MQA Users & Discussion Thread

[heman_]

The use of the word 'studio' in discussion surrounding MQA is extremely concerning. As is 'mastering'. The mastering engineers and studios are purely third parties and have absolutely no say when it comes to the completion of the track (for at leas 99% of music being released). Exemptions might be for things like movie soundtracks or on labels like 'Reference Recordings,' or when every aspect of the music is being produced in house, which is pretty much never. No engineer or studio could ever sign any track for release and only rarely could they confirm that the music product released is the product which they mastered. 

 

Anyway..

 

Gamechanger? Not for me. But I'm not desperately trying to escape typical 16bit flac. Also, most of the music I listen to is released on indie labels. I can't see MQA getting the sort of market penetration needed for it to have any real impact on my listening. 

[Guest rmpfyf]

Here is Bob Stuart 'theorising' why MQA sounds natural in his own words

 

 

Excellent. Again, Bob's taking the piss somewhat. 

 

You'd have to have particularly poor (and old) filtering in mastering processes to get that sort of pre-ringing, and extremely short sample times in the master (not usually correlated with music).

 

It'll sound 'better' in MQA... and so would a Redbook remaster using newer mastering tools.

 

There is no doubt anywhere that recreating something that's truer to the master involves higher sampling frequencies, more effective filtering etc. This allows your brain to judge what a cymbal sounds like, not a combination of psychoacoustic and Fourier principles - assuming you've the installation to have the reproduction work as effectively, or you're just effectively LP-filtering throughout (with, guess what - phase delay effects of your own). 

 

There is no doubt anywhere that certain music is more amenable to hires in general. I have a small library accordingly. 

 

The credibility of new formats isn't increased or promoted with obtuse interpretations of those sought to be replaced.

 

Won't stop me hearing it when it lands. Pack your critical hear, brain and BS detector

[New Sensations]

"There is no doubt anywhere?" X 2.

Yet MQA begs to differ.

If MQA were just another DSD-like format, I too might be skeptical until I heard either way from listening.

That said, MQA isn't just asking studios to start over.

It can allegedly correct the ringing of ADCs used in creating the world's existing music library.

Sure, a remaster might yield better results but which major label in their right mind would entertain that? Probably not many. Instead, all they have to do is run the existing digital file through the MQA and it will sound better than before, MQA DAC or not. That's the theory. We won't know for sure until we hear the darn thing. I'm reserving judgement till then.

Mqa also allows for hi-Res pcm files of any sample rate to be transmitted in 24/44.1 or 24/48. That alone is impressive is it not? It ultimately means we might see hires streaming via Tidal. (Again, cool, and *probably* the only long term future for hi-res audio).

Sent from my iPhone using Tapatalk

[Guest Eggcup The Daft]

I don't follow your logic here. Yes, everything you said can indeed happen, but there is good reason to think it probably won't. Recall one of the reason why SACD/DSD/hi-res audio can sound great is because the recording engineers took better care in the studio to produce a higher quality master from the recording. This is one key reason people put forward as to why high resolution can sound better than the redbook CD issue. The recod industry culture is fixed in this context.

 

Now if you apply the same logic to MQA it is clear the reason for quality mastering remains the same. Of course anyone could apply crazy compression like you say. But if this didn't happen for SACD/hi-res, then it probably won't happen to MQA as well. 

Compression is about commercial pressure. If MQA remains a niche product, this won't happen. If Tidal goes MQA and is becomes a major player, or if Spotify adopts MQA, the commercial pressure will be there for streaming. If MQA encoded releases replace CD or become a dominant format in commercial radio, the pressure will be there.

 

The nightmare scenario is that MQA sounds brilliant, is widely adopted, we lose the hi-res market, and only then do commercial pressures ruin it.

[New Sensations]

This picture he posts makes no sense  (when compared to knowledge on what audio does).

 

Whatever he is saying, he need to make a better picture .... otherwise nobody can be really convinced.

 

 

Why don't you ask the man himself here: http://www.computeraudiophile.com/f8-general-forum/your-chance-ask-bob-stuart-anything-about-mqa-27412/ ?

 

Also, is it not a little short-sighted to hinge MQA's worth on a theoretical conversation? That conversation might help us better understand how it works but being convinced one way or t'other must surely come down to listening....as we would surely do with speakers or an amp.

 

Why does noisefloor db in 16bit rise when above the 15000 frequency? (but original not) .... and why is '16 reference dithered' and CD frequencies going higher than 22050.

 

Even if there is the reason, they should know people will be confused as say something.

 

why?

