Meridian's MQA

From the press release:
Developed by Meridian, MQA is a breakthrough technology to reverse the trend, in which sound quality has been continually sacrificed for convenience. Vital elements of our music have been thrown away to fit thousands of songs into a pocket or millions in a cloud. With MQA there is no sacrifice; it brings us right back to the enthralling sound of live music. MQA captures and preserves nuances and vital information that current music files obscure or discard, but in a file that is small and convenient to download or stream.
MQA stands for Master Quality Authenticated and is Meridian's answer to delivering high resolution quality in a lossless compressed proprietary format that squeezes down even 24/192 files to roughly the size of CD-quality files. But that's not all.
MQA, or Master Quality Authenticated, brings together the three ideals of studio-quality sound, convenience and end-to-end authenticity. It uses a completely new concept of capturing the total essence of an original recording and conveying it all the way to the listener, assuring that what they are listening to is identical to the master recording.
Meridian goes on to explain that the MQA process "is informed by the latest neuroscience and psychoacoustic research that shows how we identify and locate sounds, and that timing details of a few microseconds are important." And "MQA can be delivered inside any lossless container, e.g. as ALAC, FLAC, or WAV." Of course you need an MQA-compatible decoder in your system which can reside in an app, software player, or hardware. The MQA "encapsulation" technology works with sample rates up to 768kHz.

Meridian also recently announced the Meridian Explorer2 ($299) which supports MQA lossless files. One interesting aspect of the MQA technology is you can play back an MQA encoded file on a non-MQA device but it will be played back at CD-quality, not the full original resolution.

Meridian's MQA technology can be used for digital downloads or perhaps of even greater interest, high resolution streaming.

For more on MQA, see Meridian's MQA website.

COMMENTS
DH's picture

I think the tech concept is interesting. I'm curious if it really sounds better than what we have now.

Eltonnotjohn's picture

Oh yeah?

"For the first time in the history of recorded music you're really hearing the artists performance"

So none of us have ever heard 192/24 WAV or AIFF if you believe his claim above?

And he's got a nice chart too, showing a steady decline in quality from the master tape though high speed open reel, vinyl, cassette, CD (a brief flurry of 'quality' with that) and then descending furthest of all to what we listen to now. Does he think we ALL listen only to low bit-rate MP3? Sure, his encoding MAY be better than that, as is AAC or ALAC. But his 'history' comment implies we have NEVER heard high res 'pure' PCM, which is simply untrue.

If he is so clever, why is all the Meridian gear "fairly good, but not brilliant or outstanding" despite its very high price?

Eltonnotjohn's picture

I've just been listening to Handel's Water Music, Music for the Royal Fireworks, Vivaldi's Four Seasons, and Clannad's Legend ripped from CDs using EAC, which says 'error free' for all of them and also checks with Accurate Rip and gave very high confidence levels. On pretty expensive equipment. Good enough to show that WAV is BETTER than FLAC, as some say (though being an 'Objectivist' I hope I am deluding myself on that).

Sure 'only' 44.1 and 16 bit, but I have never heard anything like it. I only tried it because some crank on Computer Audiophile said he could hear the difference. Maybe he is not so cranky after all.

If Meridien's coding method even comes close I will eat my hat, and yours too.

Michael Lavorgna's picture
But I'll buy a big ten gallon number just for your challenge ;-)

Obviously the proof of this concept and its implementation lies in the listening which we'll hopefully have an opportunity to do sometime in the not too distant future.

rompolompo's picture

"WAV is BETTER than FLAC"

In what way? Both are PCM, both are bit identical files. FLAC has standard support for meta data while WAV does not.

deckeda's picture

Fa fa fa fa fa fa fa fa far far better
Run, run, run, run, run, run, run away

PDQ.Bach's picture

And you may ask yourself
How do I work this?

Wavelength's picture

Really?

So you take a 24/192 file and through MQA input and out is going to be lossless, but it's the size of a 16/44.1 file. I have a degree in math and I would like to see this happen. Especially since if you play it back on non-MQA equipment it plays fine at 16/44.1???

Lossless is the key here and I don't think there is mathematically anyway that it can be lossless with this wording.

I would like to see Stereophile take a 24/192 track encode it look at the size of the file. Then play it back using the Meridian system with DIGITAL output and record that file and see if it is identical to the 24/192 original track.

I will take bets that it is not.

Thanks,
Gordon

jazz and cocktails's picture

and i love your work, but i think there's something entirely new here. bob has been working on this problem for 10 years, using entirely new research.

as Michael says, the proof will be in the listening, but i think we will all be very pleased/surprised.

this is what pono was meant to be.

deckeda's picture

It can't be. Re-read what they're saying amongst the "customers are owed the best yada yada ..."

Math is a "finite resource" if you will with regard to how much you can losslessly compress any file. Anything else is tossing data away, i.e. the psychoacoustic modeling Meridian necessarily mentions.

That's not to say it couldn't sound better than a straight 16/44 file, but it does beg several questions, namely:

1) In a world where not even lossless 16/44 is the norm for selling music online (where trying to save download bandwidth is a goal while maintaining fidelity) how is offering yet a different 16/44-ish sized file attractive to labels, to Apple's iTunes Store?

2) FLAC, ALAC, AIFF, WAV will cease to mean lossless here -- more confusion.

3) So, it's not a "Pono" solution of selling you the best-available file, it's selling you a file that sounds better than a CD for about the same size but not the best potential file available.

It's a "practical"l but somewhat cynical solution aimed at audiophiles (the only, and minor, clientele even remotely interested in sound quality) -- the very group that's not a fan of compromise, either sonically nor (and this is important) philosophically.

Which brings me to: ignore everything I just typed and let's say these files sound fantastic. And so ... it's a solution for audiophiles who won't buy more drive space? Don't care about obtaining the best-available file?

John Atkinson's picture

Wavelength wrote: So you take a 24/192 file and through MQA input and out is going to be lossless, but it's the size of a 16/44.1 file. I have a degree in math and I would like to see this happen. Especially since if you play it back on non-MQA equipment it plays fine at 16/44.1???

Hi Gordon, see the article on MQA I posted today at Stereophile.com: www.stereophile.com/content/ive-heard-future-streaming-meridians-mqa.

Wavelength wrote: I would like to see Stereophile take a 24/192 track encode it look at the size of the file. Then play it back using the Meridian system with DIGITAL output and record that file and see if it is identical to the 24/192 original track.

I have sent Bob Stuart some of the hi-rez masters for my recordings for him to encode with MQA. I will therefore be able to do some tests when I get the encoded files back.

John Atkinson

Editor, Stereophile

CincyFlyer's picture

Note that they are comparing the MQA file - which is a lossless compressed format, with CD audio, which is uncompressed. Depending on content, data cannot be arbitrarily compressed to any given size, but I'd expect the MQA to be quite a bit smaller than a [say] WAV file with the same content. Just for fun I compressed [using zip software 'cuz it's handy] some WAV recordings that I had made, and they dropped at least 30% in size, and some much more than 50%. Add to that different ways of encoding data, and it's quite possible that what they are saying is correct.

Maury's picture

Isn't Meridian just the latest in a long series of companies that have developed various enhancements to digital Redbook processing? The Sony SACD Hap Z1ES uses a "remastering" engine, there was HDCD, there have been all kinds of "filters" such as on Oppo, Esoteric to impart different sonic flavors. I'm sure I am forgetting 50 more types of enhanced processing so called that the sales forces of digital audio have trumpeted over the years. We just need to be given a new pair of ears and then we can listen to MP3 and be happy.

bobflood's picture

getting a single serious studio/production company to adopt another proprietary encoding scheme with all the rights and royalty issues involved. Why would they go to the trouble when one out of a thousand of their customers would care enough to pay the extra costs to produce and distribute a piece of music in this fashion and would have the required playback equipment and software. As another person said here, this is very much like HDCD which went absolutely nowhere.

Eltonnotjohn's picture

Like him, I've got a degree in maths too. Whether or not you think 'high res' sounds better than regular 44.1 or not (I'm not yet convinced but I would like to be) you simply won't get 192's worth of data into a 44.1 file. And as others have said, it's just another 'proprietary' coding. HDCD failed, and you don't hear much about 'Mastered for iTunes' either, do you? And if Apple can't get studios to use it Meridien stands no chance.

Meridien in general, and Stuart in particular, are always talking about 'breakthroughs' and such. Yet their actual equipment sounds no better than any other overpriced, fancy looking 'Lifestyle' audio. 6000 plus dollars for the Control 15? It's just a touch screen PC with a not particularly big hard disk. And it doesn't even contain a DAC, let alone a good one like Gordon's. And as Meridiens stuff is all 'proprietary' you get locked in. This feels like another part of that 'lock-in' process to me.

If Stuart's company produced REAL good sounding 'High End' stuff for the high end prices he charges one might pay more attention to his talks and 'white papers', of which he produces rather a lot.

Drtrey3's picture

Reads like they are using psychoacoustics to decide what bits to throw away. Mp3s do the exact same thing using the exact same science. So does 16 bit perfect sound forever.

Not interested.

Trey

PDQ.Bach's picture
  • There is but one small word in Michael’s report that carries the spell of death to MQA: proprietary.

    How many proprietary solutions have made it so far?

    You need to be the size of Apple to make a dent with proprietary, and it isn’t toasted even then.

    Assuming it’s really good, Meridian will need to sell licenses from Hoboken to Betelgeuse (via Shenzhen, naturally).

  • Just glanced at the revamped Meridian website and… …OMG.

    When a technology-minded company like Boothroyd-Stuart catches MadMen®-speak, it catches it in a very acute form.

    Supercalifragilistic.

    Personally, I find it repulsive. Not in tune with their past achievements. And I sense a marketing mishap in the making, where skeptical minds are concerned.

Maury's picture

This statement by Meridian is really an amazing assertion and worded in such a way as to be impossible to disprove:

"MQA, or Master Quality Authenticated, brings together the three ideals of studio-quality sound, convenience and end-to-end authenticity. It uses a completely new concept of capturing the total essence of an original recording and conveying it all the way to the listener, assuring that what they are listening to is identical to the master recording. ":

So this is an encoding which does not pretend to be a bit for bit copy but instead is a new way to reconstruct a digital soundfile. However, this new method uses the rhetoric of MP3. Doesn't MP3 claim that it uses psychoacoustical research on how to reconfigure the original soundfile, discarding what can be discarded, saving what needs to be saved to provide the "essence" of what the original recording sounded like. Meridian is asserting that the "essence" of say an analog mastertape is encoded in abbreviated but somehow totally reconstructable form such that the listener on whatever system is assured that they are listening to the master tape. So if they don't like it then the fault lies in the essence of the mastertape, because there is complete assurance that the essence has been faithfully transmitted.

Eltonnotjohn's picture

"WAV is BETTER than FLAC" (so I said)

From rompolomoo -

"In what way? Both are PCM, both are bit identical files. FLAC has standard support for meta data while WAV does not."

I know. It should be total nonsense that there can be any difference. 35 years designing 'big' computer, and a maths degree tells me so. Bits are bits, there is no 'jitter' in a file as there is no timing information so the timing can't be 'out', leading edges don't matter, nor do accurate reference voltages as zeros and ones are detected in the middle of the pulse, etc. etc. So 'noise' from the power supply or any other source simply won't matter. This is not 'new', it all had to be so from ENIAC on or computers would not work at all.

And, leaving the computer, USB (to the DAC) uses NRZI coding so roughly 50% of the pulses won't have leading edges to detect anyway.

FACT. There are all sorts of computer audiophile myths. All they show is that the people who believe the myths, including many 'respected' professionals in the HiFi magazine field and in the audio field simply don't have a clue how computers actually work. However, even give their lack of understanding, one of the 'myths' appears to be true. I don't know why and neither can they, though some might pretend to.

So I thought I would test. Much of my music is ripped CDs in FLAC. I normally use JRIVER 19 on Windows 7 and it's fine. I rip using it too, using Secure Mode (but not the Accurate Rip database check with other peoples rips as JRiver does not support it).

But a supposed 'audiophile crank' on Computer Audiophile said rippers vary and EAC is the best (it does do lots of sophisticated checks) and a linear power supply for the CD rip drive is better than the regular SMPS. So I tried it. As I thought it was all nonsense so wasn't going to spend time doing something I didn't believe in piece by piece I did it all in one go, EAC, linear power supply, and for good measure WAV rather than FLAC. Using three of my favourite pieces, two Handel CDs and one by Clannad, an Irish folk group.

It really DOES sound better!!! Much better, though I thought it sounded 'perfect' before. Why, I haven't a clue.

BTW - WAV metadata. As you say,there is no 'standard method'. But WAV actually supports tagging perfectly well, though many says it doesn't support it at all. They are wrong. EAC tags WAV files, as does JRiver. And they both understand the others tagging, and my two Ethernet streamers, one a Cambridge Audio Stream Magic and the other a Naim, understand them both as well, displaying, from the WAV files, all the usual metadata on their screens.

rompolompo's picture

You can tag WAV files but this is forcing the meta data on the files's header. Although non standard (as I mentioned) but works on some applications. As for the difference between the files, it may be relevant to the way JRiver handles FLAC files. A year ago or so, we had 10 audiophiles trying to hear a difference in an ABX test and no one was able to hear a difference. Of course the setup was different. A CD was ripped to WAV and then converted to FLAC. Linux was used throughout to rip, convert and play via a UAC 2.0 compatible DAC. Again, 10 audiophiles in various ages and with years of experience. No one heard any difference. Not even one.

Eltonnotjohn's picture

I don't understand it at all. What I am hearing shouldn't be possible. My non-audiophile wife picked out the difference every time (she encourages me in my radio-controlled model plane eccentricity too), as did a non-audiophile friend. Neither has any knowledge of or interest in this stuff so it's not expectation bias.

You say it may be how JRiver handles FLAC. That might be true up to and including version 18. It converted from FLAC to PCM on the fly. But I use version 19, which does the whole track conversion before playing it. So the necessary FLAC to PCM conversion is done beforehand, avoiding the extra processing while the file is playing. Not that JRiver believe all this nonsense anyway. Their view, and I agree, is that all 'bit perfect' players should sound the same. They only go along with 'play from memory', prior conversion, etc. to keep cranky audiophiles quiet.

Back to Meridian. I don't say their new process won't be any good. I just take objection to (1) All the BS about we have never heard anything as good before (2) That Stuart is very prone to writing so-called 'white papers' and giving 'expert' talks while his Meridian stuff is actually rather 'ordinary' sounding and more of a 'lifestyle' product than top end HiFi. Sort of a British Bang & Olufsen but not as good. So I don't tend to give Stuart's views much weight.

Archimago's picture

Eltonnotjohn:
You should explore this further. I bet there's something unusual in your settings if indeed the difference is so significant.

Much more likely that there's something odd happening in the decoding process than that lossless formats sounding different.

Archimago's picture

And of course no evidence of lossless file formats measuring differently...

Eltonnotjohn's picture

It should be nonsense. My setting are fine, not that there is anything in JRiver or EAC that you can set to interfere with the coding/decoding process anyway.

Yes, there is no measurable evidence. Lossless formats code and decode correctly, we all know that. Only the 'crazies' believe different. Someone, as I remember either Mitchco or Julf, even went so far as to code, decode, code again, decode, a quarter of a million times and the end result was bit-identical, as of course they knew it would be. But STILL the crazies claimed to hear a difference.

Then there is noise. Lots of people say to use linear power supplies for the computer, any external drives, even for a NAS. That's nonsense too. I can turn the volume up to maximum with the preamp input shorted, computer, external drives, and NAS all disconnected and unplugged from the wall sockets and if I put my ear right up to a speaker I can just about hear a very faint hiss from the Naim NAP250 power amplifier I use. That of course is what one expects. Then I start it all up and with JRiver doing nothing or 'playing silence' the noise gets no louder and doesn't change in character. Which PROVES, very simply, that all the work on expensive linear power supplies, 'optimising' the computer, etc, is a complete waste of time and effort. Which I expect people like you and me knew anyway.

I know why they claim all these things. When we used turntables we could have hours of fun 'tweaking', trying different cartridges, etc. and the differences were clear. Now, with computer audio, having achieved the simple goal of bit perfect output the 'tweakers' don't have anything to do. Nothing has any effect. And they don't like that but do it anyway :)

But the fact remains I do hear a difference. Which is crazy!!!

selarom's picture

To me MQA sounds like a mix between what was HDCD and MP3. Call it MP3-HD!

I say this because it uses "the latest research in psyco-acoustics" to take everything above 16/44.1 and selectively add the information they deem necessary to make it sound identical to a 24/192 master by including it as extra metadata which, upon using a compatible player, can be accessed. This is how it all sounds like to me.

So more realistically, take a 2L DXD master at 24/352.8 (which can be up to a 1GB just one track) and MQA' it. Then play it back on a player without MQA, so it should be a 16/44.1 quality track... Put it on a MQA player and it should output a pseudo-DXD file that is supposed to sound identical to the original DXD file.

I think it will output a 24/352.8 but actually the frequency and dynamic range response will be different from the original, adding the extra bits for dynamic range and the samples for ultrasonic frequency only where they think is necessary. But it has to output a sampling rate higher than 44.1, otherwise it would not reproduce the ultrasonic frequencies nor transients they plan to reproduce, etc.

And if the master is a 24/192 or 24/96... then it should playback as 16/48 so as not to re-clock the signal unnecessarily to 44.1

EternalSounds's picture

Whatever and however it gets implemented, it sounds like an industry changer. I'll wait until mid-2015 before upgrading any of my equipment/system.

drblank's picture

According to another site, both Panasonic and Onkyo have announced they are going to be adding MQA to future products and 7Digital is supposed to be offering tracks encoded with MQA.

Here's the article.
http://www.monoandstereo.com/2014/12/7digital-and-meridian-audio-global....
http://www.monoandstereo.com/2014/12/7digital-and-meridian-audio-global....

I'm wondering if HDTracks, iTunes, and other reputable download suppliers are going to be added as well as the 3rd party music players adding the decoding (or if we have to had a decoder chip AND software to play back this content). If this technology requires a decoder chip AND s/w to provide the playback, this is going to be a tough sell for the masses, but if it just requires s/w with pre-existing DACs, then it will be a lot easier to adopt providing we can get the content we want.

Wavelength's picture

John,

First happy holidays to you and Laura. Skipping CES so hope to see you sometime soon.

~~~~

The obvious problem with any type of encoding like this is that what is the difference between what is considered noise and low level detail. We have all seen this in practice with really well recorded red book material.

Second the original use of lossless is then totally untrue as the input would never equal the output in terms of numerical data. It would only come close to the original for those companies willing to pony up the big cash for the algorithm.

Third isn't this just what Bob was going to do when with the Pono camp when the deal was the record companies wanted a) DRM b) High Res c) A red book output format?

Thanks,
Gordon

CG's picture

Personally, I think it's great that they've worked out a way to compensate for the filter characteristics of the various ADCs that were used in the conversion or even in the initial recording.

But... Just how are they going to correct for the zillions of filters used in the Pro Tools, ahh, editing (?) of the original tracks? Won't each of those 32 or 64 tracks used for the mix be filtered differently?

It's too bad that the audio world didn't follow the photo world by embedding exif-like metadata in each pcm or dsd track so that the DAC end would know how to reconstruct the original recording. Every DAC manufacturer could still put their own twist on this, but at least it wouldn't be guessing.

drblank's picture

and that Tidal is going to be streaming MQA audio. Here's the article...

http://www.monoandstereo.com/2015/01/tidal-x-meridian.html#more

I just wish I could get a s/w plug-in or have one of the s/w apps that I use (PureMusic, Amarra, BitPerfect, etc.) would add MQA decoding and I didn't have to buy a new DAC to be able to listen to MQA audio through my existing system.

drblank's picture

and that Tidal is going to be streaming MQA audio. Here's the article...

http://www.monoandstereo.com/2015/01/tidal-x-meridian.html#more

I just wish I could get a s/w plug-in or have one of the s/w apps that I use (PureMusic, Amarra, BitPerfect, etc.) would add MQA decoding and I didn't have to buy a new DAC to be able to listen to MQA audio through my existing system.