The 2012 Vinyl Re-Issues Were From 44.1khz Source

[Lord Reith]

I had been under the illusion all these years that the 2012 vinyl re-issues were from 192khz sampled digital masters. Even people who ripped the discs ripped them at 192khz presumably for that very reason.

But a 2016 interview with the man responsible - Sean Magee - makes it all too plain that this wasn't the case. The highest format the remasters exist in is 44.1khz 24 bit.

http://theaudiophileman.com/stereo-beatles-vinyl

Has the vinyl been sourced from the original master tapes? Apparently not.

“We couldn’t really,” said Magee. “We have all the cutting notes left by Harry Moss (the original cutting engineer for The Beatles’ recording output) but we don’t have the same equipment. We could kind of recreate the analogue chain and kind of recreate what Harry Moss did to get that sound but it wouldn’t be the same.”

Another reason has been the demands of Apple: that amalgamation of the remaining Beatles plus the estates of the rest. Apple want any Beatles recordings to have a particular ‘sound’, a traditional presentation based upon the original recordings which, to some extent, constrained the mastering engineers at Abbey Road. To get the required sound required a considerable amount of EQ (Equalisation: boosting or reducing the levels of different frequencies in a signal), “To physically do this in real time whilst cutting from the original analogue masters would have been almost impossible to do,” said Magee.

The approved EQ shouldn’t be taken lightly, either. It took four and a half years to create it, prior to the release of the CD box sets in 2009.

Instead, therefore, the vinyl has been remastered from digital sources. These were created, before the CD box sets were released, at a rate of 24bit/192kHz. Magee found, however, that even those files were going to be a problem when remastering the stereo vinyl because of the EQ requirements. More than that, on the earlier albums, the primitive stereo processing placed vocals on one channel and instruments on the other which meant that, “There are different EQs on the left than there is on the right because the content is different on either side. Sorting all of these EQs, track by track, whilst cutting would be impossible. Also, you cannot do separate jobs at 192kHz. You can’t de-click, then EQ and so on. You have to do the lot while cutting. There isn’t the equipment at 192 to do that. Not easily, at any rate. The practicality and time of even doing that process at 24bit/96kHz would have taken about a year. You’d also need a lot of double checking.”

It so happened that the complex EQ applications had already been done for the CD version, “To use the 192kHz sources now would have entailed recreating the EQ source that we did at 24bit/44.1kHz, which wasn’t viable.”

So the decision was made, therefore, to master the vinyl at 24bit/44.1kHz


Okay, so to summarize, the analogue tapes were captured at 192khz but the remastered versions are all at 44.1khz (redbook standard). This explains at long last why the Apple USB dongle was at 24 bit 44.1khz... there is no higher version of the remasters. The only higher source is the raw dub of the master tapes.

Magee's comments also point to significant eq work on the tapes, something which I found also in my analysis of the remasters a year or two ago. There is in particular a lot of bass boost on the remasters, and some songs have very different eq from the old lps. I would imagine that an "approved eq" that took four years to arrive at is not an odd tweak here and there.

But hang on, Magee also says that the Abbey Road cutting lathe can't cut audio frequencies higher than 24 kilocycles anyway, which would make the maximum useful sampling frequency 48khz (DAT and DVD standard):

Despite the extra time that a 24bit/192kHz or even a 24bit/96kHz master would have taken to create there was, according to Magee, no real deadline for this project. So the impetus for using the 44.1kHz files was? “I was told to use these 24bits, so that’s what we used, it was the most practical.”

Practical? Because of the cutter head, according to Magee, “It has a limited frequency response and cuts off at 24kHz. There is nothing above that. As a cutting engineer, anything of significance level above 16kHz is dangerous, you don’t want that going to your cutter head because it gets very hot and can destroy it. It wouldn’t have mattered if the signal had gone to 192kHz or 96kHz, it wouldn’t have been on the record because you can’t cut it, you can’t hear it and I wouldn’t want it there anyway because a stray signal at 60kHz would destroy the lathe head. The most important part of the figure is the 24bit but not the 96kHz or 192kHz because the cutter head won’t even cut that content up to 48kHz.”

So, reading between the lines, there are four digital versions at EMI: the flat 24 bit 192khz dub of the analogue tapes; the 24 bit 44.1khz remasters without limiting (used to cut the 2012 lps); the 24 bit 44.1khz remasters with limiting added (used for the usb dongle); and 16 bit 44.1khz redbook remaster with limiting added (used for the cds and streaming).

