MQA The Truth lies Somewhere in the Middle

[Fokus]

14 hours ago, The Computer Audiophile said:

Without opening a huge can of worms, I'll ask the question - What do you think consumers should know?

 

They should know, without going into details, that there is a large amount of scientifically-founded critique on MQA's technology and on its alleged usefulness. And this critique comes from, not lab-coats, but people who actually care about sound, about music, about the industry.

 

They should be asked if they really want (to pay for) a locked-in format based on such a technology.

 

But you really don't want to go into technicalities, because that will go in no time down the rabbit hole, losing 99% of the audience in the process. It doesn't matter if it is 18 bits or not. It does not matter if it is lossy, lossless, or 'subjectively lossless'.

 

 

[Fokus]

10 hours ago, christopher3393 said:

My own consumer question is: what important statements about MQA by Bob Stuart are most misleading, and what would be a more transparent and moderate way of making these statements in plain english as much as possible.?, but my impression is that this could get ugly during Q+A, given the tensions that I think still exist.

 

I would say the unproven premise that filter ringing is significantly detrimental to subjective audio quality, and that it has to be battled at any cost.

 

And if it is really that important, where are the ear-opening before-after comparison tests and demos?

[Fokus]

TI, AD, et al. will happily include MQA if this nets them more sales. 

[Fokus]

We've been through this before. If not here then on another forum. What MQA call

'dispersion' is any widening of a channel's impulse response, be it linear phase

or not. They fight it by replacing all filters in the channel with slow roll-off filters,

in some cases even no filters, in order to obtain a more compact impulse response.

This happens at the cost of aliasing and imaging, something they claim to keep below levels

of audibility.

In principle they also provide for correction of ills of the original recording gear, but we all

know that this is not really done, and in the few cases where it is done, it amounts to

the sort of care anyone competent can (should?) apply in a restoration and remastering

project.

 

But this is off-topic and this discussion can better be held elsewhere.

[Fokus]

44 minutes ago, Shadorne said:

Sony tried to sell a new format SACD (dubious benefits),

 

When I wrote a critique on my then-website they were quick to send in the lawyers, have my provider shut down the site, and threaten me.

[Fokus]

4 hours ago, Lee Scoggins said:

 

The DRM code is not for DRM, it's for authentication so they can control the quality of the end product.  It certifies that the resulting file has had the correct MQA filters applied and has been approved by the label/producer/artist or some subset thereof.

 

The part of the encryption used for authentication is not the DRM we are pointing at, pay attention.

 

This said, the authentication scheme does not even work. It has been demonstrated here that an MQA file can be mutilated, with the Famous Blue LED still remaining lit.

 

(Apologies to Jenny and Lenny.)

 

[Fokus]

 

6 hours ago, Lee Scoggins said:

The unfolding is the clever aspect of Stuart's approach which enables the smaller file sizes.

 

Let's put this straight: the folding/unfolding is not a Stuart or MQA invention. There is a Japanese patent predating this by years.

 

And this too:  there are non-MQA approaches that similarly would reduce file sizes.

 

6 hours ago, Lee Scoggins said:

 You are essentially saying that any folding is DRM but the classic definition is that DRM controls user rights. 

 

The folding has nothing to do with DRM. It is mind-boggling that 4 years after MQA's birth you don't even grasp something this fundamental.

 

The DRM is in the system being entirely closed. The consumer never has access to the full unfolded and decrypted data. I cannot inspect an MQA file to see if it really is hi-res, rather than a fake upsample. My rights are being managed. I cannot fully decode an MQA file and then send it to some other part of my system for digital processing (such as a digital-active speaker). My rights are being managed. And MQA does not even sell a gizmo or software that would allow me to do these things. I only get invited to replace a lot of hardware with MQA-licensed hardware. My rights are being managed. Formats like FLAC and even DSD do not pose such restrictions.

 

6 hours ago, Lee Scoggins said:

2.  All the major labels and Merlin have committed to applying MQA to their entire back catalog.  I know from people working in the industry, that the MQA masterings are being done in volume. 

 

So much for the touted white-glove approach then.

 

 

 

(Is there a boldly banging on the table emoji?)

 

 

[Fokus]

12 minutes ago, mansr said:

So which label or production company does Derek work for?

You forgot 'artist'.

[Fokus]

9 minutes ago, mansr said:

Are you suggesting that the table-banging was his attempt at performing a Steve Reich piece?

 

It wasn't Cage's 4'33".

 

[Fokus]

11 hours ago, adamdea said:

incidentally can anyone (@Mansr?) tell me whether they have been able to reproduce archimago's findings with the hole in the spectrum for the bruno mars track?

 

Yes, totally, early in 2017.

[Fokus]

8 hours ago, adamdea said:

And the gap is because they were conventionally brickwalled at 20 kHz or so when originally produced? 

 

I would say unconventionally brickwalled, because most ADCs or SRC tools (used in default mode) would run up to 22.05kHz (and a bit beyond).

[Fokus]

38 minutes ago, firedog said:

We all understand that MQA files have no content above 48k and everything above that is upsampling.

