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

[MDT]

You mean the likes of Apple and Amazon are going to cave, start paying royalties to MQA and force people to buy MQA hardware? Really? What do you know that these companies don't know? That MQA value proposition is that strong that they are already defeated by offering open-format high-res content?

I think perhaps you've replied to the wrong person, or just failed to properly read my post.

 

[amirm]

I appreciate your comment here, Amir, but with respect I don’t see evidence that the more technically expert MQA critics in this forum are “on the wrong side of technology understanding.”

They freely move from their area of expertise, to areas that are not. Discussion on DRM is one. None of them understand the technology, dynamics, content owner relationships, IPR landscape, etc. They are also so emotionally driven on this topic that even the areas that they should understand are misstated.

They also quickly move from technical to business side of things with again, no professional experience in that domain.

So there is a lot to correct for me who has lived and breathed all of these format making efforts from soup to nuts.

 

[amirm]

I think perhaps you've replied to the wrong person, or just failed to properly read my post.

No I didn't. You wanted to scare people by saying MQA is going to take over the world so people should be afraid of it despite its tiny piece of land it exists on today. I answered that the big boys have already spoken and didn't adopt MQA so your prediction that the world is going to adopt MQA is dead wrong.

 

[Jimbob54]

What issue is MQA trying to solve? If Amazon and Apple can stream lossless...was this not possible previously?

You're left with time smearing and provenance. One of which no one understands and the other no one believes.

 

[amirm]

From https://bobtalks.co.uk/blog/mqaplayback/origami-and-the-last-mile/

View attachment 132366
Therefore MQA is not lossless.
In addition, IMO it is safe to assume, that the "safety margin" does not extend beyond the gray rectangle (i.e. CD spec), otherwise they would surely mentioned it, and therefore MQA is not lossless even for the 16/44 part of the hires file.

QED and no youtube video was necessary

Eh, no. MQA in that graph is saying that will encode and deliver everything that is in that music clip as defined by its peak signal at every frequency. And they even encode more than they need with that guard band.

That they don't encode empty space (CD rectangle or otherwise) that is not used by said music is of no concern to any unless you want to burn bits to hold nothing.

Does everybody understand this? It is key to their response. They are saying that their format is "lossless" because they encode the music and give it all the bits it needs including its ultrasonic spectrum.

The only argument against them then is that they are not lossless because they don't encode empty space. Which is what OP attempted to do (filled the empty space and then wondered why it couldn't encode it).

 

[BinkieHuckerback]

'Smearing and Provenance' - they're my solicitors

 

[Grooved]

...
It is for this reason that sometimes I stop posting here. It seems that people don't want better understanding of technology. They just want to hear who is for and against MQA. The concept that I don't care about that battle doesn't compute with them. The fact that I have done a video talking about how undecoded MQA damages the content doesn't penetrate. The fact that I have said MQA is likely going to die with 2 out of 10 chances of success goes in one ear and comes out the other. The fact that I predicted major players like Amazon and Apple would do high-res without MQA and that has come true, seems to be long forgotten as well. They just come out and say what you said: "Amir is strongly defending MQA." Facts be damned.

So it is clear, OP is not technical. Neither are vast majority of people opining about this discussion. So it is natural that they say things that are simply incorrect. Do you really wish these incorrect technical assessments to be left alone here without me correcting them? I sure hope not or we are wasting our time here and just seem to want political victories, not a search for understanding.

I understand this argument, we can't make affirmation on things we are not sure, and it would be great if nobody was taking a post of anyone of us out of context to bring any different conclusion on other websites, which is hard to control, at least if we don't take care of what we say here.
Just an example, someone here say "there are 16bit MQA on Hifi setting in Tidal" and you will find hours later that "Tidal provide only 16bit MQA in Hifi setting" which is wrong... this is the kind of things I try to avoid, and since there are files in 16bit MQA and FLAC, you need to always say it clearly, each time.

