MQA is Vaporware

[Paul R]

4 hours ago, crenca said:

 

Consumerism (what Samuel T referred to above) has its downsides.  However, when we know what we know we know it.  One of the things that makes the Audiophile market so eccentric is the boldness of its confidence game.

 

Well, that can certainly be one reason, but then, how is that any different from an other product's salesforce?  High Fi pushy next to an automobile agent? Might want to re think that. 

 

I suggest that a more powerful and more pervasive reason is just that audio, in particular computer based audio has gotten so much better that there is far less differentiation, and far less reason for audiophiles to be in an endless upgrade cycle. Our current audio is just too good to support the old traditional sales tactics of upgrade every week or two. 

 

That was not true as little as 10 or 15 years ago-  digital audio was changing so quickly back then, it was a new field every six months. Now, a person can buy a modest system, how an iPhone up to it, without wires or special software, and enjoy great, even stunning sound,. Technology at its finest I think. 

 

 

It is the same factor that has slowed computer and automobile sales.  Computers today are simply too good for people to worry about being on the "cutting edge" and having to have the latest and greatest new computer every few months.  Automotive sales are dragging for the same reason. I remember when my Mom and Dad would get rid of any vehicle approaching 40,000 miles. Now a days, a new vehicle is just getting well broken in and comfortable around that milage.

 

[Paul R]

36 minutes ago, Lee Scoggins said:

 

I think this is true at the low end but not true at all at the upper end.

 

I think in the 90s, you had a mass market, an affordable/entry high end market, and a luxury high end market.  Now we have seen the emergence of people creating an "ultra luxury" market.  We have folks buying $800K Wilson WAMM Chronosonic speakers, $250K D'Agostino Relentless Amps, $90K CH Precision preamps, and $40K Nordost cables.  That's good for two reasons:

 

1.  It funds the more affordable products and the business in general.

2.  The technology very often trickles down to more affordable products.

 

I'm also not sure about the computer analogy, at least in terms of the higher end segments of the market.  Technology continues to bring ever-better sound.  I just upgraded my VPI with a 3D tonearm and I can now hear way more detail from the LP grooves.  The Soundsmith cart is now dead silent between tracks.  At the higher end, some friends are buying the even better "fatboy" 3D printed arm.  There is always something better and newer to buy.

 

But I agree at the lower end, say mass market, many are just happy with a good pair of headphones or ear buds...or a Sonos system.

 

Actually, I think it is even more true with high end audio than with low end audio. And I know it is true in computing.

  

Why don't you start a new topic for the subject and we can have a grand old debate on it. I'll enter the proposition that the ultra-high end audio market is nothing more or less than sham. 

 

 

-Paul

 

[Paul R]

12 hours ago, Lee Scoggins said:

 

This isn't true.  I only want people to watch the full video which includes the discussion with Jbara as that gets to the real tenor of the meeting.  And recognize that there was discussion with Steven and others afterward that was very civil.

 

Hey Lee - as Danny and others have pointed out, the civil discussion should have been with Chris. 

 

That did not happen. It was just like when the people here attack you. 

 

I find it hard to believe you do not recognize that. 

 

I personally would have been far less pleasant than Chris. 

[Paul R]

1 hour ago, Lee Scoggins said:

 

Sure, but sometimes it's all bundled in so the consumer doesn't see things like a royalty paid to MQA or other services like XM.

 

That’s what we call “Lucy Logic” - you do not  see the direct charge so it was free? 

 

No, in every possible scenario where MQA succeeds, the costs to the consumer will increase. 

 

So so given that, what exactly is the cost/benefit to the consumer? Yes, I realize what MQA promised, and unlike most here, I think they could actually deliver what they promised. But it appears that would not be commercially viable for them. So they are not going to deliver. At least a preponderance of evidence shows they are not. 

 

Why should they not be allowed to die off, the same as any other failed technology? 

 

 

[Paul R]

33 minutes ago, Lee Scoggins said:

 

I don't believe this is accurate as the labels ran MQA through a technical gauntlet so there had to be some testing.

 

27 minutes ago, Lee Scoggins said:

 

Rob's point is that streaming is turning profitable.

 

Yes, and? Historically, when radio got popular and started making serious profit, who struggled for and eventually gained control of the radio stations? Hint, it wasn’t the artists, and Pirate FM existed for a reason. 