 

No, that it not true at all.

 

MQA has the capability of passing a 2x file through a 1x channel more or less unscathed. This has been proven, no-one tries to deny that.

 

For a 4x original MQA passes the 2x part, and then extends back to 4x by upsampling, so that is part  fake, but in their mind it does not matter because only shallow filters were used, 'avoiding temporal distortion'.

 

For 1x originals MQA cannot do much to distinguish themselves, to they revert to very lazy, NOS-like upsampling for such files. This is what the spectral hole in Mars, Beyonce, ... is about. They call it master quality, we call it a fuzz box.

 

[Fokus]

8 minutes ago, firedog said:

 My understanding is that when MQA encodes a file that is higher res than a 24-48, it only encodes material up to 48k, and throws the higher frequencies away as "perceptually lossless".  That's one of the reasons it is "lossy". So for a 88, 96,176, 192,etc., it is discarding material above 48k if it is there. 

 

Up to 48k payload, that is, 96k sample rate.

[Fokus]

Consensus is agreement of opinion. Math exists outside of opinion. Outside of ego. Outside of identity.

[Fokus]

52 minutes ago, Shadorne said:

 - so phase relationships are very important and should be preserved as accurately as possible in a high fidelity system.

 

So you are audibly aware of the attrocious phase distortion caused by your speaker’s tweeters?

 

Anyway, for 2x material and higher MQA does not introduce phase distortion in the audible band. Their filters may be MPish, but they operate at higher frequencies and are very shallow.

 

1x rate may be a different story, but that is bound to be dominated by the huge imaging caused by MQA, and subsequent system IMD.

 

[Fokus]

39 minutes ago, Shadders said:

Is the phase distortion of the MQA minimum phase filters greater than a linear phase filter ?

 

By definition, yes.

That does not make it a problem, though.

[Fokus]

14 minutes ago, FredericV said:

but do not confuse with KWAK:

 

 

That veritably is the MQA of beers:

-not very good

-marketed based on ... on ... erm ...

-hampered by totally unnecessary complications

 

And this while the world could have pure, undiluted

 

[Fokus]

28 minutes ago, Shadders said:

Can you provide the reference where MQA state ringing is blur ?. Thanks.

 

As you have been told countless times, here and in other forums,  MQA are crusading against wiggly impulse responses and wide impulse responses. You only have to pick up the earliest papers on this topic by Peter Craven, and start connecting the dots from there on.

 

I am sorry I don't know by heart which year this started, but it is a bloody while ago.

 

[Fokus]

35 minutes ago, Shadders said:

Hi,

Being told is not proof. You need to present the proof which contradicts the MQA AES paper, page 3.

 

 

 

This is from 2015, but Craven was hammering on this as early as 2004. It is the log(magnitude) of a channel's impulse response versus time. Conventional filters result in a wide response, described by BS as 'blurred' or 'dispersed', whereas short (but weak) filters have a narrow response. Whether these short filters are MP or LP is of secondary importance, although BS has a preference for MP as this shortens the attack side of the response even more.

 

Brought to you thanks to CA's copy-paste support. I really really wasn't prepared to find an external host to upload and link this image ...

 

[Fokus]

5 minutes ago, mansr said:

It bears mentioning that nobody besides MQA has ever used such a plot. It provides no useful information.

 

You mean "nobody besides the people who are now MQA".

 

Likewise 'Shannon diagram' and 'Shannon space' also are not established terms.

 

MQA really built its own playground and expects the world to join in.

 

[Fokus]

2 hours ago, mansr said:

Right, Stuart and his pals ...

 

It pains me that Michael Gerzon associated with this lot.

[Fokus]

9 hours ago, MikeyFresh said:

 almost as if the strategy there in inserting such a low-level operative is to essentially just run interference and try to lend the appearance the forums are insignificant or beneath the likes of Stuart or others

 

It's the classic "insert an apparent idiot into the discussion and drag it all down to his level".

 

Just ignore the clowns (at both sides). They are bait.

 

 

[Fokus]

8 hours ago, Shadorne said:

Yes this shows what an apodizing filter does.

 

Not at all.

 

The blue curve is a regular filter for 48kHz.

The brown curve is a regular filter for 192kHz.

The orange curve us MQA-style. It is narrow because it is hardly filtering at all.

All other curves are the response of air at different distances from the source.

 

Where do you see apodising in this picture?

 

[Fokus]

People should learn to read properly.

 

Carver is not talking about your and mine MQA files. He is talking about specially-prepared early demo files:

 

"What you hear when you download a Tidal Master-encoded song is definitely not what audio reviewers heard during initial demos and wrote about a year ago. The original sample tracks — I call them “MQA-1” — were clearly stated to be “on loan” with instructions to return them to the company or destroy the files when finished.

...

I was fortunate enough to get access to some of these files.

...

I had been struck by journalists’ exclamations of great sound and I had to agree when I critically listened to MQA-1.

...

Based on the descriptions, I concluded thatMQA-1 used a mild form of my Sonic Holography,"

 

Not that I believe him. These are more the ramblings of a paranoid, or someone desperate to get into the limelight again.