Ask AudioStream: MQA

in a men's room at RMAF 2016

"Should I sell my DAC, which I love, to buy an MQA DAC right now?"

No.

And I say no because MQA does not supersede all other DAC design factors. A $99 MQA DAC will not outperform a dCS Rossini. Sorry, life doesn't work that way.

"What would you do?"

I recently purchased the totaldac d-1 six, a non-MQA DAC. That's what I did. Why? I love listening to my music through it. All of my music.

"But, you said MQA sounds really good!"

It does. I also said "Music availability [is] the overarching determination of whether or not a new audio codec is worth investing in, for me." In other words, I would skip troubling over MQA until there's significant content available. What that means to me is Tidal HiFi streaming their catalog in MQA. I'm not a fan of the re-buy.

"But I read on a forum you said..."

The thing about forums is the people who spend the most time participating on forums, spend the least time participating in real life.

COMMENTS
bobflood's picture

require that there were several stable streaming services streaming a wide array of music in MQA. Just having Tidal do it would not be enough. Tidal, it seems, is constantly in turmoil and on the verge of going under.

Maybe MQA will be the avenue for the more stable and better funded streaming services to abandon lossy compression formats at least in some tier of service. They would also be able to use MQA to generate enough revenue to become profitable.

To spend considerable resources for a new MQA DAC without a healthy music streaming delivery system just seems foolish to me. And, make no mistake, MQA is all about streaming. I doubt that there will ever be more than a very small number of MQA releases made on physical media of any kind.

BradleyP's picture

Yes, I wish TIDAL well, but the jury is out on its viability. Shoot, the jury is out on the entire streaming universe. If Spotify goes under, where I've curated 2000+ albums, I'm back to my far smaller CD collection. At that point, I'm not re-buying, but I might buy new music in MQA if it's plentiful, not much more expensive, what I want, and I can justify the special DAC. Right now, I'll wait and see.

BradleyP's picture

Yes, there are other streaming services, but if Spotify should go under, the whole thing is a house of cards and the handwriting is on the wall.

christopher3393's picture

What is this "real life" you speak of? ;)

jazz's picture

to see the number of MQA recordings on HIGHRESAUDIO, even ECM released some (not much more than 200, but still).
https://www.highresaudio.com/studio_master.php?fids=153&cr=MQA

Vade Forrester's picture

Not all of the albums on highresaudio.com are available for purchase in the USA.

Vade Forrester
Reviewer, SoundStage! Network and The Absolute Sound
Anything I said here are my personal thoughts and ideas, not
those of the magazines I work for.

ctsooner@alumni.ou.edu's picture

Still burning in my new Ayre QX5 Twenty and loving it over my last DAC, which I loved. I hope MQA makes it, but like you said, until Tidal stream it and all the top record labels move over to it, it doesn't matter.

Tog's picture

I forget codecs - is that a picture of the new Mac Pro?

mtymous1's picture

Ha!

I think there's one actually inside the receptacle "in a men's room at RMAF 2016."

CG's picture

Gee - I thought it was a promo for "The Power of the Daleks".

solarophile's picture

I believe that the recent blog post from Archimago calling this a "partially lossy codec" and that MQA should release a software decoder is a powerful idea.

If I had a great DAC like Michael's Totaldac, I really cannot imagine myself needing to upgrade just because of a codec that operates in PCM which the DAC plays back beautifully. That's IMO absurd! It's like having to buy a new DAC because it won't play back MP3 or AAC. We all know a modern computer can already do this processing, so why link MQA decoding as a hardware only solution???

MQA I believe really is DOA and only benefits a business model rather than driven by consumer benefit.

Michael Lavorgna's picture
...pretty much from the beginning (see my Munich report from 2015). While this is still the case, I have not heard a firm date as to when MQA will roll it out.

Is a "partially lossy codec" the same as a partially relevant point?

;-)

miguelito's picture

A software decoder exists and was used by Auralic's Aries at CES 2016. MQA shut it down the moment it realized it would work with any DAC... See my other reply to your other comment...

Michael Lavorgna's picture
...we're not talking about a decoder embedded in a piece of hardware. We are talking about a software-based solution. For example, Roon is working with MQA to embed MQA in their app.