 

When streaming gets very Profitable, the same thing will happen. Then history will repeat itself. 

[Paul R]

2 minutes ago, Lee Scoggins said:

 

The consumer cost increasing as I outlined above are a result of them buying a premium tier.  So they are getting better sound but spending more from it.  The same trade-off we make on other goods and services.

 

No, the costs will not be paid out of a premium tier. They will be amortized into the cost of every vehicle, or the recording studio, or the distribution network, or many other places. The end result is everyone will pay more, and that $2k upcharge  is pure profit. At least accounting wise, and what else counts to big business of any kind? 

 

Accounting is a tricky thing, and very slippery. Almost like statistics. 

[Paul R]

3 minutes ago, Rt66indierock said:

 

Don't forget Buck Owens but he did eventually sell his radio stations to those you are talking about.

 

Buck was one of my hero’s, but as you note, even he eventually had to give in.

[Paul R]

2 hours ago, psjug said:

So you think MQA is cool because they can get some revenue selling technology that nobody will use?  Healthy ecosystem.

 

Huh - you do know they get people excited about paying extra for Bose sound systems in their cars, don't you? And amazingly enough, those people are very happy with their Bose stuff. 

Marketeers can make canned 100 year old crap "desirable." 

 

I just don't think MQA can survive in the hyper competitive market of the 21st century, which is why I don't get as excited about it as you guys.  

On the other hand, I utterly despise Tidal, and hated finding myself using it. I feel the same way about Tidal as you do about MQA, and really really wish it would just go away. Maybe we will get real lucky and the two will merge and die a quick death. 

 

-Paul

 

Qobuz all the way!  

[Paul R]

4 hours ago, Jud said:

 

Hi Paul - 

 

Here are the reasons why that's not possible:

 

- The more the interpolation filter in a DAC cuts, the more it rings, or to use MQA's non-standard parlance, "blurs."  In order to avoid the DAC filter itself being a cause of "blurring," it can't cut much.

 

- A filter that doesn't cut much doesn't remove frequencies in the near ultrasonic, which has two effects: (1) It cannot remove ringing caused by the ADC filters, which occurs in the low ultrasonic range; and (2) it allows low ultrasonic frequencies to intermodulate with themselves and with audible frequencies, causing harmonic and intermodulation distortion.

 

- If you don't want your technology copied, and don't want to spend time and money chasing down alleged patent violations, you must somehow obscure the precise effect of the technology.  Therefore you cannot use lossless compression, since simple decompression will reveal the effect of the technology.  You must use lossy compression, which obscures the effect of the technology at the cost of unrecoverably throwing away information from the original file.

 

- Since your DAC filters don't cut in the low ultrasonic, with the potential bad effects we've described, if you don't want those bad effects realized, you must remove the low ultrasonics from what you are feeding the DAC filters, and the only place you can do this, other than mics that don't have any ultrasonic response, is in the decimation filtering in the ADC.

 

- If you use an ADC filter with a strong cut, it will ring ("blur"), and we know that the DAC filter can't remove this (cannot "deblur"), or it will ring itself.  So the only thing you can do is cut gently, but start in the audible range so that you have sufficient cut by the time you reach ultrasonic frequencies.

 

- Therefore you will either be feeding your DAC filter a file with rolled-off highs, or one with ringing and ultrasonics it can't remove, which will cause distortion.  And you will be applying lossy compression to this file as well.

 

Thus what one might call the "central dogma" of MQA is itself impossible; it isn't just this particular implementation that is bad, any attempt to apply the principles MQA advertises itself as built on will inevitably, due to sheer mathematics, have the same set of problems.

 

There are ways to do filtering well, and wind up with a product that doesn't have a great deal of ringing, or IM or harmonic distortion.  And of course it isn't necessary to use lossy compression.  But it's not possible through application of "MQA technology."

 

Hi Jud  - thought I posted a Thank  you for the great summary above, but either I did not hit save, or for some reason It got deleted. In any case, thanks. 

 

I need to think about what you posted here, as I agree with everything you said, but: still have the feeling that there is something missing.  

 

My only clear thought is that MQA would work mathematically under some conditions, and that I think the idea is pretty dang cool. Obviously, I am not as worried about it as some here, and I doubt that the vicious vindictive will serve any purpose except as publicity for MQA.

