixed frequency lines as the result of a resampling for real-time sample-rate change (e.g., 48 to 44.1), but I don't think the frequency is right for that in this case.  

On LPs it can be oscillation in the drive to the cutter.  I once had two different LP cuts of a particular recording, one with a c.10kHz tone for the first few minutes, and the other without; I now have two commercial CD transfers of the same recording, and one has the 10k tone, the other not!

Paul

Here I've got  Jaques Loussier CD from the late 80's with a fixed low level frequency of 15633.  It shows in spectral view, and in quiet passages, but no in a frequency analysis scan due to its low level.  Must be a jazz thang...

I see the same artifact in the hi-fi audio on S-VHS recordings.  Some consumer brands, and most "professional" S-VHS VCRs seem to be "quieter" than others with regards to this signal -- although in truth I can't say I actually hear the noise under normal listening conditions.  I assumed it's part of the process of multiplexing the audio into the video signal on the tape... on some machines it nearly disappears when the audio signal goes quiet.

If I'm concerned with the impact on the audio,  the AA Frequency Space editor makes short work of it.  But I'm usually recording voice for lo-fi uses anyway, so I generally don't have to worry about anything above 11k.

I've been wondering a lot about this, because it doesn't altogether make a lot of sense - as I'm sure that anybody reading the thread has realised. What I've been trying to come up with is a plausible explanation for this at all, because clearly TV itself doesn't get implicated in several of these recordings.

My current thoughts are centred around Sony and some of the machines that were used for digital mastering back in the 70's, and have continued in use until quite recently in some cases. I'm just wondering whether an out-of-adjustment Sony 1630 or any of the F1 variants could do this? Considering that they are all based on video technology (and some were used for mastering stuff that was only ever released on vinyl, I believe), then this could possibly be the missing link...

Welllll.... don't know if this helps as a clue, or further confirmation.  For the live recordings I have to do, a Sony Minidisc machine receives the soundboard's signal first, then passes it out to other places including my soundcard via the SPDIF connection.  Only every once in a while will I also get this peak around 15 - 16K.  But, since it's intermittent maybe this doesn't qualify as evidence of Sony's involvement.

We use a camera and a monitor for our organist when the sight line to the conductor is blocked (which is frequent).  We also use TV monitors at the front row of the audience for some of the singers when their back is to the conductor.  Thus, even though we've never shot for TV, it is quite common to find that 15,625 signal in my recordings.  And it takes very little of it to show up in spectral view.

As far as the other instances go, I can only imagine what all has gone on in the studios.  It's reasonable to suppose that all manner of reference who-knows-what, as well as communications, could be in play here.

Can a TV camera create this as well?  A friend recorded a vocal album with one of the London orchestras about a decade ago, and his audio has "flyback whine" also.  I've no idea why a TV would have been present.

I've just spent half an hour ripping bits of tracks from commercial CDs, and the results are variable. There seems to be more of this on the left channel when it shows up, and it's usually around the -70dB mark. And you can see it on a frequency analysis, certainly. On some tracks it comes and goes, as well. I've not seen a really solid line of it, but a few that get pretty close.

I have no idea exactly what the production process was on all these random CDs, so for a comparison, I picked a commercial CD where I know exactly what the processing involved was - because I did all of it up to the point where it was glass-mastered. And what's really weird about that is that there's a very faint line at about 18k on it that's definitely not on the file supplied for the glass mastering - I checked carefully. And it's not consistent, either. This is at a level which is, quite frankly, completely immaterial; it's around the -80dB mark.

So now, I have no idea what's happened here at all. But you have to be a little suspicious about the glass mastering process - especially one where it's not doing this consistently - and this seems to apply to a lot of CDs.

Just curious, where are you guys getting 15625hz?

Original B&W NTSC Standard:
30fps x 525lines per frame = 15750hz horizontal scan rate

Color NTSC:
29.97fps x 525lines per frame = 15734.25hz

JTC

Whoops!  Just read the thread again and realized many of you are well outside the land of NTSC.  Apologies all around.  It does sound like a video related noise source.

JTC

OK, so PAL runs 625 lines at 25 frames per second, which gives you 15,625.  But you're right, since we're in NTSC land over here, I should be talking about 15,750.

I've only had the opportunity to inspect audio both before and after glass mastering five times, and have never found a single bit to be different -- the result one would hope for.  Steve, could some other mastering step be involved?  Say the digital master went through one of these places that seem to be all the rage these days, passing the signal through an analog step in the chain, and a tiny bit of interference is picked up along the way.  I'd vouch for re-digitization, in any case.

