Stereophile Tests of MQA

[watchnerd]

What does the Impulse Response of an Analog (parts bin) low pass filter look like?

What kind (Butterworth etc.) and what order?

 

[RayDunzl]

I'll plead a bit of ign'ance here...

May I order the Sampler Platter?

 

[watchnerd]

I'll plead a bit of ign'ance here...

May I order the Sampler Platter?

1st order is time / phase coherent. Higher-orders exhibit all the usual trade-offs with regard to amplitude and phase (no free lunch), with some clever compromises on the typical Butteworth via topologies like Linkwitz-Riley.

 

[watchnerd]

Both - i.e. the fact that it is not correct.

All filters are trade-offs -- which do you deem to be "correct"?

 

[amirm]

I'll plead a bit of ign'ance here...

May I order the Sampler Platter?

Here you go:

 

[Cosmik]

All filters are trade-offs -- which do you deem to be "correct"?

The filter that is a genuine reconstruction filter rather than someone squiggling around with a biro and seeing what it does.

A windowed sinc function of sufficient size to give acceptably small errors. Everything else is unnecessary FUD.

 

[watchnerd]

The filter that is a genuine reconstruction filter rather than someone squiggling around with a biro and seeing what it does.

A windowed sinc function of sufficient size to give acceptably small errors. Everything else is unnecessary FUD.

Okay, I was just talking filters in the generic sense.

Sounds like you're anti-origami-folding-fippery. About which I agree.

 

[Werner]

This is the most interesting figure in Austin's Part 1: Fig.4.

Why?

It is the replay of the MQA-folded impulse at 48kHz through an orthodox linear-phase half-band filtering DAC. As such it reveals one of the two quadrature matched filters that are used during the MQA encoding band split operation. These QMFs, so far, are not publicly documented.

I contend that these QMFs, being severely constrained due to the requirement of a lossless band split and join, are very likely sub-optimal with respect to the baseband signal in an MQA file. Or, in other words, I expect that when the nature of these filters is revealed that there is formal proof that MQA damages the signal, compared to 1x rate (i.e. CD). This contrary to MQA's claim that even an undecoded file sounds better.

If we deconvolve the Benchmark's 48kHz reconstruction filter from the above graph we will get the details of the low-pass origami filter.

 

[RayDunzl]

Part II:

https://www.stereophile.com/content/mqa-tested-part-2-fold

 

[Nathan Raymond]

I thought Arstechnica's article on MQA was interesting, though it didn't delve into the technical side as much as Stereophile. The subjective impressions, under "MQA's cracks begin to show" were what caught my attention:

https://arstechnica.com/gadgets/201...hing-you-need-to-know-about-high-res-audio/7/

"While focusing on an individual strand of a song it is easy to be impressed by the hi-fi clarity of that singer, this soloist, that instrument. But stepping back to hear how a group plays together as a band, I found myself increasingly confused by the result. Rather than hearing an ensemble fitting together to provide a result at least equal to the sum of parts, the overall pulse of the music seemed less coherent. The song's pace was often askew, sounding too fast and hurried.

Stereo soundstaging with MQA was wide and airy, but foreshortened in depth, to give the effect of musicians crowding and overlapping each other. The sound palette could at times seem enervated, lightened, lacking substance. And my concentration through the length of a song would waver, as if I had to reconstruct the sense of the music in my head, some cognitive load upsetting relaxed infusion.

Looking at other rave reviews, I may be the first and only listener to be unsettled by MQA encoding. But it is possible my subjective impressions of spoiled timing, tonal recasts, and soundstage confusion are symptoms of phase anomalies introduced by various non-linear phase filters in the MQA chain. Somewhat tellingly, I did find that these amusical effects could be more or less replicated from regular PCM audio, by simply switching from linear-phase to minimum-phase interpolation filter on both Mytek and Pioneer hardware."​

 

[allhifi]

I thought Arstechnica's article on MQA was interesting, though it didn't delve into the technical side as much as Stereophile. The subjective impressions, under "MQA's cracks begin to show" were what caught my attention:

https://arstechnica.com/gadgets/201...hing-you-need-to-know-about-high-res-audio/7/

"While focusing on an individual strand of a song it is easy to be impressed by the hi-fi clarity of that singer, this soloist, that instrument. But stepping back to hear how a group plays together as a band, I found myself increasingly confused by the result. Rather than hearing an ensemble fitting together to provide a result at least equal to the sum of parts, the overall pulse of the music seemed less coherent. The song's pace was often askew, sounding too fast and hurried.

Stereo soundstaging with MQA was wide and airy, but foreshortened in depth, to give the effect of musicians crowding and overlapping each other. The sound palette could at times seem enervated, lightened, lacking substance. And my concentration through the length of a song would waver, as if I had to reconstruct the sense of the music in my head, some cognitive load upsetting relaxed infusion.

Looking at other rave reviews, I may be the first and only listener to be unsettled by MQA encoding. But it is possible my subjective impressions of spoiled timing, tonal recasts, and soundstage confusion are symptoms of phase anomalies introduced by various non-linear phase filters in the MQA chain. Somewhat tellingly, I did find that these amusical effects could be more or less replicated from regular PCM audio, by simply switching from linear-phase to minimum-phase interpolation filter on both Mytek and Pioneer hardware."​

Hey Nathan: Great points. I was chatting (last 1-2 weeks) at 'S-phile' regarding MQA and I, as you commented that perhaps a MQA facsimile could be attained with simple filter selection (on DAC's so equipped), as you suggest:

"...that these amusical effects could be more or less replicated from regular PCM audio, by simply switching from linear-phase to minimum-phase interpolation filter on both Mytek and Pioneer hardware."