Reading through the old publicity and press releases for the 2012 set, I consistently found the misleading quote that the tapes had been captured at 192khz... but then they were telling a half truth there, because they omitted to say that they were then remastered at 44.1khz.

Now, personally I don't really care about the difference between 44.1 and 192, but I think if it had been generally known that the vinyl was made from 44.1khz files, and that the cutting lathe can't cut higher audio than 24kilocycles anyway, a lot of people would not have bothered to buy them. Because the only difference from the cds is 2 or 3db worth of transients on loud passages, which are actually impossible to hear anyway because they are so short.

Food for thought. It also means the mono vinyl has no audio above 24khz.

[WilliamShears]

There’s a good article here that includes details on the mono set which includes the following:

https://www.superdeluxeedition.com/news/beatles-remastering-team-answer-your-vinyl-mono-box-questions/

Do you think it would be possible to create new stereo LP reissues using the same all-analog process used to make the mono LP reissues? In what ways do you think all analog stereo LPs reissues would sound different that the recent stereo LP reissues?

> Sean Magee: “It would be possible, but they would not sound the same as the eq choices and repairs that were done would not be possible to do in realtime, so they would be like the originals. The point of this box is that the Mono mixes were the way the band heard them, and the way they were indended. This is a historical document and frankly, the stereo versions, excepting Abbey Road and Let it Be, and some of the White album. Don’t have that. There are plenty of stereo Lps out there, millions in fact.“

This is why I settled on the mono vinyl set since I figured the stereo vinyl would not be too different than the CDs.

[alphabeatles]

This is very interesting! LR, can you (or anybody else who knows how to read a spectrogram) tell me if downsampling the 2012 vinyl rips in our collections to 24/44 is harmful to the audio in any way? If not it would save a ton of HDD space over the seemingly unnecessary 24/96.

Also, does this new information affect any of the other Beatles (or other artists') releases in our collections? ie, is 24/96 really necessary for any vinyl release, if the cutters can't even transfer that audio info to vinyl? I want to defer to our own experts' opinions...

Using Sound Studio on the Mac I got these spectrogram results for I Saw Her Standing There (mono and stereo results from different sources looked the same), from L-R 24/96, 24/44 and 16/44. Somebody please interpret:

hxxps://drive.google.com/file/d/1vd8M9vBIK3bujbLAJcbKv5u8HDHcbKr6/view?usp=sharing

Finally, I assume based on previous and upcoming releases (such as the Lennon sets) that of course Bluray discs are not limited by a mechanical process and we can indeed receive true hi-res copies of the masters – if such masters have been prepared.

PS: I had actually seen that article before, ha!

PPS: Dive in to the controversy!
hxxps://www.head-fi.org/threads/24-96-vinyl-rip ... er.680903/

"The main advantage of 24/96 is dynamic range and extension into frequencies above the range of human hearing. A vinyl record has a noise floor roughly half the dynamic range of 16/44.1 and the upper frequencies are usually rolled off to prevent distortion due to premature wear. The only frequencies above 18kHz or so are in the surface noise. So by using 24/96, all you are improving the sound quality of is noise. A total waste of time."

"The purpose of the 24 bit word length is to provide overhead for the studio to mix and master the recording. It is what recording engineers use. There isn't any sonic advantage in the final product because 16 bits handles the dynamic range just fine. A 96 kHz sampling frequency also affords some overhead but no sonic advantage. You can do a blind test for yourself to help understand that increasing word length and sample frequency above the red book standard doesn't have anything to do with sound it simply creates a larger file with more inaudible data in it.."

[Lord Reith]

This is very interesting! LR, can you (or anybody else who knows how to read a spectrogram) tell me if downsampling the 2012 vinyl rips in our collections to 24/44 is harmful to the audio in any way? If not it would save a ton of HDD space over the seemingly unnecessary 24/96.

Afterwards I actually found (and it wasn't easy amidst literally hundreds of pages) some posts by Sean Magee on Hoffman forums in 2012 where he restates (to the horror and disbelief of everyone) that he cut from 24bit/44.1khz. He then explained that the cutting head SX74 response "drops considerably at 14khz (audio)" and can't cut anything above 24khz. Note: Can't. Someone then asked about the 60/70s cuts and he said that the old cutting heads had an even poorer response. He also pointed out that no cutting engineer ever wants supersonic audio frequencies to reach the cutting head because it overheats.

