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

[holicst]

Isn't it ironic to defend MQA by saying that simple tones are not natural sound so it is not suitable to test the performance of MQA, on a website that is all about driving sine waves through equipment and measuring the SINAD?

 

[Jimbob54]

MQA tends to have a tiny bit of house sound, or a signature (not a tonal coloration). If you listen enough you pick up on that.

Uh huh

 

[DimitryZ]

This makes HDCD look like a great format.

It is! I just got a dozen HDCDs and a proper decoder!

 

[DimitryZ]

Isn't it ironic to defend MQA by saying that simple tones are not natural sound so it is not suitable to test the performance of MQA, on a website that is all about driving sine waves through equipment and measuring the SINAD?

You are misreading. Amir is proposing a proper test with tones, which would be great.

What was stupid is to put the test tones into the upper triangle, where MQA doesn't encode.

 

[Grooved]

That second article is an analog capture of fully, hardware decoded mqa vs an analog capture of flac.

Right, but I was only replying about the test Pkane and I did recently, I thought DimitryZ was talking about that

I'm posted no more than you or anyone else and i don't care if you're irritated or not,who are you to tell me how little or how much I can post,you have a problem this site have moderators,take it to them.P.S. "Addicted to fun and learning" was never my name or "Major contributor" ,I have no control over that this is not my forum I'm just a member like you.

"Bellow" your name was the correct one, sorry for that, and yes, I know it doesn't come from you, it was some kind of coincidence to see that, and it was just a joke (see the smiley at the end) to bring it back to a bit "Fun" because this should be a funnier thread than it is.
I clearly don't say it's your fault and only yours, but I clearly say you had some posts as much annoying as some other people posts you seem to find annoying.
You changed the sense of what they said, it's a fact, I've read their answers after yours, and I think you did it to me some weeks ago.

Now, you're missing one point or I may not have explain it clearly enough : even if I posted more than you, you should have seen that I provided several tests, or just answering question if I can answer, and I even asked questions. That's sharing with people, I think it's what we are supposed to do on a forum, isn't it?

I tried to share with you by asking you what you think of the OP file getting Blue light while the file was supposed to be not acceptable ? Just logic and no technical background needed (as you said yourself you had not, it's not me supposing it as I don't know you), but you never answered.

I already said you can post as much as you want but that it would be refreshing to not see your "soft-trolling". That's this "soft-trolling" that I hope you can stop, not posting by itself. It's a big difference.

You can search in my posts if I ever had one discussion of this kind with another member except you... if someone never had a problem with everybody has one with me, I think I would start to question myself about what I wrote.
I'm doing it on our "discussion" and I will admit I could have done better by just asking "don't you have the feeling you're doing some "soft-trolling" this weekend?"
Now, you can think what you want about me, I will be back to tests and sharing

 

[Grooved]

But maybe more apples to apples comparison. Also, since MQA does some digital processing on the original digital master, even if their packaging scheme was perfect, they will not null perfectly.

I supposed they won't too. Not sure if I'm right here, but since I have devices that produce less difference by encoding an original file than what you get between FLAC and MQA, and supposing they have a better encoder than mine, it shoud mean that they add processing to create to MQA file.
What I'm not sure at this moment is about the rendering effect, if it's changing the sound more than the decoding.
I will test decoding only compared to decoding+rendering, but it will only a ear test.
If I try to record the analog output with an ADC, the problem is how to ouput with the same DAC with and without rendering, if the MQA DAC receive a MQA stream, decoded or not, it will do at least the rendering.

EDIT : I think I found a way to do it with Roon

 

[Grooved]

Absolutely better in many cases... perhaps 2/3.

This is what I thought first, but now it's changing a bit. I realize that I changed my habits and that I use Tidal for short listening sessions, and Qobuzz for longer, not sure why, maybe I get a bit of fatigue sooner with Tidal...

