Tidal and Watermarking (and MQA) ?

[mav52]

record labels were watermarking back in 2000. A few interesting reads

 

https://www.washingtonpost.com/archive/business/technology/2000/08/18/record-labels-try-to-remix-the-cd/9a0c8b8b-4a02-4de7-b99b-9a583ab2a636/?utm_term=.02823b144e34

 

https://www.wired.com/2007/08/listeningpost-0820/

 

https://www.wired.com/2008/01/drm-is-dead-but-watermarks-rise-from-its-ashes/

[PeterSt]

16 minutes ago, left channel said:

@PeterSt listen to the samples at this link provided by @Don Hills: http://www.mattmontag.com/music/universals-audible-watermark

 

Oh yes,  I know. And I did. But with that I don't know where it's all buried in. Unless you want me to recognize it all by means of this nature of sound. 

Thanks !

[left channel]

@PeterSt if you can hear it and you can use free spectrogram software to see it, what more do you need?

[PeterSt]

32 minutes ago, left channel said:

if you can hear it and you can use free spectrogram software to see it, what more do you need?

 

It is true that I loaded that into WaveLab a few hours ago, with some hope it could be recognized and from there proceed. I got distracted (shopping) and ...

And you know ... I forgot about it. 

Thank you again and more.

[Pure Vinyl Club]

On 23.11.2017 at 7:53 PM, PeterSt said:

But maybe when compared with vinyl or something but then still (digital digs out quite a bit more from a recording than LP ever can).

 

Thanks,

Peter

 

Just listen to these samples for comparison:
Official release on Acoustic Sounds '2016 (24/192):
 http://store.acousticsounds.com/d/114115/Harry_Belafonte-Belafonte_Sings_the_Blues-FLAC_192kHz24bit_Download 

VS 
LP rip (24/192) (LP production 1958!): https://www.discogs.com/Harry-Belafonte-Belafonte-Sings-The-Blues/release/1520512

here listen to vocal (sibilants and speech articulation):

07B02 - God Bless' The Child (Test, OFF).aif

08B02 - God Bless' The Child (Test, 9.52).aif

here listen to sax in right channel:

13B05 - Fare Thee Well-2 (Test, OFF).aif

14B05 - Fare Thee Well-2 (Test, 9.52).aif

DR (Off release): foo_dr (OFF).txt

DR (LP rip): foo_dr (LP rip).txt

[PeterSt]

3 hours ago, Pure Vinyl Club said:

here listen to vocal (sibilants and speech articulation):

07B02 - God Bless' The Child (Test, OFF).aif

08B02 - God Bless' The Child (Test, 9.52).aif

here listen to sax in right channel:

13B05 - Fare Thee Well-2 (Test, OFF).aif

14B05 - Fare Thee Well-2 (Test, 9.52).aif

 

Well, we have a few problems here, besides it seems to be off topic. It is NOT when you meant to showcase Watermarking. In that cae - apologies in advance (OK, sort of ).

 

First problem is that you asked me. Wrong pick. Haha.

Second problem is that I don't know what you are asking, except to compare - I suppose.

 

So I did. And only with what you provided as shown in the quote above.

Oh boy.

 

The 0702B : failed completely because colored as hell. Seriously. Unlistenable for that reason already (read on, in case you think "yeah wel, 1958").

 

The 08B02 : failed even more because of all the noise. Not sure whether this is meant to be the LP rip version, but better don't ...

It just as well can be a denoiser.

 

In both above you possibly think you hear brushes, right ? sorry, but it is noise only. NOISE.

 

13B05 : Not sure what to think of it, except that it it tolerable (for its age). But really nothing more.

 

14B05 : I really have no clue what you intend with this, but hear from me that it is severe distortion only. I did not capitalize it this time, but emphasized it. If you intended to lay out the square sax from the right speaker then ... I hope for you you didn't. Haha. The whole thing, everywhere throughout is distorted and if you don't hear that you have a serious problem of some kind. So let's hope that you wanted to point out the distortion. Of what ? the CD version ? I can't tell with all tracks at 24/192. All I know is that the 07B02 exhibits ticks which seem from vinyl and that IIRC the 14B05 compresses 6dB less (so theoretically better) but which is the worst.

