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

[awdeeoh]

I have access to and the capability to program a number of modem - FPGA/SoC plus DAC - development boards. At any bandwidth relevant here (and beyond). Mostly Xilinx (eg, large UltraScale+) with Vivado tools (and GHz-class DACs). If I'd get access to a [MQA] decoder SW (FW), I can do pretty much anything with it, very quickly and efficiently.

I have no idea what it might take (or even if it is possible at all) to get the source 'for evaluation & analysis' purposes... But if there is anyone (@mansr ?) out there who sees the point, let me know and we'll discuss...

You can get the source, by reverse engineering it.

and you can do all what you want as long at you don't do it for commercial purposes.

 

[jensgk]

Are you listening to explanations or just repeating talking points?

What conspiration theories are you implying here?

Whoever is doing the encoding is a mastering engineer with many tools.

That is not how MQA were/is used on Tidal, since @GoldenOne clearly is not a mastering engineer, and was allowed to do the encoding without getting the error codes.

If you feed it a full spectrum of random noise to 44 kHz, no, it can't encode that, nor is it broken since their customers are not asking for such an application.

There are test tracks with noise on Tidal.

 

[Raindog123]

You can get the source, by reverse engineering it.

Despite what some newcomers here say, I still have a rather demanding day job. So, looking for accelerators...

 

[amirm]

That is not how MQA were/is used on Tidal, since @GoldenOne clearly is not a mastering engineer, and was allowed to do the encoding without getting the error codes.

Look, I am going to answer this but if you ask again for something that has been answered a number of times in this thread, I will give you a Reply ban (and a refund on your donation if you need it).

The service OP used was put together so that smaller labels could quickly and cheaply encode MQA content. Such automated service is NOT the norm for vast majority of MQA content that is either encoded by MQA itself or a mastering house. Both of these groups have tools with knowledge of how to run them, what parameters to set, etc.

 

[jensgk]

The service OP used was put together so that smaller labels could quickly and cheaply encode MQA content. Such automated service is NOT the norm for vast majority of MQA content that is either encoded by MQA itself or a mastering house. Both of these groups have tools with knowledge of how to run them, what parameters to set, etc.

That might be so, but that means, that in general I cannot trust that an MQA file was encoded correctly. So the process is broken.

 

[amirm]

That might be so, but that means, that in general I cannot trust that an MQA file was encoded correctly. So the process is broken.

And you trust what instead? "Lossless" Flac? Here is what is inside that:

We used to have an expression while I was at Sony: "if you knew how the sausage was made, you would never eat it!" You have no assurances of anything in music you buy or consume. There could be total crap in there as far as loudness compression, noise, distortion, etc. You accept that but want some kind of guarantee from MQA that it did or did not give you the spectrum from 22 to 44 kHz? Get real.

 

[jensgk]

And you trust what instead? "Lossless" Flac?
...
Get real.

This is a technical discussion. Can I trust that MQA lives up to all its claims, when processing the master file submitted by the mastering engineer?
How the master file was made and what it contains is irrelevant.

 

[amirm]

Absolutely. Though your professional credentials seem ever thinner as you post.
MQA simply sounds better than LPCM counterpart.
Now go back to your golden parachute and enjoy your retirement.

This kind of language is not acceptable. Please tone it down and stay focused on technical matter.

 

[amirm]

This is a technical discussion. Can I trust that MQA lives up to all its claims, when processing the master file submitted by the mastering engineer?
How the master file was made and what it contains is irrelevant.

It is your question that is irrelevant. Content, regardless of its distribution format, comes in all levels of quality. It is your job to pick what pleases you and how much fidelity you expect in there.

 

[jensgk]

It is your question that is irrelevant. Content, regardless of its distribution format, comes in all levels of quality. It is your job to pick what pleases you and how much fidelity you expect in there.

But what about "as the artist intended" and "better than lossless"?

EDIT: I have been reply banned, so cannot participate anymore.

 

[amirm]

I'm happy to take you up on this.

Let's say the Holo May vs topping e30 if that suits? Both measure >110dB sinad with no audible concerns raised on your review for the e30, will set the may to OS so that there's no inherent FR difference.

I'll volume match to 0.01dB etc.
Any other conditions or things you'd like me to put in place let me know and I'll accommodate them if possible. How many runs, any other checks, or detail the full procedure for the test if you'd like.

The thread dedicated to this is here: https://www.audiosciencereview.com/forum/index.php?threads/blind-test-we-have-a-volunteer.23857/

 

[awdeeoh]

But what about "as the artist intended" and "better than lossless"?

same can be said to albums that have been remastered and offered in other streaming services with the marketing "as the artist intended".

