Record Label Honesty In the Download Era

[acousticsguru]

Nobody was trying to tell the difference between analog or digital master that has been used. People were and still are trying to get the best possible pressing and listen to music.

MOFI used digital copies of the master tapes and lied to customers. The thing is not primarily about analog or digital but cutting from a copy of the master tape. In this case digital copy which is adding an additional A/D and D/A conversions and some computer processes in the chain. If you can not cut from the master tape let other companies who can.

It’s not a matter of being able to. First-generation master tapes never leave the premises anymore, not at major labels. Archive vaults provide flat transfers to mastering studios. Exceptions apply where a company like Universal employ their own cutting engineer, e.g. at Emil Berliner Studios, to avoid the insurance nightmare of originals traveling by courier. Those have become rare.

Greetings from Switzerland, David.

 

[mtemur]

First-generation master tapes never leave the premises.

No, they leave depending on the mastering/cutting facility and reissue label.

 

[acousticsguru]

No, they leave depending on the mastering/cutting facility and reissue label.

Majors may provide facilities and personnel, and/or hire or employ mastering and cutting engineers. The bottom line is: as long as original tapes don’t travel, are being checked/restored under supervison, and aren’t played back more often than absolutely necessary (which in a real world scenario usually means: to make a flat transfer for the mastering engineer to work with).

Greetings from Switzerland, David.

 

[acousticsguru]

No, they leave depending on the mastering/cutting facility and reissue label.

And yes, I understand that audiophiles find it counter-intuitive that flat transfers may henceforth be referred to as “master”, but that’s what engineers consider a flat transfer to be: a straight copy.

Greetings from Switzerland, David.

 

[acousticsguru]

No, they leave depending on the mastering/cutting facility and reissue label.

Besides, common sense would have it that audiophiles favor the best available (in terms of sound quality) source, such as in the case of Solti’s Ring the Japanese production master and not the original session and master tapes that have oxides flaking off.

MoFiGate is/was about false advertising, which incidentally is what this thread originally is/was about.

Great remastering is about making the right decisions, sonically and artistically. In an ideal world, we’d have transparency. But the problem appears to be, in part, that the purists prefer to be lied to, whereas audiophiles are primarily concerned about sound quality.

Greetings from Switzerland, David.

 

[Rt66indierock]

Does anyone have a pertinent response to pleroma's thoughtful and well-evidenced opening post?

Challenge accepted.

Classical download files need to be tested by buyers.

A database needs to be created so the results of buyers’ tests can be easily uploaded and shared.

 

[Moladiego]

In the era of high-resolution digital music downloads, classical music collectors are often confronted with a confusing array of formats and sampling rates within formats. One eventually learns about the different sampling rates available for PCM releases, and does his best to make an informed decision about which one has the best sound quality for his requirements. For legacy classical releases which have been marketed as "remastered" at a certain bit-depth and sampling rate, one generally expects his purchase to reflect what was advertised. Unfortunately this may not always be the case.

Are some record labels essentially repackaging redbook-CD-resolution data as high-resolution? A friend of mine and I have consistently raised this question about a number of recent digital releases from "Alexander Bak Reference Recordings" and "Praga Digitals" on Qobuz. But most recently we may have discovered an Australian affiliate of music titan Universal doing the same. I make no accusation here, but simply lay out our analysis for you to decide.

Here is the spectrogram of the first track from Henryk Szeryng's Beethoven Violin Concerto, recently released from Universal Music Australia on Qobuz. You can see the truncation of the Nyquist frequency between 22kHz (RBCD) and 24kHz:

View attachment 116988

Beethoven: Violin Concerto (Hans Schmidt-Isserstedt Edition 2, Vol. 1), Ludwig van Beethoven by Henryk Szeryng - Qobuz

Some might argue this is could be 24/44.1kHz data, but then why use 48kHz? Attempts to contact Universal Australia for inquiry have been unsuccessful. We did contact Alexander Bak Reference Recordings, who chose not to disclose their sources (though larger labels routinely state if original sources were used). At best this is misleading, as Universal has the resources to provide true high-resolution music. We've found other questionable examples from Universal Music in Japan, who distribute some 24/192 releases which appear to have been downsampled from DSD (SACD) releases, when the mastering studio uses 24/192 for the working master (in the interest of keeping this article brief, I've omitted this spectrogram with tell-tale, out-of-band noise from a recent purchase of mine).

