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

[DimitryZ]

OK maybe I should not have said it s typical. Maybe the decoder gets updated. Maybe analog output varies all over the place depending on the DAC. So then which is the authentic sound heard in the studio?

Actually your second point is incorrect. The final analogue output should be as DAC independent as possible.

I have four MQA DACs (all right I know) and they do come close, but not quite.

 

[DimitryZ]

Here's a word for you: dilettante.

Beyond that, I'm content to let anyone who's interested read our exchanges - and your quite entertaining attempt to try to tell @lucretius what lossiness is - and decide for themselves who does, and who does not, understand.

I am a dilletante in this field, absolutely. MIT Aero/Astro '86 (course 16 for fellow sufferers). My real name is in my profile. Engineering Fellow, Raytheon - structural engineering.

You thought you were insulting me, but you didn't. I am very comfortable with my limitations and am working to improve.

 

[pjug]

Actually your second point is incorrect. The final analogue output should be as DAC independent as possible.

I have four MQA DACs (all right I know) and they do come close, but not quite.

It was a bit tongue in cheek. The decoded MQA tracked the PCM so I don't know why you were questioning MansR's decoder. On MQA DACs conforming to have the same out for MQA tracks, I wonder though. Same enough that my ears wouldn't know the difference, probably, but if you compare them with measurements I bet it looks a little shaky for MQA and the idea of authentic sound guaranteed.

 

[tmtomh]

I am a dilletante in this field, absolutely. MIT Aero/Astro '86. My real name is in my profile. Engineering Fellow, Raytheon - structural engineering.

Thanks for telling us your area of expertise - no one was aware, since you haven't already mentioned it several times.

As to the significance of a self-confessed dilettante joining a forum and within their first 24 hours posting more than 60 comments, all in a single thread, the majority of said comments claiming that others don't know what they're talking about, I am again content to let those who are interested read and decide for themselves.

And @pjug , he's questioning @mansr 's decoder because questioning the knowledge and competence of anyone who makes a comment critical of MQA is, thus far, the apparent purpose of his participation in this thread, based on the totality of his comments here.

 

[DimitryZ]

It was a bit tongue in cheek. The decoded MQA tracked the PCM so I don't know why you were questioning MansR's decoder. On MQA DACs conforming to have the same out for MQA tracks, I wonder though. Same enough that my ears wouldn't know the difference, probably, but if you compare them with measurements I bet it looks a little shaky for MQA and the idea of authentic sound guaranteed.

I stipulate that Mansr's decoder is perfectly fine, if a bit old. He is technically excellent, though he hates MQA. His technical work should be beyond reproach. Him and I agree on nothing, but I fully endorse his technical expertise and engineering honesty.

I was only questioning the rendered part. Capturing this precisely seems tricky, as the legal option would require an ADC that was previously precisely matched to the DAC analogue output.

 

[DimitryZ]

Thanks for telling us your area of expertise - no one was aware, since you haven't already mentioned it several times.

As to the significance of a self-confessed dilettante joining a forum and within their first 24 hours posting more than 60 comments, all in a single thread, the majority of said comments claiming that others don't know what they're talking about, I am again content to let those who are interested read and decide for themselves.

And @pjug , he's questioning @mansr 's decoder because questioning the knowledge and competence of anyone who makes a comment critical of MQA is, thus far, the apparent purpose of his participation in this thread, based on the totality of his comments here.

Of course. I stand behind my posts.

You overarching point is that new members shouldn't post much and generally defer to older members?

Sounds like the Russian army...

 

[tmtomh]

Of course. I stand behind my posts.

That's good, because someone has to.

 

[pjug]

I stipulate that Mansr's decoder is perfectly fine, if a bit old. He is technically excellent, though obsessed with anti-MQA emotion. His technical work should be beyond reproach. Him and I agree on nothing, but I fully endorse his technical expertise.

I was only questioning the rendered part. Capturing this precisely seems tricky, as the legal option would require an ADC that was previously precisely matched to the DAC analogue output.

So you think all MQA DACs render the same, or you think they don't?
edit: didn't @Archimago show that rendered output varied a lot with different MQA DACs? I'll try to find that.

 

[DimitryZ]

So you think all MQA DACs render the same, or you think they don't?

That's a complicated question.

My four DACs come close, but I do hear small differences.