The thing that I would have loved is Tidal adding "Random Access Memories" to their MQA list, because it contains tracks where it's so easy to hear the difference between AAC 320 and FLAC on Tidal that it would have been a perfect test for me between HiRes FLAC from Qobuz and MQA

Archimago has already analyzed MQA way back in 2017 and made perfectly reasonable conclusions. Obviously OP was oblivious to that.

As most of the anti-MQA posters here and elsewhere.

Yet, no one had tried to publish file, and pro-MQA could have done it too. The thing I admit is a good idea in OP is publishing, but I would have done music files (from different kind) and test files seperated.
And I'm still getting the fact that he got a Blue light on his file with tone test out of the triangle as an alarming fact, more linked to marketing lies than technical, as we can imagine the not acceptable file created a bug in the encoder... What's your thought on that ?

 

[amirm]

Isn't it ironic to defend MQA by saying that simple tones are not natural sound so it is not suitable to test the performance of MQA, on a website that is all about driving sine waves through equipment and measuring the SINAD?

No. It is the opposite. It shows that we/I understand how you test each audio system appropriately.

Besides, pure sine waves would have been perfectly fine and suitable to test with for OP since music can be decomposed into them. What is not proper is square waves, Impulses, noise going to 44 kHz, etc.. You do NOT see me use these to test hardware -- which doesn't care what it is told to play -- yet is used by OP with technology that absolutely cares what it is fed (MQA).

As I have explained repeatedly, what you testing matters with respect to testing. For lossy codecs for example, listening tests rule and test tones would play a tiny, tiny role (to check flatness of response and such). There are specialized objective measurements such as PSNR but they are a) not simple hardware test tones we run and b) are rarely used.

Both airplane and cars have wheels but they serve completely different purposes. Don't mix them up together. Or listen to however has coin that talking point. It is very wrong and shows total lack of understanding of how you test different components in audio chain.

 

[DimitryZ]

I supposed they won't too. Not sure if I'm right here, but since I have devices that produce less difference by encoding an original file than what you get between FLAC and MQA, and supposing they have a better encoder than mine, it shoud mean that they add processing to create to MQA file.
What I'm not sure at this moment is about the rendering effect, if it's changing the sound more than the decoding.
I will test decoding only compared to decoding+rendering, but it will only a ear test.
If I try to record the analog output with an ADC, the problem is how to ouput with the same DAC with and without rendering, if the MQA DAC receive a MQA stream, decoded or not, it will do at least the rendering.

EDIT : I think I found a way to do it with Roon

My sense is rendering is a minor effect. It consists of a filter choice (one of 32 available, but according to mansr one of 16 used in practice) and upsampling command to the frequency of the original master.

There is only one unfold and a relatively minor render operation.

Hilariously, PFM in a diplomatic move, banned me from their parallel anti-MQA thread, ostensibly in response to mansr being banned here.

 

[DimitryZ]

This is what I thought first, but now it's changing a bit. I realize that I changed my habits and that I use Tidal for short listening sessions, and Qobuzz for longer, not sure why, maybe I get a bit of fatigue sooner with Tidal...

The thing that I would have loved is Tidal adding "Random Access Memories" to their MQA list, because it contains tracks where it's so easy to hear the difference between AAC 320 and FLAC on Tidal that it would have been a perfect test for me between HiRes FLAC from Qobuz and MQA



Yet, no one had tried to publish file, and pro-MQA could have done it too. The thing I admit is a good idea in OP is publishing, but I would have done music files (from different kind) and test files seperated.
And I'm still getting the fact that he got a Blue light on his file with tone test out of the triangle as an alarming fact, more linked to marketing lies than technical, as we can imagine the not acceptable file created a bug in the encoder... What's your thought on that ?

As far as the automated decoder was concerned, it encoded the content the best it could, issued its warnings and issued an authentication code.

Any technology can be broken by uneducated use.

What do you mean "no one had tried to publish the file?"

Archimago's analyses have been publically available for four years - plenty of time for anti-MQA folks to get educated, instead of yucking it up on various forums. For the consumer, his analysis tells you how it works what its' flaws and weaknesses are.