For collectors who've bought the same releases multiple times over the decades, and believe these distinctions are significant and audible, one solution is to buy only a track and analyze it with the free Audacity editor. Ideally, labels would disclose what they're doing.

I came across a few of "Alexander Bak" remastering releases (24.96) (e.g., Wagner's Lohengrin (Solti/VPO) and Kemper/VPO's Wagner's Preludes) - they are uniformly awful, worse than original CD releases. Think they are fraudulent.

 

[Mcsnare]

If it really does sound different, there’s a chance it isn’t just a resampling, but that some sort of equalization or processing was applied. If so, this might be considered added value. One could for example improve on the sound of lesser quality RBCD files running them through e.g. Daniel Hertz Masterclass, tastefully using the five-band equalizer and applying A+ (similarity to Burwen Bobcat). There are countless tools out there, Audacity which I’m assuming you used to have a look at your file being one of them. One would need to do a so-called null test (Audacity includes that function: invert / combine / render into a new file) if in this instance you have the original CD ripped.

In other words, the fact alone that an RBCD file was resampled, i.e. rendered and saved into a larger file container, isn’t sufficient information to determine if one is looking at a bogus offering. Note I hesitate referring to such a file as an “upsampling” (versus a resampling) without determining whether the tool in question is using interpolation filters (think of a film as a succession of snapshots, multiplying them with or without interpolation software).

Years ago there was an uproar against HDtracks offering resampling files, forcing them to have a second look at their catalogue. Some of their (later) competition such as HiResAudio will check all and flat out refuse files from labels when they find them to different from advertised, possibly because in Germany the legal situation is different (I know in Switzerland repackaging / renaming / relabeling for profit is a criminal offense).

Other sites such as Qobuz appear to leave this duty (of checking file content) to their customers, which I think is bad practice (and which may point to a different situation legally, presumably referring back to the source and claiming they’re the victim).

There are worse problems in terms of diminishing the sound quality of commercial high-resolution releases, e.g. “Mastered for iTunes”: what it means is the dynamic compression level is tailored so a (re-)mastering will sound “best” (they really believe that) on an MP3 (mobile) device. At that point it no longer matters if a master with a compression level of e.g. 10 and below (check and download tool at Dynamic Range DataBase / Google “Loudness War”) is simultaneously being marketed in a high-resolution format (e.g. 24/96 PCM being the most popular for years, increasingly replaced by the more popular 24/192 - with absolutely no difference in content), yet one must wonder if the practice can be considered illegal (it could, in Switzerland, if a site sold an overtly identical product, e.g. in 24/96 and 24/192 file containers at different price levels, but not if the price difference were realistic, e.g. based on additional cloud storage cost).

The irony is that the tools are available online and for free.

Greetings from Switzerland, David.

Mastered for iTunes is the result of producer Rick Rubin noticing that his productions (more obviously than the full rez file) distorted when played back in mp3 format. This is because the data reduction of the mp3 codec results in a higher level and more intersample peaks. The answer was to render full resolution wav files at a lower peak level based on using an Apple app during the mastering process that tells the enginner exactly what the number of overshoots and clipped sample will occur when data reduced during the making of a 256kbs mp3. The Apple spec for a MFiT file is at least 24bit, 44.1K (Apple encourages higher sample rates when possible) and no or few, post mp3 conversion, clipped samples. It is simply a peak level spec to give more headroom for the dac.

 

[Mcsnare]

never tried that compare. but never liked a single MoFi record.

I feel the same way, altough I do think most of the Stan Ricker era releases are excellent,

My dislike of MoFi vinyl is all about what I consider to be poor eq choises and a poorly curated, un-musical sounding, analog signal path.
Using a carefully done DSD capture of the tape with high quality gear IMO is essentially a transparent step that is completely swamped by the sonics that result from the lacquer cutting proceess. Yes it was disingenuous, but no, its not responsible for what I consider to be the poor sound quality of 90% of the post-Ricker Mofi offererings.

 

[Mcsnare]

In the era of high-resolution digital music downloads, classical music collectors are often confronted with a confusing array of formats and sampling rates within formats. One eventually learns about the different sampling rates available for PCM releases, and does his best to make an informed decision about which one has the best sound quality for his requirements. For legacy classical releases which have been marketed as "remastered" at a certain bit-depth and sampling rate, one generally expects his purchase to reflect what was advertised. Unfortunately this may not always be the case.

