Stereophile Series on MQA Technology

[Fokus]

20 minutes ago, mansr said:

It's anything but clear what that "test" entailed or what it was supposed to show. A file cannot have MQA in one channel only. Besides, there's no reason to expect a cumulative timing error, nor any variation with a repeated input.

 

From what I understand the test impulses were in 96k MQA file format, but they were still ideal dirac pulses, i.e. totally unfiltered, except for the origami split/join. As such the individual pulses are in sync with the sampling grid, and indeed there should be no variation in output timing.

 

But this is contrary to what Austin expected: " I saw no random offsets—or offsets of any kind—in where MQA's impulses landed ". Of course he didn't. He didn't test for it. But now he thinks, and wrote, that he has debunked one item of the criticism.

 

As for the channels: he recorded the MQA and PCM outputs of a DAC in to an ADC, and afterwards put one channel of each file side by side. That should be obvious.

 

[PeterSt]

21 minutes ago, Fokus said:

he recorded the MQA and PCM outputs of a DAC in to an ADC, and afterwards put one channel of each file side by side. That should be obvious.

 

Apart from it being very easy to make mistakes and thinking errors and what not, you see one here yourself - your quote.

So how would a resampling system which tries to catch dirac pulses of 1 sample width (where they ?) NOT land randomly by this means ?

From the article :

 

while MQA's approach may improve the shape of the impulse response, its sampling method—and the resulting, presumed increase in aliasing—introduce randomness in precisely when those impulses occur.

 

Jim must be talking about a means of digital resampling, applied by MQA. During the process of decimation or whatever.

What I'd like to head for is a virtual conclusion of Jim that this was nothing much MQA'd (hence no random sample landing). Also, there *is* not such a thing as possible improvement of a dirac pulse; only worse is possible (and otherwise I am curious about an idea about how the dirac pulse can be improved upon (band limit it is not allowed for a solution, I'd say)).

So the general consensus : something ain't right here.

 

[PeterSt]

FWIW :

 

@John_Atkinson, what Jim should do is provide MQA with a dirac pulse train with varying distance.

If MQA is right with their promises, it should show what I could show you (not sure whether this is a good idea to do this in advance of matters). It will also show what a normal DAC with decent filtering will make of it : a mess (as in : no dirac pulse anywhere and one pile of distortion).

Best is to have a train of e.g. 20 pulses in a row, then a space of 50 or so "nothing" then 20 pulses again, 5 spaces with nothing - and so on for a nice 30 seconds. The "5 spaces" can be longer but must cover for the filter length to bridge the gap (and show the mess).

 

If this works out, all it will show is what I expect : anyone can do this as long as the filter has been made for it. But also : how such a filter overrules the in-DAC filter.

And regarding "starting in the middle" ... still only starting from there (the digitally generated file). The main part happens in the music file and the MQA encoding (means).

 

Regards,

Peter

[Rt66indierock]

3 hours ago, mcgillroy said:

 

Yes but to my knowledge nobody has done a proper debunking of the paper and its sources. Charles opinion mattered but he was easily framed as a competitor. Hifi News is just another mag.

 

What is needed is a upping the ante on MQAs bullshit marketing and getting some science people publicly taking a good hard look at it. The neuroscience seems to be an especially low hanging fruit.

 

The FAQ on the MQA website states:

 

"MQA is based entirely on science. Specifically, it is based on new findings in Neuroscience that have told us that the resolution of timing information is critical to our hearing and our ear/brain interaction."

 

See: http://www.mqa.co.uk/professional/for-content-producers - scroll down to FAQ.

 

These are pretty unspecific claims and perhaps it's easy to find neuroscience papers supporting it.

 

But I am pretty sure that none of the neuroscience papers they cite in their AES-paper would provide proper support for their specific use-case, not to speak of their implementation. Has MQA shown any evidence that they tested and validated their stuff in some "neuroscience" setup?!

 

I am collecting the papers over the holidays and will have a look at them. Also lets see if there is any feedback from the authors.

