INTERNET BLIND TEST: MQA Core Decoding vs. Standard Hi-Res (24/88 or 24/96)

If the subjects go into the test with a bias that there will be no difference heard, then they don't need to know which files were which in order to come out of the test with that bias intact.

 

That's true.

 

If the participant believes MQA has no benefit, they may be more likely to perceive no difference, or vise versa. Plus participants usually want to do their best and succeed so they're under pressure to be right, which may make your instrument (human) function less effectively, hence we need to know performance when their is no difference

I've said before that I give Archimago great respect and this is definitely better than people just expressing their opinion, I just had concern that some may think that we now have a scientific study with results that can be relied upon. If we had funding this would bring us to the next step in trying to improve our understanding, a start had to occur and Archimago deserves a big thanks, for doing it

 

Well, I’m not so worried about the vice versa. Even if some of the subjects were firm believers in MQA sounding better or the opposite, they didn’t know which file was which, and that would mean that if they couldn’t hear a difference, then they would be guessing.

But you’re right that subjects who don’t believe there to be a difference could have weakened any findings of a difference there might have been if they hadn’t had that bias.

But maybe the interesting bit of this discussion isn’t whether the test qualifies as science but whether it tells us anything about MQA. I think that is also what you’re saying.

In that respect I first of all trust (and that’s a personal choice, and thus again not science) that if there was a major difference between PCM and MQA—large enough that confirmation bias wouldn’t have mattered—most participants would have reported that honestly (and not just flipped a coin).

Secondly, I regard the many different systems used as a virtue of the test, since this gives an impression across a wide span of gear of cost on whether or not MQA makes a difference out there in the real world.

Just to be clear: Thank you for qualifying this discussion. I think that's what's so great about this forum--highly qualified people who express their opinions in a respectful and constructive manner.

 

I wouldn't describe myself as qualified to be confident on any of this, confession time left psychology for law 30 years ago. I was presuming any benefits of MQA are not blindingly obvious, perhaps that's the problem, I've an headache now.

 

I can only say, listening to MQA and “Hifi” versions of the same song on Tidal, it is a mixed bag as to which sounds “better” on my rig.

And for what it’s worth, I’m absolutely biased in WANTING the MQA to sound better. It just doesn’t always...

 

I have a TIDAL trial period. I don't have an MQA compatible DAC, but I have to say that in ordinary hifi mode, TIDAL doesn't sound very good compared to redbook streamed from my local disk. A friend of mine has made the same observation in his system.

 

And as such, even if the subject just guesses which one is which, they will likely guess at about chance level.

 

With respect, that indicates that you aren't comparing like for like or something isn't quite right in the comparison process.

 

Nor is there any proof for your alternatives. @Archimago did extensive work on the subject. You're offering Psyche 101 generalities that may or may not apply in the specific circumstances, but which doesn't constitute evidence of any kind because you don't have access to the subjects who participated. Simply assuming that an arbitrary number of people fudged their responses is hubris because you have no evidence that it actually happened. We can proceed on the basis @Archimago provided until someone proves his conclusions inaccurate or until someone reinforces his conclusions in a follow-up that is even more rigorous. It is that follow-on work that will substantiate @Archimago's work or dismantle it or present a related but heretofore unanticipated conclusion.

Your studies at Liverpool should have taught you all about Occam's Razor. In any choice of cause & effect, in any choice of possible reasons for a particular outcome or for a particular prediction, the best choice is usually the one that contains the fewest assumptions. @Archimago may have conducted a study that raised as many questions as it answered, but those questions should form the basis for further studies rather than a basis merely for fact-free remonstration. Just as well, nobody from MQA has offered even the slightest whiff of rebuttal with respect to @Archimago's work.

What the MQA people themselves have also not offered is any viable technical explanation of how their MQA process actually operates. Their proffered explanations have been judged to be largely rubbish by experts in the same field - essentially, a solution in search of a problem. MQA has consistently refused to allow anything other than its own, carefully managed presentations at show and exhibitions, which in the insular environment of the audio world and the audio Show world presents willing, unquestioning believers with a Hobson's choice, nothing more. Speaking of psyche, that cleverly presented sort of Hobson's choice is one of the main reasons why a number of high-end reviewers and audiophile writers have swallowed so much MQA Kool-Aid.

MQA has invited certain pro engineers and musicians to send in their music files for processing, and then (in at least two cases) simply ignored the files that were sent. No processing and return of the MQA'd files.

MQA is being supported by some hardware companies willing to not only pay MQA license fees and also willing to allow MQA to modify their firmware. That last bit is the 'rub' as they say, because what MQA says I should hear and what I actually hear are two different things. What I actually hear sounds for all the world just like a bit of midrange and high treble EQ boost. Nothing better - just a bit different in those cases where something audible is actually evident to my ears.

 

how much of an improvement in SQ would it take for you to buy all your music with DRM (copy protection)???

 

A TIDAL trial provides two options - TIDAL Premium at 320kbps ($10/month when you subscribe), or TIDAL HiFi at 1411kbps ($20/month when you subscribe). There’s no such thing as “ordinary hifi mode” on TIDAL. TIDAL does not use that terminology anywhere on the site.