 

With the notice that both (all) versions can only have been presented from some "self" of other external processing (no 24/192 exists anywhere from this unless faked / destroyed / etc.), :

 

My own 16/44.1 sounds perfect all the way. No hollow sound. No coloring anywhere and BRUSHES. Real ones.

N.b.: I have a similar ultimately good one (1959 or 1960 - I forgot) from Bill Evans. And we all spent a 300+ page topic on it (where to get this particular version).

 

Right.

So this was my unknown task. I hope you are now going to tell me what the real idea about this all was. Please.

 

Peter

[RichardSF]

A thought occurred to me, that since MQA encoded recordings already have embedded identifiers, and MQA claims to preserve integrity of the master recording, then surely they should not add the recording label's watermark on top of everything?

 

To test this idea, I listened to some albums on Tidal where both MQA and non-MQA versions of the same album are available (actually this is common – when Tidal adds an MQA album, they do not remove the older non-MQA version). In the examples I listened to, the UMG watermark is clearly audible in the non-MQA version, but the MQA version doesn't seem to have the watermark.

 

This has some important implications. Some people have been saying that they find MQA recordings to sound better, while others counter that the MQA recordings might have been produced from better quality masters. I will now suggest another possibility, that some MQA recordings may actually come from the same masters but they sound better not because of MQA encoding, but because they don't have the music label's audible watermark.

 

This would be a side benefit of MQA, completely unrelated to its encoding technology. It opens up another topic for discussion, that MQA technology itself might offer little benefit and the real gain comes from eliminating the watermark.

 

I'm interested to hear what all of you think of these ideas.

 

Here are some of my examples. I focused on a case where the watermark produces especially bad distortion — long sustained piano notes. The watermark produces a fluttering, warbling effect. Some music that came to mind for sustained piano notes are Chopin's funeral march (3rd movement of Piano Sonata No. 2), and the beginning of Keith Jarrett's Köln Concert.

 

Searching for the Chopin sonata on Tidal, I found two MQA recordings, by Martha Argerich and Seong-Jin Cho, both on Deutsche Grammophon. The funeral march is on tracks 29 and 28, respectively, of these albums. The Köln Concert is on ECM, and Tidal has an MQA version. Tidal also has non-MQA versions of these three albums. Deutsche Grammophon is part of UMG, and ECM is distributed by UMG. These recordings would be expected to have UMG's watermark.

 

Listening to the non-MQA versions, the characteristic distortion of the UMG watermark is clearly audible in all three albums. The distortion is especially unpleasant in the Argerich recording and the track has terrible sound.

 

Listening to the MQA versions of these albums, the watermark distortion is gone! The Argerich recording is wonderful! I concluded, at least for this sample, that these MQA recordings don't have the awful UMG watermark.

 

Here are the tracks mentioned. The quickest way to compare for yourself is the following. Open tidal.com in a web browser and sign in. Open the links and play the tracks. If you try to open a track but the player is stuck on the previous track, clear the play queue and try again.

 

Argerich non-MQA: https://listen.tidal.com/track/4792459 (track 29)
Argerich MQA: https://listen.tidal.com/track/77599620 (track 29)

 

Cho non-MQA: https://listen.tidal.com/track/53068878 (track 28)
Cho MQA: https://listen.tidal.com/track/77599024 (track 28)

Jarrett non-MQA: https://listen.tidal.com/track/10843698 (track 1)
Jarrett MQA: https://listen.tidal.com/track/77620366 (track 1)
 

[asdf1000]

3 hours ago, RichardSF said:

Jarrett non-MQA: https://listen.tidal.com/track/10843698 (track 1)
Jarrett MQA: https://listen.tidal.com/track/77620366 (track 1)

 

Hi, for the Jarrett comparisons can you share at which time (approx) in track 1 you hear this fluttering?

