• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

MQA Deep Dive - I published music on tidal to test MQA

Status
Not open for further replies.

PO3c

Member
Joined
Apr 27, 2021
Messages
67
Likes
123
Jisus man, how many times do we have to repeat the same? Those "repeatable"measurements are measuring the wrong things:

1- square waves, high amplitude white noise, and presumably big amplitude impulse tones completely outside the maximum amplitudes of the music the system is programmed for, as all of them contain upper octaves and ultrasonic in high amplitudes. MQA is not intended to register high amplitude in ultrasonics, because there is NO MUSIC with that profile, and because that space is better used for custom filters fixing time domain issues. If you understand what MQA does (and if both accomplished amateurs are in fact accomplished, they knew it BEFORE doing those tests), you don't need a test to know a square wave will not perform OK.

2- A bit perfect match that by definition is impossible, as MQA replaces those below the noise floor with dithered information. Again, what an accomplished MQA reviewers should know in advance.

3- To make things worse, MQA assumes the noise band is dithered; these tests omitted that basic step, and so, the algorithm is fed with wrong information. It is obvious the multitude of anomalities would occur doing that.

4- Even if all of the above were not an issue, even then you wouldn't get that pretended bit-perfect match, because the system is intended to be lossless compared with analog input; but to fix the flaws of that input (by correcting time domain issues) if instead that input is digital.

5- If MQA were performing as bad as their conclusions say, the plots above ( that even then are incomplete, since they are comparing a 352K input with a 88.2 Khz output, leaving information the system has registered unprocessed) wouldn't have the degree of coincidence they have. Then... something must be wrong with those tests, don't you think?

What you are questioning is the very definition of the MQA design (noise shaping, encapsulation of data in noise bands, limited amplitude in high frequencies, deblurring of time domain information, etc), because all of that differs with your immaculate conception of how perfect a simple Redbook file is. If so, you don't need to backup your criticism in test trying to demonstrate the things the system is not intended to do.
Why is it wrong to show these limitations? MQA intent are to store music. If we remove seatbelts and airbbags form a car and say its not intedet to be crashed should not the driver be informed what to expect if she do crash?
 

RichB

Major Contributor
Forum Donor
Joined
May 24, 2019
Messages
1,946
Likes
2,611
Location
Massachusetts
And slightly worse...
And more expensive...
And more complicated...
But other than that, MQA is just fine

You forgot:
And more restrictive...

- Rich
 

mansr

Major Contributor
Joined
Oct 5, 2018
Messages
4,685
Likes
10,700
Location
Hampshire
Why is it wrong to show these limitations?
Exactly. MQA is presented as a black box. To characterise what the box does, a perfectly reasonable first step is to throw a variety of test signals at it and observe what happens. If we're lucky, some of them cause it to break down, hopefully in interesting ways. From there, we can then narrow down the tests, gradually identifying what the box can and cannot handle. The problem in this case is that they took the box away before we could perform even a second round of tests. In doing so, they have unwittingly told us all we need to know, that MQA is a sham that can't tolerate any scrutiny. Of course, we already knew that. Still, having official confirmation is nice.
 

tmtomh

Major Contributor
Forum Donor
Joined
Aug 14, 2018
Messages
2,636
Likes
7,497
Is the most egregious thing about MQA that it is unnecessary?

Definitely the most ironic and tragic aspect, but I would say probably only the 3rd most egregious. The most egregious is that it degrades the source's bit depth (and its distortion levels and upper frequencies if you count the aliasing from the leaky filters). The 2nd most egregious is that even if you use only MQA-capable playback software and hardware, you still get only a partial escape from that degradation.
 

BinkieHuckerback

Addicted to Fun and Learning
Joined
Feb 16, 2021
Messages
716
Likes
1,055
Definitely the most ironic and tragic aspect, but I would say probably only the 3rd most egregious. The most egregious is that it degrades the source's bit depth (and its distortion levels and upper frequencies if you count the aliasing from the leaky filters). The 2nd most egregious is that even if you use only MQA-capable playback software and hardware, you still get only a partial escape from that degradation.
...true...though why bother in the first place, is my point...
 

jokan

Addicted to Fun and Learning
Joined
Feb 19, 2019
Messages
532
Likes
563
Location
Tokyo, Japan.
I just had my thread erased because I didn't know THIS thread existed.

Everything I've read tells me that MQA is another way for hardware companies to charge more and pay more for licensing fees.
Am I wrong to believe that MQA is lossy?
I'm not concerned about data storage since storage keeps becoming less expensive. I have 16tb's of files.

My simple question is it noticeably better or worse?
Does it measure better? Does it sound better. Sound is far more important than anything else to me.
It maybe subjective to say one sounds better than the other, but since our ears are analogue and we have individual hearing loss it's a credible question.
Our ears also get used to hearing music, our favourite music played back by the same equipment all the time. If you were to cut the treble a single db before you go back to listening, will you notice the difference immediately? Would you notice at all?

Any information that you can share with me about MQA and if I need to purchase another device would be appreciated!

