Tidal High Res (MQA) news...

[Axiom05]

The way I read Meridian is that software or server implementation would bring benefit of the time correction from the AD encoding even if not fully optimized for the receiving DAC.

 

I think this is what they are saying, provided that the original master had been either recorded with the MQA "filter" or "scrubbed" after the fact. MQA can be used solely as a compression scheme to allow streaming of a Hi Res file in a 16/48 package. As others have already mentioned, MQA appears to have three or four parts that can be used in any combination:

 

1) MQA encoding using special filters performed on master file either at recording session or afterward

2) Compression of the Hi Res file into a 16/48 package

3) Decoding of compressed package to give a Hi Res file, however I believe this file is lossy with frequencies above 55KHz discarded.

4) DAC specific decoding of compressed file to remove "colorations" due to DAC.

 

So a MQA file can be streamed but it may or may not contain the benefits of item 1, depends how the file was prepared. Playback of an MQA file without decoding gives you a 16/48 file, not particularly Hi Res, probably cheaper to buy the CD. I think the key point is item 4 which, I would imagine, must be incorporated into the DAC. An Aurender streamer may be able to decode the compressed file but it will not likely be able to clean the effects from your DAC. So, to get the full benefit from MQA you will need all four above to be in place. Maybe it will be a simple firmware update for the DAC's to handle this. So far I have not seen any DAC-specific makers announcing MQA updates to existing equipment. Will Mytek being updating the DAC 192 to handle MQA? If not, why?

 

All of this seems very complicated (at least to me). Too complex to succeed? What about people using free open source players such as mpd, Clementine, Foobar? I doubt that the developers will pay for a MQA license. Maybe Meridian could make available a plugin for these programs that users could purchase? Complicated...

 

Sorry if all of this is obvious at this point but I am just trying to get it all straight in my head. Feel free to correct anything that I have gotten wrong.

[jeffmudrick]

My understanding also. I think MQA holds great promise. As a Tidal subscriber, personally I wouldn't spend a lot on a new DAC that wasn't MQA certified given the state of things. Fortunately I'm quite content with my Audio-GD REFERENCE 5 and can hang tight to see who besides Meridian and Mytek steps up this year. If/when Tidal leverages MQA to offer 24/96+ streaming my interest in local playback will approach zero.

[firedog]

To me, interesting interview, elaborates on some of the topics in this thread:

 

Meridian

 

Meridian Co-Founder and MQA Co-Architect J. Robert Stuart Talks with Robert Harley

 

more info-press releases:

 

MQA Explained in Short Videos | The Absolute Sound

MQA Launches Content at CES and Teams with HTC for World's First Smartphone MQA Demonstration | The Absolute Sound

[manisandher]

To me, interesting interview, elaborates on some of the topics in this thread:

 

Meridian

 

Thanks for the link. I found this paragraph very interesting:

 

"At various times in the history of recorded music, the thing that was distributed was the same as was in the archive, and at various other times it was different. But we believe that the very, very best thing should be put in the archive even if you can’t distribute it. And MQA is about that; but it’s also about delivering the sound of the studio and of the existing archive, as authentically as possible, to the listeners today."

 

It seems to me that with the bandwidth we have today, there is absolutely no technical reason for not distributing the digital archive file. So no need for MQA in this case.

 

BUT... he also says:

 

"What we’re trying to do, conceptually, is directly connect together the modulators at both ends—the high-speed delta-sigma modulators in the A-to-D and the DAC."

 

I've personally never heard a DAC with a delta-sigma modulator that I've liked the sound of. If the MQA processes can make SDM DACs sound more like ladder DACs, then there may be something really interesting here after all.

 

Why am I constantly reminded of HDCD when I read about MQA?

 

Mani.

[firedog]

Mani-

 

What he is saying is that even if you have the so called archive/master file, you aren't actually hearing the music the way it was recorded in the studio (say the mike feed). In his view, MQA accounts for the particular distortions introduced by the A to D converter used in the studio and those particular to the DAC on the playback end, allowing something virtually identical to what was heard in the studio, certainly more accurate than what you presently hear with a "master" or "archive" file with no MQA. Not saying he's correct, but that's what he's claiming.

 

As far as sigma delta - you listen to SD modulators with virtually every recording you play back, as virtually all music that is distributed or processed digitally runs through at least one - if not several - before it gets to you. Even if you have a ladder DAC.

[mansr]

As far as sigma delta - you listen to SD modulators with virtually every recording you play back, as virtually all music that is distributed or processed digitally runs through at least one - if not several - before it gets to you. Even if you have a ladder DAC.