And to that I'd add the Gustard X-20pro filter selection -and the same "softening/stretching" (phase-messing) is observed in all MP filter choices. The best SQ (to my mind) are the LP/or 'Brickwall' settings.

MQA is one 'hell-of-a-mess' and complicated, signal-impairing "system" that surely must only be wondrous in the minds of its creators. And to think everyone must pay (dearly) for the privilege of this 'Monetarily Quenching Algorithm' -that makes a mess out of our beloved music.
Did Stuart and company not foresee any of this? Did they feel they were 'smarter' than everyone else ? Nobody would ask, inquire or or resist ? Incredible.

pj
P.S> Tip of the Day: Unload MQA stock shares ! lol

 

[allhifi]

I thought Arstechnica's article on MQA was interesting, though it didn't delve into the technical side as much as Stereophile. The subjective impressions, under "MQA's cracks begin to show" were what caught my attention:

https://arstechnica.com/gadgets/201...hing-you-need-to-know-about-high-res-audio/7/

"While focusing on an individual strand of a song it is easy to be impressed by the hi-fi clarity of that singer, this soloist, that instrument. But stepping back to hear how a group plays together as a band, I found myself increasingly confused by the result. Rather than hearing an ensemble fitting together to provide a result at least equal to the sum of parts, the overall pulse of the music seemed less coherent. The song's pace was often askew, sounding too fast and hurried.

Stereo soundstaging with MQA was wide and airy, but foreshortened in depth, to give the effect of musicians crowding and overlapping each other. The sound palette could at times seem enervated, lightened, lacking substance. And my concentration through the length of a song would waver, as if I had to reconstruct the sense of the music in my head, some cognitive load upsetting relaxed infusion.

Looking at other rave reviews, I may be the first and only listener to be unsettled by MQA encoding. But it is possible my subjective impressions of spoiled timing, tonal recasts, and soundstage confusion are symptoms of phase anomalies introduced by various non-linear phase filters in the MQA chain. Somewhat tellingly, I did find that these amusical effects could be more or less replicated from regular PCM audio, by simply switching from linear-phase to minimum-phase interpolation filter on both Mytek and Pioneer hardware."​

Hi Nathan: Your listening observations are near identical to mine (regarding MP filter selection).

I had/have my MQA suspicions in that it will sound eerily similar to what we can achieve now with MP/Apodizing filter selection.
Since I've never heard MQA (never any rush was/is required), I can't say what the 'encoding' component does to the SQ characteristics we can (easily) detect via Minimum-Phase/Apodizing filter choice, but I suspect it's simply a complimentary "end-to-end" type deal.
Even so (and as you -and so many others rightfully point out) MP filter choice alters (distorts) the signals time-phase/amplitude characteristics. In which case (full MQA) SQ should remain virtually the same as it would when such a filter choice is employed at the DAC end only -or perhaps worse ?

I truly enjoy hearing someone speak openly and honestly -and one possessing some listening skill !
I've yet to read of a 'global' listener consensus on MQA, oddly enough. But then again, if it's floundering, then there's not much opportunity.

I hope this serves as a final 'cautionary tale' for anyone attempting to disrupt a (more-or-less) sensible, open audio recording and playback specification. And to demand payment for such a flawed system that can be (and is) easily outperformed by existing/readily available 18-20/96-192 FLAC/Uncompressed makes it (MQA) even more bizarre.
I'm confident others, previously explored/considered the so-called "end-to-end" approach of recording/playback filter compliment, but likely foresaw (or even identified/heard) the issues currently discussed (and indeed over the past near-four years now).

peter jasz

 

[Beershaun]

Part II:

https://www.stereophile.com/content/mqa-tested-part-2-fold

Adding the manufacturer's comment to this section of the article.
https://www.stereophile.com/content/mqa-tested-part-2-fold-manufacturers-comment

 

[Beershaun]

The follow up commentary from the author.

https://www.stereophile.com/content/mqa-contextualized

 

[Beershaun]

Continuation in the series...
https://www.stereophile.com/content/mqa-drm-and-other-four-letter-words

 

[Beershaun]

More continuation of the series...
https://www.stereophile.com/content/mqa-aliasing-b-splines-centers-gravity

 

[mocenigo]

Yeah, the remote would be a pre-req for me to use it in my living room, but the lack of analog input kills it for me since I also have a TT.

You could get a Puffin and feed its digital out to the DAC…

 

[EB1000]

There are a few things in there. Let's start with the basics.

In practice it can never be generated because it requires infinite energy and infinite bandwidth. Such a thing does not exist in nature.


.

Just a small correction:

An impulse signal in continuous time (Heaviside function) has infinite power and a finite energy (an instantaneous change from 0 to infinity and back to zero).

An impulse in discrete time (Dirac function) has both finite power and energy (a change from 0 to 1 and back to 0 in three consequential samples).

 

[watchnerd]

You could get a Puffin and feed its digital out to the DAC…

what is this in reference to?

I already have a Puffin.

You necroed me

 

[mocenigo]

what is this in reference to?

I already have a Puffin.

You necroed me

This is a reference to your “ the lack of analog input kills it for me since I also have a TT.”

There is a Puffin with digital out - do you have that version?

Also: what does “you necroed me” mean?