 

[JustAnandaDourEyedDude]

But, as said, I did not make this statement based on solid evidence, and so I apologise for this and it should not have been said.

The point of the statement was that a company which incorporates MQA is going to be more defensive of it. Including representatives of said company, who are active on this forum, and those with friendly ties to said company.

I suppose the easiest option would be to simply ask you straight. Do you or any of your companies have any relationships or ties to Aoshida or any of its sub-brands?
If not, I sincerely apologise and should not have said what I did. And will edit the posts if you wish me to do so.

In the US at least, a person is considered innocent by default unless proven guilty. What you posted on SBAF as seen in one of Amir's posts here is implying unethical behavior by Amir, for which by your own statement you lack hard evidence. The default process should not be to quiz Amir to get him to affirm his innocence and await his request that you edit the posts. My take on it is that you ought to edit the posts without delay, retract the defamatory-type aspersions on Amir's character and ethics, and apologize, but of course I am nobody to tell you or anyone else what they should or should not do.

The posts in question cast you in a bad light. My impression is that you were connecting dots that do not actually have a connection between them, perhaps induced by the stress you must feel from multiple attacks on your YT videos by multiple players and random internet denizens. JohnYang1997, a Topping rep here has posted on ASR multiple times that based on his experiment he believes MQA is messed up, and has also provided a plausible explanation for why many DAC manufacturers are hastening to add MQA decoding capability to many of their products. Amir has always (I have only read many of his posts on ASR, not on previous forums) been extremely transparent. He has already affirmed multiple times in the past that ASR, and he on ASR, have no hidden financial ties to the companies whose products he reviews. Where his company Madrona sells speakers or other equipment he may review, he has stated that up front in his reviews. He has also recused himself from awarding a panther in any cases where there may be what could be viewed as a conflict of interest. I really see no room for casting aspersions on Amir's ethics as a reviewer or leader of ASR. Amir has every right to a personal opinion about what he views as the elegance of the MQA algorithm or his respect for Bob Stuart's research contributions, and to try to present facts on what he believes are wrong various people's posts. And of course, posters including you and I are free to argue back with facts if we disagree with him or anyone else. Also, just to clarify, I have no ties to Amir or the mods, other than being an ASR member. I have never spoken with any of them or exchanged any e-mails, or submitted any items for Amir to test. They do not know my personal identity (I hope ), nor I theirs, other than some of what is known publicly about Amir.

My testing showed that MQA was not lossless. The counter argument is that because my tests didn't represent natural sound they unfair to MQA.

It is not. It is simply to call out misleading marketing. MQA claimed to be lossless.

I agree with you completely here. The term "lossless" with regard to digital data has an unambiguous meaning (mathematically lossless, bit-for-bit reversible transmission, compression, etc.) to informed lay people and experts (in data compression, computer science, math, etc.) alike. MQA marketing has been dancing around the term, using it to imply overall losslessness by careful wording, and in at least one case outright calling MQA lossless. The traction that your videos have gained has seemingly caused MQA marketing to walk back their use of the term lossless. Wait, they did not walk it back with explanations or apologies, it simply transformed magically overnight into "better than lossless". Many kudos on your achievement here. Your experiment and videos on it succeeded where previous analyses by Archimago, Mark Waldrep, and a couple of others did not get to the impact level of exposure for that to happen. The change in MQA marketing wording will hopefully help prevent some music consumers from being lured into MQA, who might otherwise have been drawn in by the dazzle of "lossless".

