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

wooff, wooff!

 

Let's get this straight, once and for all. The package contains 2 discs, one is MQA encoded, the other is straight Redbook CD (albeit manufactured i.e. pressed using higher quality materials and cut using a blu-light laser).

Both discs will play in a conventional CD player. but only the former will yield hi-res (of sorts) when played back with MQA-enabled equipment. If this disc is played with non-compliant equipment, or in bypass mode, it will sound worse -- certainly in comparison to fully decoded, maybe also in comparison to the other disc, the plain vanilla CD (that's what remains to be proven here). And yes, all comparisons need to be at the same loudness level.

 

You're right about the single layer. Wasn't thinking and defaulted to comparing layers of a hybrid SACD vs MQA's decoding process. My primary interest is what actually caused the level change. Since I don't foresee purchasing a MQA disc player any time soon, hopefully someone who does can provide some insight.

 

I think you're mistaken. Only MQA CD samplers have 2 discs: one - MQA-encoded version, another - standard CD version. UHQCD album in question (DS - BIA) has 1 disc only (i.e. MQA-encoded CD).

 

Aha, well in that case, he should have use the samplers! Elementary my dear Watson, that is what samplers are for... He says as much in the video. Now he had the Jazz sampler in his hand - don't want to imply any ulterior motives, but have wonder why he didn't do the test in such a way as to provide genuine value?

Incidentally, I have ordered the Jazz and Pop/Rock samplers (at a cost of around $25 incl. shipping to NYC) and should be able to post my own findings in about 3 weeks time as my Oppo 205 should be here by then

 

Totally agree. Just struggling to understand why he is so determined to "spread the MQA gospel" in spite of all the obstacles and erroneous conclusion

 

I think that Techmoan got so balled up with the subject - struggling just to get it going in any sort of listenable way - that the pyrrhic victory he finally extracted from the process appeared to him far more significant that it actually was.

 

He did explain the reason. 13:33 in the video.

 

Except, that’s still not much of an explanation by Techmoan because he never got the MQA datastream to work without it flickering on and off constantly during play. All the video does is demonstrate something that doesn’t work properly. I think it was a mistake for Techmoan to post the video in the first place.

 

Not quite, what he says is nonsense. The point is that he should have ripped the conventional CD to the computer hard disk and played it back via USB as normal. The external CD-drive with the MQA disc in it remains connected as he describes. Now he can synchronize start of both playback methods and toggle the input selector for A-B comparison. Elementary, my dear Watson!

p.s. at the least that's how I would have done it. Maybe there is a reason as to why this wouldn't work but seems not to have been contemplated at all

 

But then he would get shredded by those believing that playing a FLAC fail to capture the essence of a good CD playback.

Anyway, the series of hoop jumping he did on the video which finalized in an MQA playback that can't even lock the decoding in place was rather hilarious. He is a tech savy guy and if he can't manage to spread the MQA sauce correctly I seriously doubt if MQA is a dish for the general public.

 

"you can please some of the people some of the time .... yada, yada ... but not all the people all of the time"!

Your second point is absolutely spot-on and pretty much irrefutable

 

I got an answer recently. He does not agree with your centiment. He also says a later more rigurous study shows the exact same results.

Now state the reason why its flawed without stawmaning or asking me to contact others.

 

As @Leonthepro noted above, we have been in contact around the test and results. As with any test there are limitations and especially for an internet distributed survey like this, there are obviously going to be uncontrolled variables!

Regarding the study Leon mentioned, it's the McGill University listening test... I don't think the paper is available freely, but I summarized the results here at the end of the blog post:
Archimago's Musings: MUSINGS: Why Do People Equate High End Audio with Snake Oil? [And McGill MQA Study Summary]

Basically as I found on the internet test, there was no significant difference. At best even in a controlled lab setting with a high quality listening room, Mytek Brooklyn DAC, B&W 802D speakers, Sennheiser HD800 headphones, trained listeners, 24-bit MQA was about the same as the original hi-res. That's good from the perspective of same/similar audio quality in a smaller filesize for streaming (IMO 16/88 or 16/96 FLAC would have been just as small filesize and sound at least the same also without nonsense DRM-like crypto signatures, low quality filters and such). But it also implies that whatever "deblurring" done didn't equate to a better sound as MQA had initially claimed.

Ultimately, IMO, the technical limits are clear for MQA and the listening tests both informal and more formal are pointing to the same thing - there's just no real benefit and there's no reason at all to call MQA "high res".

 

Did you miss the part where I wrote "But it's not a controlled study and there are many problems with basing research on answers from questionnaire"?

 

Did you miss the part where I said "There is no inherent issue with scientific papers. The problem is ones flawed interpretation of data."

I asked you what the flaws of this study was, not if questionnaires are the best methods.

Its like someone talking about evolution and all the good evidence and data, and then you come in and critisize science by saying there are no absolute truths anyway. Its poisoning the well.

 

I honestly don't know where you go and find these ideas but it's very entertaining nevertheless. I wish you had a newsletter I could subscribe to.

 

Extra, extra! There has still been no flaw found regarding the MQA study!

In other news, audiophile socks are now the new hot trend to get your audio to sound that little bit better.

 

Where do I get those audiophile socks? Do they have Shakti stones sewed into them?

 

FYI, the Richard King who co-authored the McGill paper is this Richard King:

Home

He is one of the world's elite classical balance engineers (he is usually recording engineer for producer Steve Epstein, who is also affiliated with McGill). Richard is the very definition of a golden ear! He is also an exceedingly nice guy. If you require more information, or want to discuss the paper, you could try reaching out to him, or Mariane Generale (as indicated in the paper). I'm sure they would appreciate your interest.

 

Now - to see if you need to be buying any kind of "hi-res audio" at all, first downsample the 96k source to CD audio quality with a good converter, and apply the same rigorous listening comparison.

If one can't readily tell the difference between 44/16 and 96/24, any test of MQA is either testing for something you know you can't hear anyway, or testing to see how MQA detectably alters the original sound.