 

 

 

Quick comment my profession public accounting in the United States has professional standards for research in tax. These have the force of law. So I get paid to do research and to plug myself earlier this I did some good enough I was asked to write a schlarly article about it. 

 

Part of what I looked at was the actual activity in the papers cited and what were the conclusions.

 

Another thing I looked at can others duplicate the results?

 

Steven Stone in his latest post about hi-res cites a paper that says   It is unclear if this relates to hi-res music in the discussion.

 

Bob's paper makes a lot people of people involved in the AES nervous  when I mentioned standards in a back and forth with Michael Ritter of BAD in The Absolute Sounds articles about MQA.

 

But carry on I'm not a member of the AES because they don't have a chapter in the Valley of the Sun.

 

[#Yoda#]

The comprehensible concerns about MQA are now elevated by Michael Lavorgna to be a religious statement : "And for those people who want to avoid MQA for religious reasons, Qobuz's Sublime+ is MQA-free".
Read more at https://www.audiostream.com/content/qobuz-prépare-son-arrivée-aux-etats-unis-yes-us#qkoRMIP2FX0d57W6.99

[asdf1000]

20 minutes ago, #Yoda# said:

Qobuz's Sublime+ is MQA-free

 

For now.... there is always potential for the labels to force all streaming companies to only stream MQA.

 

I guess that's the fear that's been discussed for a long while now on this site - the potential for DRM going to the next level.

 

My gut can't see that happening in 2018 but who knows - that's all crystal ball stuff.

 

[Albrecht]

1 minute ago, Em2016 said:

 

For now.... there is always potential for the labels to force all streaming companies to only stream MQA.

 

I guess that's the fear that's been discussed for a long while now on this site - the potential for DRM going to the next level.

 

My gut can't see that happening in 2018 but who knows - that's all crystal ball stuff.

 

FWIW, - i appreciate the wisdom in your posts. IMO, - you are asking/implying excellent questions that no one else is asking. Namely, what good is MQA without Tidal? And, how powerful are the labels now that the industry has devolved & is collapsing: and is even more exploitative and punishing of its artists? Hopefully, - Tidal will be gone very soon, and take those Meridian people down with them. I have defended the reviewers of Stereophile in the past, but their cursory and superficial analysis, with barely a mention of the really bad aspects of MQA, and streaming is most troubling. People are fighting each other to get down on their knees in front of Tidal/MQA without even questioning the value of its core process. I find it funny that as horrible as Netflix is, people are clamoring all over themselves to do the same thing to music, - where there (is/was) far more, and far more varied content.

[rickca]

2 minutes ago, Albrecht said:

People are fighting each other to get down on their knees in front of Tidal/MQA without even questioning the value of its core process.

Maybe MQA has some fantastic junkets for journalists.  It's curious that I've never seen an account of a single one.

[asdf1000]

8 minutes ago, Albrecht said:

Hopefully, - Tidal will be gone very soon, and take those Meridian people down with them.

 

But Deezer (with larger market share) has MQA coming:

 

http://www.digitalaudioreview.net/2017/09/deezer-ends-sonos-exclusivity-joins-hands-with-mqa/

 

10 minutes ago, Albrecht said:

FWIW, - i appreciate the wisdom in your posts. 

 

You must have me confused with someone else... 

[Albrecht]

7 minutes ago, Em2016 said:

 

But Deezer (with larger market share) has MQA coming:

 

http://www.digitalaudioreview.net/2017/09/deezer-ends-sonos-exclusivity-joins-hands-with-mqa/

 

 

You must have me confused with someone else... 

Very interesting about Deezer.... thanks...

[asdf1000]

1 hour ago, Albrecht said:

Very interesting about Deezer.... thanks...

 

Use these 2016 figures as a guesstimate only.  Obviously Dezzer is still a small piece of the pie overall, but it shows Deezer having over 6x the subcriber numbers of Tidal. 

 

And that was when Deezer Elite (CD quality streaming) was exclusive to Sonos. That exclusivity with Sonos has now ended.

