MQA: A Review of controversies, concerns, and cautions

[Thomas savage]

My point was this:

WHAT DOES MQA BRING TO THE TABLE FOR THE CONSUMER (THAT WASN’T ALREADY THERE)?

Unless they do a great job remastering the files , nothing.

 

[amirm]

My point was this:

WHAT DOES MQA BRING TO THE TABLE FOR THE CONSUMER (THAT WASN’T ALREADY THERE)?

Streaming of > 16/44.1 kHz through Tidal in US.

 

[amirm]

And potentially hand-tuned new masters created by MQA team. Some of the MQA content I have heard compared to non-MQA sound remarkably better, indicating better mastering.

 

[amirm]

I don't think the fact that there is a demand for something indicates anything positive about a product other than perhaps successful marketing. There is still a healthy (groan...) demand for cigarettes despite the fact that they are likely to give you cancer, emphysema etc and kill you. I take my hat off to the MQA people for creating a demand, and for getting the hifi press onboard but as a consumer I still haven't seen any good reason to buy into it.

MQA people did not create demand for high-res. The rest of the industry had done that prior. MQA provides one solution to that problem. Flac distribution is another.

There are a ton of independent labels distributing high res content which they create themselves for example and sell at > CD prices. I have bought a lot of it.

 

[Thomas savage]

Normal people( 99.9% of everyone) just wants convenient streamed audio , they won’t know anything about bit rates etc if you ask them about getting something better they will start talking about TT’s .

That’s whats actually happening in the real world. Iv no clue about this world where there’s huge public demand for digital hires recordings , I don’t live on that planet. Even if there becomes a demand due to some cynical marketing strategy and record and distribution companies making coin, it still remains unnecessary from a objective ‘scientific ‘ POV so why are we trying to champion that .

There’s a demand for dynamically compressed music because it sounds good on crappy ear buds before that maybe radio, why suddenly is that going to change? Are people going to go back to having hifi’s at home ? No so they will continue to want compressed recordings ..

Find some 20 somethings , play them a MQA file on their phone or sonos or whatever then put the regular 320 compressed file on..

Then tell them 1s free the others £20 a month lol

 

[watchnerd]

I secretly hope that he does (become your hero for aformentioned reasons) but suspect he is only after people's money.

If Watchnerd is correct* he may be his hero for that particular album already .


*
MQA vs FLAC:
100% success rate (so far)

P.S. I take it most have read Archi's response to the video already

I didn't say that I preferred the MQA versions, only that I could reliably differentiate MQA in an A/B test.

 

[watchnerd]

Normal people( 99.9% of everyone) just wants convenient streamed audio , they won’t know anything about bit rates etc if you ask them about getting something better they will start talking about TT’s .

I keep nudging ASR to measure cartridges or turntables, but instead we keep testing yet-more-cheap-but-hi-rez DACs. Have we hit 100 DACs tested yet?

 

[watchnerd]

I have no interest in MQA's complicated mechanisms designed to fit high resolution audio files through internet pipes that are now so fat they can stream HD video.

MQA's claims about "deblurring" are also of little interest to me, as I can already adjust the filters on my DACs to different settings.

I find MQA's partial compatibility / incompatibility with DSP room correction, crossovers, and convolution engines to be on the wrong side of history and my personal preference for software-defined audio architectures.

However....

From a philosophical POV, I do like the *concept* of an ADC (at recording time) and a DAC (at playback time) to be two known, certified entities following a particular standard or method. Filter choices, for example, are always a compromise, but it would be cool to know if the ADC used in recording was, say, a minimum phase filter, and thus the playback DAC should be, too, and the mixing and mastering were done using DACs that followed a particular implementation, too, so I can replicate it at home.

It would be a small step to reduce the circle of confusion.

 

[mansr]

From a philosophical POV, I do like the *concept* of an ADC (at recording time) and a DAC (at playback time) to be two known, certified entities following a particular standard or method.

We already have that. It's called "as linear as physically possible." Of course, software may be used to compensate for unavoidable errors in the hardware. That does not, however, necessitate a special format. All it requires is that an agreed upon intermediary is used. This intermediary is linear PCM.

it would be cool to know if the ADC used in recording was, say, a minimum phase filter, and thus the playback DAC should be, too,

No, it shouldn't. That would only exacerbate the phase distortion.