 

There's also the fact that sigma-delta DACs are demonstrably more accurate than ladder DACs at typical audio bit depths.

[manisandher]

As far as sigma delta - you listen to SD modulators with virtually every recording you play back, as virtually all music that is distributed or processed digitally runs through at least one - if not several - before it gets to you. Even if you have a ladder DAC.

 

Yep, fully aware of that. That's why I spent a fortune buying a Pacific Microsonics Model Two with which to make all my PCM needle drops.

 

But you're right, the vast majority of the music I listen to has been run through an ADC with SDMs. Not sure why I don't like the sound of DACs with SDMs - I've tried so many, but always come back to DACs with Ultra-Analog 20400A or BB PCM63/1704 chips. If MQA can make SDM DACs sound like these ladders, I'll be happy.

 

Mani.

[manisandher]

There's also the fact that sigma-delta DACs are demonstrably more accurate than ladder DACs at typical audio bit depths.

 

Yeah, maybe. Depends what you're measuring. In any event, I still don't like the sound of them.

 

Mani.

[miguelito]

I'm a bit confused as to the format of MQA files and where decoding takes place.

 

Couple of questions:

 

1- MQA files are supposed to play back at (approx) CD quality on any dac without decoding. What format is it played in? Is it a FLAC 24/48 file?

 

2- When I play a TIDAL MQA file, what is coming out of the TIDAL app, or Roon for that matter? The undecoded MQA data stream or a decoded higher bitrate/bitdepth stream - ie do I need an MQA decoding DAC?

 

3- When something like an Auralic Aries supports MQA, is a decoded stream coming out or an MQA encoded file - ie do I need an MQA decoding DAC?

 

If a new DAC is required (which would make no sense to me since MQA should be decodable in the streaming app) then I would give MQA a failed.

[miguelito]

Mani-

 

What he is saying is that even if you have the so called archive/master file, you aren't actually hearing the music the way it was recorded in the studio (say the mike feed). In his view, MQA accounts for the particular distortions introduced by the A to D converter used in the studio and those particular to the DAC on the playback end, allowing something virtually identical to what was heard in the studio, certainly more accurate than what you presently hear with a "master" or "archive" file with no MQA. Not saying he's correct, but that's what he's claiming.

 

As far as sigma delta - you listen to SD modulators with virtually every recording you play back, as virtually all music that is distributed or processed digitally runs through at least one - if not several - before it gets to you. Even if you have a ladder DAC.

In other words, MQA is modifying the file - not necessarily for the better.

 

Frankly I am a bit shocked by the claims of "inaccuracies on the recording chain". If you're recording with a high quality chain, one that never downsamples below 24/192 or DSD etc, then I would very much like to see proof of these supposed inaccuracies.

 

I attended an MQA demo in NYC. One of the demo files was a 24/192 recording followed by an MQA-processed version of the same file. The MQA version sounded markedly better. I am all for euphonic filtering and such, but I have trouble again and again with dubious claims from Meridian.

 

I wish they would just be honest.

[james45974]

from reading through the various threads on MQA in my opinion Meridian has done a poor job communicating critical details about the technology. Some of this seems to be Meridian being cagey, for instance not doing proper comparison testing, and possibly trying to build up some anticipation/interest by slowly leaking info. But as of now I feel that their is some railroading being done by Meridian, for what is so far a questionable technology. You can prove me wrong on that point Meridian but you have your work cut out, I don't like the feeling that I am somehow being toyed with and manipulated! Maybe Microsoft will buy it and basically kill it off like they did with HDCD!

[jeffmudrick]

I'm bemused by the cynicism on display. Sure Meridian could do a better job but as I really don't think, for example, Morten Lindberg enthusiasm about MQA stems from any expectation that he's going rake in millions on downloads of Norwegian folk music or is being taken in by mumbo jumbo euphonic filtering of stuff he has personally recorded. The assumption that either or both are true doesn't wash with me. Whether it's worth a $10 premium per download may be an issue but if it comes with my Tidal subscription I won't be complaining if it sounds better.

[miguelito]

Found some answers...

1- MQA files are supposed to play back at (approx) CD quality on any dac without decoding. What format is it played in? Is it a FLAC 24/48 file?

The MQA encoded file can be encapsulated in any format - so you can have it in ALAC, FLAC, AIFF, WAV.

 

2- When I play a TIDAL MQA file, what is coming out of the TIDAL app, or Roon for that matter? The undecoded MQA data stream or a decoded higher bitrate/bitdepth stream - ie do I need an MQA decoding DAC?

