News

Lenbrook Acquires MQA

Jason Victor Serinus  |  Sep 19, 2023

Note: This is a developing story. Updates will be posted as they unfold.

Lenbrook Corp, the privately owned Canadian enterprise whose holdings include NAD electronics, PSB speakers, and Bluesound (the maker of the BluOS music operating software system) has acquired the assets of MQA, Ltd, including MQA technology and the SCL6. The press release announcing the acquisition, which went public September 19 at 8am EDT, notes that the deal "further solidifies Lenbrook's commitment to excellence and innovation in the evolving landscape of audio technology."

The announcement ends months of speculation that began in April, when MQA entered receivership. An accompanying FAQ affirms that "As one of MQA's most significant licensees and also the owner of the award winning BluOS high-res content platform . . . Lenbrook is in the business of providing high resolution audio experiences [to] informed customers who appreciate innovation and value having options . . . We believe MQA fits this mission as the research that makes up the foundations of the technology are based on neuroscience and cutting-edge digital sampling. Although MQA is a digital technology, it is an analog-to-analog conception and not simply a digital codec. Put simply, the MQA Encoder corrects for the A/D converter, 'deblurs' that signal and then uses a package that is much more efficient than regular PCM. Fans of MQA speak to its improved transparency, noise stability and temporal effects."

MQA has attracted many critics since the technology's release several years ago. Lenbrook's FAQ addresses the controversy head-on. "We have always found it unfortunate that the core attributes of what we understand MQA to be seemed lost in a distorted narrative around some of the technical nuances in its implementation," the FAQ states. "In this fray, the artist-first origins of MQA and the sheer technical elegance of its handling of the entire audio signal path got muddled. We are excited to have the opportunity to clarify the narrative and build on the technology in ways that can better demonstrate their true value, while also promoting innovation in a specialty and premium audio industry that thrives on healthy discussion, subjective views, and debate.

"Lenbrook's position is that anyone doing work to advance audio processing and sound reproduction is positively contributing to the vibrancy of the industry. The vitriol directed towards innovations like MQA and what it means to those creating, delivering, and listening to better sounding music has always disappointed us when the technology and the patents that underpin it are so novel.

"We prefer instead to build off the fact that many influential content creators and reviewers absolutely understood that MQA was not really about 1s and 0s. We also believe that differing opinions is what makes this industry healthy—for example, we do not believe in one way to design a speaker and carefully approach product development in ways that offer differentiation and respect for individual listening preferences. A specialty hi-fi industry where there is no debate or new ideas would be commoditized far too quickly."

Lenbrook affirms that MQA "was born of a vision that a group of like-minded musicians and audio engineers had to give musicians the tools they needed to capture their works in high resolution . . . We have listened extensively to MQA content and believe in the results of what we actually hear."

The affirmation of MQA as a vital, high-resolution codec that honors the intent of artists and engineers was echoed by prominent Grammy winning producers and engineers. Reached in Norway, 2L's Morten Lindberg stated via email, "I've had the great pleasure and privilege to work with Bob Stuart since the early days of his development of MQA. Thru hundreds of critical listening hours, I have really come to appreciate this tool brought to our sonic craft. I'm very optimistic to the future of MQA. And for the record: I have absolutely no business interests in any of the companies."

In the press release, Lindberg stated, "For 2L, using MQA has allowed us to enhance the experience of our recordings, beyond the raw capture, with increased access to sonic details, transparency and lower listening fatigue."

George Massenburg, Grammy and Academy of Country Music Award winning producer and recording engineer, also lent his endorsement. "I'm so relieved that MQA and SCL6 will continue under Lenbrook," he stated for the press release. "MQA's technology, with its faithful rendering of detail, complexity, and soundstage, gave us the reason to go back into the recording studio and reverse a 20-year decline in the quality of audio delivery methods."

A "select group" of MQA's UK-based employees are joining the Lenbrook team. These employees will remain in the UK. While Bob Stuart, MQA's inventor/founder, will not join Lenbrook as an employee, he will serve in an advisory capacity focusing on MQA and SCL-6 product development. The licensing model for MQA and SCL6 will not change. The press release noted that record labels, artists, and producers continue to encode and upload new music in MQA to Tidal daily. "We also support consumer choice, and [Tidal's] current 'Max' labelling does not allow consumers to search for content in their preferred format easily and that is where our reservations about it come in." Recent changes to the Tidal app blurred this distinction, but Tidal appears to be stepping back this change.

Lenbrook describes SQL6, a more recently developed technology from the MQA team, as a "time-domain optimized scalable codec" with applications in wireless audio. "The technology is versatile and also suitable for applications in streaming and broadcast. SCL6 provides studio-quality sound even at low data rates and can be scaled rapidly and without audible interruptions. It is also worth noting that SCL6 is source agnostic, supporting PCM audio as well as MQA."

In a May 2023 Industry Update that appeared in the print edition of Stereophile, Julie Mullins discussed SCL6, which at the time was being marketed as "MQair." Billed as "an advanced codec created to offset wireless streaming's bit-depth and sample-rate transmission limitations," she described it as "the equivalent of a Bluetooth audio codec that utilized core MQA ideas." SCL6 is said to be scalable from below 200kbps to20Mbps, covering Bluetooth, Ultra-Wideband (UWB), and WiFi connections. SCL6 supports MQA and PCM datastreams with a sample rate up to 384kHz.

