Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: >20kHz content found in vinyl? (Read 46219 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Re: >20kHz content found in vinyl?

Reply #25
DVDdoug and Cavaille, thanks for your comments. If you only read my last post and left out the rest of the thread, it might not have been so clear what I was asking - or maybe I wasn't being very clear to begin with.
So, I'll spell it out here:
The wiki on vinyl myths includes the two bullet points I mentioned in my last post. We should strive to only include something in the wiki that can be backed up by actual proof of some sort. What I'm advocating for is that either someone, preferably the person who submitted those two points to the wiki, should show us some realiable source for this (as mentioned earlier, I hadn't been able to find any by a quick Google search), or those two points should be removed from the wiki altogether.

As for your vinyl albums, Cavaille, as I've mentioned earlier, it's my belief that most vinyl albums are cut using a low-pass filter that cuts off the highest frequencies. So the source material might contain supersonic frequencies, but a low-pass filter cuts it off. It varies where the filter starts cutting off, but usually between 15 to 18 kHz is my impression (just an impression). Even the world's biggest CD hater (Michael Fremer) admitted to the use of this filtering (when pressed).

Quote
* Vinyl records commonly (not occasionally) contain musical content (not noise or distortion) up to 23-24 kHz
If you can't hear it, it's not "musical".... It's not even sound...

Without having the master tapes (or digital masters) you can't possibly know if any noise/harmonics were generated as part of the vinyl production/playback process or if they were present acoustically or generated somewhere else in the production chain.  It could also be helpful to know what filtering was used ahead of the cutting lathe.

I made more or less the same point previously. We need actual proof that what is on the vinyl records at 23-24 kHz is actually music and not noise. But even though we can't hear it, it is possible that it is actually musical content - just like the supersonic musical content on certain hi-res files. Personally, I have no interest in hi-res at all (and can't hear the difference), but I'm striving for an accurate wiki on this topic. Just to be clear: If somebody can actually show us proof of the claims, then they should of course be allowed in the wiki.
"What is asserted without evidence can be dismissed without evidence"
- Christopher Hitchens
"It is always more difficult to fight against faith than against knowledge"
- Sam Harris

Re: >20kHz content found in vinyl?

Reply #26
What do you make of this? Does this show that vinyl records include content above 20 kHz, as has now been added to the wiki (it wasn't in the wiki a while back).

1. You cannot be sure, that the signal in that part of the spectrum is actually encoded on the Vinyl, it might be coming from the pickup system.

2. You can make whatever that is that signal, by shifting all frequencies downwards, essentially running the vinyl at a slower speed, however this would also reduce the dynamic of the signal, and the tempo of whatever content. If you want to investigate using your own set of ears, you can bandpass the range you're interested in, and then shift the frequencies downwards, into a region that is audible. Frequency shifting is implemented in several audio programs, I think Audacity has one, too - and if not, I'm sure there's a plugin available for that. It shifts the frequencies, but preserves dynamic range and tempo. If you want to go the mathematical route, you could simply work on that signal samples in Matlab and manipulate them there, etc.

For instance to make 20kHz - 25kHz audible, I'd bandpass below 20kHz, at a fairly steep roll-off, and then freq-shift by a factor of 1/2 or so. so 20kHz becomes 10kHz and 25kHz becomes 12.5kHz, kinda squished together, but certainly in an audible range. Now listening back to that, you should be able to discern if the signal is nothing but noise, or if it is actually content which is discernible. As I said before, it would be just about enough to check if it is noise or "something else", actually understanding what sort of signals (or part of those) are encoded there, would be next to impossible.

I did something like this a couple times with "tuning" into sections of radio transmissions within the modulated audio signal, which are data etc. It makes it possible to check if something is different from noise, but that's about it. Techniques like these are frequently employed by HAM radio operators to identify what sort of signal comes through, whether it's AM, FM, SSB, PSM, QAM, or something else, like digital data. It's essentially using a radio set to SSB to "walk the frequency band" and see whatever comes through the noise.

supersonic [...]

...goes at the speed of over Mach 1.0...

FYE: Below audible range: "infrasonic", above audible range: "ultrasonic". When talking about sound the terms "infrasound" and "ultrasound" are used.

"Infrasound" is of course used by the snail people to control our minds and mating behavior, while "ultrasound" is used to impregnate women...

As a side note: one might argue, that once frequency-shifted the signal loses meaning because it only might make sense with its other bands, which have been also shifted, or suppressed. However if all you hear is noise, I'd argue it's just that: noise. When the noise isn't changing at all throughout the recording / sample, I'd argue it has no correlation with any other signal and is simply a very high frequency noise lacking any pattern, adding nothing to the original recording. If you really want to go into details how noise and spectral banded noise floor interacts with signals, especially signals that are below the noise floor, we're getting into technical details of signal processing. However I don't think any of that is really applicable for audio.