And for those with the whine more on the left channel, well, I guess the video monitor was on the left side!  But why would this be so consistent?  (Ours would be heavier on the right, for what it's worth!)

Remember that this was a higher frequency - something strange is happening in the transfer process, I think. With the particular one I mentioned, I watched the courier depart to the place where they only have glass mastering and pressing - the only thing that happened before it went was that we picked the master copy that had the lowest BLER - and that's a read-only process. What I was looking at is, I suspect, something different.

But as far as the others are concerned, what is quite intriguing is the CDs where this comes and goes. I haven't yet seen one where the spectrum line is as well defined as PQ's example - all the ones I've seen have close sideband noise, which has the effect of blurring the line. That could be the result of the FFT spectrum though - I'll have to try altering it and see what changes. One of the things that puts this firmly into line rate territory is the absence of other spurii - the harmonics of this would be outside the audio band. The other thing to note is that this would have to be an effect where the spurious frequency was added to the original at the point where it was digitised - which makes any of the earlier digital mastering machines rather more suspect, because they were all based on modified video recorders. Or in some cases, unmodified video recorders. The world standard machines were all NTSC ones, though - that's what a Sony 1630 is. PAL ones weren't used for this at all. Since the digitised audio was spat into these things in video frames (that's how you got away with unmodified machines) and the resulting video was reclocked into a continuous stream sent to the glass mastering machine, it's quite easy to see how spurious signals could be injected by a machine that wasn't set up particularly well.

The slight variations in frequency also make slightly more sense if this is the scenario, as well. The U-matic VCRs have some lattitude in what they record, and effectively derive their clocking sources from the incoming signals. You can drift the sync signals quite a bit before the machine refuses to lock up, and in theory this doesn't matter, because the digital stream is reconstituted in a separately clocked FIFO afterwards anyway. What this means is that the frequency you see might appear to drift slightly - even though in most cases, it won't.

Anyway, this is my currently preferred (theoretical) source - unless somebody's got another they wish to postulate!

Just in case, I heaved out the my first ever CD (double) that I recorded for Philips back in April '84 - on an F1/SLF1 system, edited on the Sony Umatic systems then in use (at HHB when they were working out of a garage).  No sign of the mystery tone.  I also took the opportunity of checking the 1997 re-release / repackage of the CD with other works, and it inverts nicely, which shows they didn't monkey with it at that point.

Aha - but in the non-Ozpeter tracks, which are on the second half of the second CD of the re-released set, there is a faint trace of the tone - I nearly missed it at first glance.  The date of that recording was Sept 85 - I know not how it was done, apart from it being an in-house Philips production in a Swiss studio.

Note that no TV equipment other than the F1/SLF1 was present at my original recording.

Make of that evidence what you will!

I don't know if any of you susbscribe to Gramophone magazine, but the opening track on this month's cover disc, from a new DGG release of Beethoven Piano Concertos, suffers from 4 mystery tones.

Spectral view in AA shows a continuous line at 15625Hz, an intermittent line at about 15585Hz and two higher solid lines at 19235Hz and 19335Hz.

I had ripped this track (which is from a live recording) because there was so much rustling noise at the start I wondered whether AA could get it out. I'll admit I'd not heard the tones, but of course I couldn't miss them in spectral view...

Only FWIW because it's consistent, whereas what others are reporting here seems to be INconsistent:  When dithering is performed, at least on Triangular 48kHz, there is always a low-level peak from 20500 - 21000kHz.  It's never heard, but always there, and always in those frequencies.  In spectral view it shows up as a very similarly solid line as what Pawel showed.

I think that these tones are so low in amplitude that they're just not heard, just as dither isn't heard.  But Cal, I hope you're not inferring that this is a byproduct of dither.

When you get into the current generation of recording and start seeing tones at frequencies other than 15,625 or 15,750, we need to take into consideration other electronic devices that may be present.  In particular we can blame computer monitors, which can have many different possible horizontal frequencies.  I don't know about PDAs, Segas, Gameboys, or other toys the percussion section might have stayed amused with...

Aha... and that's why you won't find my name in any of the technical threads.  Like Sergeant Shultz says: "I know nothinkkkk!!!!"

My technical knowledge about all this renders me hard-pressed to make any knowledgeable informed statements that could actually mean anything.  But by accident?  For sure, there.  