Lenbrook brand PSB, which has long made Bluetooth-based noise-canceling headphones, has announced that it will release a headphone in the first quarter of 2024 that incorporates the SCL6 codec via Sonical's CosmOS, an "ear-computing platform" incorporating a microchip designed for wireless headphones and earbuds. Sonical says that CosmOS uses UWB radio technology that provides a higher data rate and very low latency for more accurate sound and performance, offering potential advantages to headphone manufacturers and users worldwide.

Check back for updates.

Footnote: Stereophile's coverage of MQA can be found here.

COMMENTS
cgh's picture
Submitted by cgh on September 19, 2023 - 5:44am

Wait a few months, rebrand it with a new name, and re-launch it as something new and shiny. Certain facts that are always conveniently left out of this story is the marketing hype around the original launch. Something like this will always have its detractors, but the roll out of v1.0 was nonsense talk, and you can't fault an incredulous group of people salivating for the next best thing to kick the tires and call BS. Good luck to Lenbrook and thank you if this creates a path forward that buries the MQA name in the history books.

miguelito's picture
Submitted by miguelito on September 19, 2023 - 5:49am

the vitriol came from blatant mischaracterizations of what MQA was. "As the artist intended" Really? What artist validates the CHANGES MQA makes to the sound - and when? And why? Are you telling me high end digital workstations cannot get the sound right? I could go on and on... But I will also say I am in favor of new tech, but not in favor of tech that is trying to fix a problem that does not exist.

mns3dhm's picture
Submitted by mns3dhm on September 19, 2023 - 9:13am

+1

JRT's picture
Submitted by JRT on September 19, 2023 - 6:28am

Like MP3, MQA is lossy compression, while FLAC is lossless.

While lossless FLAC is free, lossy MQA is proprietary and promoted due to profit motives.

Stereophile more usually promotes perfectionist audio, and lossy MQA does not advance that agenda, rather advances the agenda of the rights holder promoting the non-free proprietary lossy compression technology.

Poor Audiophile's picture
Submitted by Poor Audiophile on September 19, 2023 - 6:11am

Looks like it didn't generate much profit.

JRT's picture
Submitted by JRT on September 19, 2023 - 6:44am

Do you really think Bob Stuart did not generate any significant personal income from MQA?

Poor Audiophile's picture
Submitted by Poor Audiophile on September 19, 2023 - 7:12am

I don't have a clue and I doubt you do either. My point was, if they were in "receivership" I think that means they were broke.

cgh's picture
Submitted by cgh on September 19, 2023 - 7:36am

It might be more complicated than that. I have no info and am speculating. Perhaps the name has value (was was involved in a contingent capital deal that collateralized the name of a well known large cap company, the name having a measurable $ value) and perhaps the company itself has future value, which is common in bankruptcies. It's a liquidity problem. They may have done analysis and determined that there's a decent multiple on ebitda internationally. They may have use cases for any tech, MQA or other skunkworks, at the other companies under their umbrella. It could be a play on patents. Who knows. While they were most certainly broke, I assume that means that they couldn't keep the lights on, not that there wasn't some non-zero trailing residual value that the bean counters figured out. Or it could have simply been a bail out to help a buddy mash the company through the legal sausage making process of receivership. Who knows. I've tried to not contribute residuals when I play music on Tidal so hopefully my meager contributions didn't factor into the math. Like many of us I was thoroughly annoyed by the whole fiasco.

hfvienna's picture
Submitted by hfvienna on September 19, 2023 - 6:39am

Nonsense does not get right if repeated many times. Look into details of MQA solution to find out why the "lossy" claim is pure nonsense.

Kursun's picture
Submitted by Kursun on September 19, 2023 - 7:22am

@hfvienna

If you don't know MQA is a lossy format that's purely your problem.

Master Quality Authenticated (MQA) is a proprietary audio coding standard for lossy digital audio compression.

Not only that, it's a scheme to cash in from all sections of the audio industry.

They had claimed that it's better than the original! Ha Ha Ha!

Glotz's picture
Submitted by Glotz on September 19, 2023 - 12:09pm

The package containing data is lossy, not the data. It has no lossy data audio compression. The delivery package vs the data itself is what you fail to understand.

See below at JA's statement.

John Atkinson's picture
Submitted by John Atkinson on September 19, 2023 - 7:00am
JRT wrote:
Like MP3, MQA is lossy compression, while FLAC is lossless.

While it is true that the bits in an MQA-encoded file are not the same as those in the original hi-rez file, this does not necessarily mean that the format is "lossy" in the manner that MP3, AAC, etc are lossy.

MQA takes advantage of the fact that all recorded music has a random noisefloor that is higher in amplitude than the LSBs. This allows a technique called "steganography" to be used to create a buried data channel - see www.researchgate.net/publication/45949372_Steganography-The_Art_of_Hiding_Data .

John Atkinson
Technical Editor, Stereophile

DH's picture
Submitted by DH on September 19, 2023 - 11:36pm

Lossless: means the file can be uncompressed and produce a bit perfect copy of the original file. That's the definition.

MQA cannot do that.
Ergo: lossy

MQA was forced to stop claiming it's product was lossy and then use the claim of "perceptually lossless".

That claim is quite debateable.

And neither side of the argument claims that MQA files sound identical to the source file.