Your re-telling of what happened at CES is not what I heard and I spoke to both Auralic and MQA at CES.

miguelito's picture

But ultimately the code implemented in the Aries is a piece of code that could sit in Roon. In fact I'm pretty sure that code is in Roon's possession, but MQA has decided not to allow it (Roon is no longer an MQA partner).

2_channel_ears's picture

I hope the record company execs are hearing the customer sentiment to do MQA via streaming media. What I heard at one of the demos is they're trying to figure out how to deliver it. Re-buy ain't gonna work this go around.

miguelito's picture

"...why link MQA decoding as a hardware only solution???"
Auralic had the Aries decoding MQA to PCM at CES 2016. When MQA (the company aka Bob Stuart)realized that, they pulled the license from Auralic.

The argument is that there is additional tailoring to the DAC and it's output stage. I personally think that's complete BS. You cannot possibly tell me that a multi-thousand dollar DAC requires some sort of adjustment in it's output stage. Additionally the restriction means that no post-processing can be done on the decoded PCM (such as room correction).

In my opinion, if this restriction is not relaxed it will spell the demise of MQA.

Michael Lavorgna's picture
...demands that you know what you're talking about. First off, to say that MQA adds "tailoring" to a DACs output stage is not correct. Secondly, if we have a "dumb" software decoder, i.e. one that does not know what DAC it's talking to, then your premise of a "multi-thousand dollar DAC" flies out the window since *any* DAC can be used.

The thing that gets me a bit riled up is just how easily people can call into question someone's character and work with what appears to be an offhand comment.

miguelito's picture

A high quality DAC will decode PCM in a way that it reproduces what the PCM encodes. The DAC manufacturer already took great pains to reproduce in analog the exact PCM or DSD signal coming in. You're telling me that no, they are stupid and MQA profiling of the DAC is better. That is BS.

Michael Lavorgna's picture
...how do you explain DACs that offer different digital filters as a user-selectable option?

Again, if you read what MQA has said and written about this, you'll see that your as misunderstanding the key points they make. One key point they make is the anomalies they address reside in the DAC chip used. So the DAC manufacturer that uses an off-the-shelf chip is reliant up on that chip's performance, especially in those cases where the DAC manufacturer utilizes all of the chip's features.

miguelito's picture

Couple of points:

1- Yes I agree that off-the-shelf DACs might have signature sounds MQA might want to correct. With more expensive designs this tends to be less relevant, but every component has a particular sound.

2- If we were to be ultimately puritanical about it, the preamp/amp/speakers/room have a vastly bigger impact. Some people like to address this impact with room corrections, like Dirac, which cannot be applied if you cannot decode MQA to PCM.

Bottomline: Why not allow software decoding? Why not have the ability to decode a generic PCM, and on top of that to tailored PCMs for each DAC chip?

Michael Lavorgna's picture
As I mentioned, Roon is just one of the companies working on this.

That being said, I have not heard a date from MQA as to when software decoders will be available. It is my understanding that a) a hardware implementation is ideal, and b) MQA chose to start with a focus on the ideal implementation.

miguelito's picture

And BTW Roon is NOT a partner anymore: http://www.mqa.co.uk/customer/our-partners

TIDAL is still purportedly a partner but I cannot imagine TIDAL getting on board in reality, across the board, given that:
1- Actual streaming bandwidth is 2x bigger (compressed 16/44 vs 24/48)
2- No adaptive streaming is (currently) possible with MQA
3- It makes no PR sense for TIDAL to go into this with the caveat "Oh BTW you need special hardware"

Michael Lavorgna's picture
...about this so I'll see if can get some information on the status.

Tidal is still a partner along with other streaming services.

1. Actual streaming bandwidth for *high res* is lower.
2. Correct
3. I'm not a PR guy so I don't know about "PR sense" but let's assume high res streaming was on Tidal's road map (I understand it was). MQA solves any number of issues related to high res streaming and for people who are not interested in MQA, they will get what they get today for the same price. One could also argue that the MQA encoding process can improve the sound quality of playback even without decoding.

miguelito's picture

1- True if you can actually make use of it - very few people will if it requires special hardware
2- Related to '1' bc if I can't easily downsample, then TIDAL is forced to always stream the bigger file - this means dropout rates (especially over cellular) will increase - not workable
3- Related to '1' and '2'

(yes it's all related - ha!)