A TIDAL HiFi subscription (trial or full) gets you a 1411kbps stream (identical to CD/Redbook quality). So of course you weren’t hearing the same quality if you signed up for a TIDAL Premium trial. You have to compare identical file quality, otherwise the comparison is basically useless.

If you signed up for a TIDAL HiFi trial and you’re still getting inferior sound quality, you’ve got a problem with your DAC or your system. Check again.

 

Assuming that's a non-rhetorical question, my answer is that any significant improvement in SQ would justify all my purchases being copy-protected. The question is a no-brainer for me; I'd accept that trade-off in a heartbeat.

 

most seem to think that MQA is not a large improvement in SQ

and engineers who have examined MQA think it opens the door to DRM

 

Yes, but that wasn't the question you asked.

 

But these engineers are not the one who lives in your house and use your equipment to listen to your music. A technical opinion can't by itself be the only criteria by which you decide on your purchases.

To answer your question on DRM vs SQ: There will have to be an audibly change in sound and then I will have to decide if the change is for the good. To put things in my perspective: Are you using headphones? if you do, have you noticed how different one set can sound from the other, even after matching volumes and fiddling with the headphones over your ears? Well that's in my book what an audible difference is and you are not going to hear anything in the same order with @Archimago 's MQA blind test. Not even close. In fact the test results suggest that taking a guess between the samples will net you the same results.

Then you also have to decide if a perceived change is a good thing. Going back to my headphones example - I do have two sets that are very different in their sound signature and I'm still not sure which one gives a better and closer representation of the artist performance.

Bottom line: FLAC and PCM are perfectly satisfactory for me at the moment and I see no compelling reason to migrate to MQA, actually I see at least one major reason why not to do it.

 

I'm not making assumptions but trying to exclude possibilities, which may cause an assumption and I'm certainly not thinking MQA has any merit. I also can't understand why you're so agitated when we agree this could be a starting point.

 

It would have to be pretty significant.

This is still the main concern I have with MQA in general - encoders and decoders are licensed tmk so there's no way to use this as I see fit versus something like MP3 or FLAC. Or even a licensed codec like AAC or WMA is something that at least can be used via your OS and decoders are generally available for a large amount of hardware.

Probably fine in my case for a closed ecosystem like Tidal where the setup is completely handled by the vendor. But in my home system I'd still use with FLAC for archives and MP3/AAC etc for portable usage.

 

That might make a good poll. I know my answer.

 

I've eaten my breakfast and lunch so I can now give a more reasond response.

• I intend to prove MQA has no benefit to SQ.
• To do this I'm going to show no difference can be detected, this would avoid all the murky water involved with perceived improvements in SQ
• I first need to know how well we can identify that two clips are the same
• If the findings don't show any significant deviation, we can conclude MQA is a waste of money, I think Sherlock would approve.

I may have just given you more ammunition but that's OK negative assessments welcome.

 

I have found that most statements about "not good sound quality" when referring to streaming, CD or even MP3 stem from misunderstandings about settings, what actual resolution is being used or some other technical "issue", and the end user is not actually getting full sound quality that is possible.

With that being said, I have heard streaming and MP3 sound different in quality at others homes. Ranging from nearly excellent and fine sounding, to fairly poor sounding.
There are many variables in MP3 encoding and streaming, enough to make comparing difficult, when one is not sure they are even getting the full quality it is capable of.

 

More than is possible. Good PCM is already too close to audibly transparent for MQA to make a significant difference. For me at least, the advantages are not worth the downsides of DRM. The main problem with MQA is that it requires specific hardware. What happens when/if MQA dies in the marketplace and DAC manufacturers stop paying the license fee and drop support in their new DACs? There have already been several cases where DRM schemes have folded, leaving people with music they can no longer play.

And for those who say that you can still play the undecoded MQA, remember that the currently encoded MQA files use the maximum quality setting. The undecoded quality can be reduced to a level where you can barely recognise the music being played. The encryption scheme used is very good (and expensive). You can bet that MQA included it to make MQA more attractive to the record labels. And you can also bet the labels will be tempted to use it if MQA becomes the dominant format.

 

Bingo!

 

The best way to find facts or to add to a body of facts and evidence is to avoid bias. If you start off trying to prove MQA is worthless to audiophiles and music lovers, there’s a very high likelihood that’s exactly what you’ll prove and to the exclusion of all contradictory data or listening tests results you might put together.

Your earlier premise, I think implicit in some of your posts, led me to think you might be prepared to set up an objective-ish home-based listening test of your own.

See? I actually want MQA to be a boon to audiophiles and music lovers. Problem is, my own little listening test groups couldn’t support MQA, @Archimago’s more extended and more technical effort couldn’t support MQA, and an impressive list of engineers that have looked at MQA and shaken their heads aren’t supporting MQA.

That’s not to say I haven’t missed anything and @Archimago hasn’t missed anything and the engineers haven’t missed anything. The question is, what? The more direct comparative listening test sessions the better, IMO, because the more there are, the more the pendulum will swing one way or the other.

Other engineers have already dismantled some MQA files, so the technical side has begun to speak (balefully) as well.

But none of that means we throw objectivity away. I’d personally like to, but I can’t.

 

Well that's it I'm not setting out to prove it but putting my hypothesis to the test, I would be equally be looking to disprove, as prove it. Perhaps my lengthy absence from science has fogged my mind about basic principles?