Are some record labels essentially repackaging redbook-CD-resolution data as high-resolution? A friend of mine and I have consistently raised this question about a number of recent digital releases from "Alexander Bak Reference Recordings" and "Praga Digitals" on Qobuz. But most recently we may have discovered an Australian affiliate of music titan Universal doing the same. I make no accusation here, but simply lay out our analysis for you to decide.

Here is the spectrogram of the first track from Henryk Szeryng's Beethoven Violin Concerto, recently released from Universal Music Australia on Qobuz. You can see the truncation of the Nyquist frequency between 22kHz (RBCD) and 24kHz:

View attachment 116988

Beethoven: Violin Concerto (Hans Schmidt-Isserstedt Edition 2, Vol. 1), Ludwig van Beethoven by Henryk Szeryng - Qobuz

Some might argue this is could be 24/44.1kHz data, but then why use 48kHz? Attempts to contact Universal Australia for inquiry have been unsuccessful. We did contact Alexander Bak Reference Recordings, who chose not to disclose their sources (though larger labels routinely state if original sources were used). At best this is misleading, as Universal has the resources to provide true high-resolution music. We've found other questionable examples from Universal Music in Japan, who distribute some 24/192 releases which appear to have been downsampled from DSD (SACD) releases, when the mastering studio uses 24/192 for the working master (in the interest of keeping this article brief, I've omitted this spectrogram with tell-tale, out-of-band noise from a recent purchase of mine).

For collectors who've bought the same releases multiple times over the decades, and believe these distinctions are significant and audible, one solution is to buy only a track and analyze it with the free Audacity editor. Ideally, labels would disclose what they're doing.

Up until about 2010, the vast majority of mastering engineers capured any eq, level adjustments, and other processing at 24 or 16bit 44.1K. This was so an additional step of sample rate conversion did not have to be used owing to the CD standard. It didnt make sense to play mixes whether off tape or digital and record at some high sample rate, only to have to use another sonically diminishing process to get it to 16/44.1
There is so little money to made in classical releases, that for a lable to spend for a complete remaster simply to get a higher-tha-CD- rez file (to sell to an even smaller group of consumers) doesnt make economic sense.
It might sound marginally better but the lables aren't in the business of losing money to do a complete remaster cause 5000 people want a past classical release in 24/192.
Its a slightly different matter in pop/rock although I have my doubts that Peter Gabriel's - So was originally mastered at 24/192. Maybe the lable paid for a redo. Maybe not.

 

[acousticsguru]

Mastered for iTunes is the result of producer Rick Rubin noticing that his productions (more obviously than the full rez file) distorted when played back in mp3 format. This is because the data reduction of the mp3 codec results in a higher level and more intersample peaks. The answer was to render full resolution wav files at a lower peak level based on using an Apple app during the mastering process that tells the enginner exactly what the number of overshoots and clipped sample will occur when data reduced during the making of a 256kbs mp3. The Apple spec for a MFiT file is at least 24bit, 44.1K (Apple encourages higher sample rates when possible) and no or few, post mp3 conversion, clipped samples. It is simply a peak level spec to give more headroom for the dac.

Of course, but the only question that is relevant here is this: is it or is it not fraudulent labeling to sell the exact same compression level digital file in different file format containers at different price levels? In Switzerland at least, a well-known supermarket chain had to retract products when it turned out (as an independent laboratory proved and reported to the price supervisor) they sold the exact same content (in this instance it may have been foodstuff, but the law is the same for any article) in several product lines at different prices. We have a term for this: "moon pricing" (think of offering an article at an unrealistic price at a discount, same thing). So yes, you might find it understandable labels/remastering engineers don't want to do a separate high-resolution remastering of e.g. classical music expecting to sell only a small number of copies - but that is their problem, not the buyer's (who's purposely led to believe the upcharge for the more expensive high-resolution release over e.g. the cheaper MP3 is indicative of a qualitative difference). Provided we agree on legislature and regulation being at all necessary, sales offers must be neither counter-intuitive nor misleading.