I agree with Amir that we should be aiming for better than CD quality. As Amir pointed out, sticking with CD quality makes it pointless to rank DACs and amps with SINADs higher than 96. From reading most of the info in this and previous MQA threads, I would opt for properly encoded 24-bit 48kHz FLAC, even though it may be lossy relative to modern digital masters. Even though my old ears could not reliably distinguish between 320kbps mp3 and CD quality in the online NPR 2015 test . I see no substantive advantage for the music consumer of MQA over 24/48 FLAC. I see only significant disadvantages such as the cost of MQA and its DRM-like nature that hampers my convenience of use. FLAC, on the other hand, thanks to the generosity and hard work of the good folks over at xiph.org, is free and universally available now, except perhaps in Apple's walled garden where ALAC is an option instead. Let's not kid ourselves about the cost. MQA wants to eat at the same table as the rest of the audio industry, and in a big way. If they succeed and thrive in a big way, ultimately consumers will be paying the MQA tax that I view as bringing no proven benefit with it. I agree with others that MQA is a solution to a problem of more than a decade ago. I also prefer to stick with 24/48 FLAC that can represent loud high-frequency tones in the audible band that rise above the MQA orange triangle than get low-amplitude ultrasonic components of music that audibility experiments tell us would be below the threshold of hearing for most of us. Yes, MQA can probably handle music above the orange triangle by configuring the encoder for it, but then the MQA file size will increase too. Unless the MQA time-deblurring apodizing filter is proven to improve audio transparency very significantly, but even that is unlikely to sway my preference. MQA has tracked the envelope of music for a million or a few million tracks. Apple and Amazon claim libraries of 70 to 75 million tracks, and who knows what music may be released in the future.

Well, I think the conversation in this thread has moved beyond the use of the term lossless in MQA marketing material, and is diving into demystifying what it is that MQA actually does. Several expert people are participating, and I think you (and I) need to dig deeper into the technical aspects if we want to understand further posts on it.

 

[Grooved]

As far as the automated decoder was concerned, it encoded the content the best it could, issued its warnings and issued an authentication code.

Any technology can be broken by uneducated use.

What do you mean "no one had tried to publish the file?"

Archimago's analyses have been publically available for four years - plenty of time for anti-MQA folks to get educated, instead of yucking it up on various forums. For the consumer, his analysis tells you how it works what its' flaws and weaknesses are.

My thought on it is that if the encoder can generate error log, if he does, he sould also never give a Blue light in this case. Bad programming.

Regarding the "publish" part, I was meaning sending a file to a publisher to get the file on Tidal service. If I'm not wrong, no one did such a thing before the OP

 

[DimitryZ]

In the US at least, a person is considered innocent by default unless proven guilty. What you posted on SBAF as seen in one of Amir's posts here is implying unethical behavior by Amir, for which by your own statement you lack hard evidence. The default process should not be to quiz Amir to get him to affirm his innocence and await his request that you edit the posts. My take on it is that you ought to edit the posts without delay, retract the defamatory-type aspersions on Amir's character and ethics, and apologize, but of course I am nobody to tell you or anyone else what they should or should not do.

The posts in question cast you in a bad light. My impression is that you were connecting dots that do not actually have a connection between them, perhaps induced by the stress you must feel from multiple attacks on your YT videos by multiple players and random internet denizens. JohnYang1997, a Topping rep here has posted on ASR multiple times that based on his experiment he believes MQA is messed up, and has also provided a plausible explanation for why many DAC manufacturers are hastening to add MQA decoding capability to many of their products. Amir has always (I have only read many of his posts on ASR, not on previous forums) been extremely transparent. He has already affirmed multiple times in the past that ASR, and he on ASR, have no hidden financial ties to the companies whose products he reviews. Where his company Madrona sells speakers or other equipment he may review, he has stated that up front in his reviews. He has also recused himself from awarding a panther in any cases where there may be what could be viewed as a conflict of interest. I really see no room for casting aspersions on Amir's ethics as a reviewer or leader of ASR. Amir has every right to a personal opinion about what he views as the elegance of the MQA algorithm or his respect for Bob Stuart's research contributions, and to try to present facts on what he believes are wrong various people's posts. And of course, posters including you and I are free to argue back with facts if we disagree with him or anyone else. Also, just to clarify, I have no ties to Amir or the mods, other than being an ASR member. I have never spoken with any of them or exchanged any e-mails, or submitted any items for Amir to test. They do not know my personal identity (I hope ), nor I theirs, other than some of what is known publicly about Amir.