Well, as you can imagine this got all sorts of people in a tiz because their spectrograms showed what appeared to be frequencies right up to 48khz. He said he didn't know why that was but it was probably something to do with the tone arm (ie: distortion). This was 2012, four years before that interview I posted so the general response was that he was an imposter and lying. A few people believed him though.

Okay, so the thing to take away from all this: 192khz for vinyl is complete overkill, because there are no audio frequencies even remotely approaching the 96khz audio cutoff that would provide. For any vinyl album, 48khz would capture all the musical information on the disc. You could sample at 96khz if you want but this would just give you an extra octave of distortion too high to be heard any way. 48khz is the DAT standard which was arrived at because the powers that be felt that while 44.1 captures all the audio info, the sharp cutoff may adversely affect the treble content lower down the spectrum. So for vinyl rips, 48khz seems eminently sensible, and what's more 48 is exactly one quarter of 196 so if you downsample you'll get a better result than the odd 44.1 value. And your file will be 4x smaller.

Also, does this new information affect any of the other Beatles (or other artists') releases in our collections? ie, is 24/96 really necessary for any vinyl release, if the cutters can't even transfer that audio info to vinyl? I want to defer to our own experts' opinions...

According to Sean (who is a cutting engineer) even the latest and most sophisticated cutting head can't cut above 24khz audio. So capturing at 48khz is eminently sensible.

He posted the workflow of the 2009 remasters. The analogue tapes were captured digitally at 24 bit/192khz, and declicking was then done with Cedar at 96khz then upsampled back to 192khz and inserted back into the main files.

These were then played back through an EMI TG mixing desk so they could get a bit of the old sound and do some broad eq. This was then converted back to digital but at 24bit/44.1khz (and some spot eq was done according to some of the other engineers). They then applied a Junger limiter at 24/44.1 and then did the conversion to 16 bit for the cds. The vinyl used the files prior to the Junger, but he did some extra de-essing patching in Cedar (not on any other version).

Finally, I assume based on previous and upcoming releases (such as the Lennon sets) that of course Bluray discs are not limited by a mechanical process and we can indeed receive true hi-res copies of the masters – if such masters have been prepared.

The only 24/192 masters are the raw dubs of the analogue tapes with the click repairs it seems. Don't think we'll ever get those. What was on the USB key was the highest rez that exists for the remasters.

"The main advantage of 24/96 is dynamic range and extension into frequencies above the range of human hearing. A vinyl record has a noise floor roughly half the dynamic range of 16/44.1 and the upper frequencies are usually rolled off to prevent distortion due to premature wear. The only frequencies above 18kHz or so are in the surface noise. So by using 24/96, all you are improving the sound quality of is noise. A total waste of time."

Yes, I've heard people like csnyfan say that before but this is the first time I've heard it come from an actual cutting engineer. He should know!

"The purpose of the 24 bit word length is to provide overhead for the studio to mix and master the recording. It is what recording engineers use. There isn't any sonic advantage in the final product because 16 bits handles the dynamic range just fine. A 96 kHz sampling frequency also affords some overhead but no sonic advantage. You can do a blind test for yourself to help understand that increasing word length and sample frequency above the red book standard doesn't have anything to do with sound it simply creates a larger file with more inaudible data in it.."

A lot of people confuse bit depth with bitrate. Higher bitrate means higher quality. That's why a 320k mp3 sounds better than a 128k one. But bit depth (either 16bit, 24bit, or 32 bit) has nothing to do with that. It provides greater dynamic range. It's like a bigger envelope to put your letter in. If I'm working on some audio and I did everything in 16bit, some digital noise would gradually build up in the background. If you do everything in 24 bit, that doesn't become a problem. But it doesn't make the audio sound clearer, or more vibrant, or more dynamic, or anything like that. It just gives you more dynamic range because the noise floor is much lower. 32 bit gives you even more room because you can have audio clip over 0db and it doesn't matter. 16bit gives a dynamic range of 96db. To put this in perspective, a symphony orchestra has a dynamic range of about 40 or 50db. A Beatles record has a dynamic range of about 10 decibels.