 

 

 

 

[psjug]

On 12/16/2017 at 6:17 AM, mansr said:

What we're seeing here is the convolution of the linear phase filter with something that isn't a simple impulse. The honest thing would be to provide those test files for all to look at in whichever way they choose. Obviously, that's never going to happen.

 

Yes it seems that the MQA encoded test signal is quite different from an impulse. There is a manufacturer comment that was posted: https://www.stereophile.com/content/mqa-tested-part-1-manufacturers-comment 

I am wondering what some of you make of this.  The test input is shown there and below.

Bob Stuart comments:

"If we look at the 48kHz MQA test signal waveform (fig.1 in this comment), there is no pre-response. It is elegant that the decoder "unfolds" it back to a perfect impulse as we can infer from Jim's fig.5. Fig. 4 shows us the convolution of the signal with the linear-phase response of the particular chosen DAC (which is contributing the pre-and post-ringing). The result, as would also be the case with non-MQA files, will be different with other converters, according to the filter type (linear- or minimum-phase), rate and user settings." 

 

Fig.1 MQA-encoded impulse response sampled at 48kHz (50µs/horizontal div.).
 

[mcgillroy]

16 hours ago, Em2016 said:

 

Use these 2016 figures as a guesstimate only.  Obviously Dezzer is still a small piece of the pie overall, but it shows Deezer having over 6x the subcriber numbers of Tidal. 

 

And that was when Deezer Elite (CD quality streaming) was exclusive to Sonos. That exclusivity with Sonos has now ended.

 

 

 

 

 

Do you have the whole Midia slide-deck and can you share it?

 

Thx!

[semente]

3 hours ago, psjug said:

 

Yes it seems that the MQA encoded test signal is quite different from an impulse. There is a manufacturer comment that was posted: https://www.stereophile.com/content/mqa-tested-part-1-manufacturers-comment 

I am wondering what some of you make of this.  The test input is shown there and below.

Bob Stuart comments:

"If we look at the 48kHz MQA test signal waveform (fig.1 in this comment), there is no pre-response. It is elegant that the decoder "unfolds" it back to a perfect impulse as we can infer from Jim's fig.5. Fig. 4 shows us the convolution of the signal with the linear-phase response of the particular chosen DAC (which is contributing the pre-and post-ringing). The result, as would also be the case with non-MQA files, will be different with other converters, according to the filter type (linear- or minimum-phase), rate and user settings." 

 

Fig.1 MQA-encoded impulse response sampled at 48kHz (50µs/horizontal div.).
 

 

Have you noticed that Bob Stuart uses the same term to describe MQA as John Atkinson?

 

The Elegant MQA

 

 

 

This reminds me of The Strong And Stable May...

[psjug]

If this is the test signal, shouldn't it be a digital series?  Why does Stuart present it as an analog waveform?  Am I misunderstanding what he is saying?

 

[mansr]

35 minutes ago, psjug said:

If this is the test signal, shouldn't it be a digital series?  Why does Stuart present it as an analog waveform?  Am I misunderstanding what he is saying?

 

That looks a lot like the impulse response of filter 7 in the MQA renderer.

[asdf1000]

4 hours ago, mcgillroy said:

 

Do you have the whole Midia slide-deck and can you share it?

 

Thx!

 

Unfortunately nope - those 2016 figures are the only free ones I could Google. There are up to date 2017 figures but needs a paid subscription.

 

[#Yoda#]

23 hours ago, Em2016 said:

 

For now.... there is always potential for the labels to force all streaming companies to only stream MQA.

 

I guess that's the fear that's been discussed for a long while now on this site - the potential for DRM going to the next level.

 

My gut can't see that happening in 2018 but who knows - that's all crystal ball stuff.

 

 

I agree, but at least, the streaming services are the "gate keeper" and it seems to be that there are other real HiRes streaming options on the horizon that wouldn't support MQA like HighResAudio.com or maybe XStream some day.

 

Streaming services are not the issue for the labels, IMO. They just don't want to sell the real master quality, their "crown jewels" to the customers, finally except to a "reasonable" price in their perception, but "streaming" is not "download" and no steady ownership. 