 

[PeterSt]

8 hours ago, RichardSF said:

A thought occurred to me, that since MQA encoded recordings already have embedded identifiers, and MQA claims to preserve integrity of the master recording, then surely they should not add the recording label's watermark on top of everything?

 

Hi Richard - Maybe it wasn't the mos obvious in the OP but this was exactly my own idea about the matter :

 

On 23-11-2017 at 5:53 AM, PeterSt said:

The other two we can compare to see whether the phenomenon is present in the MQA version of it as well.

 

So, great that you took the effort and came up with examples (I did not have any yet). I will listen to them later today and I will try to perceive differences to this regard.

 

8 hours ago, RichardSF said:

I will now suggest another possibility, that some MQA recordings may actually come from the same masters but they sound better not because of MQA encoding, but because they don't have the music label's audible watermark.

 

IMO this goes one step too far. Why ? well, because I mostly tend to compare MQA in absolute sense. Thus, for example I have a Doors album which I know for my lifetime and which I play regularly, and then suddenly there's the MQA version of it, I play that and next it comes to me that it comes across completely different. What is in the equation is that I hardly ever play any non-MQA from Tidal, just because I already have it locally. Counts for e.g. Keith Jarrett just the same (not the other two you mentioned) so for the occasion I will of course play the non-MQA from Tidal as well. But actually the comparison needs to start with the local Keith Jarrett vs. the Tidal non-MQA one.

 

8 hours ago, RichardSF said:

Searching for the Chopin sonata on Tidal, I found two MQA recordings, by Martha Argerich and Seong-Jin Cho, both on Deutsche Grammophon.

 

If you are interested, a few more of her are there :

 

 

 

[RichardSF]

18 hours ago, Em2016 said:

Hi, for the Jarrett comparisons can you share at which time (approx) in track 1 you hear this fluttering?

 

The watermark distortion starts from the beginning of the track, and you can listen for it especially during the first couple of minutes while Jarrett is playing slowly. After a piano note is struck, listen as it sustains and fades away. There is a bit of what can be described as a flutter, buzz, or modulation where it should be a smooth even tone. The distortion isn't as obvious on this album as on some others, but the character of the distortion is recognizable.

 

When comparing the MQA and non-MQA tracks, you can switch quickly between them if you have both of them open at the same time in separate tabs in the browser.

[AudGuy]

Very interesting topic, and more than a little disconcerting also. Sort of explains a lot, in an indirect fashion.

Listening to Tidal Hi-Fi ... I find I cannot listen seriously, but it does make for pretty good casual music.

But, MQA seems better, although the choices need expanding, and something is certainly less irritating in playback.

 

I do find though that for me, Tidal is an exploration tool that can ultimately lead to the purchase of music I enjoy.

 

However, I am also finding that I am drifting back towards spinning more vinyl, something more analog and less digital.

Not without it's share of problems records can be awful, and they can be superb and I think lacking any watermarking.

 

Now that this little bug of an idea has been planted, it will be quite hard to not listen to everything with a jaundiced ear.

[Steenbz]

If Tidal finally succeeds in getting all the tracks from UMG without the watermark artifacts through the MQA versions, we shall finally after long long waiting have a proper alternative to CDs. I have contacted their support several times over the last year, and they’ve kept on promising that soon the watermarked titles would be substituted but it is being done so slowly with the bad versions still there even when the non watermarked ones are in place. 

Hopefully they will find a way, too, to make track transitions without pause (especially important when listening to opera) when playing over Apple airplay and from the new Apple TV 4K app (as the quality inferior Apple Music has been able to for years), and not just when playing from computer. 

Hoping for the best!

steen

[Samuel T Cogley]

25 minutes ago, Steenbz said:

If Tidal finally succeeds in getting all the tracks from UMG without the watermark artifacts through the MQA versions, we shall finally after long long waiting have a proper alternative to CDs.

 

I will begrudgingly agree that MQA is a more palatable alternative DRM than audible watermarking.  The watermarked classical titles on Tidal are unlistenable to me.  I'm thinking the reason is headphones is my primary listening environment, and the watermarking is quite noticeable with headphones.