I remember doing this sort of "upper band investigation" once (by this time this is about ten years ago), but since my ADC caps at 48kHz, the highest frequencies capable to be represented are 24kHz. So, essentially I was looking at 22kHz - 24kHz coming from the vinyl. When frequency shifting a 5min sample, which included one and a half songs and the silence between the two songs, the noise coming from that had pretty much no discernible pattern. I remember being able to make out the silent section, though. In that section the noise got a bit lower in volume. However, I don't remember hearing the noise "following" the dynamic of the song, or something. It simply stayed at a higher volume than in the pause between the two songs.

Re: >20kHz content found in vinyl?

Reply #27
I do not want to take a stand on how common it is, but here is a test with an 80s digital recording vs. an early 70's synth, made to address the claim that the ultrasound is merely distortion from the medium: https://www.youtube.com/watch?v=_DdxdBNd5Ew
There seems to be way more in the old recording than what can be explained by noise. So as long as the master tape does not limit it, and there is no low-pass filter applied in the cutting process ...

By the way: http://www.positive-feedback.com/Issue2/mastering.htm , catching a 122 kHz tone by cutting at half-speed.

Re: >20kHz content found in vinyl?

Reply #28
From 20kHz it's at around -60dB and falling to -108dB as frequency increases, is it possible to hear that?



Re: >20kHz content found in vinyl?

Reply #31
The thing with the video is that you can see a very obvious cutoff with the first album that doesn't exist in the second.  Without getting into what is going on at the far end of the frequency range, it can be clearly deduced that the vinyl medium itself contains information at the transition band.

While I think the wiki article should reference evidence for many of the claims being made, some of them aren't really controversial, so in the edits that I made, I left them.  It should be noted that my edits were only focused on a few portions of the article.  I didn't bother with the 1000x playback, though I am skeptical and think it is at least deserving of a "citation needed" type of annotation.

Re: >20kHz content found in vinyl?

Reply #32
Did some searches concerning wear, and think I found something. It is obvious that too heavy weight can damage a record (here: https://www.youtube.com/watch?v=J4_GYfPZfq0 ), but I it is uncontroversial that vinyl records are not immune to abuse, the question is  the converse.

A couple at  https://www.vinylengine.com/turntable_forum/viewtopic.php?t=12931
- The posting from "Klaus R" with 13 literature references (none later than 1980). Reference #12 indicates that wear can be practically eliminated. Reference #13 indictes that dust in the groove is an efficient sander ...
- Same link, posting by user flavio81. Referring to something archived at https://web.archive.org/web/20110713112415/http://www.johana.com/~johana/dorren/cd-4paper4.pdf : the 30 kHz carrier tone for quadrophonic CD4 records survived 500 plays at 4.5 grams, as long as one cleaned the record.

Then I tracked down the source of various claims that an Ortofon OM40 could play for a tens of thousands times with negligible wear: http://www.faktiskt.se/phpBB3/viewtopic.php?f=9&t=3018&start=330
Search for 2006-12-13. Ingvar Öhman, former president of AES Sweden. They engraved a record with one single locking groove, so the same groove would repeat 33 1/3 times a minute, listened to it against the master tape, and then they left it to play repeatedly for a day and night (that is 48000 plays!) without any chance to cool off. Some styli did cause wear in two or three minutes (sixty to hundred spins then).


As for the frequency response, I agree with Greynol (the latest post) here: there is an obvious cutoff when the master has one, and not on the other. Now, whether that is audible, being in a range generally thought to be inaudible, is a different question.

Re: >20kHz content found in vinyl?

Reply #33
Let there be no mistake: the extension of harmonic content up to Nyquist in that video coming from the master and being "stored" on the medium is also a different question.

Re: >20kHz content found in vinyl?

Reply #34
Greynol, I agree that there should be a "citation needed" flag for those points I brought up. If you know (or can find) the person who put those things in the wiki, maybe he can shed some light on it.

The things Porcus brings up seems to say that certain equipment will damage a vinyl record quickly, whereas other equipment won't damage the record, but it was still a bit unclear to me if those sources verify the claim that a record can be played 1000 times without showing distortion.
"What is asserted without evidence can be dismissed without evidence"
- Christopher Hitchens
"It is always more difficult to fight against faith than against knowledge"
- Sam Harris

Re: >20kHz content found in vinyl?

Reply #35
I'm bumping this topic after a long break, as I forgot about the thread.