"better than lossless" that's exclusive to MQA.

 

[UliBru]

The service OP used was put together so that smaller labels could quickly and cheaply encode MQA content. Such automated service is NOT the norm for vast majority of MQA content that is either encoded by MQA itself or a mastering house. Both of these groups have tools with knowledge of how to run them, what parameters to set, etc.

I wonder about how Roon is handling the MQA content. Especially when Roon decodes the MQA track, applies some DSP treatment and encodes the result back to MQA for the connected MQA DAC. As this is an an automatic process anyway.
IMO it clearly breaks the end-to-end condition.

 

[mSpot]

The service OP used was put together so that smaller labels could quickly and cheaply encode MQA content. Such automated service is NOT the norm for vast majority of MQA content that is either encoded by MQA itself or a mastering house.

How is Warner Music doing their MQA encoding of many millions of tracks? Surely it must be an automated service that runs quickly and cheaply.

 

[abdo123]

I wonder about how Roon is handling the MQA content. Especially when Roon decodes the MQA track, applies some DSP treatment and encodes the result back to MQA for the connected MQA DAC. As this is an an automatic process anyway.
IMO it clearly breaks the end-to-end condition.

MQA created a special condition for this, this sort of content will be upsampled by the MQA DAC and get the MQA reconstruction filter, but it won't show MQA it will show something else, in the case of the Gustard X16 it is 'OFS'.

 

[JoshF]

First message, long time lurker.

After reading and trying to understand almost all messages what I see is a soon to resolve tension, so let's assume for the sake of argument that we have perfect reproduction systems and perfect ears:

1- MQA as a technology that enables record labels to retain the masters (propierty remains very important) while giving consumers the illusion of getting them
2- In a world where disk space and network costs are increasingly negligible, consumers paying a subscription (not getting files forever) want the real thing to play with

Technical problems or advantages aside, sorry guys, but for me the real battle is what I described. Like it or not MQA has a function for labels.

Disclaimer: former Tidal user here, Qobuz user for 2 years now. Veni, vidi, vinci.

Please forget my message if it's too obvious.

 

[awdeeoh]

dear all,

we’ve been provided with an MQA Real Time Encoder, and along along with other necessary tools to measure MQA. in tests thus far, we’ve been able to verify almost all of MQA’s claims, except i wish they would qualify the use of the word “lossless” as it’s more “lossless” in the analog domain.

as an aside, i’ve been listening to MQA-encoded music for over 3 years, and i’m very enthusiastic — among other things, i find listening to MQA-encoded streams & files with a good D/A much less stressful for extended listening.

testing continues, but it must be supported by data, preferably llarge numbers of subjects.. we hope to do more livewebcasts in the fall using MQA. we’re still waiting for an A-V streaming service (other than YouTube, of course, who don’t give a fiddlers fart about audio).

meanwhile, i’ve bought 20 MQA DACs to distribute to our faculty and SR2 students this fall, the better to be able to listen better remotely if that becomes necessary.

George Massenburg
McGill University

 

[UliBru]

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

says:

It is clearly stated: All the salient music-related information exists inside the Orange triangle, which MQA protects very striictly.
The upper orange line has a -2dB/kHz falling slope.
So e.g. the peak level at 24 kHz is not allowed to exceed at least -48 dB.

The paper "A Hierarchical Approach for Audio Capture, Archive, and Distribution", by Bob Stuart shows a similar picture

Here we have a picture of the peak levels represented by an orange line with a slope of -1dB/kHz. According to it the max. allowed level at 24 kHz would be -30 dB. This clearly exceeds the allowed MQA level.

So does natural music already hurt the MQA conditions?

 

[firedog]

i believe i have already pointed this out, but i think its a given that marketing is going to embellish the product in some way. after all everyone wants to sell their product and claim their product is the bestest.

i hardly think MQA is unique in this regard.

There's embellishing and there's straight out and out lying. MQA lies. About lossless, about "authenticated", etc. It's as simple as that. Therefore, I wouldn't take anything they say about their product on faith. The opposite: a good chance they are lying there, too.

 

[firedog]

Question has nothing to do with Tidal. It was about mastering MQA CDs. Physical media at mandated 16 bits, not the normal 24-bit files MQA produces for online distribution in Tidal.

As to Neil Young, I don't care what he said. He didn't conduct any controlled test to show that his content had degraded, or even changed.

He didn't need to. He simply claimed they weren't "authenticated" and weren't his masters. Simply by running them through the MQA process they are changed, and aren't "masters". As he apparently has control of the masters, Tidal had to remove the MQA versions when he asked..