Michael Lavorgna's picture
...about the importance of file size while others argue that file size is not important.

From my perspective, the more important issue is *the fact* that streaming services (and record labels) are embracing MQA. The idea that we can second (and third) guess the motivations of all players involved, including DAC manufacturers and recording engineers, based on second and third-hand information (and guessing) strikes me as a waste of time.

miguelito's picture

Is it verified to be better? Especially when no remastering is involved? That is: Same PCM source, one downsampled to 16/44, one MQA encoded and played undecoded... Is the latter one truly better? Some studies by the maker of HQPlayer seem to indicate this is not the case, but what about listening tests? Just curious.

anupmc's picture

MQA tracks on the $249 MQA-capable Meridian Explorer 2 doesn't even sound as good as a non-MQA capable $599 Chord Mojo! IMHO, MQA is a non-starter, the basic premise of why MQA is necessary is as old as Bob Stuart's MLP/DVD-A vs. DSD/SACD arguments :)

Michael Lavorgna's picture
...to offer such disrespectful comments without offering anything of substance?

I do not get the sense that you know what you're talking about from your brief comment.

miguelito's picture

They are certainly skeptical... The burden of proof is on MQA!

Michael Lavorgna's picture
Since anyone can buy an MQA-enabled DAC today, as well as MQA encoded music, the burden of proof is within anyone's grasp. To suggest that MQA needs to continue addressing every skeptic, no matter how, um, un-rigorous their skepticism, is un-reasonable, imo.
anupmc's picture

Quote: ...to offer such disrespectful comments without offering anything of substance?
I do not get the sense that you know what you're talking about from your brief comment.


Disrespectful?? How is it disrespectful when I offer my opinion? You offer your opinion all the time, how is yours less “disrespectful” than mine???

That said, to be clear, anyone can go try this for themselves. When a friend and I compared a (2L) 24/352.8K PCM non-MQA track on the Chord Mojo versus the MQA version of that same track on the Explorer 2, the Chord was far better.

Same track, 24/192K PCM non-MQA version on the Chord Mojo against the same MQA version on the Explorer2. The Chord Mojo was STILL better.

With Broadband being so pervasive (both fixed and wireless), the compression benefits MQA provides is practically prehistoric and completely unnecessary :)

By the way, try any 24/192K PCM non-MQA’ed track on the Explorer 2… it sucks big time. I wouldn’t be surprised if Meridian intentionally cripples non-MQA PCM playback on the Explorer2 just to make MQA versions sound better.

Michael Lavorgna's picture
That bit is disrespectful. Not to mention that it misses the point.

With MQA, streaming services need *one* copy to serve their customers. Think about that ;-)

I've compared MQA on 3 different MQA-enabled DACs and have heard a number of A/B comparisons on others. If you read my review of MQA, you'll see that we agree on the importance of the DAC.

miguelito's picture

"With MQA, streaming services need *one* copy to serve their customers. Think about that ;-)"
They could have one copy of the high res file and downsample on the fly... Really not a good argument... :)

Michael Lavorgna's picture
...between storing and serving 1 copy of 50 million+ (and growing) tracks versus storing and serving 2 copies?
anupmc's picture

Quote:
"the basic premise of why MQA is necessary is as old as..."
That bit is disrespectful. Not to mention that it misses the point.
With MQA, streaming services need *one* copy to serve their customers. Think about that ;-)
I've compared MQA on 3 different MQA-enabled DACs and have heard a number of A/B comparisons on others. If you read my review of MQA, you'll see that we agree on the importance of the DAC.

Nothing disrespectful about that. It’s FACT.

The most foundational aspect of MQA is to fold a wide-band audio signal into a smaller footprint. i.e. compression! It’s very little more than a modern day Meridian Lossless Packing (except it’s lossy this time!!).

As for MQA’s “timing” related corrections and preservation, anyone with even basic engineering knowledge knows thats just an updated version of Craven’s Apodizing filter technique, applied both at the encoding and the decoding stage for MQA.