Thank you kindly.
 

voodooless

Grand Contributor
Forum Donor
Joined
Jun 16, 2020
Messages
10,229
Likes
17,811
Location
Netherlands
My simple question is it noticeably better or worse?
Does it measure better? Does it sound better. Sound is far more important than anything else to me.

The quick answer is: no it does not. Does it sound worse? Probably not either. The few blind tests that have been done were rather inconclusive. That in itself could be seen as a testament if the whole thing actually had a valid purpose.
 

muslhead

Major Contributor
Forum Donor
Joined
May 28, 2020
Messages
1,560
Likes
1,718
The quick answer is: no it does not. Does it sound worse? Probably not either. The few blind tests that have been done were rather inconclusive. That in itself could be seen as a testament if the whole thing actually had a valid purpose.
Oh, it has a valid purpose, isn't an ability to control a portion of what we listen to and financially reap any benefits therein, a purpose? Maybe not to you but surely it is to its designers.
 

voodooless

Grand Contributor
Forum Donor
Joined
Jun 16, 2020
Messages
10,229
Likes
17,811
Location
Netherlands
Oh, it has a valid purpose, isn't an ability to control a portion of what we listen to and financially reap any benefits therein, a purpose? Maybe not to you but surely it is to its designers.

I obviously meant a purpose to the consumer, not to our MQA overlords :facepalm:;)
 

tmtomh

Major Contributor
Forum Donor
Joined
Aug 14, 2018
Messages
2,636
Likes
7,497
I am not sure where you get the impression that 3 bits of "quality" are lost. The container still has the original bit depth. But there is now a hidden data channel in which data encrypted as pseudorandom noise can be buried without reducing the original resolution of the audio data,

This is not snake oil or handwaving. Without getting into the specifics of what MQA does, creating a buried data channel takes advantage of the spectral nature of the analog noisefloor present on all music recordings.

I have made several choral recordings since the turn of the century. I always try to make the recording in a quiet hall and spend time chasing down and eliminating sources of noise before the sessions start. My microphones have low self-noise and I use low-noise microphone preamplifiers from Millennia Media. Nevertheless, there is always noise present in the recording.

View attachment 126313

As you can see from this graph, made from a 24-bit recording of the “room tone” in the Oregon church where I made some of these recording, the spectrum of the noise is not flat or “white.” Instead it is closer to pink. The peak level is close to -70dBFS in the low bass (thanks to distant traffic noise) and slopes down at around 24dB/decade to 1kHz and with a somewhat shallower slope in the treble.

An FFT-derived spectrum of dithered 16-bit silence with the same number of FFT bins would produce a flat spectrum with all the components lying around -130dBFS. As the music is always higher in level than the noise, you can see that the only part of the spectrum that would need to be encoded with >16 bits lies between 2kHz and 30kHz. 13 or even 12 bits would be sufficient in the bass.

What this means is that a 24-bit recording of music made in this church that peaks at 0dBFS has spectral space available below the analog noisefloor. If I encode low-bit-depth data of some kind as pseudorandom noise – much easier to write than do - and add it to the 6 or 7 least-significant bits of the original 24-bit audio file, I have created a buried data channel. As the spectrum of that buried data channel is identical to the noisefloor of the recording, I haven’t reducing the resolution of the music data and there is a negligible rise in the overall noisefloor. I haven’t truncated the original 24-bit data to 17 bits or 13 bits or whatever, as has been stated elsewhere in the thread.

You don’t get something for nothing, however. As the noise floor now includes real information, albeit in encrypted form, I have increased the entropy of the file. The data can’t, therefore, be compressed as much by FLAC etc, as the original data.

This is not a new concept. Alan Turing did something somewhat similar in WWII to allow encrypted communication between Winston Churchill and FDR. Turing encoded the voice message as, IIRC, 8-bit audio data then buried it in a recording of random noise. When the message was transmitted, anyone listening would just hear noise. Decoding the message depended on the receiving station having exactly the same recording of random noise. Subtracting the noise signal from the received transmission reconstructed the original voice recording. (For transmission the noise was played from a 78rpm disc and for the system to work, a copy of that disc had first need to be flown across the Atlantic.)

In the 1990s, the late Michael Gerzon worked with Peter Craven (now with MQA) on a similar subtractive dither scheme with a buried data channel intended to increase the resolution of digital audio recordings – see https://www.aes.org/e-lib/browse.cfm?elib=7964

John Atkinson
Technical Editor, Stereophile

I appreciate you entering this conversation and your comment is, as always, detailed in length and courteous in tone.

That said, I'm stunned - not at all surprised, but still stunned - at how your otherwise clear and detailed comment completely obscures what you're actually saying: MQA-style encoding doesn't impact the resolution of the recording because "real world" recordings don't need more than 12-13 bits to fully capture them (except above 2kHz in this example).

Surely you must know that any reasonable person familiar with (a) your writings, (b) Stereophile, (c) the audiophile hobby and (d) the hi-fi industry will immediately respond that "13 bits is just fine" has most certainly not been the guiding principle of any of those things for the past 30+ years.