 

[watchnerd]

We already have that. It's called "as linear as physically possible." Of course, software may be used to compensate for unavoidable errors in the hardware. That does not, however, necessitate a special format. All it requires is that an agreed upon intermediary is used. This intermediary is linear PCM.


No, it shouldn't. That would only exacerbate the phase distortion.

What filter should be used, instead?

 

[NorthSky]

And potentially hand-tuned new masters created by MQA team. Some of the MQA content I have heard compared to non-MQA sound remarkably better, indicating better mastering.

Amir, is MQA vaporware or is it real?

* Did you read this article from SoundStage! HI-FI last year?
• https://www.soundstagehifi.com/inde...uencies-is-mqa-better-worse-or-just-different
_____

Extra:
• https://www.whathifi.com/news/tidals-hi-res-masters-tracks-surpass-one-million

 

[mansr]

What filter should be used, instead?

If you want to undo the phase distortion of a minimum phase filter, you need a filter with the inverse phase response.

 

[SIY]

If you want to undo the phase distortion of a minimum phase filter, you need a filter with the inverse phase response.

If anti-imaging filters were standardized, it's trivial to phase pre-correct the source material.

 

[Wombat]

And potentially hand-tuned new masters created by MQA team. Some of the MQA content I have heard compared to non-MQA sound remarkably better, indicating better mastering.

Better mastering can be done without MQA.

 

[pkane]

A quick and dirty Audacity frequency spectrum analysis of the first track on the 2L site, "POLARITY - an acoustic jazz project." Spectrum plots taken from 2:30 - 3:30 marks on both files because that's where a lot of action is happening and buffer size limits in Audacity.
...snip...
Unfortunately, I didn't normalize the plots.

But even without that step, there are visible differences, especially in the high end.

Here's a similar analysis of the same project files (POLARITY from 2L). Using the entire file, except for the first and last 5 seconds. 1M point FFT.

Match results MQA and 24/192:
Volume diff= -0.041dB; Phase offset=0.06ms (10.607 samples)
Difference (rms) = -55.93dB
Correlated Null Depth=63.28dB
Phase drift before correction: -0.0008 ppm

Blue is the the 192KHz PCM file, red is the MQA encoded one. Note that 44.1KHz MQA file was upsampled to 192KHz. Anything beyond 22.05KHz on the red line is a byproduct of upsampling:


Here's what it looks like where the two appear to start to diverge, around 16KHz:


Match results with 16/44.1 and 24/192:
Volume diff= 0.001dB; Phase offset=0.06ms (11.957 samples)
Difference (rms) = -65.31dB
Correlated Null Depth=64.32dB
Phase drift before correction: -0.0001 ppm

Comparison between 16/44.1 file and 24/192 file. 192 is the blue line, just like before.

And zoomed in to see where they diverge:


I also matched 24/96 with 24/192. The 96KHz file was upsampled to 192KHz. Here are the results for this comparison.

Match results with 24/96 and 24/192:
Volume diff= -0.200dB, Phase offset=-0.18ms (-34.690 samples)
Difference (rms) = -72.08dB
Correlated Null Depth=70.78dB
Phase drift before correction: 0.0000 ppm



If someone can capture a quality PCM file using, say 24/192 ADC, from the analog output of an MQA-equipped DAC playing the MQA-encoded file with full MQA decoding, I could run a similar comparison with that one, also.

 

[amirm]

Blue is the the 192KHz PCM file, red is the MQA encoded one.

Has the MQA one been decoded or is it in the base layer?

 

[pkane]

Has the MQA one been decoded or is it in the base layer?

Base file — at least that’s what I think is posted on the 2L download link.

 

[amirm]

Base file — at least that’s what I think is posted on the 2L download link.

2L has the fully encoded file but of course if you don't decode the MQA layer, then you just have the base (in the clear) layer.

 

[pkane]

2L has the fully encoded file but of course if you don't decode the MQA layer, then you just have the base (in the clear) layer.

Right. That’s what someone would get when playing this on an non-MQA capable DAC. If someone could ADC this file playing through a full MQA-decoding DAC, I could run the same comparison.

 

[amirm]

Right. That’s what someone would get when playing this on an non-MQA capable DAC. If someone could ADC this file playing through a full MQA-decoding DAC, I could run the same comparison.

Some pro cards have loopback capabilities. That would be the best way to captured the decoded MQA so that we don't have to convert to and from analog.