Jump to content
Rt66indierock

MQA is Vaporware

Rate this topic

Recommended Posts

9 minutes ago, mansr said:

I have looked at rips of a few MQA CDs. All of them had non-PCM data in the LSB only. The data in this bit did, however, look a bit different from 24-bit MQA files in that much more of the bitstream was used for some kind of data rather than metadata or holes

[snip]

 

Ok, this makes sense. But rather than being a touch-up stream (which would only exist in a 24-bit format), this data channel must be the "approximation stream" encoded alongside other data in the LSB.

 

it is this data, when suitably decoded by the MQA Decoder (this has nothing to do with upsampling/rendered), that will reconstruct an approximation of the hi frequency content.

Share this post


Link to post
Share on other sites
19 minutes ago, abrxx said:

Ok, this makes sense. But rather than being a touch-up stream (which would only exist in a 24-bit format), this data channel must be the "approximation stream" encoded alongside other data in the LSB.

 

it is this data, when suitably decoded by the MQA Decoder (this has nothing to do with upsampling/rendered), that will reconstruct an approximation of the hi frequency content.

There seems to be a bit of both kinds of data in both places. If I decode an MQA file, then zap the 8 low bits and decode that, the spectrum of the difference looks like this:

image.thumb.png.b8684f84ceeba73b8b6a262ead76ba00.png

Share this post


Link to post
Share on other sites
44 minutes ago, mansr said:

There seems to be a bit of both kinds of data in both places. If I decode an MQA file, then zap the 8 low bits and decode that, the spectrum of the difference looks like this:

image.thumb.png.b8684f84ceeba73b8b6a262ead76ba00.png

 

Fascinating. Bob has hinted several times that the patents are "outdated" etc, and this sort of change aligns with that.

Share this post


Link to post
Share on other sites
9 minutes ago, abrxx said:

Fascinating. Bob has hinted several times that the patents are "outdated" etc, and this sort of change aligns with that.

The patents are probably only vaguely applicable to the actual product. Enough to make a compatible system infringing, but not enough to enable anyone to make one based on them.

Share this post


Link to post
Share on other sites

Was just looking up the terms of using the FLAC container:

 

What FLAC is not:

Lossy. FLAC is intended for lossless compression only, as there are many good lossy formats already, such as Vorbis, MPC, and MP3 (see LAMEfor an excellent open-source implementation).

DRM. There is no intention to add any copy prevention methods. Of course, we can't stop someone from encrypting a FLAC stream in another container (e.g. the way Apple encrypts AAC in MP4 with FairPlay), that is the choice of the user.

 

Seems to me that someone has not read these terms.

 

Have written to Monty Alexander of Xiph.org to get his two cents

Share this post


Link to post
Share on other sites
On 1/22/2020 at 1:56 PM, FredericV said:


Bob even confirms it ;)

image.png.a371f0ef0fbe967912a786d78ea4cef6.png

 

For a 24-bit file, MQA reconstructs the high frequency "remarkably close". This is because most likely, if it is anything like the patent, there is space for both an approximation stream and a touch-up, The "math" is basically approx + touch_up somewhat equals original.

 

For a 16-bit file, there is a low bandwidth approximation stream (only 1 bit in the file that @mansr looked at), that MQA decoder can use to construct an approximate waveform for the high frequency.

 

According to Bob, if you truncate the 24-bit file to 16-bits, it should behave in a similar way to the MQA CD. Perhaps identically?

 

 

Share this post


Link to post
Share on other sites
53 minutes ago, abrxx said:

For a 16-bit file, there is a low bandwidth approximation stream (only 1 bit in the file that @mansr looked at), that MQA decoder can use to construct an approximate waveform for the high frequency.

 

According to Bob, if you truncate the 24-bit file to 16-bits, it should behave in a similar way to the MQA CD. Perhaps identically?


I did run stripped MQA (24 bit with bottom 8 bits blanked or replaced by an ascii string so MQA discards or lacks the required data for the first unfold ultrasonic content) through an MQA decoder and also through a leaky sox upsampler (so no MQA involved), and do not see such a big difference.

They both generate fake aliased content.

And then Bob can claim there's audio at 30+ Khz ;)

To generate this with squeezelite, try this:

 


Designer of the 432 EVO music server and Linux specialist

Discoverer of the independent open source sox based mqa playback method with optional one cycle postringing.

Share this post


Link to post
Share on other sites

Hi,

Not sure if this has been covered before, or whether it is a valid discussion in respect to MQA.

 

For my active speaker project, i had to calculate the delay required between the drivers to ensure time alignment of the signal. The times required between woofer, mid and tweeter were 120us and 54us.

 

Given that the majority of speakers do not implement any time alignment between drivers, then surely any gains from MQA re-aligning the digital filter delays (dispersion) are thoroughly f*cked up by the speakers which do not have time alignment ?

 

Is the time alignment error prevalent in so many speakers dominant in any timing issues of a recording ?

 

Regards,

Shadders.