So let's stop using the false MQA newspeak version of the term lossless.

mieswall's picture
Submitted by mieswall on September 23, 2023 - 5:16am

Yes, it is lossy in the noise ground present in every recording (to replace it with data) and the blank space that music dynamic content is physically unable to fill in upper octaves (to replace it with the slope of extended filters).
Then it fixes the time domain problems of the music content present in every other pcm format (except dxd masters, from which MQA is an exact, not bit by bit, but *sonic* copy) .
You tell me which of both is more important…

DH's picture
Submitted by DH on September 23, 2023 - 9:21am

What's important is not to fall into the trap of believing the falsehoods promoted by MQA. It doesn't fix time domain problems: MQA claims this, but they've never actually presented evidence of what it does. Others have demonstrated how the filters MQA uses actually increase smearing of transients, adding in massive amounts of anti aliasing artifacts.
What's good it that it's a failed format. Tidal is dropping it, and soon there will be no legally available source of any significant amount of material in MQA.
Don't think that Lenbrook will save it: the market doesn't want it, and it's proved itself NOT to be a money maker.

John Atkinson's picture
Submitted by John Atkinson on September 23, 2023 - 11:36am
DH wrote:
Others have demonstrated how the filters MQA uses . . . adding in massive amounts of anti aliasing artifacts.

This is only the case if the spectrum of the signal has top-octave content at or close to 0dBFS in level. The MQA codec operates on the assumption that the spectrum of the music signal is similar to that of pink noise, ie, one where the energy decreases as the frequency increases. This is almost always the case with music, in which case, there aren't "massive amounts of anti aliasing artifacts."

The test that produced the result to which you are referring used signals that didn't conform to the spectrum of real-life music, so they "broke" the encoder.

John Atkinson
Technical Editor, Stereophile

cgh's picture
Submitted by cgh on September 20, 2023 - 8:35am

Reminds me of something I did a long time ago hiding data in images using DFT.

adamdea's picture
Submitted by adamdea on September 21, 2023 - 3:49am

The problem with characterising MQA at "lossless" is that it involves dismissing the information it loses and the spuria it introduces as irrelevant. Ultimately this is based on what boils down to a form of psychacoustic model of what counts. This is ultimately no different from a perceptual codec. The rest is sophistry.

miguelito's picture
Submitted by miguelito on September 19, 2023 - 9:30am

I don't necessarily have a problem with lossy compression. Every part of the chain in our systems compresses and mangles the signal. "Perceptually lossless" is a term Bob Stuart used and I don't necessarily have a problem with it. But why do it? If you had a bandwidth problem to fix, just consider FLAC at 96KHz/18bit or something along those lines.

But MQA is many things. Amongst a bunch it is a "condiment" added to music (always sounded to me like a sprinkle of autotune was added to the music). I like the effect but after a while it seems it is the same across the board.

The part where it became damaging in my opinion was when redbook 16/44.1 files get replaced by 16/44.1 MQA files. In this case there is no unfolding possible, and some bandwidth is required for MQA authentication, so you end up with something that is LESS than redbook in resolution, albeit with some umame sprinkled on... Good? Not good in my opinion.

I will add that White Glove MQA releases are generally great. In particular, I purchased two of Aretha's albums (Lady Soul and I Never Loved..) because I think these are the best digital transfers I have heard (and I own most if not all of them). Is this something to do with MQA per-se, or is it careful remastering, or is it the beautification MQA puts into the music? I don't know.

Jason Victor Serinus's picture
Submitted by Jason Victor Serinus on September 19, 2023 - 9:28am

I've compared some of Warner's 16/44.1 files to the 16/44.1 MQA files and have consistently felt that latter sound superior. I've also compared Peter McGrath's original 16/44.1 files to their MQA remakes. No contest.

miguelito's picture
Submitted by miguelito on September 19, 2023 - 11:44am

Herb has mentioned to me the McGrath files, but they are un-obtainium so I cannot say. I presume what you hear is coming from the beautification it introduces to the sound. Sound engineers have said the sound after MQA processing is different than the original sound - and that is with a digital master delivered to the MQA team. So there's clearly some very precise DSP applied, which I will say I often like, same as I like MORE soy sauce on my sushi.

Also, I used the term "damaging" and I meant a less-than-redbook file is something I dislike even if the condiment makes it sound better. It is NOT the original sound. And in most cases it is not validated by anyone as the vast majority of MQA files are done without validation.

Indydan's picture
Submitted by Indydan on September 20, 2023 - 3:15am

You also claim to hear differences with ethernet switches. That makes anything you say hard to believe...

hb72's picture
Submitted by hb72 on September 20, 2023 - 9:03am

not difficult to do that with hifi audio. And there are probably good reasons for that, although they work indistinguishable wrt to e.g. netflix data stream.

miguelito's picture
Submitted by miguelito on September 21, 2023 - 5:03am

I think anyone criticizing a review that has "unexpected" conclusions should listen for themselves. When fancy USB cables came about it was not understood why they made a difference, but they do. I don't understand why an Ethernet switch would make a difference, but I am willing to try it and listen for myself. There is always a physical/technical reason for differences even though they might be hard to measure or we might not know what to measure.

cognoscente's picture
Submitted by cognoscente on September 19, 2023 - 7:10am

Why lose anything at all with lossless if you don't have to lose anything with AIFF, the maximum quality through uncompromised music files? Another disadvantage of streaming, after all, you get lower quality, data costs are the largest cost item for the providers of streamed music, they have a vested interest in keeping data costs as low as possible and therefore do not provide you with the best quality, in addition to the disadvantage that owning (music) gives you more freedom and less dependence. Less (music at your disposal) is more (quality).

prerich45's picture
Submitted by prerich45 on September 21, 2023 - 7:05am

I'd rather use RIFF aka .wav files myself.

supamark's picture
Submitted by supamark on September 19, 2023 - 8:39am

but it seems to me that they're finally going the direction (try to replace Bluetooth for wireless audio connectivity) they should have gone years ago. Regardless of your opinion about MQA, ya gotta admit it sounds better than Bluetooth audio (and can do hi-rez).

If they could have implemented a version for satellite radio (which sounds awful) that would have been cool but it's way too late for that.

Bluetooth's real advantage is that it was made an open standard in the late 1990's and is also a general data exchange protocol, but there are still fees associated with implementing it.

Mark Phillips,
Contributor, Soundstage! Network

Anton's picture
Submitted by Anton on September 19, 2023 - 9:20am

I am agnostic regarding MQA's proper place in the pantheon of audiophile playback. In general, the fewer arbiters telling us what artists and listeners want is likely best.

miguelito's picture
Submitted by miguelito on September 20, 2023 - 4:52am

How are you comparing these two? There are no devices that use SCL6 to date.

supamark's picture
Submitted by supamark on September 20, 2023 - 12:17pm

Whether SCL6 products are on shelves or not, do you honestly think that there won't be? That they're, like, lying about developing them? Just threw away a bunch of money because reasons Elon Musk style?

And if you don't understand why I would compare two ways to wirelessly transfer audio to playback devices I don't think I can help you there - that's all on you. I honestly hope something comes along to replace Bluetooth for wireless audio soon, and I don't care if it's the remnants of MQA or some currently unknown entity. I want progress and improvement, higher fidelity.

Mark Phillips,
Contributor, Soundstage! Network.

miguelito's picture
Submitted by miguelito on September 20, 2023 - 12:29pm

There might be in the future although the Bluetooth Consortium has already rejected SCL6 - which in all likelihood triggered the investor pulling off funding and resulting in MQA filing for bankruptcy.

But the key point is SCL6 is not MQA per-se so you (or anyone else) don't know how it sounds compared to Bluetooth.

DH's picture
Submitted by DH on September 23, 2023 - 9:26am

Isn't MQA. It's something else.
Seems as if the new owners of MQA - Lenbrook - are going to use SCL6 in their products. I don't see any future for MQA. It's proven itself to be a commercial failure. It doesn't make money, and Tidal hasn't gotten any serious amount of customers from it. They are dropping it.
Especially now that a competitor owns it, other hifi manufacturers won't want to have it in their products, as it requires revealing tech secrets to MQA.

miguelito's picture
Submitted by miguelito on September 25, 2023 - 5:09am

MQA as a distribution format will not survive as there are no streaming services committed to it. And as you say, no manufacturer will further develop MQA given one of the competitors owns the patents now.

I wonder how much information is included in the purchase (I would expect every single piece of info is included) which could give Lenbrook a lot of information about how various manufacturer's DACs work. In all likelihood this is irrelavant information for someone like Lenbrook as they will never develop their own DAC chips/circuitry.

jimtavegia's picture
Submitted by jimtavegia on September 19, 2023 - 9:29am

I know that in terms of file sharing, the size of the file was always the issue. Once you get above redbook, 24/48 24/96, and 24/192 are large files and took up great space. Now that HD space is cheap this is no longer the issue.

I am surprised that over the years since 24/96 files could be burned on DVDs as DVD-V, that 24/96 did not become a high-rez standard has they could be manufactured and sold as DVD's and are playable on most DVD and Blu-ray players like I have done to my own files for the last 15+ years. I know that the industry did not want those high rez files out on the lose for folks to rip and steal, which is the main reason I would think for an MQA format as you must buy a decoder to extract THAT data.

Is this not the same reason that DSD did not go over well as you needed a dedicated player and many were over $1k early on while the usual CD player was half that and seemed adequate for most music lovers.

Sadly now you can buy a DAC that with USB can play the DSD files from your computer, but you can't use that same DAC to upgrade the SACD sound in that player. SACD died due to all the hardware problems, mixing and mastering. Plus, the UHF noise was always an issue.

I have long loved recording in 2496 and am now using 2448 to keep file sizes down and the sound is as good as I need it to be. I think most of us are over the format war. I do wish NAD well.

Kursun's picture
Submitted by Kursun on September 19, 2023 - 9:11am

I believe flacs use the same amount of file size for silence and music, making it a wasteful format in terms of hdd space. Late Ken Ishiwata had mentioned this problem and pointed out DSD being a better choice.

I have a DSD image music archive and am very pleased with its audio quality.

UHF noise? Shouldn't it be more of a problem for the bats? :)

miguelito's picture
Submitted by miguelito on September 23, 2023 - 10:03am

FLAC compression does compress silence effectively. This is actually why it cannot compress the 8 LSBs in MQA and results in a file size equivalent to 'true' 24/44.1 or 24/48 because those 8 bits seem random.

MFK's picture
Submitted by MFK on September 19, 2023 - 9:09am

Sigh. When MQA entered receivership I hoped that would be the end of it. In my opinion it is a solution for a problem that no longer exists. To my ears it degrades the sound. A few years ago here in Vancouver, the hi fi stores were really pushing it. I listened to a number of demos (same tune with and without MQA) and was never convinced. It's a proprietary format and in the end it's about $. Qobuz rules!

cognoscente's picture
Submitted by cognoscente on September 19, 2023 - 9:21am

Qobuz rules only if you buy the music and download it uncompressed in the highest quality (however, beware of incorrect remasters), Qobuz rules not if you stream music, see my previous contribution.

koblongata's picture
Submitted by koblongata on September 19, 2023 - 9:45am

Don't know, but PCM is a pretty shitty technology that's for sure

jimtavegia's picture
Submitted by jimtavegia on September 19, 2023 - 10:03am

I have found that even my own 2496 recordings are better than most R2R music I have heard in my 76 years on this planet. You may like the sound of tape, but to say it is better than high rez pcm is a real stretch.

I don't find 24/192 to be anymore beneficial than 2496 and I have strained to hear it in my younger years. I would think that anyone who owns a Halo May, Benchmark, or any thing in the A or B categories of DACs would think differently. I would think anyone's speakers or the room would be a weaker link than 24/96 used as the carrier of a well engineered recording.

Archimago's picture
Submitted by Archimago on September 20, 2023 - 7:11am

Barring the fact that basically all music has gone through PCM because you can't really do much editing in DSD, MQA is PCM. MQA just applies some lossy encoding and decompression when played back with weird filters that allow ultrasonic distortions to seep through to make it look like it has higher frequency content. (They shamefully bamboozled engineer Bob Ludwig with these tricks if you look up his testimonial video on YouTube for MQA.)

Not sure what the argument is. MQA is NOT some special "post-Nyquist" system. It just ignores proper bandwidth filtering. It's all based on PCM but hyped.

DSD is wasteful of data storage and tends to have too much high-frequency noise unless you start at least at DSD128 and apply some good low-pass filtering.

koblongata's picture
Submitted by koblongata on September 20, 2023 - 7:43am

I think MQA can be seen as Bluetooth, packeted, but wired and lossless
It's pretty easy to get great playback from Bluetooth, but with PCM, oh man, you have to have really clean power, clean circuit, clean pathway to maintain its timing information from the source to DAC, to reveal the true quality of PCM, not easy.

John Atkinson's picture
Submitted by John Atkinson on September 20, 2023 - 8:18am
Archimago wrote:
MQA just applies some lossy encoding and decompression when played back with weird filters that allow ultrasonic distortions to seep through to make it look like it has higher frequency content.

The use of slow rolloff reconstruction filters was pioneered by Pioneer and Wadia in the late 1980s and slow rolloff anti-aliasing filters were introduced by dCS in the 1990s. The late Charley Hansen, who was a fierce critic of MQA, also supported the use of slow-rolloff filters. As long as the music doesn't have high energy in the top octave, which is almost always the case, there are no ultrasonic distortions.

What you get with such filters is optimized time-domain behavior with both the A/D converters and D/A converters. I examined this subject, supported by measurements, in an article I wrote for Stereophile in 2018: www.stereophile.com/content/zen-art-ad-conversion.

John Atkinson
Technical Editor, Stereophile

CG's picture
Submitted by CG on September 20, 2023 - 9:01am

So, then, just how much of MQA's appeal to some listeners (some being a percentage that I can't identify, so I won't) is because of the filters MQA DACs are required to use and how much is due to the process itself? Does anybody know?

cgh's picture
Submitted by cgh on September 20, 2023 - 9:28am

[In deep DJ voice] Welcome to to the smooooth sounds of MQA. B-Spline and recline, relax while you let us interpolate your favorites.

miguelito's picture
Submitted by miguelito on September 21, 2023 - 5:57am

There are two steps in MQA decoding: unfolding (this is DAC independent) and rendering (this is upsampling with a specific filter choice - so depends on the DAC).

Example: The Dragonfly implementation of rendering only touched the controller firmware - not the ESS DAC firmware. What the update did is add code to determine what upsampling filter (of a panoply of 32) is to be used by the ESS DAC as indicated by the unfolded MQA PCM stream coming from the source.

CG's picture
Submitted by CG on September 20, 2023 - 5:38pm

Well, that's my exact question.

It's hard to imagine that the folding process actually improves the sound quality. No matter how clever you are, the best you can hope for when getting rid of data is that it has no perceptible bad effect on the reconstructed waveform. Even breaking even is a heroic act. The folding was apparently aimed to allow for smaller files and a reduction in the associated transmission bandwidth.

So, is the sound quality that some people enjoy from MQA processed recordings because of the filters?

Or, is it because a lot of those MQA processed recordings have also been remastered to 2020's tastes?

Both?

When I've looked at the impulse responses of DACs using MQA filters that have been published right here and elsewhere, they look very similar to what you see for the same kind of upsampling filters from Ayre and before that from Wadia. That's not the entire story, of course, but it's something to consider. I think, anyway.

This is all pretty much academic now, I guess.

miguelito's picture
Submitted by miguelito on September 21, 2023 - 5:15am

improvement. Say you have a 24bit/48KHz MQA file. The first 16 Most Significant Bits (MSB) are "true to source". The remaining 8 Least Significant Bits (LSB) are an encoding of the sample rates from 48KHz to 96KHz. If you don't unfold these 8 LSBs, as you play the file the result is they average out to zero, effectively playing a 16bit/48KHz file (in reality, these 8 LSBs introduce some high frequency hash noise, but lets ignore that). If you "unfold" the data, you get information to recover the musical information (albeit at less than 16 bit resolution) for the sampling rate region 48KHz-96KHz (ie the music region 24KHz-48KHz). So long story short, unfolding gives you MOST of the information in the MQA file. Rendering as the last step is supposed to improve things further, but it IS upsampling so whether rendering is a real improvement or an extra tailoring of sound is up for debate.

miguelito's picture
Submitted by miguelito on September 21, 2023 - 6:03am

Take the 24bit/48KHz MQA data stream. Strip the 16 MSBs out, leaving you with an 8bit/48KHz PCM stream. Fourier transform this into frequency space - you will get a spectrum that describes a signal with frequencies up to 24KHz (Nyquist). Now take that realtime FFT spectrum and do an inverse Fourier transform back to the time domain, BUT when you do that, DOUBLE the frequency of the harmonic to convert back. So for example, if the spectrum says the amplitude of the harmonic at f=15KHz is 10, take that 10 and when you reconstruct the time domain signal, apply it to a harmonic at 30KHz. You now have a signal with 8 bit resolution that gives you music in the region 24KHz-48KHz. This sounds very complicated but I am sure it can be coded up pretty neatly.

I will add this is all speculation, I don't know how MQA actually does unfolding.

CG's picture
Submitted by CG on September 21, 2023 - 5:57am

I should've been clearer.

I fail to see how cleverly hiding bits in a smaller file can be better than using the larger original file with nothing removed.

It certainly can reduce the file size, which reduces the transmission bandwidth. The hidden/folded bits can also be used for some sort of DRM, if that's what you want.

You could even argue that those LSBs don't matter. But, then, why record them in the first place?

Answers? I'm not proud and am very willing to be shown where I am wrong on this.

miguelito's picture
Submitted by miguelito on September 21, 2023 - 6:09am

You will undoubtedly loose information. MQA works on the assumptions that:
1- Cutting those 24 bits to 16 bits is perceptually lossless - ie you cannot hear better than 16 bits
2- High frequency information in the 24KHz-48KHz region does not need as much resolution as the 0-24KHz region

I don't necessarily disagree with these two assumptions.

There is of course more to MQA than just this... As mentioned elsewhere there's the DSP applied to the music which DOES change the sound - often in ways I find (at least initially) pleasing.

CG's picture
Submitted by CG on September 22, 2023 - 6:32am

So, basically, it's much like a remaster by another name.

That could be good. Or, not.

Many of the re-masters I've heard have not been as realistic to listen to, for me, as the earlier versions. Based on glowing reviews, I've bought more than one remaster of albums I never heard first time around and found the music itself to be great, but the sound not so good. So, I found earlier versions that ended up sounding much better. YMMV, of course.

Ansel Adams was famous for sitting and waiting for the light to be just right - for him - at his photo shoots. Then, he'd apply his own darkroom secrets to the photo. That ended up being his interpretation of what he saw. All great.

I don't really think that after passing through a couple Photoshop sessions by somebody else a half century or more later any remastered photographs are better renditions of what the artist had in mind. Same for remasters of music recordings.

(To me, one of the real virtues of vinyl is that often the versions you buy at the used record shops have had fewer, ahh, "improvements" made by somebody further down the recording chain.)

But, then again, if somebody likes the sound of these newly updated tracks, who am I to dispute that? Same for people who love the sound coming from audio systems that give me a headache. It's a hobby for most of us and supposed to be enjoyable and fun.

adamdea's picture
Submitted by adamdea on September 21, 2023 - 10:10am

Time-optimised is just more meaningless jargon (not least because the rough equivalent it is also applied for the opposite method by Rob Watts). You cannot be right in the time domain and wrong in the frequency domain.
And if there was anything in the leaky antialiasing and reconstruction filters then the whole thing could still happily be contained in a 16/96 non-proprietary flac wrapper.

miguelito's picture
Submitted by miguelito on September 21, 2023 - 10:56am

I never really understood what the "deblurring" and "time optimized" jargon was all about. I have assumed the point is to use a minimum phase filter in the upsampling of the signal - phase is the frequency-domain equivalent to time-shifts in the time-domain.

cgh's picture
Submitted by cgh on September 21, 2023 - 11:26am

I heard (and used) the term blurring in physics long before I heard it being used in (EE signals applications) to audio. In its most generic form it meant that time information is smeared out in 1/t space (just like the frequencies are "smeared" out in t space) and any monkeying or optimizing you do in one space gets smeared or blurred when you FT^-1 the signal back. I assume the terms descend from this usage.

CG's picture
Submitted by CG on September 22, 2023 - 6:34am

Succinct! Bravo.

hollowman's picture
Submitted by hollowman on September 25, 2023 - 3:31am
Quote:

JA wrote:

"The use of slow rolloff reconstruction filters was pioneered by Pioneer and Wadia in the late 1980s "

Yes, but I believe Wadia may have gotten that idea first in their "Spline" algorithm.
http://marlene-d.blogspot.com/2012/01/pioneers-legatolink-wadia-spline.html
Wadia used DSP/FPGA and a few years later, Pioneer, incorporated Legato into their PD2026/28/29 chips-- roughly mid 1990s. It is very difficult to obtain any datasheets on the Pioneer chips, so I suspect they are hiding a rip-off of Wadia's original design in the late 80s.

miguelito's picture
Submitted by miguelito on September 21, 2023 - 5:24am

In my opinion...

MQA was NOT purchased to launch a new streaming service, or to get license payments from the rare leftover streams from TIDAL, or anything of the sort.

The reason is the patents held:

-- SCL6 which could possibly, eventually, become a hi-res Bluetooth standard (although it has already been rejected by the BT association, but with some repackaging maybe would work)

-- Streaming protocol from NAD amps to wireless powered PSB speakers

-- Some other patents in the 100+ patents MQA-Ltd held

CG's picture
Submitted by CG on September 20, 2023 - 9:10am

In the high tech corporate world patents are like another form of currency. Often, technologies are patented and never used by the patent holder. Instead, the associated claims are dragged out when Company A wants to sue Company B for patent infringement. Then, some arrangement in terms of licensing is worked out.

For a classic example, go back a decade or so ago when Google purchased Motorola Mobility. Yeah, Google thought they might like to be in cellphone business. But, mostly they wanted the IP because they were getting hammered by some big competitors for patent infringement. Within two years, they had divested themselves of the companies actually doing the engineering and manufacturing but they kept the IP. Problem solved.

Plus, for smaller companies, a large patent portfolio looks attractive to investors or potential buyers.

miguelito's picture
Submitted by miguelito on September 20, 2023 - 12:32pm

that buys patents and makes money enforcing them. That is all they do. I forget the name. A notable case was the Blackberry patents a few years back.

CG's picture
Submitted by CG on September 20, 2023 - 5:42pm

There's various patent troll businesses all over the world. Most of them know exactly how much it costs for a company to protect their IP and they offer to settle for less than that amount. It then becomes a business decision for the company being sued.

It's essentially like a protection racket.

I spent way too much of my working years trying to help the attorneys craft proper technical responses to these kinds of lawsuits.

Glotz's picture
Submitted by Glotz on September 19, 2023 - 12:03pm

And MQA largely solves that. MQA is not traditionally lossy, it simply changes the enormous package the data arrives in. A big package filled with nothing equals wasted space. Change the package and make the delivery package more efficient, not less data or lost data.

JA is once again right, and everyone else still wants to crap on MQA, largely because the haters are mis-informationists in every way. If everyone loves JA (me included), why isn't there more trust for what he says??

To fully understand MQA, read Robert Harley's book and the chapter on MQA. There are excellent articles here as well, and in TAS.

MQA assumed too early that Tidal users wanted to pay for their top tier of playback. It was fatal.

DH's picture
Submitted by DH on September 19, 2023 - 11:28pm

MQA files are larger than equivalent flac files in some instances.
Plus, you can encode a hi-res flac file as dithered 18/96 and it will be no bigger/smaller than an MQA file and be less lossy.

Archimago's picture
Submitted by Archimago on September 21, 2023 - 2:59pm

"If everyone loves JA (me included), why isn't there more trust for what he says??"

What is this? Some kind of cult that we "love" JA and thus accept his words without pointing out inconsistencies or issues?!

That doesn't make him divine or know everything. When it comes to MQA, he has at least been way too forgiving of its limitations and the lack of necessity for this codec especially among "perfectionist" audiophiles who want the best. MQA is not the best in an era when we can stream 24/192 losslessly.

I appreciate what JA has done over the years in his leadership over Stereophile and find his measurements particularly useful. His unfettered appreciation for MQA however has been a particularly sad sore point from the beginning. Nobody's perfect.

Glotz's picture
Submitted by Glotz on September 21, 2023 - 3:25pm

Or is it you're butt-hurt that JA proved you wrong in just a few posts above? After your claim of HF garbage by MQA...

"As long as the music doesn't have high energy in the top octave, which is almost always the case, there are no ultrasonic distortions.

What you get with such filters is optimized time-domain behavior with both the A/D converters and D/A converters. I examined this subject, supported by measurements, in an article I wrote for Stereophile in 2018.."

Hmmm. You as well are far less than perfect. But please keep bullshitting yourself and everyone else.

He may not be divine, but you hardly show up for church...

Trusted audiophiles and reviewers that have heard MQA full rendered feel it's as good as it gets and many of those run audio magazines like Robert Harley and John Atkinson. You, on the other hand, post a lot. I haven't read your magazine yet.

Yeah, I'm sure you at least believe you're right about everything. For me, I'll stick with the professional opinions.

Glotz's picture
Submitted by Glotz on September 21, 2023 - 4:03pm

as a cultist.

Archi-MAGA?

Archimago's picture
Submitted by Archimago on September 22, 2023 - 12:16pm

I'm Canadian, dude - like Lenbrook! I don't think the acronym applies.

On second thought, if you meant "Make Audiophilia Great Again" instead of the usual snake oil over the years we've been treated to. Then sure, I can get behind that!

Glotz's picture
Submitted by Glotz on September 22, 2023 - 12:00pm

That explains a lot, eh?

Jk

Archimago's picture
Submitted by Archimago on September 22, 2023 - 11:55am

JA didn't say anything that proved my claims wrong.

What you get with these filters are phase shifts and poor suppression of imaging distortions. Whether a person hears these or care is up to them. But these are not ideal digital filter designs and many high-end reputable DACs do not use these - go ask Chord or Benchmark if they would ever use the parameters MQA uses and think they're in any way optimized for best time-domain performance.

"He may not be divine, but you hardly show up for church..."

How would you know? Maybe I just go to a different kind of church? ;-)

Glotz's picture
Submitted by Glotz on September 22, 2023 - 11:58am

Nt

Indydan's picture
Submitted by Indydan on September 20, 2023 - 3:17am

I hope Lenbrook goes bankrupt!
Lenbrook is now the Mike Lindell of the audio world...