 

Ask Bob using the thread above.

 

I agree, we need better explanations. 

 

Here is a description of how they applied MQA to 'white-glove' an old digital recording from 1993. http://www.2l.no/pages/album/120.html

 

 

Again, why not ask Bob directly in the CA thread link provided? Here it is again: http://www.computeraudiophile.com/f8-general-forum/your-chance-ask-bob-stuart-anything-about-mqa-27412/

Edited February 3, 2016 by J_o_h_n

[New Sensations]

Compression is about commercial pressure. If MQA remains a niche product, this won't happen. If Tidal goes MQA and is becomes a major player, or if Spotify adopts MQA, the commercial pressure will be there for streaming. If MQA encoded releases replace CD or become a dominant format in commercial radio, the pressure will be there.

 

The nightmare scenario is that MQA sounds brilliant, is widely adopted, we lose the hi-res market, and only then do commercial pressures ruin it.

 

I don't understand how the digital encoding mechanism changes the way masters are made hot or not? Assume for a moment there is no MQA: do those same mastering pressures not exist still? The commercial pressure remains no matter what Tidal or Spotify use, no?

[Guest Eggcup The Daft]

I don't understand how the digital encoding mechanism changes the way masters are made hot or not? Assume for a moment there is no MQA: do those same mastering pressures not exist still? The commercial pressure remains no matter what Tidal or Spotify use, no?

Yes.

 

I don't use Tidal or Spotify, but am all too aware that Spotify has plenty of obviously compressed tracks: so did iTunes when I checked in the past.  I'm sure there are Tidal users reading this thread who can check for compressed music there.

 

Of course MQA could work in our favour here, if uncompressed  or less-compressed MQA files replace the older compressed files. But will the streaming companies just mass-process what they have already?

[New Sensations]

Yes.

 

I don't use Tidal or Spotify, but am all too aware that Spotify has plenty of obviously compressed tracks: so did iTunes when I checked in the past.  I'm sure there are Tidal users reading this thread who can check for compressed music there.

 

Of course MQA could work in our favour here, if uncompressed  or less-compressed MQA files replace the older compressed files. But will the streaming companies just mass-process what they have already?

 

Hang on a sec. Just to clarify: we're talking about DYNAMIC RANGE compression here? So - If an album is dynamically compressed as heard on Spotify and unless Tidal source a different master, it will play back on Tidal with the same dynamic range compression, no?

[gcgreg]

Of course MQA could work in our favour here, if uncompressed or less-compressed MQA files replace the older compressed files. But will the streaming companies just mass-process what they have already?

I can't see Tidal choosing to instantly increase their bandwidth requirement by around fifty percent if they don't believe it delivers a marked improvement in some regard. It is too much overhead just for a marketing gain, especially when people both here and more importantly in the US still experience stream issues with HiFi.

Having said that, I suspect "hot" tracks will likely remain so to a large extent until the industry culture shifts away from the practice.

[New Sensations]

I can't see Tidal choosing to instantly increase their bandwidth requirement by around fifty percent if they don't believe it delivers a marked improvement in some regard. It is too much overhead just for a marketing gain, especially when people both here and more importantly in the US still experience stream issues with HiFi.

Having said that, I suspect "hot" tracks will likely remain so to a large extent until the industry culture shifts away from the practice.

 

Why would they need a 50% bandwidth increase?

[gcgreg]

Why would they need a 50% bandwidth increase?

That's a rough estimate I made based on what I've read so far. Currently the Tidal HiFi FLAC stream appears to be around 900Kbps, my understanding is that the MQA packaged FLAC would be about the size of Redbook audio (or slightly larger, around 1.5-ish Mbps.

If I have that wrong please correct me.

[Guest Eggcup The Daft]

Hang on a sec. Just to clarify: we're talking about DYNAMIC RANGE compression here? So - If an album is dynamically compressed as heard on Spotify and unless Tidal source a different master, it will play back on Tidal with the same dynamic range compression, no?

Yes, and yes.

[a.dent]

I'd urge people to check out http://archimago.blogspot.ca/2016/01/measurements-mqa-master-quality.html#more to get an early assessment of an MQA file.

[aechmea]

That's a rough estimate I made based on what I've read so far. Currently the Tidal HiFi FLAC stream appears to be around 900Kbps, my understanding is that the MQA packaged FLAC would be about the size of Redbook audio (or slightly larger, around 1.5-ish Mbps.

If I have that wrong please correct me.

 

No - not quite.

 

A CD in original form (16/44 PCM bundled as a WAV or as AIFF)  is 1.411 Mbps (= 16 x 44.1 x 2).  The same data when data file compression is used means that the file size is smaller.  The actual saving is dependent on the algorithm and settings used.  FLAC has several compression settings which will produce smaller files than the original WAV.  So a raw CD will be seen as 1.4M and a FLAC version of the same thing will be something like .9M depending how much they have squashed it.  The receiving end of course has to de-compress the FLAC, but at the end of the day the WAV and the FLAC are identical in PCM content, it's just that the FLAC is smaller to transmit.

[New Sensations]

That's a rough estimate I made based on what I've read so far.

Where did you learn this? Got a link?

Sent from my iPhone using Tapatalk

[gcgreg]

Where did you learn this? Got a link?

Sent from my iPhone using Tapatalk

One for starters: "MQA keeps all this timing information, and folds it into a file that's only a little bigger than a CD. "

http://www.mqa.co.uk/customer/how-it-works

[gcgreg]

No - not quite.

A CD in original form (16/44 PCM bundled as a WAV or as AIFF) is 1.411 Mbps (= 16 x 44.1 x 2). The same data when data file compression is used means that the file size is smaller. The actual saving is dependent on the algorithm and settings used. FLAC has several compression settings which will produce smaller files than the original WAV. So a raw CD will be seen as 1.4M and a FLAC version of the same thing will be something like .9M depending how much they have squashed it. The receiving end of course has to de-compress the FLAC, but at the end of the day the WAV and the FLAC are identical in PCM content, it's just that the FLAC is smaller to transmit.

Isn't that exactly what I said about FLAC compressed CD audio size? It comes to about 900kbps. MQA have indicated that their format will package the hi-res into something around the size of CD, not compressed CD.

I will come back with some more sources that have lead me to believe this, again I may be missing something here and would honestly be happy to learn that Tidal will be able to stream this new format with zero increased bandwidth.

[gcgreg]

Looking at the track length in minutes on the list below, many of the MQA file sizes are around 10Mb/minute. This is very close to uncompressed WAV 16/44.

There are some notable exceptions on the list, by quite some margin too which is curious.

http://www.2l.no/hires/index.html?

Edit: I mean 10 megabytes above, should have used MB to be precise.

Edited February 4, 2016 by gcgreg

[aechmea]

Yeah but hi-rez reorganised/folded to 16/44.1 can then be FLACed to even smaller size.

 

The problem is that the 'folded' stuff in the lower significant bits seems to be more like random noise/dithering and could be hard/impossible for algorithms like FLAC to compress.  The result is that FLAC'd MQA could be larger than FLAC'd 16/44 WAV.  Compressed but not as much.

 

Anyway, I am a glass-half-empty sort of person so I can't see this getting any more traction than previous attempts such as SACD, HDCD, DVD-A (all of which I have tried).  An initial blip followed by a steady decline.  How many of the music buying/listening public are in the slightest bit interested?

[Guest Eggcup The Daft]

I'd urge people to check out http://archimago.blogspot.ca/2016/01/measurements-mqa-master-quality.html#more to get an early assessment of an MQA file.

All this seems exactly as advertised for a non-MQA DAC, apart from the subjective report about the piano sound seeming a touch artificial rather than the promise of processed 16/44.1 being slightly improved.

 

As I understood it, the data in the low eight bits is packed, so will look like noise from a non-MQA DAC as described. Is that right?

 

I guess we need data with an MQA equipped DAC in place!

[New Sensations]

Flac is the file wrapper, MQA the codec.

Sent from my iPhone using Tapatalk

[thoglette]

So far, I'm not impressed,  It walks, talks and smells like half a dozen other "improvements" which have attempted to rebuild the walled garden that the studios used to own.

 

Bandwidth being basically free there's no joy in adding yet another "secret" coding scheme - just use more bits and keep it as compressed PCM.

 

The pre-ringing removal sounds interesting but given the mess usually made in "remastering" by the studios I'm not expecting them to do anything with it.  Frankly releasing that as a plug in for standard tools, coupled with a database of "these settings work for track x on recording y"  would  be much more useful.

 

I'd love to be proved wrong.

[Guest Bonobo]

Yeah but hi-rez reorganised/folded to 16/44.1 can then be FLACed to even smaller size.

 

it is folded to 24/44 .... but seem like the last 8 bits don't compress (noise).