 

What you wrote up however - that deserves an answer from MQA, one with clarity and honesty.  I wish you would polish it a little more with an eye to getting it published here and everywhere else that would print it. If people are really serious about opposing MQA, it would serve as a manifesto. 🗣🤯

[Paul R]

1 hour ago, mansr said:

What is this cool idea?

Speaking only of the technology, not the  company or their license practices -

 

I like the idea that MQA appears to use sub-band ADPCM, and that the lossy compression is not perceptual, but based upon bit rate reduction.  That is cool. I believe that could be implemented in such a way that could be effectively lossless in the audio band, though I have not done any proof on concept. Just kind of played around with the math. 

 

In other words, I think dismissing the technology’s potential  out of hand without more and  careful consideration is an emotional overreaction. The kind that always seems to find answers that are neither complete or to our hobby’s best advantage.

 

Given that, I no more agree with say, Lee’s point of view about MQA than I do with the “burn all the bastards at the stake” group. In terms of the benefit to our hobby, and to each of us individually, I think a calm, reasoned approach, backed with unassailable facts is the only way to realize any benefits at all. 

[Paul R]

53 minutes ago, mav52 said:

 So in the scheme of things its not about people actually buying access to Tidal to hear MQA music, its about how many devices MQA capability resides in is how MQA is saying its "a success" !

Gosh Chris -  yes. That seemed so obvious to me I just assumed everyone else saw it as well. 

 

Do note however, MQA is software, and can run acceptably well on very modest hardware. 

[Paul R]

1 hour ago, Shadders said:

Hi,

If the lossy coding is NOT perceptual lossy coding, then it WILL be perceived.

 

Are you sure you meant to say that ?

 

If you arbitrarily drop bits based on an algorithm that is NOT perceptual, then how can you state it will NOT be perceived ?

 

Regards,

Shadders.

 

Umm,  No? Whatever gave you that idea?

 

Or that it is it arbitrary?

 

The same type of compression that MQA uses has been used in telephony for ages. It is common in all VoIP that I know of.  Where do you think MQA got the idea from anyway? It seems to be pretty common knowledge. After reading your post I did a google search that returned several references in just a few seconds. For heaven’s sake, even the Wiki reference! (Which had I read before replying, I could just have referenced you to,) 

 

MQA’s implementation is their own, of course

 

I am not defending them mind you, but they of course built upon existing work, speaking only of the technology. 

[Paul R]

3 minutes ago, John Dyson said:

All compression (space/storage reduction) schemes have an attribute of bit rate reduction, just as a matter of physics and math.  Any scheme that does bit rate (space) reduction without consideration of perceptual attributes is most likely a very inferior scheme.  Just throwing bits away (even in an organized/mathematically reasonable way) isn't going to be as 'nice sounding' as a scheme which considers perception as its most important requirement.

There ARE schemes which can compress effectively, playing games to hide their artefacts because of choices that mitigate perceptual distortion, but most all schemes (all that I know of) have negative perceptual attributes (even mpeg at 320k or opus at 256k.)

By packing information into encoded 'noise bits', and recovering the information from encoded 'noise' will be less effective than a well designed full perceptual scheme.  How can 'partial data compression' work as well as 'data compression of all of the data' given the same bit rate?  Doesn't make any sense.  Even though the sound 'sucks' for high fidelity applications -- listen to opus at 64k or even lower -- it is surprising.  Usually, data compression of the higher energy data (lower frequencies) ends up with MUCH better compression than data compression of higher frequency data (that is a simple matter of the associated statistics and what DCT/MDCT and associated KLT theory implies.)

 

The approach of encoding the high frequency information in a few bits per sample -- especially with complex material (mixed vocals for examples) seems to be impossible to do without artefacts.  This is one situation where opus or worse -- mpeg, have troubles distinguishing multiple near-time coincident MF/HF sources.  Mpeg tends to totally hide the fact, while opus does a notch or two better.  This very subtle information is going to be encoded in a few bits? -- don't think so. For my own project, just a few little differences in some of the time domain sensitive processing can EASILY destroy the subtle details.

 

Anyway -- ANY *compression* scheme which doesn't consider perception in the data reduction choices is going to sound inferior.  (Well, except certain exceptional circumstances.)