 

3- When something like an Auralic Aries supports MQA, is a decoded stream coming out or an MQA encoded file - ie do I need an MQA decoding DAC?

Meridian's site mentions hardware and software decoders - so clearly it is possible TIDAL will do MQA decoding prior to the output.

 

But will it? How about Roon?

[mansr]

I'm a bit confused as to the format of MQA files and where decoding takes place.

 

Couple of questions:

 

1- MQA files are supposed to play back at (approx) CD quality on any dac without decoding. What format is it played in? Is it a FLAC 24/48 file?

 

Everything remotely authoritative I've read (including Meridian patents) suggests MQA files look like 24/48 PCM but with the low 8 bits containing the compressed high frequency bands and whatever else it provides. A non-MQA DAC will play this as if the low bits were noise. A 16-bit DAC can simply be sent the top 16 bits.

 

2- When I play a TIDAL MQA file, what is coming out of the TIDAL app, or Roon for that matter? The undecoded MQA data stream or a decoded higher bitrate/bitdepth stream - ie do I need an MQA decoding DAC?

 

3- When something like an Auralic Aries supports MQA, is a decoded stream coming out or an MQA encoded file - ie do I need an MQA decoding DAC?

 

If a new DAC is required (which would make no sense to me since MQA should be decodable in the streaming app) then I would give MQA a failed.

 

In principle, the decoding can happen anywhere before the actual conversion to analogue. I can see the point in both MQA-aware DACs (for use with "dumb" software/streamers) and software MQA decoders (for use with "dumb" DACs).

[crenca]

I attended an MQA demo in NYC. One of the demo files was a 24/192 recording followed by an MQA-processed version of the same file. The MQA version sounded markedly better. I am all for euphonic filtering and such, but I have trouble again and again with dubious claims from Meridian.

 

I wish they would just be honest.

 

This is the first I have read of a demo of the same recording (in any resolution) side by side with the same recording MQA encoded. What was the recording? What was the basic setup of the playback chain? What was "better" about the MQA encoded file and what DAC was used for the non-encoded playback and what DAC was used for the MQA encoded playback?

 

Sorry for all the questions...

[sdolezalek]

This is the first I have read of a demo of the same recording (in any resolution) side by side with the same recording MQA encoded. What was the recording? What was the basic setup of the playback chain? What was "better" about the MQA encoded file and what DAC was used for the non-encoded playback and what DAC was used for the MQA encoded playback?

 

Sorry for all the questions...

 

At the demo I attended, they were using a Diana Krall album (sorry can't remember which one) that had been MQA encoded. I actually had a hard time hearing significant benefits over the non-MQA encoded version of the same album. But they also played newly recorded material where MQA had been integrated from the start and I thought that sounded stunningly good.

 

In this case, they were using an all-Meridan system including the latest powered speakers incorporating MQA decoding. It didn't make me like their speakers any more (I don't particularly go for their sound), but the nature of the improvement was very much the same as I've seen over the last couple of years in implementing upsampling and better filtering through A+ and HQP. By that I mean that "lifting of veils" sensation, better cleaner transients, better instrument separation, ... so I felt as though it wasn't some euphonic filtering as much as it was making the D/A process work better and perhaps (when directly MQA encoded from the start) making very exact matches between the filters that best recreate the original mike feed. We have previously speculated that it might be as if you tried to separately optimize your iZotope or Jussi's filter settings for each recording depending on the A/D devices used.

[james45974]

I'm bemused by the cynicism on display. Sure Meridian could do a better job but as I really don't think, for example, Morten Lindberg enthusiasm about MQA stems from any expectation that he's going rake in millions on downloads of Norwegian folk music or is being taken in by mumbo jumbo euphonic filtering of stuff he has personally recorded. The assumption that either or both are true doesn't wash with me. Whether it's worth a $10 premium per download may be an issue but if it comes with my Tidal subscription I won't be complaining if it sounds better.

 

I put the comments more down to skepticism rather than cynicism (although there is some of that too), as the saying goes in Missouri "Show Me!"

[barrows]

I read your earlier posts and I believe you are doing a good job describing it (though I think most technical descriptions break MQA into a 3 part beast). I was responding to the description of MQA being an "upsampling" technique or something similiar.

 

By the way, when you say that:

 

"one is the compression aspect which places a high res file into a smaller package by folding the high frequency information into the (supposedly unused) lower bits, this first aspect would not give any improvement in sonic performance, it is just a new way to compress a high res file."