[Guest Bonobo]

Meridian says to consumer:

 

2. At the lowest level the keys verify that the stream is genuinely MQA. This is important for the full benefit of Authentication to be realised and we hope that facility will ignite new and enriched ways for artists to communicate with fans and for listeners to appreciate ‘the real thing’. MQA is neither a DRM nor conditional-access system; listeners can still enjoy the music without a decoder in a variety of legacy playback scenarios, in actual CD quality. However the keys protect the ecosystem.

 

 

 

Meridian says in patent:

 

 

 

In some embodiments the method further comprises the steps of analysing a spectrum of the captured audio, and choosing the decimation filter responsively to the analysed spectrum. The method may then further comprise the step of furnishing information relating to the choice of decimation filter for use by a decoder.

 

 In some embodiments the method further comprises the steps of analysing the noise floor of the captured audio and choosing the decimation filter responsively to the analysed noise floor. In that way both the decimation filter and a corresponding reconstruction filter in a decoder can be optimally matched to the noise spectrum or other characteristics of the signal to be conveyed.

 

....

 

wherein the combined response of the encoder and decoder produce a total system impulse response having a duration for its cumulative absolute response to rise from 1 % to 95% that is less than the characterising duration of the impulse response of the encoder alone and the characterising duration of the impulse response of the decoder alone.

 

This aspect may be useful when special characteristics of the material being encoded require extra poles or zeros in the encoder frequency response to address spectral regions with high levels of noise in the captured audio. Corresponding zeros or poles in the decoder response cause the special measures to have no effect on the passband of the complete system, and also lead the complete system impulse response to be unchanged by the special measures. The individual encoder and decoder responses are however lengthened by the measures and may both be longer than the combined system response.

 

 

 

... and many more (it is long)

 

 

This should be scary!!!!

 

 

Meridian chose some encoding filter... and then they tell which encode filter (secretly) encoded in the file.

 

When play back with MQA decode, you get the decode filter which match encode filter, to make output which is right.

 

When play back with no MQA decode, you have sound distorted by encode filter .... that it has no matching decode filter.

 

 

 

Meridian can make system select encode filter which when listened without corresponding (secret) decode ..... it will sound bad....  and with the right decode filter chosen by MQA system, it will sound better   (maybe even improve sound over original recording).

 

Other thing Meridian says, tells about the encryption which is used so none can be reverse engineer.

 

 

 

We can do these benefits for audio already.... with out need new format  (upsample audio, and convolve with new filter)

 

Meridian want a system where all music is encode, and then only them can allow you to decode the good sound....   they get power to make audio sound good if you pay, and sound bad if you don't pay.

 

Over time, they can make sound without MQA decode get worse and worse and worse ..... until everybody believe MQA magic.

 

 

Everybody pay for light to come on the box.   Bob Stuart says himself, it makes a smile.      Maybe you don't even notice.  Not much, $3 for plater, $0.1 for music .....   but Meridian get rich, and everybody get confused.

 

 

 

Warning.

 

 

 

 

[Guest Eggcup The Daft]

Meridian says to consumer:

 

 

 

Meridian says in patent:

 

 

 

 

... and many more (it is long)

 

 

This should be scary!!!!

 

 

Meridian chose some encoding filter... and then they tell which encode filter (secretly) encoded in the file.

 

When play back with MQA decode, you get the decode filter which match encode filter, to make output which is right.

 

When play back with no MQA decode, you have sound distorted by encode filter .... that it has no matching decode filter.

 

 

 

Meridian can make system select encode filter which when listened without corresponding (secret) decode ..... it will sound bad....  and with the right decode filter chosen by MQA system, it will sound better   (maybe even improve sound over original recording).

 

Other thing Meridian says, tells about the encryption which is used so none can be reverse engineer.

 

 

 

We can do these benefits for audio already.... with out need new format  (upsample audio, and convolve with new filter)

 

Meridian want a system where all music is encode, and then only them can allow you to decode the good sound....   they get power to make audio sound good if you pay, and sound bad if you don't pay.

 

Over time, they can make sound without MQA decode get worse and worse and worse ..... until everybody believe MQA magic.

 

 

Everybody pay for light to come on the box.   Bob Stuart says himself, it makes a smile.      Maybe you don't even notice.  Not much, $3 for plater, $0.1 for music .....   but Meridian get rich, and everybody get confused.

 

 

 

Warning.

I'm not sure that you are absolutely right in your understanding of this. The "secret" filter is only used by an MQA equipped DAC; it can't be used by an non-MQA setup; so unless they change the base part of the encoding what we get today on non-MQA DACs should hold into the future.

 

The information currently appearing (see a.dent's link above to the Archimago blog) should therefore remain accurate with non-MQA DACs.