[Golden Ring, Happy Records] Incognito Five - "Invitation", "Music in the Ear"

[John_Fred]

Many thanks John Fred again! But you don't have "Happy Records - HR 2231 - Music in the Ear"?

Sorry, I thought you only wanted the Invitation title.  Here's the other one:-
...

Please note that this rip of "Music in the ear" is a fake.

As in?

[likedeeler]

...
As in?

I do not understand the question. Could you clarify?

[nidostar]

...
As in?

I do not understand the question.

I believe John_Fred is asking why you think Music In The Ear is a fake. I'd also be interested to know.

[likedeeler]

...
I believe John_Fred is asking why you think Music In The Ear is a fake. I'd also be interested to know.

Ah! Thanks for jumping in. Just take a look at spectral analyses. I've deleted the download and cannot describe the clues in detail anymore. But it should be fairly obvious. These FLAC files were transcoded from lossy sources.

"Invitation from Incognito Five" on the other hand is legit. Compare their spectrals with those of "Music in the Ear", and you'll see the difference immediately.

(For spectral analyses I recommend Spek.)

[likedeeler]

Let me perhaps add that the aspect of the veracity of a sound file in a lossless format is (almost) completely independent of sound quality. Lossless formats only have the potential to attain a quality that is unattainable with lossy formats. How they really sound on a capable reproduction system is another question.

Therefore it is imprecise to say of a 16-bit, 44.1 kHz FLAC file that it has a "higher quality" than an MP3 file, or of a 24-bit, 96 kHz FLAC file that it has a "higher quality" than a normal FLAC or a CD rip. Too often, people speak like that, also here in this forum. Please remember: that alleged higher quality -- which can only mean sound quality -- exists foremost as a potential that needs to be realised to be of any value.
 
It is very easy to produce a truly lossless 24-bit, 192-kHz rip of a vinyl record that does no justice to the actual recording. A different rip, even when it's an MP3 or a transcoded MP3, may sound much better. There are a lot of lossless vinyl rips around that sound mediocre to abysmal, despite not being lossy transcodings.

Even if the record to convert to a digital file is in a good state (and I don't mean the record-grading "good", which rather stands for "don't even think about buying this") -- which regularly is not the case --, and even if the record is properly washed -- which regularly isn't done --, and even if the ripper has a decent record player with a good pick-up and a high-quality phono preamp (and not some USB-whatever shit) -- which regularly is not the case --, and even if he does not use a crappy hobby-DJ mixer -- which he will usually do --, and even if he has a hi-fi soundcard or (preferrably) a high-quality external A/D converter -- which regularly isn't the case --, even then ...

... he still has to resist the urge to put the result of the ripping process through automatic de-clickers and de-noisers. But many people don't resist, and then it all may have been for the birds in the end.

So, that much-loved "lossless upgrade" could well be a downgrade, soundwise.

Do note that I'm not saying the faked files of Music in the Ear sound bad -- I did not listen to them before I threw them away. All I'm saying is that they are not actually lossless.

[nidostar]

Thanks for the clarification likedeeler. I have run each track through Spek but whereas a lossy file can often show an obvious frequency roll-off at 16kHz all these tracks appear to show frequencies rising up to 24kHz. I think it unlikely but could it be that my converting the FLAC file to ALAC has anything to do with that? Otherwise, and I don't doubt your suspicions, to my relatively untrained eye I can't see anything on the Spek graph to suggest the file is originally from a lossy version.
As you say, on occasion I have found a FLAC rip to appear muddy when compared to a lossy version I am looking to replace. It all depends on the equipment used for ripping, the amount of click and noise suppression used and other factors including the quality of the playback equipment. Hence your point that "a much-loved "lossless upgrade" may well be a downgrade, soundwise". I believe I can notice the improvement in quality between FLAC and MP3 from a CD rip. In another thread today Retronic referred to compilations he made back in 2010 with MP3s at 128kbps. It prompted me to listen to one of them and even though I know the sound is compressed it is still acceptable to my ear when listening through speakers. And believe me I have a good musical ear. Which raises the question whether our obsession with FLAC rips is justified. But that may be a discussion for a different thread rather than here.

[likedeeler]

Thanks for the feedback, nidostar.

I've downloaded the offending files again and looked at the issue a bit more closely. I will qualify my previous assessment of their content having been subjected to lossy compression.

MP3 does indeed only go up to 20 kHz and a bit, so that speaks against a direct transcoding (I converted the tracks from FLAC to MP3 to make sure). The spectrals look as if the files may have been ripped from a record which itself was produced from MP3, because record players are music instruments and add frequencies of their own that are not present in the grooves.

Having been produced from MP3 sources, however, isn't possible for a record from 1981 -- whereas it is being done nowadays. But I maintain that something is wrong with these files. Consider the apparent cut-offs at 16, 20 and 21 kHz. This simply cannot happen with a record produced at a professional studio -- which in 1981 was fully analogue -- if it's ripped properly.

There are many potential causes that may cause these cut-offs. One possibility could be a crappy mixer or some other component in the signal path that acts as a (partial) low-pass filter. But then again, spectrals of rips produced with crappy mixers that I have seen so far are more likely to have certain frequencies or frequency bands suppressed or attenuated, which turn up as darkened horizontal stripes in the spectral analysis -- if you're used to checking spectrals you'll be familiar with that phenomenon. But this looks a bit different.

Maybe it is a rip from minidisc. I have never seen a spectral analysis of a minidisc file, but they have their own lossy compression algorithm which may result in such a spectrum. Would be interesting to know.