The apple usb key was piointless because it had a 44.1khz sampling freqency, the same as the cds. So any true supersonic content that may be on the master tapes and could perhaps be subliminally "sensed" by the listener is not there. The 24 bits doesn't make any difference at all, because to hear the difference in the noise floor from the cd, you'd have to turn the volume up so loud it would be like having a 747 taking off next to you when the song started. You might hear a small difference for very dynamic music like an orchestra or music with extremely quiet passages like soft piano. Magee actually said that it made no difference if you mastered an lp from a 16 bit file because the noise floor of vinyl is so high in comparison. That didn't go down too well I can tell you!

[C90]

Thanks LR, a very interesting read, but I get confused between sampling rates, and audio frequencies, which are both measured in cycles-per-second or herz.
Am I right in thinking that a sampling rate of 192khz means each second of audio content is sliced into 192,000 pieces which are individually digitised. But the 24khz limit on the cutting lathe's performance is nothing to do with sampling rates, but refers to the physical movement of cutting head, and the audio waves themselves as heard by the human ear? Two different things.
Apologies if this is basic principle stuff to many here.
Let's not start discussing the maximum frequency response of the human ear!
(Or of dogs, and the Sgt Pepper inner groove!)

[K_A_P68]

This is a fascinating read for me, thank so much, I was unaware of any of this.

[alphabeatles]

OK, so for downsampling LP rips, LR recommends 48khz (on the back half, due to the clean math/division). What about the front half (16 or 24?).

FWIW, DESS was distributing all of his vinyl rips on CD (16/44), so...

Regardless of bits 'n sh-tz, I understand one main attraction of a vinyl rip is to have music that still retains natural dynamics, as the format cannot handle full loudified brickwalled audio anyway. A digital equivalent would be an unlimited hi-res release, yes?

Another advantage of a vinyl rip is to own an original mix of a release that may have been altered for CD or hi-res release later on.

[alphabeatles]

Doing some tests with Spec, as if I understand the results, other than the extra sound info does not seem to be useful...

Source: Buckingham Nicks (DLedin 2012, 24/96), track Crying In The Night

Screenshot shows 24/96, 24/44 and 16/44
hxxps://drive.google.com/file/d/1MLgHs2NOXwn2e_LwsD4vLxdQD-_oWqeX/view?usp=sharing

Source: Sgt. Pepper (1982 MFSL UHQR, PBTHAL 2012, 24/96), track Lucy

Screenshot shows 24/96, 24/44 and 16/44
hxxps://drive.google.com/file/d/1o3CXbDcumdf0qvkxMdTKH_4DblhUQsoj/view?usp=sharing

Abbey Road (2012 remastered EU vinyl, Dr. Robert 2018, 24/96) tops out at 48k in the spec (not uploaded); everything above 15 or so kHz is purple.

Info on hearing:
hxxps://en.wikipedia.org/wiki/Hearing_range

More geeky stuff:
hxxps://www.mojo-audio.com/blog/the-24bit-delusion/

Bottom line seems to be, you create and edit the music at hi-res and distribute at low(er) res. Those of us at home most likely can't play back full hi-res or even hear the difference if we can.

[DrRobert]

Hi everybody glad I found this board! Had totally forgotten about BZ and went hunting and here we are!

In general, bit depth (more information per sample) is more important than sample rate (capture rate) for vinyl cutting.

Most of the vinyl I've had pressed has varied in terms of what the record cutter has asked for. Last two were 44.1/24 and they're fine. The mastering guys all want 96. When digitising tape that's a different ballgame as you have no real limitations other than what is on the source. Pressing vinyl is rife with limitations.

[Lord Reith]

Thanks LR, a very interesting read, but I get confused between sampling rates, and audio frequencies, which are both measured in cycles-per-second or herz.
Am I right in thinking that a sampling rate of 192khz means each second of audio content is sliced into 192,000 pieces which are individually digitised. But the 24khz limit on the cutting lathe's performance is nothing to do with sampling rates, but refers to the physical movement of cutting head, and the audio waves themselves as heard by the human ear? Two different things.

It is confusing you're right. The sampling rate refers to the numbers of digital samples per second. The more there are the more high frequencies it captures. The sampling rate should be twice the highest frequency you desire. For example, on cds the highest frequency is 22.5khz (that's 22.5 thousand vibrations per second of your eardrum, not 22.5 thousand samples). Like I said for DAT they decided to make the sampling frequency a bit higher (48K) to avoid any possible detrimental effect caused by having it at 44.1K. You may have heard of the phenomenon of "digital strings" - the often unnatural sound of string sections on early cds. That is why classical people are obsessed with high sampling rates - and for good reason. I too can hear an improvement with high sampling rates for this kind of delicate music.