Archimago's picture
Submitted by Archimago on September 20, 2023 - 7:25am

I think Lenbrook made a major advertising error by announcing their MQA acquisition like this, knowing the negative brand recognition.

I think what they wanted was SCL6/"MQair" because they were already developing headphones using the codec.

IMO they should have just acquired MQA Ltd.'s IP's, announced it without all this fanfare nonsense about end-to-end this or that which MQA never really was, or remind everyone of the name Bob Stuart and Lyndberg's pronouncements. Especially bad to have it splashed on the pages of places like Stereophile; get it away from audiophiles! And over time incorporate the SCL6 codec into their products and hope it competes and sells well next year. Sort of like Dolby grabbed the MLP codec and rolled it into TrueHD; most consumers/audiophiles would not have known about the role Meridian played and that's for the better given that it was used in the failed DVD-A.

MQA codec which has no role in modern streaming or "MQA-CD" (horror!) is dead. It should never have been brought up to tarnish Lenbrook's products.

funambulistic's picture
Submitted by funambulistic on September 25, 2023 - 8:40am

So, you are advocating for a large, respected corporation, including all of their brands (Bluesound, NAD, PSB) go bankrupt, thereby unemploying hundreds, if not thousands of employees all because you do not like a certain codec?

Why don't you show us on the doll where the oh-so-terrible MQA hurt you. Might as well point to where Lindell hurt you while we're at it...

David Harper's picture
Submitted by David Harper on September 22, 2023 - 10:28am

The amount of stupidity and nonsense that this subject results in on audiophile forums is testament to how silly and irrelevant audiophile concernes are. Nobody with a life cares about any of this.

bhkat's picture
Submitted by bhkat on September 23, 2023 - 9:29am

Are proponents really saying that adding MQA processing to lossless files improves them somehow?

Anton's picture
Submitted by Anton on September 23, 2023 - 11:14am

I guess it's time to take my MQA decoder ring down into the cellar where I keep my Betamax, Elcassette, cassette, VHS, CD, Mini Disc, DAT, Laser Disc, and 8-track players.

The only way I can see saving MQA is to price MQA recordings above 600 bucks and sell them in the reel to reel aisle.

hollowman's picture
Submitted by hollowman on September 25, 2023 - 3:15am

Were there ever any official (or non official) studies conducted -- preferably, double-blind, placebo-controlled -- in which subjects were fed MQA/non-MQA audio over, say, headphones, using ABX methods, or even longer duration sessions (whole songs, tracks) .... and these subjects' reactions were officially recorded/documented and presented in a credible Journal paper?

cognoscente's picture
Submitted by cognoscente on September 25, 2023 - 4:08am

For everyone in Europe who has ARTE TV (app), watch the documentary about how music listening has changed over the past 30 years, yes with streaming you have more music at your disposal than ever, and yes also music from other regions (Asia and Africa) and what they telling you over and over again, but still, the algorithm ensures that you only hear what they want to present to you (unless you're looking specifically). The already famoes musicians ones become more famous (and they are only ones that earns), the unknown ones remain unknown (and they earn nothing) just as like all algoritne on all social media platforms works. Okay, off topic, I know, but this is in addition to the sound quality issues of streaming music and all this discussion about formats. I actually know nothing about all these techniques and that is why I trust my common sense that files that are done as little as possible with (no packing and unpacking, as less calculations / conversions as possible) sound the best, that is why I trust that uncompressed purchased AIFF files (in my Apple environment and purely as storage) sound the best. Heavy, large files, yes, for sure. And I need the Onkyo music app on my iPhone and a cable to get the HiRes versions to my dac. All true. Back to topic: Lossless literally means "loss less", so losing something, less but still something, otherwise it would not be called lossless but lossnothing. And I don't want to lose anything. That's why I trust on AIFF.

hollowman's picture
Submitted by hollowman on September 25, 2023 - 4:38am

What is the "digital flywheel"? It was mentioned in a Stereophile review back in 1995. As were: "Jitter Jail", "Time Lens", etc. I'll let you guys look for it.
Point being ... many technologies and paradigms have come and gone (and have come back -- NOS, multibit, R2R, etc). You can check out my DIY thread in the Stereophile forum where I mention some major-manuf projects (from Sony, JVC, Kenwood, Philips) that were based on large R&D budgets and scientific know-how from those corps. And got very good reviews. And continue to attract attention and praise on forums like Audiokarma.
So, given all the wealth of past NOVEL and non-trivial research investments, it's unclear why MQA continues to attract so much attention given the many equally credible efforts -- past and present -- for improving digital sonics. Did we simply get bored by those past effort?
Stereophile might be largely "guilty" in setting such a high precedence for MQA, in the 2018 post about "Post-Shannon" sampling:
https://www.stereophile.com/content/mqa-aliasing-b-splines-centers-gravity
That sets a very high bar, Mr. Austin.
But it does keep the Comments section active ;)

cognoscente's picture
Submitted by cognoscente on September 25, 2023 - 6:03am

I believe if the recording and type of file is good (maximum), then it is good, and therefore leave it as it is. All that packing and unpacking is not for quality, but for quantity and to reduce costs. Or making money. That's why my motto, less (music at your disposal) is more (quality). And yes I have a NOS R2R dac. But as I said, I know nothing about technology and I am only a music lover (who also listens to the sound, an audiophile), and who relies on his common sense. Leave what is already good and do nothing with it. You can only ruin what is already good.

X