No... my only reason to mention dither in this thread was only as a by-the-way type comment, since I'd seen similar displays of the dither signal that looked even remotely like what Pawel showed.... just in case you guys who can talk the talk might have missed anything.  But that's why I qualified my comment by stating I recognized there probably is a difference between dither and what is being discovered in some of this audio data.  And you've just confirmed that.

So I can offer this Dither Disclaimer: Any mention of dither by me in this thread is probably spurious and totally confusing the issue.  Please disregard.

(Whewww.... off the hook now), and I'm puhlennnty content to keep these type subjects as READ-ONLY by me.

 :)

.. and Steve tentatively blamed older equipment

I would like you to know that the Charlie Haden's CD, where I first noticed the mysterious tone, was released in 2004 and, apparently,  recorded in 2003 or 2004 ("Land of the Sun  had its genesis in 2003 when Haden met with..." -  from the Verve record website). The signal was stronger here than on any other album I've checked, and the other albums were older.

BTW, I realized that the apparent strength of the signal as seen in the Spectral View might be misleading because it depended on Display parameters. As far as I remember, these were: 120 dB range, logarithmic energy plot, FFT:1024, Hanning window, 50% Window width.

For a more precise view of the signal please see the current attachment. This is the same track, bass intro - same as before, but other part of the intro.

The reason that this is tentative is that I believe that in some places, it is still being used to present masters to pressing plants for glass mastering. It's certainly currently listed as an acceptable transfer mechanism, along with DAT, and there may well be material still being presented today in this format. It's not one I ever used for the purpose, though. This might account for why the artefacts are there. Certainly, the older and further out of adjustment one of these machines gets, the better the chances of additional spurii.

The only way you could find out easily would be to ask the CD pressing plant, I'd think.

So far no-one has mentioned SSL consoles which, apparently, were well known for injecting these tones from their video display. In my younger days I used to hear 15kHz or so on many records and I just put it down to having better hearing than the mastering engineers at the time. Using a Decca FFSS cartridge which probably resonated at around that frequencydidn't help either.

Cheers

James.

I think that I can explain these signals :idea: .  The governments work together with record companies and transmit subliminal messages to our brains using the 15625 signal as a carrier  :shock: . This signal is exactly the thousandth harmonic of brain beta waves' frequency and therefore allows to control our minds by interfering with these waves  :roll: .  That the frequency of the signal is very similar to the frequency emitted by TV sets is not just a coincidence. The plot includes the makers of audio-video devices as well  :!:  :!:  :!:

 :wink:  :wink:  :wink:

This is quite true. It's no coincidence that we tend to fall asleep in front of the TV. This puts our minds into a state where we are much more receptive to these insidious messages. How else can we explain the re-election of George W and Little Johnny Howard? Removing these signals from our TV sets would have the twofold advantage of allowing us to enjoy our programs through to the end and freeing our minds of the evil propaganda of our politicians.

Sounds like a cogent argument for removing all signals sent to TV sets to me. It would probably give rise to better radio programs to enjoy - assuming we could keep the politicians off those too.

You'd think that there was enough trouble in the world already without inflicting politicians on us via the media as well, wouldn't you?

Ah, but I would miss the cricket & Aussie Rules.

I prefer the real thing... Okay, I'm a bit spoiled here - I can get to the Oval and Lords on the tube, Rosslyn park and Twickenham in a short car ride, and I can go and watch the Derby, or go to any one of 3 other racecourses within easy reach. And that's just sporting events - when you start on the cultural stuff - well, I almost don't know where to begin.

And because it's all so easy, I hardly ever bother with the sports stuff at all. If there was less of it on TV, then perhaps I might. I do still go to live concerts, though - you don't get quite so many spurious tones that way!

Ah yes, so do I, and it's accessible here too so I do, even though we're not quite so spoiled. But we manage to attend a variety of sports, plays & concerts. From our suburb built on what used to be a vineyard Sydney CBD is 40 min away via the expressway.

But I need that TV sport fix. :?

Careful, make it sound too good and, just as soon as a good airfare deal is found, you might have a family of 5 on your doorstep looking for a tour guide!



I second this!

Dunno about spurious tones not being heard at concerts - what about all those mobile phones?!  :(

Deja vu!

Wouldn't exactly be the first time. Sometimes we've had to accomodate them as well...


I suppose that you wouldn't believe me if I told you that London audiences were better behaved?  :wink:

Hmm.... I thought not!

Good then!  We'll be on our way shortly.
 :shock: