Jump to content
IGNORED

MQA at CES


Recommended Posts

Since they don't allow producers to encode MQA files on their systems, I can imagine similar reasons for being very tight on decoder side too.

 

A fairly common business model is to charge plenty for the content creation tools but provide reproduction for little or nothing. See for instance Adobe PDF. This creates a large demand with which to motivate producers to pony up for the creation tools. MQA could do something along those lines. They could also not do that.

Link to comment
I compared the MQA tracks without a MQA decoder and that might not have revealed the true benefits of MQA. Conducted the listening comparisons solely because it was said MQA's improvements could still be heard via non MQA DACs. [...] I have listened to more MQA tracks from 2L and they lose out to DXD.

 

I did a quick FFT analysis of a couple of the MQA samples vs their non-MQA counterparts.

 

First 2L-064, piano and solo voice:

2L-064-MQA.png

The obvious difference in the undecoded MQA is a small hump at 19-22kHz. Moreover, around 7kHz the MQA spectrum starts falling off ever so slightly compared to the DXD file. It's probably not audible, but it's a difference nonetheless.

 

Next 2L-120, an old 44/16 recording of Nielsen piano music that allegedly sounds better as MQA even without a decoder:

2L-120-MQA.png

As evidenced by the graph, the original recording contains nothing of value about 16kHz, only sigma-delta modulator noise. The MQA encoding has filtered this out and replaced it with ... something. Here the MQA version has lower noise level well into the (somewhat) audible band, so it's no surprise if it sounds better. However, it would probably sound better still if it was simply filtered with a cutoff at 16kHz. Also of interest is that the difference at lower frequencies seen in the first sample is pretty much absent here.

 

From this I would say the claims that MQA preserves full CD quality even without a decoder are clearly bunk. What an MQA decoder might do is anyone's guess at this point.

Link to comment
I tried again just now.

When I clicked this link with the registered MacBook Pro I was directed to the test bench without MQA.

When I used the iPad to visit computer audiophile and click this same link above I was directed to the test bench with MQA tracks!

 

Have you tried clearing your browser cache?

Link to comment
Fees? Tidal has said they will provide MQA at no extra cost to users who in any case have a subscription for lossless streaming. As I'm not required to buy into MQA on the HW side to continue streaming - it costs me nothing unless I decide I want an MQA HW device.

 

It is quite possible Tidal have been given a good deal or even been paid to implement MQA in order to drum up demand for DAC support. In MQA's position that probably wouldn't be a bad move.

Link to comment
I did a quick FFT analysis of a couple of the MQA samples vs their non-MQA counterparts.

 

And here's 2L-111, the same one Miska looked at in his blog post. The DXD for this one is quite clean, the modulator noise seems to have been filtered out (why bother with DXD then?). No idea what those spikes at 115-120kHz are.

2L-111-MQA-full.png

 

Zooming in on the low region:

2L-111-MQA.png

Again a slight drop in all frequencies starting around 7kHz and a hump of noise in the top few kHz.

 

To claim at the same time that the barely-existent high-frequency content in the original is audible and that these alterations to the low frequencies are not strikes me as contradictory.

Link to comment
You could also run the same source content through conversion to 44.1/16 using SoX and then compare that against the MQA one.

 

Average power spectrum of 2L-111 downsampled to 44.1 (using sox "rate -u 44100" and default dither):

2L-111-MQA-CD.png

Nothing remarkable there.

 

Since this is average (?) spectrum of the entire thing, it doesn't give very clear picture of high frequency content, because the HF content is typically on transients (since that is sort of definition of a transient) and thus has very low contribution to the average. So there are two good ways to get a picture of that. One is to use spectrogram and another if you prefer to simple plot, is to use "peak hold" mode that plots the highest peak level of every spectrum bin throughout the recording.

 

Yes, those plots used average spectrum. Taking the maximum reveals brief peaks better but doesn't show the difference with MQA as well since both have peaks exceeding the constant level of the MQA noise.

Link to comment
AFAIK, they don't support DSD content. Any rate higher than 88.2/96 kHz is decimated to such rate for encoding.

 

DSD could obviously also be downsampled to 96k and MQA-encoded. Many ADCs start out with 1-bit sampling after all.

 

It really technically needs a change in the DAC to support this sampling, similar to how you need to have DAC to specifically support DSD.

Just that for example the Mytek DAC that currently supports MQA uses bog standard ESS Sabre DAC chip... And the decoding output produces PCM...

 

I'm just waiting for someone to capture the input to that DAC chip, i.e. the output of the MQA decoding. I'd do it myself if I had one.

Link to comment

Here's my analysis of 2L-048, originally recorded in 96/24.

 

Spectrogram of the 96k master, left channel only:

2L-048-96k-spgram.png

Nice and clean, no visible modulator noise and music content nearing 40kHz in places.

 

Spectrogram of the MQA file, left channel:

2L-048-MQA-spgram.png

Here we clearly see the tell-tale noise band at the top and also slight haze over the entire image.

 

Average power spectrum of the entire track:

2L-048-average.png

Looks familiar.

 

Average power spectrum of a 4 seconds near silence at end of track:

2L-048-silence.png

Here we see the elevated noise floor of MQA from 5kHz upward. At 15kHz the noise level starts rising sharply and peaks at 21kHz.

 

Finally, average power spectrum over a 50ms high-frequency burst starting at 224.6s:

2L-048-hf.png

The MQA noise is visible above 20kHz but naturally not as obviously as when there is no signal at those frequencies.

 

Undecoded MQA is probably psychoacoustically transparent for music like commercial rock that's pretty loud all the time. Anything with a decent dynamic range is likely to suffer.

Link to comment
You guys should read the AES paper below from a bit over a year ago. Only 16 pages, and 3.5 pages are references. Pictures and graphs so really not that much reading. It describes the why, how and other particulars of how this works.

 

Some of these wonderings will just disappear. Like the sample rate issue.

 

http://www.aes.org/tmpFiles/elib/20160114/17501.pdf

 

Still doesn't answer how Meridian plans to handle this upon their botched rollout. But at least some of the areas of contention will be clarified with some real facts about MQA.

 

That file doesn't exist.

Link to comment
I'm not sold on it because, even though it may be the greatest thing since sliced bread, it is an endpoint to endpoint solution that doesn't leave anything open for other standards.

 

If all music is encoded on MQA and one needs MQA certified hardware to decode it, then it's pretty much a locked system that will not let anybody else play in their ground.

 

My thoughts exactly.

 

I have heard that if you don't want MQA you will simply receive a FLAC stream. However, that's not really the truth. I don't get to pick between 2 different streams (for instance I can pick the resolution in Netflix). With Meridian how do we even know that they are not messing with the FLAC stream to make it sound bad just so folks are forced to upgrade to a Meridian certified DAC and get locked into their system.

 

A system/service like this opens a can of worms that I would rather do without.

 

PS: It's all about freedom, we should get to choose. While we don't get to choose how music is recorded and/or mastered, we do however get to choose how we listen to it. In an all MQA encoded music world, that choice is going to go away.

 

My worry is that some portion of music will be distributed exclusively as MQA, leaving those without MQA-enabled hardware listening to the undecoded version, which contrary to the claims by Meridian does not match up to 44.1/16, let alone 96/24 or higher resolutions.

 

A better way to package the MQA data in a format playable by existing equipment would be as a new FLAC subframe type that older decoders would simply skip over. There are other ways additional data could be hidden in a FLAC container as well. Of course that wouldn't give them same iron grip on the deployment they so clearly desire.

Link to comment
I do not see any evidence in my readings, for example, that MQA "does not match up to 44/16" in sonics.

 

Undecoded MQA is clearly inferior to 44/16 as demonstrated by Miska and myself.

 

We have a prototype for MQA in the old HDCD, which Stuart did not develop but which he acknowledges had some considerable influence on the MQA approach. I think it is very hard to prove that HDCD, while it was not a huge success in the marketplace, actually degraded the sound of a non-HDCD-processed but HDCD mastered CD in any noticeable way. I never heard any engineer or listener claims that it actually did, though it would have been difficult for a typical listener to make a meaningful comparison between a recording in its HDCD and non-HDCD mastered versions. Both versions were typically not available to consumers.

 

Unprocessed HDCD, i.e. played back on a regular player, sounds louder and overly compressed compared to decoded playback. How big the difference is depends on the specific track.

 

But, there were no insistent complaints that HDCD mastered recordings sounded bad when not played back with HDCD encoding disabled.

 

HDCD didn't really amount to more than a curiosity and most people probably never encountered it. Here's an easy to follow explanation of what HDCD does and how it damages playback on a regular player: The HDCD Enigma

Link to comment
Not sure I am following. What interest does a manufacturer like Wilson Audio have in whether a market is open or closed by proprietary/DRM format? It is complex, but in the long run they might benefit just a little, and being the type of guys who would sell their firstborn for a SQ tweak they are not going to have the "perspective" regular old music consumers like you and I should be looking to when we think about MQA.

 

Producers and engineers (and labels, artists, etc.) would just LOVE to "authenticate" us all, as they are piping mad at the changes digital has brought to their world. They BLAME the very consumers of their art for their fortunes (who all admit have not been good lately) because these very consumers have access to digital technology. They would like nothing more than to shackle us in the way that DVD/Blueray does for video (even though this is not a full solution - it helps however in their perspective).

 

These repeated assurances that MQA is primarily about SQ strike me as naive at best, and do not take into account what is really happening in the industry - something you point out (around CD sales) just a few posts back...

 

If there's one thing labels like, it's making us repurchase our music in a new format. With high-resolution PCM in standard containers, they lost this ability since DACs are likely to keep supporting PCM input for the foreseeable future, and few people would see any worthwhile benefit in upgrading a 192/24 (or even 96/24) copy to something even higher. If MQA becomes the new de facto standard for music distribution, they'll have that leverage back. A few years down the line, they might introduce MQA 2.0 and drop support for the original ("legacy") MQA (they can refuse to license it for new gear). I'm not saying this is what's being planned, but I can certainly see some people seeing such an outcome as desirable.

Link to comment
OK, I'll calm down and make one more attempt by focusing the conversation. Talking about the industry IMHO is too amorphous. I'm focused on the individuals lining up in support of MQA and their motivations. Do you really think Vandersteen and McGrath are enthusiastic about the technology simply because it will sell speakers? You've suggested not only but maybe somewhat (??) If that's not their primary motivation (and I question whether that even figures) rather than sound quality then ascribing moneygrubbing motivation to "the industry" while avoiding discussion of the individuals is rather both unfair and unconvincing.

 

Different people may have different motivations for supporting MQA. Some may also have been misled into it.

Link to comment
Right now 2L has more content available and more of a market with its DXD format than Meridian does with MQA. Many DACs will not accept DXD (352.8KHz) input. Shall we be very, very afraid? In some speculative future, after all, it has the ability, because it is a format, to cause the industry to change over to a format your DAC won't accept.

 

DXD can be converted to a lower rate using freely available software. The details of MQA are secret, so you must use an approved playback system. Well, someone is bound to reverse engineer it eventually, but it's the thought that counts.

Link to comment
Where did you get the idea that the "details" of MQA are secret? It's been the subject of white papers, and patent, with its legally required disclosures.

 

Those resources provide only a high-level, and at times contradictory, picture of the format. To decode it into PCM, you need to know the actual formatting of the bits in the file, and that information is nowhere to be found. If a software decoder becomes available, it is possible to disassemble it and painstakingly recreate a format spec. A firmware file for a DAC might also be used in this way. This is how formats like ALAC have become known to the public.

Link to comment
Or you can filter out the noise and work to make the rest of the file sound like something, but as I mentioned above, why would you, when you can produce a better sounding result from a non-MQA-encoded file?

 

Suppose in the future you will only have the MQA-encoded file.

Link to comment
Very very good points you raise here in my opinion. In particular, what will happen to the people that like to perform digital room correction?

 

From your points: I can see that "If done after the decoding this would exclude decoding in an MQA-certified DAC", but why "If done before the MQA decoding it would destroy the MQA format"?

 

I would think that if digital eq is done before MQA decoding, then this eq would be applied only to the 44.1/16 part of the undecoded container, leaving untouched the folded higher frequency content. Anyway, my speculation

 

Filtering the 44/16 stream would destroy the bit patterns required by the MQA decoder. Those low bits look like noise, but they actually follow a strict format. If they didn't there's no way the MQA decoder could know when to perform its magic.

 

They'll probably tell you to buy an MQA-enabled DAC with integrated digital room correction. Or to buy a better room.

Link to comment
I am not an attorney either, but I know there is no small amount of truth in what you say. One thing I hate about this whole issue is what you said in your last sentence. You did not mean it this way, but the implication is I "want" to do something illegal with "content" simply because I can. There is a sort of "presumption of guilt" when a person who actually $paid$ for something wants to say, back up that file - or play it on multiple devices, or ones that do not have DRM built in.