Now, even if the OP test was not perfect, can we agree on some things like confirming that you can get a Blue light (it could have been Green which is supposed to be less confirmed) even on a file that were not technically acceptable and got an error the encoder log ?
Yes, we can still say "hey, it's was not a real music file, it's created a bug, that's why it got the Blue light", on which I will quote Brian Lucey who confirmed that part of the masters he created are in MQA with Blue light in Tidam while he never confirmed these files. As it may be done by the label if they are owner of the master, there's a more important point : he confirmed it was not souding like the masters he created, and he explained it clearly in an article that summarize perfectly the main thing around MQA, and I would encourage anybody to read this article if it was not done yet, and even if explains more technical details outside of this article :
http://fairhedon.com/2017/11/05/an-interview-with-mastering-engineer-brian-lucey/

He had for example said that used differently it could provide an interesting thing, and was labeled pro-MQA because of that by some people.
He also got words from Meridian employees at this time telling him that the goal was to find something that can, first, generate a lot of money, but it's not a problem for me by itself, everyone company wants to generate money, but I don't like doing with lies in the marketing materials that consumers can access to.

 

[dc655321]

I'm curious why the OP did not consult anyone with technical chops to advise on a better, or at least less controversial, test plan?

IMO, he had the right idea, just poor execution.

 

[Grooved]

What issue is MQA trying to solve? If Amazon and Apple can stream lossless...was this not possible previously?

The problem is what they say is different depending to who they talk to
- to consumers, they say that they provide what the artist wants you to listen, which they can't by encoding a PCM master if their technology changed the sound a bit. If the mastering engineer is listening through a MQA system while creating the master, it may it possible, but it's not how it's done right now. As it changes the sound, the mastering engineer would have to use different processing than what he did.
- to labels, they say "you should never give your real master to streaming providers as it's your crown jewels

 

[pjug]

The only argument against them then is that they are not lossless because they don't encode empty space. Which is what OP attempted to do (filled the empty space and then wondered why it couldn't encode it).

How is this lossless or without coloration, even the way you are looking at it?

Figure 13: This graph shows both the impulse and frequency responses of the MQA system in a complete end-to-end configuration. Note the dual scales.
[from https://www.soundonsound.com/techniques/mqa-time-domain-accuracy-digital-audio-quality]

 

[amirm]

Yes, we can still say "hey, it's was not a real music file, it's created a bug, that's why it got the Blue light", on which I will quote Brian Lucey who confirmed that part of the masters he created are in MQA with Blue light in Tidam while he never confirmed these files. As it may be done by the label if they are owner of the master, there's a more important point : he confirmed it was not souding like the masters he created, and he explained it clearly in an article that summarize perfectly the main thing around MQA, and I would encourage anybody to read this article if it was not done yet, and even if explains more technical details outside of this article :

I don't encourage anyone to read anything from Brian Lucey on any listening tests results. I know because I have discussed it with him. See this starting post for example on what he has written: https://gearspace.com/board/mastering-forum/1171365-mqa-discussion-denver-rmaf-10.html#post12952542

"Where as mastered for iTunes is harmonically cold and loses some low volume/low end information, actually altering the groove to make everything sound like a nerdy white wedding band, MQA brightens the high-mids in the Mid section while thinning the low-mids on the Sides. "

You believe this?

Harmonically cold? MQA brightens the high-mids? Thinning low-mids? You really believe this kind of subjectivist nonsense? I sure hope not.

He needs to start learning to use blind controlled test before he speaks on any topic like this.

This is the problem with MQA. People become completely different when it comes to this topic. If I had brought up those quotes in any other context, people would throw stones at Brian Lucey. Have him fight with MQA and all of a sudden he is a beacon of audio science when it comes to matters related to MQA. Don't do that please.

 

[amirm]

How is this lossless or without coloration, even the way you are looking at it?
View attachment 132457
Figure 13: This graph shows both the impulse and frequency responses of the MQA system in a complete end-to-end configuration. Note the dual scales.
[from https://www.soundonsound.com/techniques/mqa-time-domain-accuracy-digital-audio-quality]

Huh? An impulse signal is not a real one and certainly doesn't represent any music. We use impulse response as a theoretical construct because it has infinite spectrum of equal strength. It has little to do with reality. This is why you don't see me posting impulse response measurements.