I agree with you completely here. The term "lossless" with regard to digital data has an unambiguous meaning (mathematically lossless, bit-for-bit reversible transmission, compression, etc.) to informed lay people and experts (in data compression, computer science, math, etc.) alike. MQA marketing has been dancing around the term, using it to imply overall losslessness by careful wording, and in at least one case outright calling MQA lossless. The traction that your videos have gained has seemingly caused MQA marketing to walk back their use of the term lossless. Wait, they did not walk it back with explanations or apologies, it simply transformed magically overnight into "better than lossless". Many kudos on your achievement here. Your experiment and videos on it succeeded where previous analyses by Archimago, Mark Waldrep, and a couple of others did not get to the impact level of exposure for that to happen.

I agree with Amir that we should be aiming for better than CD quality. As Amir pointed out, sticking with CD quality makes it pointless to rank DACs and amps with SINADs higher than 96. From reading most of the info in this and previous MQA threads, I would opt for properly encoded 24-bit 48kHz FLAC, even though it may be lossy relative to modern digital masters. Even though my old ears could not reliably distinguish between 320kbps mp3 and CD quality in the online NPR 2015 test . I see no substantive advantage for the music consumer of MQA over 24/48 FLAC. I see only significant disadvantages such as the cost of MQA and its DRM-like nature that hampers my convenience of use. FLAC, on the other hand, thanks to the generosity and hard work of the good folks over at xiph.org, is free and universally available now, except perhaps in Apple's walled garden where ALAC is an option instead. Let's not kid ourselves about the cost. MQA wants to eat at the same table as the rest of the audio industry, and in a big way. If they succeed and thrive in a big way, ultimately consumers will be paying the MQA tax that I view as bringing no proven benefit with it. I agree with others that MQA is a solution to a problem of more than a decade ago. I also prefer to stick with 24/48 FLAC that can represent loud high-frequency tones in the audible band that rise above the MQA orange triangle than get low-amplitude ultrasonic components of music that audibility experiments tell us would be below the threshold of hearing for most of us. Yes, MQA can probably handle music above the orange triangle by configuring the encoder for it, but then the MQA file size will increase too. Unless the MQA time-deblurring apodizing filter is proven to improve audio transparency very significantly, but even that is unlikely to sway my preference. MQA has tracked the envelope of music for a million or a few million tracks. Apple and Amazon claim libraries of 70 to 75 million tracks, and who knows what music may be released in the future.

Well, I think the conversation in this thread has moved beyond the use of the term lossless in MQA marketing material, and is diving into demystifying what it is that MQA actually does. Several expert people are participating, and I think you (and I) need to dig deeper into the technical aspects if we want to understand further posts on it.

Replying to your last paragraph only, there is no mystery of what MQA technically is or does. Archimago's analyses from 2017 have long answered this question definitevly.

That this continues to be an active "mystery" is itself astonishing

 

[Raindog123]

Besides, pure sine waves would have been perfectly fine and suitable to test with for OP since music can be decomposed into them. What is not proper is square waves, Impulses, noise going to 44 kHz, etc.. You do NOT see me use these to test hardware -- which doesn't care what it is told to play -- yet is used by OP with technology that absolutely cares what it is fed (MQA).

As I have explained repeatedly, what you testing matters with respect to testing. For lossy codecs for example, listening tests rule and test tones would play a tiny, tiny role (to check flatness of response and such). There are specialized objective measurements such as PSNR but they are a) not simple hardware test tones we run and b) are rarely used.

@amirm. It also was pointed out continuously that what you suggest is NOT how good testing works. Instead, you create test cases that intend to break the system. Put it out of the comfort zone. This is how you find (or confirm) the operability envelope. Eg, ask our new friend @DimitryZ regarding his aerospace industry. Though I suspect you know that already.

The real question is, why are the true interested parties not stepping in to collaborate on the tests and drive them in the 'right' direction?

 

[DimitryZ]

My thought on it is that if the encoder can generate error log, if he does, he sould also never give a Blue light in this case. Bad programming.

