MQA disappointing

Well I guess I meant in general terms. Of course not everyone will grasp what MQA is consistently doing to alter and distort the original file.

However, let’s face it - most people hear a difference - the challenge and diffficulty is to recognize that the difference is in fact distortion and that the original file is higher fidelity.

I have several DACs.

PS Audio Directstream Sr and Benchmark DAC 3 are the ones I use most.

Roon settings are everything off. No DSP. No volume control. No upsampling. Nothing. Nada. Only MQA processing allowed in order to get the first unfold (with DAC 3) and both unfolds with the Directstream.

The MQA added distortion is consistent between all the DACs I own. There is a clear fingerprint or watermark effect of MQA processing most likely from apodizing and minimum phase filtering.

Over the past two and a half years my research team and I have carried out more than 500 (!) scientifically reliable listenings tests that prove such claims are untenable. Before the test many people were convinced they could hear a difference, afterwards ALL of them had to admit they couldn’t… (BTW, not even in informal tests at home - with their own audio gear, at least not when we were present :grinning: )…

I’m not surprised given that 256kbs mp3 has been shown to be indistinguishable from lossless in properly controlled studies involving the general public.

Aural memory and acuity is incredibly sketchy when making comparisons. The well-known experiment involving concert violinists being unable to pick out a Strad from a modern instrument under blind conditions is particularly eye-opening, and not an isolated example.

Thankfully MQA has that light, so you know it’s on. :smiley:

3 Likes

On my setup it is obvious as long as you can A to B switch fairly quickly (does not need to be instant switching). After A to B many times then you can learn to identify the telltale signs of MQA. Knowing what to listen for requires some effort and self-training (I would guess several hours of comparisons of known files MQA vs non-MQA may be needed - up to a day or more for some ). Somebody who hasn’t made this extensive effort would not know what to listen for and might be lost if suddenly thrown into a blind listening test. That said, throw a random file on and ask me if it is MQA or not and I would probably be challenged without having something to compare to (an A to B)

Perhaps it depends on system and hearing acuity. I really thought anyone should be able to hear it. A possibility is that if a DAC uses minimum phase filters normally for non-MQA files then perhaps the difference from apodizing with the MQA processing is too subtle to hear: since both files are played with minimum phase filter distortion then the difference would be much less and possibly inaudible to some folks.

The approach you’re describing does not work. Extended A/B listening sessions always lead to the same result: People UNLEARN to hear any differences - even if they’re there. We’ve done that. We asked people to compare two tracks with clearly audible differences (more bass, less treble etc.). The longer people were listening, the more they got confused and couldn’t concentrate any more. I’m talking about “trained professionals” (sound engineers, musicians, audio manufacturers etc.)…

Try real blind tests – you will be surprised…

I’ve had a listen to the 192K MQA and a 16bit CD version from Tidal.

I am not picking up any sense of distortion on either or I am not listening enough all the way through.

However, I also do not think they are the same master unless MQA encoding also includes some compression - I think a little over 3dB boost in loudness relative to peak which I feel may be a tiny bit too much of a boost.

As for pitch - the CD version is a tiny be slower - really tiny.

I also agree it has lost a bit of focus as well - I suspect whatever compression was applied was done with slightly differing mid/side settings (which is not usual to try to widen a sound a bit, but can also defocus a bit if not careful).

Either way - I blame the compression rather than MQA. I even suspect it may have been automatically applied rather than manually applied - so probably blame Tidal.

Wat?

Are you saying that Tidal are DSPing audio on their platform?

Who actually did it and/or requested it - of course I have no idea, however I would not be surprised if Tidal were commissioning these MQA versions encodings, OTOH maybe the publisher decided to do this if there are many consumers of MQA releases.

I have noticed quite a few MQA version that are louder and seen several posting from people mentioning this, but this is the first one I have bothered to do comparatively listen to.