 

There is too much snake-oil out there, with the marketeers hooking the vulnerable recipients of pseudo-technical gobbledygook.

 

 

 

There is a specific set of compression techniques used in, for example, MP3. These are labeled perceptual because they choose data to modify in the audio band based upon some specific ideas oh how humans perceive changes in sound. Nothing in the term indicates or implies that the changes will not be perceived. 

 

What you are arguing is a straw man here, as I never claimed or even implied  that the MQA compression was imperceptible - merely that they choose a different method than MP3, choosing instead to use a part of the spectrum they believe will be unheard. I hardly see where that is arguable.

 

And MQA apparently does sound different. So either their compression is at fault, or the way they interpolate the signal during reconstruction, or...

 

As for encoding useful information into higher frequencies in a signal, that will depend upon the bit rate and the information you are encoding now, will  it not? 

 

Again, I am not defending them!

[Paul R]

5 minutes ago, Shadders said:

Hi,

As per my post, and John Dyson above, any coding that is lossy, which is based on an algorithm which is NOT perceptual (not a perceptual coding scheme such as MP3), then it WILL be perceived to be different/inferior.

 

Not sure about the reference to VOIP or telephony - what do you mean ?

 

ADPCM is lossless - just bit rate reduction for a channel. The reconstituted bits at the end of the channel are bit perfect to what actually was transmitted.

 

MQA is lossy, and uses perceptual lossy coding.

 

Regards,

Shadders.

 

Okay, Whatever you say. This is an argument I am not going to have, so you totally win. 

 

✌️🙂

 

 

[Paul R]

43 minutes ago, Shadders said:

Hi,

No one is trying to win anything.

 

MQA is perceptual coding with an algorithm. MP3 is perceptual coding with an algorithm. The algorithms differ, but they are still perceptual coding. It is just the classification of the coding schemes.

 

Regards,

Shadders.

 

No, you misunderstand. Most if not all of your (and Dyson’s) points I agree with, so why argue over terms? I just think MQA has unrealized potential.

 

Discussions about the shape and historical background of software algorithms  are fun for me, but off topic here. 

 

Yours 

-Paul

[Paul R]

14 minutes ago, mansr said:

That makes no sense. Telephony uses voice codecs, typically of the CELP variety. Such codecs aim to reproduce speech using very low bit rates. They generally prioritise intelligibility over absolute fidelity. Because they rely heavily on patterns typical of human speech, these codecs are ill suited for anything else. Anyone who has been subjected to hold music on a mobile phone knows why. Basing a codec intended for music on such technology would be a rather bizarre thing to do.

 

VoIP in specific my friend, and including hi res multipoint video conferences, and real time live shows in theaters, and so on. 

[Paul R]

1 minute ago, mansr said:

The V in VoIP is still voice. Anything that isn't voice-only uses one of the usual audio codecs. Telephony, including conferencing, also has latency requirements that one-way transmission does not. For these applications, low-latency codecs exist, but there is always a price to pay in either bandwidth or sound quality. Again, using them for distribution of pre-recorded music would be bizarre.

 

Where are you getting your information about MQA?

We will just have to agree to disagree here, because what you judge as bizzare is rather commonplace and not bizzare at all for me. And yes, packet queuing delays do seem to always be a thorn in my side. 

 

[Paul R]

7 minutes ago, Shadders said:

Hi,

I do not see any potential for MQA as benefiting the consumer at all.

 

A very simple statement :

 

An MQA file or CD is 16bit/48kHz, or 16bit 44.1kHz with added noise for the MQA upper frequency coding. Unless you have an MQA capable product (which costs), then as a consumer you have a degraded sound file. If you want to remove the degradation from the file you have to purchase an MQA capable product.

 

So, as a consumer, MQA means to have at least, a clear/unadulterated 16bit sound file which has lossy coding, you have to purchase an MQA capable product, which adds the HF, which is lossy, causes blur, and aliases.

 

There is absolutely NO potential in MQA. It is a complete and utter abortion of a scheme.

 

Only MQA Ltd gains.

 

Regards,

Shadders.

 

You are talking here about MQA the company.  

 

[Paul R]

6 hours ago, ddetaey said:

You have a number of supporters on personal level here on the forum.

But honestly , if you are going to make a technical link between MQA and Voip and Videoconferencing, you really are undermining yourself at the technical level.

