Just heard some MQA today...........

John Willmott, I have been reading some of the MQA article, are from I have seen MQA uses psychoacoustic modelling to take advantage how generally here sound impulses and sound blurring so as to remove mostly uneccessary data in the higher frequency domain ... thetefore this is lossy.. as with all models people conform to varying degrees. Meridian I am sure argue that for the vast majority this loss of information is not detectable as it is mostly redundant ... It would appear this modelling with MQA is applied predominately to the upper frequencies... so perhaps variances in people conforming to the model will be less noticeable... that is of course my supposition. Where I am not sure how MQA would work is with spectral timing which is where some of the very latest thinking and research is with high definition sound... MQA appears predominately based on frequency and analogue to digital and digital to analogue reconstruction accuracy rather than timing from what I can see... If I have this right then this could be the lossy compromise that will limit MQA. However perhaps it's this improved reconstruction that is quite revealing and striking ? Theoretically I can see why that might be so and an improvement with MQA over current traditional digital to analogue reconstruction filtering.

However without talking to a very knowledgable individual on the matter it's hard for me to develop my understanding further at this time.

Simon

Simon:

Last year I attended T.H.E. Show in Newport Beach California and had an opportunity first hand to hear MQA.  After the demonstration I peripherally took part (I listened) in a discussion with Bob Stuart (co-founder of Meridian and one of the architects for MQA) about the technology.  He was very forthcoming on some aspects of MQA, specifically that it was lossless and that they folded the higher frequencies under the noise floor of the lower frequencies in order to make a compressed file.  The smaller files were specifically suited to streaming.  As this packing occurs within already existing lossless formats (FLAC etc) any MQA file can be played back on any piece of equipment and the folded under frequencies are just ignored.  It effectively is "just" a CD quality file.  However, if the playback equipment is MQA aware the unfolding takes place and a lossless file is then played.

However, Mr. Stuart was a lot less forthcoming on what actually happened and how the folding was done.  In fact he assumed quite a Cheshire cat look when pressed on this topic.  I assumed as he didn't know the occupations of his audience he was being professionally coy.

I heard MQA and I liked what I heard.  They did not A/B any of their recordings which left me puzzled as it did other members of the audience.  The music they played, Frank Sinatra and Roberta Flack if I remember correctly, was being streamed in via Tidal from somewhere in Scandinavia on the hotels' less than stellar wi-fi.

The literature I've read states that the upper frequencies, outside of the audible range, carry information regarding timing, phase and temporal blur .. all of which are important to try and recreate the original analog recording.  I imagined these data as akin to metadata informing the actual audible frequencies.  Again, how that is reapplied I do not know.

An article I read just the other day by Doug Schneider in Soundstage Hi Fi mentioned the patent applied for by Meridian back in 2013/2014 titled "Doubly Compatible lossless audio bandwidth extension" 

=====================

ABSTRACT
An encoder for digital audio signals at a higher sample rate creates a stream for consumer distribution at a lower sampling rate, with compatibility for standard PCM players without a decoder. In conjunction with a suitable decoder, two enhanced playback options are supported, the first option allowing full lossless reconstruction of a noise-shaped higher sampling rate signal, the second option allowing lossy bandwidth extension even if an intervening transmission chain has truncated the least-significant-bits of the encoder's output signal.

====================================================================================

Which seems to confirm that in an end-to-end MQA environment the files are indeed lossless but where the environment is not MQA aware the files are played at CD quality.

John W - thanks. I think much talk about MQA in the consumer press appears to be about 'hidef' in a smaller file type opposed to focussing reconstruction authentication which appears to therefore focus on the hidef lossy option. I am personally more interested on the reconstruction authentication - as to me that is a compromise - possibly quite a large compromise in current digital music encoding and replay chain..

Massimo Bertola posted:

What's MQA?
Thanks,
Max

Google meridian mqa I think I can not publish the link.

Regards. Erich

John Willmott posted:

 

ABSTRACT
An encoder for digital audio signals at a higher sample rate creates a stream for consumer distribution at a lower sampling rate, with compatibility for standard PCM players without a decoder. In conjunction with a suitable decoder, two enhanced playback options are supported, the first option allowing full lossless reconstruction of a noise-shaped higher sampling rate signal, the second option allowing lossy bandwidth extension even if an intervening transmission chain has truncated the least-significant-bits of the encoder's output signal.

====================================================================================

Which seems to confirm that in an end-to-end MQA environment the files are indeed lossless but where the environment is not MQA aware the files are played at CD quality.

 

John, my understanding of MQA is that, in simple terms, MQA preserves only a small part of the high frequency information in a 24bit file; it 'redistributes' this information as noise in a 16bit file, through dithering (but not truly random; this is where the reconstruction comes in). This explains why any device would be able to play back the file; without the reconstruction, the difference between an MQA file and, say, a CD, would be subtly different noise behind the same 'musical' bits. Probably inaudible.

Now, with hardware that has the key to reconstruct the noise, the noise in the 16bit file is turned back into a music signal. Whether the word length increases I'm not sure, but the effect should be the same. But the only part that is added back, is the small part that was converted into noise in the first place.

Meridian may be right in claiming that no actual music is lost, just random noise, but we're talking lossy compression all the same.

EJ

Just to add, It doesn't help that MQA themselves seem utterly incapable of explaining what MQA is or does - take a look at their website for laughs.

EJS posted:

Just to add, It doesn't help that MQA themselves seem utterly incapable of explaining what MQA is or does - take a look at their website for laughs.

Incapable or unwilling? I would understand why they might be unwilling, but not incapable, surely? 

As a complete dunderhead on streaming, this highlights my first post here on the Forum. I always assumed cd was superior to vinyl, seeming obvious because of the technological advancement.  I was gobsmacked when I learned this was not the case, and indeed technology had made things worse. How on earth can average joe buy with confidence when promises are made from men atop the mountain saying "trust me'" Why can't acoustical physicists and engineers use their science to verify the facts and blind tests be peer reviewed to ensure AND assist in everyone's ultimate goal, better sound. Maybe I'm just naive, but I know enough to know that scientists cannot get away with bullshit, they are caught out.  Proprietary issues understood but aside, a new technology SHOULD actually bring audio forward, not sideways. Again, I accept if I am just a dullard, however I am the dullard with the wallet.

I am sure MQA is not bullsit - in fact from what I can see there is some great engineering and signal processing mathematics behind it ... but perhaps once could  blame the marketeers for creating the hyperbole and stretched claims .. but then to be fair if they stated exactly what it did - would 99% of the population understand or care... I suspect not.. Things need to be whiter than white to stand out from the noise..  I don't think Merdian's marketing / PR department are not the only one who do this....

Simon:

I think this is a very astute statement:  "Things need to be whiter than white to stand out from the noise..  "

.. and whilst it's a rather interesting intellectual exercise for us to speculate what this technology does and how it does it, until such time as it "stands out from the noise" that's all it will be .. an interesting discussion topic.

However, companies are starting to adopt it .. Onkyo, Panasonic and others have products either in the market or ready for release .. these are admittedly in the mobile arena (phones etc.) where the small file size is better understood and appreciated.

It's a subject that fascinates me for a number of reasons .. one being the endless pursuit for a replay as close to the original source as possible, the second being an alternative to the endless upgrade of boxes and cables.  We, as consumers of the music, need to encourage all these avenues of research ..