These are at or below reconstruction filter differences in each DACs - ie really small.

 

[pjug]

That's a complicated question.

My four DACs come close, but I do hear differences.

These are at or below reconstruction filter differences in each DACs - ie small.

If you hear differences don't you think the differences must be very large compared to what MQA is trying to sell us on the authentic bit? More likely though you are imagining it.

 

[awdeeoh]

That is misleading. A cd release from a DXD master is not lossy as far as 22kHz and below are concerned.

Same can be said to MQA.

 

[DimitryZ]

If you hear differences don't you think the differences must be very large compared to what MQA is trying to sell us on the authentic bit? More likely though you are imagining it.

Well, in my extensive listening experience, DAC filter differences (small but audible according to peer reviewed AES paper by Stuart/Craven and my listening experience) are more audible than MQA-compatible DAC variations - ie the latter is very small.

Their goal is to make the DAC invisible. In my very high resolution system (Emotiva and Eminent Technology speakers), DAC differences are perceptible, but barely so.

But is your question about the difference between full MQA reconstruction vs. LPCM version? And it's complicated. At PFM, I have consistently stated that MQA is within personal/system preference to LPCM. That has been my experience over several years

In audiophile terms the difference is similar to changing interconnects - ie extremely small in ASR dictionary.

 

[amirm]

Yes you add dither (various different alghorithms including noise shaping) -- there are fewer steps available to map the amplitude levels to -- 65,5346 steps (16 bit) vs 16,777,216 steps (24 bit). Do you think you can tell the difference between those 16M + steps? Are you going to tell me that the 16 bit file is not lossless?

Me telling the difference? Is that the definition of "lossless" to you? But yes, I have passed double blind tests of 16 bits versus 24 bits. And psychoacoustics says the dithered 16 bit noise floor will be audible. The undithered version creates distortion which is worse.

The filtering of the spectrum is non-ideal and as I have explained, paper published shows that it can be audible.

So don't say CD version is lossless conversion of high-res file. Say that it is good enough and you would be fine. Otherwise you are far worse than anything MQA marketing plan has said.

 

[Newman]

@amirm how about “CD doesn’t apply a lossy masking algorithm to the audible music spectrum”? Because TBH that is the convention when distinguishing CD from “lossy compression”.

Where does MQA sit against that criterion?

 

[amirm]

@amirm how about “CD doesn’t apply a lossy masking algorithm to the audible music spectrum”?

What is a lossy "masking" algorithm?

 

[restorer-john]

What is a lossy "masking" algorithm?

One of these:

 

[DimitryZ]

It was a bit tongue in cheek. The decoded MQA tracked the PCM so I don't know why you were questioning MansR's decoder. On MQA DACs conforming to have the same out for MQA tracks, I wonder though. Same enough that my ears wouldn't know the difference, probably, but if you compare them with measurements I bet it looks a little shaky for MQA and the idea of authentic sound guaranteed.

And that's why I question MQA rendered provenance. That's always tricky, by design..

 

[Mulder]

So don't say CD version is lossless conversion of high-res file. Say that it is good enough and you would be fine. Otherwise you are far worse than anything MQA marketing plan has said.

Maybe you can then say that CD quality is the best "compression" we have of High-Res? As far as I know, there are very few correctly performed and documented listening tests where listeners can distinguish "High-Res" from CD quality. And then I think of all these old analog recordings that have been transferred to CD. Already there, should the CD surpass the original in terms of capacity in terms of noise and other things?

 

[ergre]

They have announced they are going to offer high-res and as such are ingesting it as we speak. It takes a long time to do that and catalog it before turning on the service. Amazon is there already by the way. No way did someone reach back into label archives to modify their bits that they send out because someone is distributing the MQA versions.


What "companies?" Tidal has signed such a contract but they don't own the original bits, the labels do. Major Labels have not licensed MQA since they are not releasing MQA content under their own name (they have agreement with MQA to "support" MQA but that is different than being a licensee). Even if they had licensed it, there is no universe under which they would agree to such terms.

The terms of such agreement is confidential which means whoever has started that rumor doesn't know it to be true either.

Ok. I think you are right. I am still never going to buy a MQA DAC or music though. CD quality is good for me and the secritivness and marketing MQA uses are not for me.

 

[Mulder]

What about MQA if you need DSP functions? Can I apply DSP to MQA files in the same way as with PCM files?