 

This is being used in the marketing speak of the various manufactures/promoters and their lackey's in the audiophile press as "an improvement in sonic performance" simply because it allows high res to be delivered in spaces it has not before (such as Tidal) and as everyone knows, "high res" is in-of-itself an improvement in "sonic performance". True, the point is debatable depending on many things, but all else being equal most would say that "high-res" is a "sonic improvement" - certainly if they are trying to sell you something.

 

I am obviously with you on the "dumb down" aspects. I have manyquestions, like will there be any "digital artifacts" or other sonic effects when I play a Tidal MQA stream through my non MQA DAC? As many who are more evenhanded about this have pointed out, this has as of yet to be demonstrated as all the "listening demo's" have so far been with MQA enabled playback chain. Then there are the true believers in the audiophile press (such as Audio Stream's Michael Lavorgna) who will quite literally insult you if you even dare ask relevant questions, because the creators of MQA are "geniuses" and the unwashed masses should simply trust them. It is all getting quite ridiculous and I suspect that Meridian are setting themselves up for a fall...

 

OK, we are pretty much not he same page then. Personally, I can play any hi res I want and have no need of compression in order to do so, hence I do not consider the compression aspect as a sound quality benefit.

My understanding of how the sound quality improvement of MQA works (the supposed improved time domain performance) is that it would be incorporated both in the ADC at the time of digitization, and in the DAC at the time of of conversion to analog. To my understanding of digital audio conversion, this would require that it be applied during the oversampling process by the configuration of the digital filters. For example, things like Meridian's Apodising filter removes time domain artifacts caused by most ADCs, and my understanding of what they are doing with MQA is an extension of that, developed for each DAC design.

For use, by what Meridian has stated so far, the exact technical nature of what the are doing is unclear, but it has been noted that MQA is not "bit perfect" by definition, and is a psycho acoustically based set of choices (as is the design of any digital filter, especially for lower sample rates).

 

Anyway, my main point was to show that it is very unlikely for any playback software program, whether in a server, or running on a streamer, etc, to implement the full benefit of the MQA encoding process, at least I feel we have confirmed that is true. If people are just getting all excited to be able to stream 24/192 from subscription services like Tidal then, OK, I see how that might be a benefit for some. Personally I avoid subscription services, as I do not like large corporations controlling what music I have access to.

[barrows]

At the demo I attended, they were using a Diana Krall album (sorry can't remember which one) that had been MQA encoded. I actually had a hard time hearing significant benefits over the non-MQA encoded version of the same album. But they also played newly recorded material where MQA had been integrated from the start and I thought that sounded stunningly good.

 

In this case, they were using an all-Meridan system including the latest powered speakers incorporating MQA decoding. It didn't make me like their speakers any more (I don't particularly go for their sound), but the nature of the improvement was very much the same as I've seen over the last couple of years in implementing upsampling and better filtering through A+ and HQP. By that I mean that "lifting of veils" sensation, better cleaner transients, better instrument separation, ... so I felt as though it wasn't some euphonic filtering as much as it was making the D/A process work better and perhaps (when directly MQA encoded from the start) making very exact matches between the filters that best recreate the original mike feed. We have previously speculated that it might be as if you tried to separately optimize your iZotope or Jussi's filter settings for each recording depending on the A/D devices used.

 

Right on, thanks for sharing your experience. Your observations make sense to me, so then we need to ask ourselves as audio consumers, would we rather pay for licensing fees of MQA, which will only fully benefit us when we purchase (or stream) new music which is MQA encoded, or would we rather choose a DAC (or computer based oversampling program like HQPlayer, etc) which has advanced digital film;tearing capabilities bringing us the same type of sonic advantages?

 

Licensed and proprietary stuff always rubs me the wrong way, but I do accept that if things were originally converted to digital with a much better ADC and more advanced filtering (or at DXD rates with no filtering at all) it would be a sonic advantage.

[miguelito]

Well, I have a few questions. Does this mean that Tidal is abandoning it's current "hi res" or "hi fi" offering of 16/44 Flac (with some DMR protection of course) service?

TIDAL doesn't use DRM per-se because you're not downloading anything, you're just streaming. My expectation is that the "hi fi" subscription will include MQA - and additionally that most of that content will be processed through the MQA beautification engine but otherwise will be CD-quality.

 

Do they REALLY expect all of us to get into yet another DAC/hardware situation when all of us have spent significant time and $ on our current rigs?

From what I have read, my interpretation is that the TIDAL app will decode MQA prior to sending it to the DAC, thereby sending a higher res than CD to the DAC.

 

Is MQA truly loss less, or is it yet another lossy format, and can anyone really even answer that question with any real conviction at this point? For those of us who don't jump on the MQA bandwagon, does this mean we are shifted to Tidal's "normal" lossy level service?