[asdf1000]

42 minutes ago, #Yoda# said:

I agree, but at least, the streaming services are the "gate keeper"

 

Just another angle to this. I formed the opinion a couple years ago that the labels set the minimum monthly subscription fee after the labels rejected Apple’s reported initial plan for a $7.99 monthly plan. Apple fell in line with pricing and they ended up charging the same as others.

 

We don’t know all the politics involved though so I only guess and am likely wrong.

 

@Rt66indierock may have more informed knowledge if the labels effectively set the minimum monthly subscription price ? 

 

https://www.billboard.com/articles/business/6494979/fight-between-apple-and-spotify-could-change-digital-music-labels-said-to

 

PS: don’t ask me how the labels can reject Apple’s $7.99 plan but allow Spotify’s freemium plan   There must be seperate terms and conditions agreed to, which we will likely never know publicly.

[Rt66indierock]

3 hours ago, Em2016 said:

 

Just another angle to this. I formed the opinion a couple years ago that the labels set the minimum monthly subscription fee after the labels rejected Apple’s reported initial plan for a $7.99 monthly plan. Apple fell in line with pricing and they ended up charging the same as others.

 

We don’t know all the politics involved though so I only guess and am likely wrong.

 

@Rt66indierock may have more informed knowledge if the labels effectively set the minimum monthly subscription price ? 

 

https://www.billboard.com/articles/business/6494979/fight-between-apple-and-spotify-could-change-digital-music-labels-said-to

 

PS: don’t ask me how the labels can reject Apple’s $7.99 plan but allow Spotify’s freemium plan   There must be seperate terms and conditions agreed to, which we will likely never know publicly.

 

I have confidentiality in my profession but if you look at the numbers companies have charged for streaming you will have your answer. Also notice charging more doesn't seem to work either.

[asdf1000]

Just now, Rt66indierock said:

 

I have confidentiality in my profession but if you look at the numbers companies have charged for streaming you will have your answer. Also notice charging more doesn't seem to work either.

 

Yep it's not too hard to connect the dots with what's been reported (negotiations between streaming services and labels) - if the labels don't set the price, they certainly have a big influence.

[psjug]

15 hours ago, mansr said:

That looks a lot like the impulse response of filter 7 in the MQA renderer.

Yes, I misunderstood.  I should have paid more attention to the caption of Bob Stuart's Figure 1. Stuart is showing the 48kHz sampled response to the MQA encoded test signal.  Whereas Jim Austin's Figure 5 is the fully unfolded 96kHz response.  Sorry to have added to the confusion with my misunderstanding!  And so it seems we remain are in the dark regarding the MQA encoded test signal.

[firedog]

https://www.stereophile.com/content/more-mqa

 

https://www.stereophile.com/content/mqa-tested-part-2-fold

 

Interesting that they seem to think the first MQA article showed that the time response of MQA is doing what it claims-I’m not sure it did.

 

At least JA does understand that there are DRM and consumer choice issues involved.

 

 

[GUTB]

Jim Austin's latest article seems to admit that MQA is lossy, as per CA's analysis. In that case, displaying the sample rate of the original in an MQA stream is both literally and subjectively deceitful. If MQA doesn't offer lossless, than don't suggest you do. We can content ourselves with "very good lossy" and if it sounds better than it sounds better.

 

Also. "Lossless compared to what?"....compared to a FLAC, ALAC, APE, or any other lossless format. Why is Jim tying himself into a rhetorical knot over this topic? Just say: "It's lossy" and then move on to discuss the nature of the lossy format.

 

What's "spectral components above 48kHz"?

[tmtomh]

Austin's justification of MQA's lossy nature is ridiculous - and saying it's an "ethical" and "aesthetic" thing for Stuart is, well... not something that can be taken seriously.

 

Atkinson's piece is more balanced and useful in my view - although again, the "technical elegance" claim for MQA is just silly. It's far less elegant than pre-existing solutions.