I think Greynol is mainly the person editing the vinyl myths wiki entry, so hopefully he reads this. In any case, I still find the following quoted section a bit confusing and contradictory (I highlighted what I found most confusing/contradictory), and I didn't see any references to these claims (which are fairly recent). A quick Google search didn't reveal anything that backed up the claims of "common 23-24 kHz audio content at significant amplitude on vinyl records" (but a more thorough search might reveal this). I still haven't found any concrete evidence that shows that vinyl records have actual musical content to 23-24 kHz, where it is shown that the peaks above 20 kHz are actually music present on the master it was cut from, rather than just noise or distortion.

Quote
(from source sans formal citation)
Quote
tests have been conducted which deonstrate that a record can be played up to 1000 times before there is any measurable increase in distortion as a result of record wear


I think that it is possible to do tests where low frequency tones on vinyl retain some semblance of their original integrity over multiple playings. 

One problem is that the original integrity of even low frequency signals on vinyl is not that good. About a decade ago when I had clients who wanted vinyl transcribed I assembled a vinyl playback system and bought some of the best test records could find as new products and also legacy test LPs that were NOS.  I don't recall ever seeing nonlinear distortion that was much below 0.2% under ideal conditions.  While that may be hard to hear, by modern digital standards it is piss-poor.  Things were worse as the frequencies rose.

Playing test records provided technical indications of other problems that are likely to be audible, including noise and FM distortion (Jitter or if you will Flutter and Wow). There will be audible low frequency FM distortion if the record is not nearly perfectly physically flat and centered. Neither are universal absent  and one or both are common.

There is additional FM distortion due to the friction of the needle in the groove varying its drag as the program material changes.

There is additional FM distortion that is due to the nearly universal  use of offset (angled) tone arms.  Note that near the end of the mainstream vinyl era a number of straight line tracking record players were sold by mainstream audo manufacurers, but as the LP became a niche product, these disappeared off of the market to a very large degree.

Many cartridges have magnetic circuits that fail to be sufficiently  uniform over the area where the coil or magnet travels and there will also be amplitude modulation as the magnet or coil moves about due to off center punching and records that are insufficiently flat.

I have also listened to presentations by collectors of legacy audio gear who experimented with CD4 records and decoders. The presenter described assembling a modern vinyl playback system with a modern cartridge and stylus that was designed to optimize ultrasonic response. They were successful in obtaining an indication of ultrasonic carrier detection on first playing. After something like 10 playings, the indication was lost. The presenters conclusion was that the ultrasonic content had been worn off in just a few playing's.  I believe that this carrier was in the range of 35 KHz. 

I understand that depending on stylus shape there is a phenomenon called "Pinch Effect" where the groove requires an impossibly narrow stylus to be tracked accurately  because modulation causes the groove to turn which reduces its cross section as seen by the stylus  in the plan view. For modern elliptical designs the frequencies where pinch effect becomes significant may be above 12-13 KHz. Pinch Effect reduces media life, causes nonlinear distortion, and reduces stylus  life.


As far as high frequency content goes, with a few notable exceptions almost all LPs were cut from either analog tape or digital masters.

The best of the digital masters of the day generally  had 48-50 KHz sample rates, so obviously content above 24-25 KHz would be impossible.

The overwhelmingly most common way in the day  to produce  recordings and cut a  lacquer involved 15 ips magnetic tape. Ultimately the high frequency band pass of these tapes was limited by the playback head gap. Making this gap smaller required precision but also increased the probability of introducing  amplitude variations above 10 KHz and also drop outs or loss of useful output for brief periods. Both of these effects are easy to see using test tones and an oscilloscope, and can be audible. Drop outs of any significant duration are so audible as to ruin the work. For these reasons professional tape machines had playback heads that restricted the high frequency band pass to about 24 KHz @ - 3 dB.

So a LP high frequency power bandwidth started being significantly limited in ways that were never practical to correct or compensate for  above 12 or 13 KHz, and completely died above 24 or 25 KHz for lack of program material coming from the master recording, whether digital or analog.

Re: >20kHz content found in vinyl?

Reply #36
I think we're all getting a pedantic here.

I have pressed a number of vinyl records over the course of ~10 years.  Most of the master are from 96Khz/24bit or 192Khz/24bit recordings (don't freak out over the high sample rates, they are exceedingly high in case I wish to further process the final recordings).

So I am in a position where I can validate a master digital file against a vinyl pressing (all are 33RPM unfortunately).  Now, if we could reach some sort of consensus about how I post results of measurements, I am happy to do so.  The music does contain frequencies will above 20Khz, as they are all made from rather old and dysfunctional analogue synthesizers.

Also, please let me be clear that I am not implying that I am interested in frequencies above 20Khz, merely that they can be represented in a vinyl disc.

Re: >20kHz content found in vinyl?

Reply #37
That won't be necessary. The video Porcus linked in reply #27 clearly demonstrates that vinyl is capable of preserving frequencies that extend beyond what a typical adult is capable of hearing.