Have you started a personal competition with Lee Scoggings?

Dirk

 

Thank you, I appreciate that. Honestly was not trying to polarize any one or any group of people, but apologize for doing so. 

 

Here's what I said, in response to Mansr's question of what I thought was cool about MQA. Answering that question was probably a mistake, but at least it was an honest one. 

 

Quote

I like the idea that MQA appears to use sub-band ADPCM, and that the lossy compression is not perceptual, but based upon bit rate reduction.  That is cool. I believe that could be implemented in such a way that could be effectively lossless in the audio band, though I have not done any proof on concept. Just kind of played around with the math. 

 

In other words, I think dismissing the technology’s potential  out of hand without more and  careful consideration is an emotional overreaction. The kind that always seems to find answers that are neither complete or to our hobby’s best advantage.

 

I thought I recognized sub band ADPCM which is well used in telephony, and the compression algorithm they use *is* significantly different than the compression used for MP3. I could have saved myself a lot of time by reading the Wiki page on MQA, which would have told me it wasn't a new discovery. Aw well, it wasn't a waste of time as I had to dust off some math skills and that's always an activity worth doing.

 

I just think that trashing MQA is the current fun thing to do, though I admit, with plenty of  justification.  The principals in that company perhaps have called down a hellish rain of vengeful computer savvy audiophiles on their own heads. Lord have mercy on em,  cause very few if any computer audiophiles will.  

 

 (Would you want this man angry with you?) 

 

The damning summary that Jud wrote up is what they should be very afraid of though. I think that alone could do more damage to MQA the company than all the howling outraged demons here. If they cannot or do not answer it with clear, honest, and technical answers, they are fools. 

 

 

 

 

Unknown

[Paul R]

Very well said, thank you. 

 

I wonder if the concept  could be extended well into very high speed / high bandwidth transfers?  I mean, outside of audio, we already have transmissions so high that multiple Nyquist zones must be dealt with. Probably just daydreaming in any case, and not sure there would be any practical applications in the audio world. Thank you again for that very well written summary. I am copying that into my journal. 😁

 

Yours, 

-Paul

 

[Paul R]

FLAC was the freeware open source version of ALAC, so the technology is even older. 

[Paul R]

29 minutes ago, The Computer Audiophile said:

Kind of funny. Even the labels have rejected using MQA as a master because it isn’t lossless. Yet, MQA and Tidal push “Tidal Masters” hard. I guess we shouldn’t be surprised. 

 

That is very ironic - and insightful as well...

 

[Paul R]

3 hours ago, Samuel T Cogley said:

Scoggins turned a lot of people off to Shunyata products by his relentless shilling.  But somehow, he thinks he's helping.  I honestly don't get it.  MQA is now getting Scoggins' Shunyata treatment.  And I can't say that it's helping MQA.

 

If this is so, then why for heaven’s sake are you guys complaining?

 

 Just could not resist pulling your leg a teeny little bit there...

[Paul R]

2 hours ago, John Dyson said:

So many times in the past -- I have had to rescind a claim or simply accept the fact that I am wrong.  It shouldnt' be a matter of ego, but simply an acceptance of reality.  The only thing that really bothers me is that the snake-oil just might hook a few people in the meantime.

 

On the other hand about making a false or erroneous claim -- someone has to step out and innovate, and sometimes the innovations are bogus or problematical.  The frustrating thing in this case is that the mistaken claim hasn't been accepted as a mistake by the originator.

 

John

 

 

Lots of fire leveled at MQA, but what about the real shills selling it?   Perhaps the reasons have to do with some folks secretly (or not so secretly) listening to MQA on Tidal? (grin)

 

Okay okay, I know, low blow. But the actual question is, why is there not more pressure on Tidal to stop supporting the product that is supposedly so universally hated? 

 

[Paul R]

25 minutes ago, Hugo9000 said:

 

*Idle speculation and attempted humor, I have no inside information, honestly!  If I stumbled on the truth, please don't add me to a hit list, whatever secret agency is monitoring these threads!

 

How did you learn that MQA has joined the ranks of certain other three letter agencies?  Well, no matter. MQA secretly encodes mind control frequencies, and you will soon not remember this... listen to this nice Tital Track...   Mwwahhh haaa haaa!!!.... 🤯.

 

Untitled - 3:11:19, 10.29 AM.mp3