Greetings from Switzerland, David

 

[Mcsnare]

Of course, but the only question that is relevant here is this: is it or is it not fraudulent labeling to sell the exact same compression level digital file in different file format containers at different price levels? In Switzerland at least, a well-known supermarket chain had to retract products when it turned out (as an independent laboratory proved and reported to the price supervisor) they sold the exact same content (in this instance it may have been foodstuff, but the law is the same for any article) in several product lines at different prices. We have a term for this: "moon pricing" (think of offering an article at an unrealistic price at a discount, same thing). So yes, you might find it understandable labels/remastering engineers don't want to do a separate high-resolution remastering of e.g. classical music expecting to sell only a small number of copies - but that is their problem, not the buyer's (who's purposely led to believe the upcharge for the more expensive high-resolution release over e.g. the cheaper MP3 is indicative of a qualitative difference). Provided we agree on legislature and regulation being at all necessary, sales offers must be neither counter-intuitive nor misleading.

Greetings from Switzerland, David

100% agree.

 

[PYP]

Challenge accepted.

Classical download files need to be tested by buyers.

A database needs to be created so the results of buyers’ tests can be easily uploaded and shared.

Good suggestion.

I would think that the major focus for labels now is the potentially bigger lack of transparency that occurs by AI-generated music. How do the labels embrace the technology while preserving their business and the rights of musicians? How does the consumer know if a very good imitation of their favorite artist is "real" or not? What is owed the musicians that the AI is trained on? Probably for another thread if it hasn't already occurred.

 

[Mcsnare]

Good suggestion.

I would think that the major focus for labels now is the potentially bigger lack of transparency that occurs by AI-generated music. How do the labels embrace the technology while preserving their business and the rights of musicians? How does the consumer know if a very good imitation of their favorite artist is "real" or not? What is owed the musicians that the AI is trained on? Probably for another thread if it hasn't already occurred.

You have the major lables confused with people who care about music and artists. The labels will be the ones to release AI generated content of artists under contract to them.

 

[PYP]

You have the major lables confused with people who care about music and artists. The labels will be the ones to release AI generated content of artists under contract to them.

@Mcsnare You are in the business, so have an insiders view that most of us do not. I was indeed thinking of the majors, but also the independent labels. As I listen to a lot of music via Roon/Qobuz, I am often surprised by the excellent quality of the recording (mostly, fairly recent recordings). Then I get curious about the label and often find names that are completely unknown to me. I realize that some of these are actually parts of the major labels or partially owned by them or the major labels distribute for them. The sheer number, however, is a surprise for a music lover who knows very little about the process of making music in today's marketplace. I'm assuming that independent labels are closer to the artist's intent and will be concerned about their well-being.

Perhaps I'm surrendering to the buzz around AI and thinking the music business could change radically and very quickly.

 

[Mcsnare]

@Mcsnare You are in the business, so have an insiders view that most of us do not. I was indeed thinking of the majors, but also the independent labels. As I listen to a lot of music via Roon/Qobuz, I am often surprised by the excellent quality of the recording (mostly, fairly recent recordings). Then I get curious about the label and often find names that are completely unknown to me. I realize that some of these are actually parts of the major labels or partially owned by them or the major labels distribute for them. The sheer number, however, is a surprise for a music lover who knows very little about the process of making music in today's marketplace. I'm assuming that independent labels are closer to the artist's intent and will be concerned about their well-being.

Perhaps I'm surrendering to the buzz around AI and thinking the music business could change radically and very quickly.

Two things:
The consolidation in the nineties and beyond has resulted in there only being 3 major labels, with UMG being the largest and most significant player. The major labels no longer make money from selling music, but rather from licensing, publishing, owning a large piece of Spotify, and as of the early 2000’s a percent of every revenue stream that was traditionally the artists. Concert tickets, T shirt sales, etc.
Furthermore, they are very open about the fact they no longer do artist development. So the fact they only sign artists that have gotten a lot of traction on their own, means very little of quality is now on the major labels. Many artists totally unknown to anyone but their fan base can make a lot more money NOT signing away everything to UMG. There are of course exceptions.
Truly independent labels have basically no market penetration and are akin to clueless audiophiles thinking that small boutique hifi manufacturers actually make money!

AI will absolutely change the game but only at thr major label level as they will generate more profit by emulating artists they own the rights to without having to deal with pesky artists and their desire to get paid.
I fully expect going forward that major label contracts will be negotiated with an AI yes or no clause.