IMHO, this whole MQA thing is nothing more that Bob’s old work repackaged :)

No one has to take my word for it, go listen for yourself and read the MQA AES papers.

Michael Lavorgna's picture
As for MQA’s “timing” related corrections and preservation, anyone with even basic engineering knowledge knows thats just an updated version of Craven’s Apodizing filter technique, applied both at the encoding and the decoding stage for MQA.
Anyone with even basic knowledge of the MQA process knows this in not correct.

I have listened for myself, more than most.

anupmc's picture

Quote:
Anyone with even basic knowledge of the MQA process knows this in not correct.
I have listened for myself, more than most.

Speak to Bob Stuart yourself and ask him whether MQA's "timing" preservation is in any way related to Craven's Apodizing filter technique, and tell us what he says (obviously he's not going to tell you the exact details because its proprietary confidential).

If I'm COMPLETELY wrong, I will gladly admit so and apologise to you here in public. But if I'm not completely wrong, if it is in fact a related technique, then I expect you to admit you're wrong, and apologise here in public.

Deal? :)

Michael Lavorgna's picture
..."thats just an updated version of Craven’s Apodizing filter technique"

Now you are saying, "... MQA's "timing" preservation is in any way related to Craven's Apodizing filter technique"

These are two different statements with two different meanings. I do not disagree with the later but the former is clearly wrong.

anupmc's picture

Quote: You said...
..."thats just an updated version of Craven’s Apodizing filter technique"
Now you are saying, "... MQA's "timing" preservation is in any way related to Craven's Apodizing filter technique"

These are two different statements with two different meanings. I do not disagree with the later but the former is clearly wrong.

Knowing Bob (figuratively), even if MQA's timing capability is a direct derivative of Craven's apodizing technique, he's hardly going to admit it and will likely only suggest that it's related (because of it's proprietary nature) - that's also a clue to the so called "tailoring" that they do for 3rd party DACs, it's just playing with filters! :)

By the way, I notice you're no longer insisting I'm being "disrespectful" (about MQA's compression technique being just an updated MLP)... Why? Did I demolish your defence so easily? :)

Michael Lavorgna's picture
You were being disrespectful - I have not said otherwise.

I also said you appear to not know what you are talking about and that opinion has also not changed.

I know Bob Stuart as well and we have discussed MQA in-depth on numerous occasions. He has also answered every question I've asked to my satisfaction.

Odd that you would use the phrase "demolish your defense". Do you think we're playing Battleship?

anupmc's picture

Quote: You give yourself too much credit.

You were being disrespectful - I have not said otherwise.
I also said you appear to not know what you are talking about and that opinion has also not changed.

I know Bob Stuart as well and we have discussed MQA in-depth on numerous occasions. He has also answered every question I've asked to my satisfaction.

Odd that you would use the phrase "demolish your defense". Do you think we're playing Battleship?

Counter technical argument, bring it on. Merely using adjectives to describe my comments is rather childish.

You may know Bob, but he's smart enough to mask MQA's real technical details under a cloak of marketing that audio journalists seem to fawn all over :)

Michael Lavorgna's picture
But you are clearly smarter...

I've already countered your not-very-technical argument but you may have been too busy with Battleship to notice.

You have a nice day.

miguelito's picture

I DO NOT believe that Meridian would intentionally cripple a non-MQA PCM stream just to make MQA sound better. Seriously, that's BS. What I do expect is that the "DAC tailoring" actually benefits a cheap DAC like the Explorer2 quite a bit.

anupmc's picture

... If not, please give it a listen and tell us what you hear :)

miguelito's picture

But this could be a good example of how DAC tailoring improves a cheap DAC. On the other hand, frankly for $249 I can buy a bunch of very nice PCM DACs. I got a Dragonfly Red recently and it is truly amazing for just $199. iPhone listening has never been this good.

miguelito's picture

The jury is frankly still out. There are no widespread comparisons of whether MQA is a benefit or not. I listened to MQA at Meridian in NYC in March 2015. The MQA versions of the file (a total of 6 I think) did sound markedly better. But this is not a way to assess... Did the files sound better thanks for an MQA beautification? Or where they actually remastered? Impossible to tell. Only widespread comparisons will resolve the issue. But MQA is really running out of time with this...