They have to get people hyped somehow :wink:

To be clear though, Tidal are providing a platform to deliver music passed to them by the label. The label provides the release version. There is no (or there shouldn’t be) processing by the streaming provider.

Yes - but 3dB is a bit blatant!

0.5 dB is a better more subtle amount to net in the chin scratching crowd :wink:

Oh they’re “all-in”, for sure.

The most likely explanation is a loudness war remaster. Nothing to do with MQA, which is being applied as batch processing to myriad masterings, new and old.

AJ

1 Like

Yes it does work. You listen multiple times to learn the difference between one track and another. Eventually you learn to recognize a form of distortion. Ultimately you can learn to recognize the distortion with A and B comparisons. Of course not every amount of distortion is audible but MQA is significant enough to be quite clear. Of course people can get confused or lack concentration after a while.There is always going to be some errors and confusion.

Are you also saying that all DACs sound the same - because I can also hear differences between DACs - significant differences in many cases and more subtle in others.

Some remasters are louder indeed.

The effect of an apodizing filter will be similar to more audio compression. This form of filter broadens the width of transient peaks and flattens them. MQA are using apodizing so perhaps that is what you noticed quite a few times.

No, I’m not saying that. And I’m not saying all digital formats (or amps, CD players etc.) sound the same, either. However, what I am saying is that the vast majority of people (including “trained professionals”, see above) can’t prove they hear what they claim they hear when taking part in properly conducted double blind ABX tests. Most of the time the differences people describe as “easily discernible”, “dramatic”, “night and day” (and so on) prove surprisingly small…

In October I carried out an (informal) experiment with my music students at university. I played the same recording twice (Beethoven’s Fifth, 1st Movement, Karajan 1960s), but told them the “second recording” was from Karajan’s 1970s cycle. You can probably imagine what happened… :joy:

It’s funny (in both senses of the word) how people can “hear” and vividly “describe” differences that aren’t there… (And no, I’m not saying the 1960s and 1970s cycles “sound the same”…)

8 Likes

I’m not able to hear the distorsions you’re referring to on the Norah Jones album. While i do hear more of the detailing from the plucking of the strings i am beginning to wonder if you’re not hearing what i would call “increased resolution”?
The two MQA versions available on Tidal sounds pretty much the same, as does my own DSD rip in DSD64… My own, and Tidals CD versions all sound similar, the strings vibration noise is less audible, i would say because of less resolution.
I listen with Roon performing unfolding into an sMS-200 and a Chord Hugo TT+Hifiman HE-560v3. They all sound nice, but i deffo prefer my DSD rip!

The initial pitch of the string pluck is lower than normal. Additional audio compression has been applied. What you think of as increased resolution is just compression - probably from an apodizing filter. So you are indeed hearing distortion but you are fortunate in that you prefer the sound.

It turns this isn’t just a cast of a different master. You may actually be right about the filter or more generally the MQA render phase as being the culprit (or at least the render phase as done by the pro-ject pre-box S2 DAC).

I also tried another track with CD and MQA versions from the same digital master (they perfectly sample align). That was enough to confirm it is the render phase where the level boost and dynamic manipulation is going on which is annoying and this throws out Roon’s auto levelling as it cant account for the level boost.

CD and Roon decode only play back at close enough levels with auto levelling on.

I agree with you. There are audio cables for large sums of money and all manner of dubious tweaks - largely because of the power of suggestion that you demonstrated with your students.

This is not the same as careful controlled listening where a difference is identified and then repeatedly demonstrated.

I think most people hear a difference between MQA vs non-MQA. It is just that all the marketing tells us to expect that the MQA version is better and many people will believe or buy into that. The truth is MQA is just added distortion or a form of DSP “enhancement” that disrespects the fidelity of the original. Critical listeners can tell that the original has become distorted and less high fidelity. Casual listeners will simply believe that the new razor blade is better than the old one because the package says 3x closer shave!

1 Like