My spontaneous, generic guess at the moment is that the ripping setup was somehow bad or that it's from minidisc. I maintain the statement that these files cannot be the result of a proper rip taken from a real record or master recording.

And, no, converting files from one lossless format to another does not cause such artefacts. You can repeat these conversions as long as you like without any changes to the musical content.

(If you read this and have minidisc, it would be great if you took a few spectrals and let us know what they look like! My parents have minidisc, so when I visit them next time I may be able to bring samples. But it will take a while. Lamentably I don't have a player myself.)

[likedeeler]

PS >> This is getting more interesting by the minute.

So, now. Thing is I had overlooked the catalogue number and the fact that the cover picture is not identical with the 1981 record cover. Turns out there exists a 1993 CD reissue with exactly the cover picture of the rip in question and the catalogue number that is contained in the folder name.

But then again, the files have a sampling frequency of 48 kHz and the sound reaches 24 kHz. Which means they cannot come from a CD. Therefore, they likely are a digital download provided by Happy Music.

The publishers themselves fucked it up! First thing I should have suspected. It's such a standard: some library labels kept just lossy files of old productions and threw away their master tapes. There's a lot you can do with a recording studio to spruce up low-quality files when you regret having lost the originals. Using so-called exciters to add overtones and stuff is one option (see https://en.wikipedia.org/wiki/Exciter_(effect)).

It would be interesting to know if a rip taken directly from the CD reissue has the same spectral defects or not. Perhaps someone who owns this CD will enlighten us.

The sound is actually quite good, by the way. I think I'll keep the files this time. Thanks!

[likedeeler]

PPS >> I would encourage you to take a look at the infamous 'reissues' of I Marc 4 on Nelson. I have rips of GLP 1001, 1002, 1007, 1012 and 1013 (these are the catalogue numbers of the original records; the files will be digital downloads or ripped from CD reissues, published not by Nelson but by another label). Spectrals look fairly similar to those of Music in the Ear.

There is no way that these rips come from original records or properly produced files or CDs, much less from master tapes.

Same story, very probably: ditched masters and fabricated "reissues" from lossy or otherwise low-quality sources. Note that this is a guess. An informed one, but a guess.

[John_Fred]

Having read all of that, you may have a case that the recording in question doesn't fit the criteria of what's deemed to have a full-spectrum of frequencies associated with what we know as a lossless rip recording.  However, the wording used when claiming the rip in question is, 'fake', can often take on an accusatory tone, if not too careful!     But, as you go on to explain, the fault isn't with anyone here, but with the record company that produced the rendition of the original recording who may have omitted/neglected to do what us non-experts might call a less than, 'proper job' of transferring it to a CD or on-line media, thus having less than expected results when scrutinised via a spectrum analyser.  One can only speculate as to why. 

[likedeeler]

Hey John_Fred, I regularly let it be known when I download a "lossless" rip that bears obvious signs of lossy compression or similar tampering, or if files are corrupted. This is never meant as a complaint and much less as an accusation of the uploader. The large majority of shares here do not originate with the uploader anyway but were previously downloaded from somewhere. I myself am always grateful if someone tells me that an allegedly lossless rip he got from me isn't actually lossless, or that it is corrupted. "Fake" seems an apt term for the former. I hope you won't take it amiss if I keep using it.

What I said above about labels putting out fraudulent so-called reissues -- and it is a swindle plain and simple to represent them as "reissues" -- is indeed more than speculation. It is well known that this has happened and continues to happen. There are more than just a few (official!) LP and CD releases out there containing MP3 material. Let that sink in for a moment.

The reasons are obvious: vinyl craze, ditched or lost master tapes, laziness, and the profitable desire to mislead combined with the fact that people want to believe. The marked deterioration of domestic sound reproduction that has taken place over the past decades may also play a role. For many people the question whether to listen to a gramophone record, a 24-bit, 192 kHz lossless file or an MP3 is of no consequence -- even if they think it were. In reality it makes no difference as the sound coming out of their boxes or headphones is virtually the same.

[[(Sub)]]

My vinyl rip of HR 2231

pixeldrain.com/u/8n2tQw6z

[Ice8]

Many thanks [(Sub)] for the rescue of HR 2231!

[likedeeler]

My vinyl rip of HR 2231
...

Haha! What the fuck? I'm still loading, very much looking forward ... Cheers, mate!

[likedeeler]

Sounds great, [(Sub)]!

We could probably mark this up as a seminar on "What is lost in lossy compression?" or "Why fake 'lossless' rips should be rejected" (or lossy rips in general). At least for me this has been a very enlightening exercise. As I do not own ripping gear there is at the moment no convenient way for me to contrast a record I have with a rip or a digital download of it. So I've rarely made the comparison (I need to get up, start the record player, switch over the amp etc). With these two rips of Music in the Ear, the fake and the genuine one, I can perform a direct switch-over from the easy chair.

Night and day, is it not?

Particularly interesting is the fact that first listening to the fake version (which resulted from Happy Records trying to improve their lossy files by way of exciters and similar devices**) you think the sound is alright, perhaps even good. And then comes [(Sub)]'s rip and you realise that it's not good. At all.

(If the difference doesn't seem striking to you it means your stereo wants an upgrade.)
_______

** This will be the reason for the sound going up to 24 kHz in the spectral analysis of the fake. A plain, unaltered MP3 file does not exceed 21 kHz. The partial cut-offs at certain frequencies -- which are typical for lossy encoders -- have remained, however.