From what I have read, MQA has two components to it:

 

1- An encoding process of beautification by which the signal is changed - purportedly to remove ADC and other artifacts.

 

2- A lossy compression mechanism - purportedly not losing human audible information

 

These are Meridian's claims - not mine. I did listen to an MQA demo in NYC and it sounded great. Especially a 24/192 recording and the beautified version (ie run through MQA) was remarkable - the MQA version really sounded a lot better. So clearly not only not lossless but intentionally modifying the file.

[mansr]

2- A lossy compression mechanism - purportedly not losing human audible information

 

That's what they said about MP3 too.

 

These are Meridian's claims - not mine. I did listen to an MQA demo in NYC and it sounded great. Especially a 24/192 recording and the beautified version (ie run through MQA) was remarkable - the MQA version really sounded a lot better. So clearly not only not lossless but intentionally modifying the file.

 

Are you sure it wasn't just 0.5dB louder?

[sdolezalek]

Right on, thanks for sharing your experience. Your observations make sense to me, so then we need to ask ourselves as audio consumers, would we rather pay for licensing fees of MQA, which will only fully benefit us when we purchase (or stream) new music which is MQA encoded, or would we rather choose a DAC (or computer based oversampling program like HQPlayer, etc) which has advanced digital film;tearing capabilities bringing us the same type of sonic advantages?

 

Licensed and proprietary stuff always rubs me the wrong way, but I do accept that if things were originally converted to digital with a much better ADC and more advanced filtering (or at DXD rates with no filtering at all) it would be a sonic advantage.

 

Barrows: I'm right with you! By definition these hardware/firmware solutions both represent trade-offs necessary for the broader consumer market and are time limited by the quality of the hardware available today. By moving as much of this as possible into standalone software we benefit: a) by being able to choose a more complex/expensive/requires better computing power/etc solution than what gets sold to the general public; b) we enable folks like Jussi to design new custom filters that can get better over time; and c) we can continue to implement the latest in software without having to buy upgraded hardware or firmware.

[miguelito]

That's what they said about MP3 too.

Trudat...

 

Are you sure it wasn't just 0.5dB louder?

The difference was night and day. I was shocked. And this beautification was done on the 24/192 itself - amazing. But my takeaways are:

 

1- This is a massive change - for the better in this case, but is that so in every case?

 

2- The claim that "ADC and recording chain errors" have/can be fixed is extremely dubious - a 24/192 file that has been recorded with a high quality chain and digitized with a high quality ADC does not need any fixing - no matter what BS Meridian wants to argue about.

 

I am all for better sound and not at all against a high quality lossy compression that improves sound and allows high res streaming. I am pissed off beyond belief by utter deception and bull frankly. When I asked the demoers at Meridian NY technical questions I got contradictory yiberish. That is unacceptable.

[Jud]

Trudat...

 

 

The difference was night and day. I was shocked. And this beautification was done on the 24/192 itself - amazing. But my takeaways are:

 

1- This is a massive change - for the better in this case, but is that so in every case?

 

2- The claim that "ADC and recording chain errors" have/can be fixed is extremely dubious.

 

Actually, it's pretty prosaic stuff, as Miska pointed out. Many filtering algorithms used to record CDs concentrate on eliminating aliasing, so they ring. The MQA algorithm includes an apodizing (Latin, "remove the foot") filter, which stops a good part of the ringing. So yep, the ringing is indisputably a recording chain error, and yep, an apodizing filter indisputably does fix it.

[mansr]

The difference was night and day. I was shocked. And this beautification was done on the 24/192 itself - amazing.

 

A lot of filters can make a substantial improvement. One type that comes to mind is room correction. Depending on personal preference, a gentle equaliser can also have a pleasing effect. That doesn't necessarily make the better-sounding version more correct.

 

But my takeaways are:

 

1- This is a massive change - for the better in this case, but is that so in every case?

 

2- The claim that "ADC and recording chain errors" have/can be fixed is extremely dubious - a 24/192 file that has been recorded with a high quality chain and digitized with a high quality ADC does not need any fixing - no matter what BS Meridian wants to argue about.

 

I am all for better sound and not at all against a high quality lossy compression that improves sound and allows high res streaming. I am pissed off beyond belief by utter deception and bull frankly. When I asked the demoers at Meridian NY technical questions I got contradictory yiberish. That is unacceptable.

 

Once both hardware and content become more readily available, someone will hopefully do some proper measurements on it. Until then, I'll be treating secretive demos as though they are rigged in some way. Not saying that they are, but they might as well be (and rigged demos are hardly uncommon in general).