Michael Lavorgna's picture
I like that.

There now exist any number of reviews and reports on comparisons where the source files are clearly identified. For example, in John Atkinson's review in Stereophile, some of the files he used for his comparisons were his own recordings, taken from his source files. You can't get a better A/B than that.

At the much maligned CES 2016 demos, I heard an A/B using a recording of Peter McGrath's and Peter was there so I spoke to him about the process. I also followed up with Peter in a phone call after CES to get more detail about his experience with his recordings and MQA.

If you read my review of MQA, or my CES 2016 show report, you'll see that the Mytek Brooklyn DAC allows you to enable/disable MQA decoding with the push of a button. This A/B obviously uses the *same source file*.

Also in my review of MQA, I was sent a number of MQA-encoded files of my choosing along with the non-MQA source files (which I already owned). So my comparisons were between music I knew very well from known sources.

I'd also recommend reading John Darko's MQA review as he addresses many of the technical points we are discussing here.

miguelito's picture

To be clear: I am not against MQA. It does bother me that they are requiring hardware decoding - in my opinion for less than altruistic reasons as I explained elsewhere. I think they would do themselves a favor by relaxing this requirement.

As for the A/B you're talking about, if I understand correctly you mean: play the MQA-encoded 24/48 file and switch between not-decode/decode. Of course that should be better - if it weren't this would not be a topic. Is it better than the true original PCM, assuming it was done properly? You say it is, and I believe it, and I would attribute that to MQA deblurring of the file rather than tailoring to the output stage of the DAC.

When I said "the burden of proof is on them" I was unclear... I meant that I don't believe that final-stage tailoring to the DAC chip used can be all that important - at least for high quality DACs. BTW... Do they tailor this to the chip alone or the chip+analog stage in the DAC? Bc the analog stage will also have an important impact on the sound...

Ultimately it seems to me this is all about licensing.

Michael Lavorgna's picture
...who has written about these A/B comparisons and in every report I've read there is a clear preference for MQA.
I don't believe that final-stage tailoring to the DAC chip used can be all that important - at least for high quality DACs
How can you prove this to be true? I heard an MQA demo at RMAF with the MSB Select DAC...

Again, software decoding is on the MQA road map.

So you agree that MQA can improve sound quality while decreasing the file size of high res yet "it is all about licensing"? I don't follow.

miguelito's picture

"So you agree that MQA can improve sound quality while decreasing the file size of high res yet "it is all about licensing"? I don't follow."

I agree that decoded MQA sounds better than the original file in all examples I have listened to and believe the reviewers's conclusions as well. Not sure whether this is due to remastering, deblurring, or DAC tailoring.

What I am asserting is "all about licensing" is the denial of software decoding to a standard PCM stream I can play on any DAC.

Michael Lavorgna's picture
The "remastering" thing has been addressed, and discounted, in reviews that have used the *same source file*.

You keep saying that MQA is not going to offer software decoding. I keep saying they are going to offer software decoding, based on conversations with MQA. If you can point to a statement from MQA that supports your claim, we can stop going 'round and 'round.

miguelito's picture

The only thing I have heard of is that software decoding would only exist when connected to a "know DAC" - there will be no "generic DAC" decoding. If I am wrong then by all means correct me.

Michael Lavorgna's picture
...that the ideal MQA implementation is end-to-end, that software decoding when connected to a known DAC is not-as-ideal, and that software decoding to a generic DAC is the least ideal. I have never heard that there will be no generic DAC decoding.
miguelito's picture

Then I'll stop harping about this! Please do ask your sources specifically about this, I am very interested.

As a side note, the loss of Roon in the MQA partner list is very sad, especially since these folks came from Meridian. For me at home Roon is an absolute must - I will not put up with anything else (Roon + HQPlayer to be precise).

Michael Lavorgna's picture
Cheers.
miguelito's picture
miguelito's picture

This makes sense - I would not be surprised to see that hardware decoding is best, followed by software, followed by generic. But the flipside of that is the use of post processing such as Dirac, which for some folks is a must.

Steven Plaskin's picture

Check out the videos for more information on MQA:

http://www.msbtechnology.com/mqa-has-arrived/

X