 

Perhaps someone can explain exactly what the differences in "ownership" is between a book that I own, a downloaded high-res music file and/or PCM encoded CD, and say a shovel. When I buy a shovel, or discuss a shovel and what I can do with it, folks are not inclined to say "your ownership of said shovel does not legally entitle you to do anything you want with it" (e.g. walk over to your neighbor and bash his skull in ;) ).

 

All this circles back around to the question as to why Meridian is encrypting data within their products, when the content of that data is open format and thus (as far as I am aware) not subject to the normal DRM practices...

 

MQA isn't an open format. Music buyers rejected closed formats and DRM a long time ago, and although there's no evidence of piracy being a problem (many audiophile releases are nowhere to be found on the torrent sites), there are no doubt people grumbling over it somewhere. MQA could be an attempt to entice the consumers with promises of better quality (which remains to be seen) and the produces with promises of content protection (which will be broken) into a realm where they are both under fully controlled by Meridian/MQA. Perfect profit forever.

Link to comment
Meridian digital speakers readily accept unencrypted open format PCM, that's not their issue.

 

The problem has always arisen if they wished to provide high resolution digital channels from proprietary sources/standards of which there have been many over the years: for example DVD-A (which they were involved in), SACD (which they were not involved in), and more recently HDMI.

 

That's true. However, the majority of quality music labels seem quite happy to sell unencrypted high-resolution PCM today. If they switch to MQA only, it will be a serious step back. In fact, the tight DRM controls probably played no small part in the failure of SACD and DVD-A to gain much popularity.

Link to comment
Thanks for the explanation, makes sense. Reminds me of yet another reason to stay away from HDMI.

 

HDMI for pure audio is a bit weird, but as soon as there is a picture involved it becomes incredibly convenient. Any HDMI sink (AV receiver, display) will accept unencrypted data. It is up to the source to enable HDCP if it wants to. The organisation behind AACP (the encryption used on Bluray) won't give you the keys unless you promise to use HDCP whenever you play an encrypted disc. Even if you're not intending to play Blurays, there is no reason to avoid HDMI entirely. I don't know of any other cheap interface that supports 8 channels of high-resolution audio.

Link to comment
  • 2 weeks later...
If you belong to the school that says that noise is OK, but that ringing is bad, then Miksa has pointed out that you can get equally good results by encoding fewer bits at a proportionally higher bit rate. It appears to work better (for these subjective tradeoffs between noise and ringing) because it is used with a lossless CODEC like FLAC which works well under these circumstances. Lossless compression is necessarily variable bandwidth compression, due to information theory. This raises another problem with MQA, which is that it is a fixed rate CODEC. Already with MP3 it is known that one can get the same quality at a lower bandwidth by using a variable bandwidth CODEC (VBR vs. CBR).

 

MQA could (but probably doesn't) use a variable bitrate with the added constraint that all the input affecting a given output sample be available when the corresponding PCM sample arrives. Even classic MP3 allows this kind of variable rate through its "bit reservoir" feature. That said, a true variable bitrate (possibly with an upper bound) is obviously better.

 

It is so 1970's and 1980's to be thinking about CBR in this Internet age. VBR compression works much better with packet switched communications networks where bandwidth is shared dynamically between users, unlike traditional circuit switched (telephone networks) which dedicate bandwidth to a stream.

 

That's not limited to packet switched networks. Digital TV has always (more or less) used variable bitrate to fit more channels into a shared fixed-rate carrier. This is known as statistical multiplexing (statmux).

Link to comment
Proprietary formats are as strong as ever with HEVC/H.265

 

The HEVC specification is freely available to anyone, as are open source encoders and decoders.

 

and pretty much all Dolby and DTS stuffs.

 

Some of the Dolby and DTS formats are available as ATSC or ETSI specs although you need a licence to use the trademarked names.

Link to comment
It seems that when those "besserwisser's" are afraid of confronting Bob directly.

 

Not sure who you're referring to, but I have better things to do with my time than ask questions that will be side-stepped with a non-answer anyway. Once the software/firmware gets out, people can start looking at the actual output or even reverse engineer the algorithms and we'll have real answers, not handwaving and technobabble.

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...