But if you want to, feel free.  I'd like to see how much of the ultra high frequency content is real, rather than harmonic distortion, though a band-limited master would be the best way to check for that.

Re: >20kHz content found in vinyl?

Reply #38
I'd like to see how much of the ultra high frequency content is real, rather than harmonic distortion, though a band-limited master would be the best way to check for that.

I can also do that this evening...

Re: >20kHz content found in vinyl?

Reply #39
To see whether any of this ultrasonic stuff is just noise, or if it's actually produced by the music, would we not need to see the spectrum of the master tape?  If the master tape doesn't go this high, then it's obviously the production process that is introducing these ultrasonic frequencies.

Re: >20kHz content found in vinyl?

Reply #40
Full disclosure: It doesn't seem like I'm going to be in a very good mood today and should probably not be posting.

It really annoys me when people post without paying attention to what has been said previously. I mentioned post #27 once already, but it seems you couldn't be bothered.

https://hydrogenaud.io/index.php/topic,113365.msg940575.html#msg940575

Re: >20kHz content found in vinyl?

Reply #41
That won't be necessary. The video Porcus linked in reply #27 clearly demonstrates that vinyl is capable of preserving frequencies that extend beyond what a typical adult is capable of hearing.

But if you want to, feel free.  I'd like to see how much of the ultra high frequency content is real, rather than harmonic distortion, though a band-limited master would be the best way to check for that.

I can post an audio sample and a list of my gear (turntable etc.) but I don't see that as necessary.  An FFT of a vinyl master cut from a 44.1 render is attached to this post.  You can see a harsh cut at 22.05Khz, but also some sharp spike above that during transients.  I found this to be very interesting.

I can do the other one too, I mean show you a master of a 96Khz file, and compare it to a vinyl FFT cut from that.  It holds the high frequencies remarkably well...


Re: >20kHz content found in vinyl?

Reply #42
It isn't so much that it "looks" like stuff is preserved; it's the fact that the medium and playback chain put stuff there that didn't exist before.

You could look at it like SBR in AAC.

Am I saying it's all discarded and replaced with garbage?  No, of course not.

Thanks for doing this.  Hopefully it will lead to further inquiry.

If you haven't already, check out the video Porcus linked.  You might also want to take a look at a dynamic spectrogram as was done in that video in addition to the spectrograph.

Do you have any square waves (properly synthesized and band-limited, of course) that were pressed to vinyl?

Re: >20kHz content found in vinyl?

Reply #43
It isn't so much that it "looks" like stuff is preserved; it's the fact that the medium and playback chain put stuff there that didn't exist before.

You could look at it like SBR in AAC.

Am I saying it's all discarded and replaced with garbage?  No, of course not.

Thanks for doing this.  Hopefully it will lead to further inquiry.

If you haven't already, check out the video Porcus linked.  You might also want to take a look at a dynamic spectrogram as was done in that video in addition to the spectrograph.

Do you have any square waves (properly synthesized and band-limited, of course) that were pressed to vinyl?

I don’t, but send me exactly what you want and I’ll shoehorn it into the next thing I press. And also, I’m reasonably sure that it’s the payback chain that added the noise rather than the medium- not that it makes much difference either way

Re: >20kHz content found in vinyl?

Reply #44
The content above 20 Khz would have to be very quiet. Considering the limited dynamic range of vinyl vs other music formats (60-80 dBs), I don't think it would capture anything that quiet and still be able to get the loudest part of the recorded song.

But I will admit I am no expert in this area.

Even if it does somehow capture those frequencies, you can't hear them anyway.

Re: >20kHz content found in vinyl?

Reply #45
60-80 dB of dynamic range is actually very good.  Granted, it's dwarfed by a good digital system, but 80dB of dynamic range is still overkill for music (in my opinion, of course).

Re: >20kHz content found in vinyl?

Reply #46
According to Wikipedia “Vinyl microgroove phonograph records typically yield 55–65 dB, though the first play of the higher-fidelity outer rings can achieve a dynamic range of 70 dB.”
Rhythmbox, Flac + Vorbis, Sennheiser HD650 + Sony MDR-XB1000

Re: >20kHz content found in vinyl?

Reply #47
I agree that the dynamic range of vinyl is good.  But is it good enough to capture the super quiet frequencies above 20 khz?

Re: >20kHz content found in vinyl?

Reply #48
I agree that the dynamic range of vinyl is good.  But is it good enough to capture the super quiet frequencies above 20 khz?

What does it matter?  We can't hear them. I certainly can't.

Re: >20kHz content found in vinyl?

Reply #49
While some people may be able to hear above 20 kHz, none can hear the super quiet frequencies above 20 kHz.
Of those that can hear frequencies above 20 kHz, none can hear them in the presence of louder lower frequency sounds.
Kevin Graf :: aka Speedskater