 

[levimax]

You mean the likes of Apple and Amazon are going to cave, start paying royalties to MQA and force people to buy MQA hardware? Really? What do you know that these companies don't know? That MQA value proposition is that strong that they are already defeated by offering open-format high-res content?

Who do you think has the power to decide MQA's fate, Apple, Amazon, etc. or the publishers? I think the fact that Warner Music is all in on MQA and only provides many titles in MQA is telling. I am sure the giant streamers don't want MQA but for me the Warner catalog is important and I would not subscribe to any service that did not have those titles. If MQA gets a few more big labels to go all in and continues with it's consumer marketing and creates consumer demand I am not sure the big streamers will have a choice... especially if the large publishers only offer their titles in MQA. Obviously I have no idea how it will play out but clearly MQA knows what it wants and seems to be gaining a lot of momentum with the publishers, who are the real customers of MQA... not the consumer or streaming services.

 

[pjug]

Huh? An impulse signal is not a real one and certainly doesn't represent any music. We use impulse response as a theoretical construct because it has infinite spectrum of equal strength. It has little to do with reality. This is why you don't see me posting impulse response measurements.

I was talking about the rolloff in frequency response from the rendering.

 

[amirm]

Who do you think has the power to decide MQA's fate, Apple, Amazon, etc. or the publishers? I think the fact that Warner Music is all in on MQA and only provides many titles in MQA is telling.

No it isn't. It is spotify, apple, amazon and maybe google. Record labels lost their power when Steve jobs forced them to a) break up a lucrative album into individual songs and b) give up on DRM. Today, you can show up as Joe nobody, write a check for advanced royalties, agree to MG (minimum guarantee of royalties) and you too can have a music service. They simply don't care what you do with their content. You give them the check, they give you content.

You are way, way behind times in understanding of how the music industry works today.

Now, major movie studios, yes they held on to their power although Netflix has weakened that some with their leverage and original productions. Record labels are nothing today compared to who they were.

As to Warner, they will give you MQA, or anything else you want. They made a commitment to the technology and got some warrants in return. Standard deal that has gone bust for many, many start-ups who have made similar deals with labels.

 

[amirm]

I was talking about the rolloff in frequency response from the rendering.

Oh sorry. I can't defend their choice of slow roll off filter. It is a different matter than how the bits are encoded and delivered which is what the OP post is about. He performed digital analysis, not analog that would involve that filter.

 

[amirm]

BTW, making a deal with one label will upset their competitors and they will then refuse to help you. Why would Sony Music for example commit to MQA make them famous and go public, only to enrich Warner music? I have been in those chairs and heard the very warnings and arguments. Don't assume the Warner endorsement is a good thing beyond the initial PR push. Naive start-ups think it is big deal but soon find out that the label won't do much to help them. Has Warner come to defense of MQA on any of these food fights? They have not.

 

[pjug]

Oh sorry. I can't defend their choice of slow roll off filter. It is a different matter than how the bits are encoded and delivered which is what the OP post is about. He performed digital analysis, not analog that would involve that filter.

Isn't the rendering the meat of the matter (considering just sound fidelity)? Otherwise it is just a different way of encoding the same information as ordinary PCM, just more complicated and a little less efficient compared to something like 17/96.

 

[BinkieHuckerback]

So what is the point of MQA? What is it jeopardizing? Is this a 'storm in a teacup'?

 

[jensgk]

I am when there is ambiguity by using the term "mathematically lossless." ....
I suggest adopting this terminology so that it is clear what is being talked about.

No.

When the term "lossless" codec or compression is used on this site or in every other audio discussion forum, it is by default meant to be "mathematical lossless", absolutely no reason to specify "mathematical".

So off course this is also what everybody thinks, when MQA claims "losslessness".

The term "perceptually lossless" is very rare, and on this site, it has almost exclusively been used in this thread, in all other instances lossless of course means "mathematically".

If in doubt about the common understanding of a term, it is useful to refer to Wikipedia. In the Wikipedia article about "Lossless compression" the term "perceptually" is not found.