The 24khz limit of the cutting lathe is audio frequencies, yes. It means that a source mastered at anything over 48K (DAT) sampling rate offers no benefit. All the extra audio frequencies are ignored by the machinery.

Apologies if this is basic principle stuff to many here.

No apologies necessary. It is very heady stuff, which is probably why most people have a poor grasp of it.

If you want an educational demonstration of the difference between 16 bit and 24 bit, you can do this in Audacity:

- load up one of the 24 bit tracks from the Apple usb key
- export it as an ordinary redbook 44.1khz,/16 bit file
-load that exported file back into audacity and invert it ("invert" in the EFFECTS menu)
- highlight both and "mix and render" under the TRACK menu

What is left is the difference between 16 bit and 24 bit. At first it just looks like a blank file, but if you play it and turn the volume up REALLY high, you'll hear some faint hiss (or you can just use the "amplify" command in EFFECTS). That hiss is the noise floor of the CD format. And that is the only difference between the two: 24 bit has less noise, which means for music with tremendous dynamic range (like a grand piano or orchestra) the faintest parts will never ever approach the noise floor of the format. That is why it is used by sound engineers too, so that no matter how many times the signal is processed it won't accumulate any noticeable digital noise.

Some people on the Hoffman board (hey, I'm not knocking Hoffman, there's a lot of valuable info there too) opine that the USB key sounds superior to the cds. It doesn't. It is exactly the same, except for the aforementioned noise floor. The only difference is the playback equipment for each. If you do the above experiment using the usb key and the actual cds, you get the same result (although it is a lot more fiddly to line them up correctly).

Let's not start discussing the maximum frequency response of the human ear!
(Or of dogs, and the Sgt Pepper inner groove!)

I can hear up to 17khz still. I'm very conscious of my hearing because of what I do here. But hearing varies greatly depending on your lifestyle, and the damage doesn't show up for decades. Anyone who regularly attended loud rock concerts in their youth will have some appreciable hearing loss now, but it usually isn't too bad. If you were in a touring rock band then you will basically end up with the hearing of a 90 year old by the time your 50. George Martin said in one of his books that he could no longer hear past 10khz by the 1980s - thanks to decades of high volume playbacks in the control room. Probably half the crowd who are into hirez audio can't hear anything past 12khz.

The other day my old dad was having a hearing aid put in, and I asked the boffin programming it to show me the graph from his hearing test. I noticed straight away that if you were to invert the graph it would closely resemble the kind of eq curve applied to the current Apple remixes. Those mixes sound excruciatingly bright to me, and I've also noticed that the younger crowd also find them to be excessively bright. Is this a result of them being auditioned by Ringo and Paul, who are in effect asking for a deaf-aid curve to be applied so that they can hear them clearly? Food for thought.

OK, so for downsampling LP rips, LR recommends 48khz (on the back half, due to the clean math/division). What about the front half (16 or 24?).

First let me say be real careful about what you use to do the downsampling! There's some shit software out there. I have only done the odd downsample because I don't have that many vinyl rips and I used Audacity (with all the resampling options set to highest in settings - not the default!) but that would be too cumbersome for large batches of files. So you need to do some research before starting on that.

I would keep the bit depth at 24, because it's only a 50% increase in filesize and it's better to retain a professional format for archiving. But for listening on your phone or whatever... for sure make a 16 bit copy and you will not hear any difference whatsoever unless your in a dead quiet room with the volume turned up to the level of a 747 taking off. In that case you might hear a faint bit more hiss during the final second of A Day In The Life, but your speakers would have exploded by then anyway.

Most of the vinyl I've had pressed has varied in terms of what the record cutter has asked for. Last two were 44.1/24 and they're fine. The mastering guys all want 96. When digitising tape that's a different ballgame as you have no real limitations other than what is on the source. Pressing vinyl is rife with limitations.

Yeah I'm all for 96k on digital files as there are no mechanical limitations. But the stuff about cutting lathes not accepting anything supsersonic was news to me. I assumed like most people that there was no limit. But apparently a stray supersonic audio signal at say 60khz will fry a cutting head.