Share this post


Link to post
Share on other sites
11 hours ago, FredericV said:

Still waiting for proof from MQA that MQA does not cause a frequency dependent phase shift.

The render filters are not linear phase, so they definitely cause phase shifts. That said, since they are so shallow, the phase shift isn't all that great.

Share this post


Link to post
Share on other sites
On 1/30/2020 at 1:23 AM, FredericV said:

I don't believe MQA is fixing phase issues

Hi,

I thought the purpose of MQA was to correct the dispersion in the existing recordings. Dispersion is due to non linear filters (example) which do not have a constant group delay.

 

AES Paper : A Hierarchical Approach to Archiving and Distribution

 

So, in MQA claiming that they solve the timing issues, what they are neglecting to recognise is that the missing time alignment in the majority of speakers swamps any correction they claim by MQA in a recording.

 

As such, MQA claims to solve a problem that we can never benefit from from the majority of hifi setups.

 

Regards,

Shadders.

Share this post


Link to post
Share on other sites
3 hours ago, John Dyson said:

MQA is hierarchical in the sense that the consumer is on the low end of the hierarchy -- it's main purpose is to control distribution in one way or another.  There has been no plausible technical advantage given the pre-existing defects and given the technology available to the consumer today.

 

When streaming from a streaming service, how can a consumer be assured that he/she is receiving an unadulterated stream, that is, how can anyone be sure that the streaming service isn't somehow mangling the bits before they are transmitted and arrive at the consumer's equipment?  At lease with MQA, my blue light lets me know that Tidal is not messing with the files, even though the MQA files are slightly lossy and have some "MQA noise".  (I no longer subscribe to Tidal.)

 

Another benefit for me is that I can use captured MQA files to test audio drivers and player/recorder/studio software to ensure that the driver/software, when all DSP options are turned off, is not mangling the bits delivered to the DAC.  Of course, once I am convinced that the driver/software is working as it should, in this case, I no longer have any need to suffer listening to the MQA files.

Share this post


Link to post
Share on other sites
53 minutes ago, The Computer Audiophile said:

I do this with HDCD files and a Berkeley DAC. It’s nice to know nothing is being mangled. However, the total cost of MQA is overbearing compared to this tiny benefit. 

I understand the reassuring idea of having a 'checksummed' file -- but...

 

Gotta recognize also, that just because something has the MQA label and someone has signed off on it, it does NOT guarantee the best available quality.

 

Interestingly, everyone reading this knows about my project, and  about the travesty that had been eventually accepted by most audiophiles and non-audiophiles soon after CDs came out...   I was one of the few who left the hobby partially because the digital technical quality problems did *not* meet my informed sense of how recordings should sound (this was back in the late 1980s, when I could really hear well.)  I wasn't amused by the fidget of turntables, nor the lack of convenience of the higher quality tape.  After initially believing the promise of quality when using the CD medium, then being disappointed -- I felt *terminally* disappointed.

 

I *loved* listening to music, and still do.  Just felt very cheated, felt most audiophiles were also cheated, eventually developing a negative 'attitude' about the situation.

 

My new hobby/avocation/vocation did not result from any kind of 'crusade' or lingering agenda, I truly just happened into the solution to the problem, it was easier for me to diagnose because of highly technical background and discipline.  This has NOT been a lingering agenda since the 1980s, and frankly, enjoyed, but seldom listened to music in the 2012 timeframe when a major CD audio problem was finally correctly diagnosed...

 

So, the question is:  Given that the standards would be established by the typical perptrators of the original problem with CDs, what is the actual benefit of an *additional* arbitrary judgement of perfection?  Signing off by another person isn't going to guarantee improved or even equivalent quality.  My expectations of quality audio getting to the consumer are now very low -- by serendipity, there JUST MIGHT be a new alternative for getting good sound to the consumer.

 

I just don't see that MQA is a sufficient, adequate or even significantly helpful scheme to get good audio to where it should go.  MQA is an obfuscation, with a slight incremental benefit with lots of downside.

 

The streaming services can fix the problem, basically sell the servic as superior by providing superior quality audio.  Lighting up an LED based upon yet another person's opinion doesn't really fix the problem -- just adds complication..

 

 

John

 

Share this post


Link to post
Share on other sites
15 minutes ago, John Dyson said:

Interestingly, everyone reading this knows about my project, and  about the travesty that had been eventually accepted by most audiophiles and non-audiophiles soon after CDs came out...   I was one of the few who left the hobby partially because the digital technical quality problems did *not* meet my informed sense of how recordings should sound (this was back in the late 1980s, when I could really hear well.)  I wasn't amused by the fidget of turntables, nor the lack of convenience of the higher quality tape.  After initially believing the promise of quality when using the CD medium, then being disappointed -- I felt *terminally* disappointed.

 

I *loved* listening to music, and still do.  Just felt very cheated, felt most audiophiles were also cheated, eventually developing a negative 'attitude' about the situation.

 

My new hobby/avocation/vocation did not result from any kind of 'crusade' or lingering agenda, I truly just happened into the solution to the problem, it was easier for me to diagnose because of highly technical background and discipline.  This has NOT been a lingering agenda since the 1980s, and frankly, enjoyed, but seldom listened to music in the 2012 timeframe when a major CD audio problem was finally correctly diagnosed...

 

 

 

 

Can't let this one go, John 😉 ... you refuse to accept that the quality of the playback chain is critical to getting the SQ that appeals subjectively; so you're determined that "something must be wrong with CDs" that displease you - your solution is to 'remaster' them ...

 

And that works, of course ... if you have a car with lousy suspension, then go and 'fix' every road you want to drive on that makes the journey unpleasant - from my POV, this is not an efficient solution, 🙂.

 

I worry that your "crusade" is as lopsided as is MQA - another workaround that allows people to get away with using equipment that's casually assembled, and that manifests its shortcomings when asked to replay anything that's more 'testing' ... a two wrongs make it right solution.


Frank

 

http://artofaudioconjuring.blogspot.com/

 

 

Over and out.

.

 

Share this post


Link to post
Share on other sites
37 minutes ago, fas42 said:

 

Can't let this one go, John 😉 ... you refuse to accept that the quality of the playback chain is critical to getting the SQ that appeals subjectively; so you're determined that "something must be wrong with CDs" that displease you - your solution is to 'remaster' them ...


🙂

 

* I regret needing to make a correction about an alledged position that I take.   Positions should not be purposefully misconstrued -- I don't take a strong opinion on playback equipment other than it is obvious that poor quality equipment can make it more difficult to discern quality issues, or even play material in the highest quality way possible.  For example, trying to make equipment play a .flac file by cheating and calling it a .wav file will not make the file work very well.  Likewise, Dolby or DBX encoded material cannot be played perfectly by equipment which doesn't know how to decode it.  This is true NO MATTER the quality of equipment.

 

Sometimes something is wrong with the equipment being used by the consumer, in my case that isn't true.  I precise in what I do and use.   Please do NOT misquote or purposefully misinterpret my position.  I am very picky about things, never giving up, and perhaps that is one reason why this digital recording problem has been diagnosed and a technical way to provide a solution is coming into existence.

 

The historical digital recording situation is analoguous to listening to undecoded DolbyB -- it will always have that 'distortion' in the sound unless it is properly handled (term of art: decoded.)  The very best equipment, but doesn't know how to process the DolbyB signal (or DBX or whatever) will simply not sound *correct.  (No qualitative claim about better or worse is being made here -- just speaking of relative accuracy.)   A filtered DolbyA signal has similar effects as a filtered/statically EQed DolbyB or DolbyC ssignal would have.  An EQed DolbyB vs EQed DolbyA signal has similar, but not identical sorts of defects.  Many/mostl CDs made and/or mastered back in the 1980s through early 1990s have this kind of defect.  Some recently made CDs still have this defect.

 

* I use the term 'correct' as in an informed knowledge of how recordings actually sound, and with a knowledge of the kinds of distortions or manipulations being used in the recording arnea.

 

Take a flac file and try to play it as uncompressed data as in a signed integer or floating point .wav file -- similar thing, but worse consequences (that is, bypassing whatever file type detection that is used in the software.)  The very best equipment, with the misused data from a flac file will sound like garbage.


To make things clear:  Many/most CDs, at least I know the early ones, were not in the form that a normal CD player and consumer playback chain could optimally play.  This is the extent of my claim, and no other inferrence about other aspects of the system or listener are implied or should be construed.


Any claims that some magical 'great quality' consumer equipment using many/most normally available consumer CDs could accurately reproduce the mixed/recorded signal from the studio are untrue.  It is however, very possible to decode many of the CDs, thereby coming much closer to what was recorded onto tape in the studio.

 

John

 

Share this post


Link to post
Share on other sites

 

Quote

To make things clear:  Many/most CDs, at least I know the early ones, were not in the form that a normal CD player and consumer playback chain could optimally play.  This is the extent of my claim, and no other inferrence about other aspects of the system or listener are implied or should be construed.

 

 

How do you know this, John? ... Who else says this, who is in the business of creating CDs, etc, for release?


Frank

 

http://artofaudioconjuring.blogspot.com/

 

 

Over and out.

.

 

Share this post


Link to post
Share on other sites
4 hours ago, John Dyson said:

I understand the reassuring idea of having a 'checksummed' file -- but...

 

Gotta recognize also, that just because something has the MQA label and someone has signed off on it, it does NOT guarantee the best available quality.

 

I'm not sure you understood.  The case discussed was not about guarantying that the file played was the best quality. Rather, it was about using an MQA stream/file and an MQA decoder strictly for (1) testing sound drivers and player/recorder/studio software, and (2) proving that a streaming service is not mangling bits (limited to Tidal in this case -- for other streaming services, we have no way of verifying that we are receiving an unadulterated stream).

 

 

 

 

 

 

 

Share this post


Link to post
Share on other sites

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...