And that's where @mansr 's point about the disingenuousness of your comment comes in: it's deeply insulting and disrespectful to peddle this notion when you know full well what the fundamental flaw in it is. And it's even more deeply insulting and disrespectful to respond in turn with something to the effect of, "I did not say 13 bits is enough; I only noted that in a recording like the one I mentioned in my comment, 12-13 bits was sufficient in the spectrum below 2kHz."

This is exactly what MQA and its apologists have tried when confronted with evidence that MQA is lossy: "Well, it's partially lossy and and lossiness isn't really an issue so who cares." Now we get, "well, it does reduce the bit depth, but we never needed that bit depth, so who cares."

For shame.
 

John Atkinson

Active Member
Industry Insider
Reviewer
Joined
Mar 20, 2020
Messages
165
Likes
1,022
I'm stunned - not at all surprised, but still stunned - at how your otherwise clear and detailed comment completely obscures what you're actually saying: MQA-style encoding doesn't impact the resolution of the recording because "real world" recordings don't need more than 12-13 bits to fully capture them (except above 2kHz in this example).

With all due respect, you are misunderstanding and misrepresenting what I wrote.

John Atkinson
Technical Editor, Stereophile
 

tmtomh

Major Contributor
Forum Donor
Joined
Aug 14, 2018
Messages
2,636
Likes
7,497
With all due respect, you are misunderstanding and misrepresenting what I wrote.

John Atkinson
Technical Editor, Stereophile

I'm sorry, but the question of whether I'm misrepresenting what you wrote is probably the one area where I've no interest in giving any credence to your claims.

As to misunderstanding what you wrote, that very well might be true. I am always interested in learning and will be happy to amend or correct any demonstrably technically incorrect statement I made in my prior comment.

But as I'm confident that many others here share my "misunderstanding" of what you wrote - and that still others with more technical expertise than I have are not ready (to say the least) to agree with you that what you wrote has been misunderstood, I think the onus is on you to explain what has been misunderstood (and, if you think you can, what's been misrepresented).

Absent that explanation, your mere claim is, by itself, either meaningless or a power play. Either way, it has no value without elaboration, which I hope you will provide.

Thank you.
 
Last edited:

levimax

Major Contributor
Joined
Dec 28, 2018
Messages
2,349
Likes
3,462
Location
San Diego
Mr., Dorsey may already have had a chilling effect on the already tenuous audiophile media.

Can they really afford to alienate him?
Loss of access to the social media is at stake, and not just twitter, the tech giants have cooperated to preserve each others interests.
Those who are vocal threaten the financial interests of powerful people may start to wonder if direct opposition is unwise, though they may find a way to justify it. Just pay the royalty, and move on.
- Rich
Unfortunately I think we are seeing this chilling effect right before our eyes. Maybe a little compassion is in order though. People need to make a living and taking on internet monopolies is a good way to get cancelled. The bigger problem is that some day any one of us may find ourselves in a similar situation.
 

John Atkinson

Active Member
Industry Insider
Reviewer
Joined
Mar 20, 2020
Messages
165
Likes
1,022
I'm sorry, but the question of whether I'm misrepresenting what you wrote is probably the one area where I've no interest in giving any credence to your claims.

As to misunderstanding what you wrote, that very well might be true. I am always interested in learning . . .

Although you say that have no interest in giving what I write credence, if you are really interested in learning, I will try again.

But as I'm confident that many others here share my "misunderstanding" of what you wrote . . . I think the onus is on you to explain what has been misunderstood (and, if you think you can, what's been misrepresented).

You're correct that others share your misunderstanding, typified by your earlier comment that "it's even more deeply insulting and disrespectful to respond in turn with something to the effect of, 'I did not say 13 bits is enough; I only noted that in a recording like the one I mentioned in my comment, 12-13 bits was sufficient in the spectrum below 2kHz.'"

I don't understand why it is insulting to offer further clarification. But even if you feel insulted, here goes:

The 24-bit recording of room tone had a spectrum that had an amplitude close to the 13-bit level in the very low bass. However, the specturm sloped down as the frequency increased and In the treble the amplitude of the noise was closer to the 18-bit level. That does not mean, as you falsely claim me as saying, that "MQA-style encoding doesn't impact the resolution of the recording because 'real world' recordings don't need more than 12-13 bits to fully capture them." I haven't made any comment on MQA.

The music recordings I made in the church where I recorded the room tone peaked just below 0dBFS. As it is not possible to record the music without the analog noise, and not feasible to record frequency bands individually, a recording of that music will need to be made with a resolution of at least 18 bits, if information is not to be lost.

18 bits, not 13 bits.

And as I explained, the spectra statistics of the analog noise floor allow a hidden data channel to be inserted in that analog noise floor. And as long as the spectrum of the data embedded in that channel is encrypted as pseudo-random noise, has the same spectral statistics as the noise floor in the recording, and has an amplitude sufficiently below the noise floor at all frequencies, there will be no loss of bits/information/resolution.

Noise is noise.

. . . and will be happy to amend or correct any demonstrably technically incorrect statement I made in my prior comment.

I look forward to you doing so.

John Atkinson
Technical Editor, Stereophile[/QUOTE]
 
Last edited:
Status
Not open for further replies.
Top Bottom