Regarding the "publish" part, I was meaning sending a file to a publisher to get the file on Tidal service. If I'm not wrong, no one did such a thing before the OP

Sure, agreed.

I am pretty sure manufacturers who are against MQA have long done that, because it's pretty short money. The fact that they never published it hints that the results may have been very good.

But we will find out ourselves!

 

[amirm]

Isn't it ironic to defend MQA by saying that simple tones are not natural sound so it is not suitable to test the performance of MQA, on a website that is all about driving sine waves through equipment and measuring the SINAD?

Let me show you what *is* ironic. OP is a subjectivist and puts his sighted unreliable listening way ahead of any measurements. Just check out this review of Schiit Magnius which he did in February of this year. I gave it a glowing review and this is his take on it. The title is, ready for it? "
Schiit Magnius Review - Measurements don't tell the whole story.....

"the dynamic range feels quite compressed...it doesn't come across in measurements"

"... this is a fatiguing amp to me..."

"there is slight graininess and lack of separation...."

"there is not a lot of texture [in bass]..."

"timbre is not great in this amp..."

"mid texture and detail is not good..."

And this is in the first 7 minutes of this video!

It ends with some nonsense about how Schiit was "forced" into designing Magnius to please people who look for good measurements. And that this can't be their fault, i.e. it is our fault!

Now he says measurements matter with MQA? If measurements are not important, why didn't he do a listening test of MQA instead? Let's see if it has or has not any bass texture and speed. What it is timbre doing. Does it have texture? Or does it not? We need to know!

My suggestion: be very careful who you are holding up as your expert witness. And making snide remarks. When the truth is not on your side, it can backfire big time!

 

[MDT]

@amirm. It also was pointed out continuously that what you suggest is NOT how good testing works. Instead, you create test cases that intend to break the system. Put it out of the comfort zone. This is how you find (or confirm) the operability envelope.

In the case of MQA, this is possibly even more relevant. Since we don't have access to the full encoder/decoder stack, the whole thing is just a black box. Figuring out what's in the box often requires pushing it to extremes.

 

[DimitryZ]

@amirm. It also was pointed out continuously that what you suggest is NOT how good testing works. Instead, you create test cases that intend to break the system. Put it out of the comfort zone. This is how you find (or confirm) the operability envelope. Eg, ask our new friend @DimitryZ regarding his aerospace industry. Though I suspect you know that already.

The real question is, where are the true interested parties are not stepping in to collaborate on the tests and drive them in the 'right' direction?

There many different test methodologies for different purposes.

For example, we typically will test to the system operational envelope for prescribed length of time to simulate lifetime of vibration exposure. After this test, the system is considered qualified for vibration.

A different test methodology is used to define system weaknesses. Here we increase test levels until we get failures. This method is useful in understanding system failure modes.

I am generally not a big fan of the latter. In a modern high end design house the process should work well enough to consistently produce high quality designs adequate for performance in specified environment. Their design envelope and failure modes should be well known before qualification testing.

Testing to failure in most cases in my line of work (defense) is not needed. Fun though.

 

[DimitryZ]

In the case of MQA, this is possibly even more relevant. Since we don't have access to the full encoder/decoder stack, the whole thing is just a black box. Figuring out what's in the box often requires pushing it to extremes.

It is common practice to test stuff under it's normal environment before breaking it.

 

[Grooved]

Sure, agreed.

I am pretty sure manufacturers who are against MQA have long done that, because it's pretty short money. The fact that they never published it hints that the results may have been very good.

But we will find out ourselves!

Or not bad enough to make a real difference, let's not forgot this possibility to stay neutral.

Anyway, my feeling is that I get a different master, because after thinking it twice, the main thing I hear can look like adding a transient exciter, and if you add such a thing to a master, it's not the same master anymore. It can technically be a different thing, but the listening result gives me this feeling. There are other things but it's the main difference I see. And won't say it's better or not, because it should be an artistic choice.
Even subtil, it can be heard, an enough to not match a small difference like we see in null test of the files, that's why I'm not sure anymore there's more change on the file itself than on the rendering. Need more test.