Why do you refuse to believe the end game is to eliminate all MQA unprotected content?
I think your heads in the sand no matter how long you worked in the industry.
https://www.onkyomusic.com/US/home
-
Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
MQA: A Review of controversies, concerns, and cautions
- Thread starter mitchco
- Start date
- Status
I think your heads in the sand no matter how long you worked in the industry.
Well defined, already: http://www.dictionary.com/browse/blur
Whatever, even if there is a definition particular to audio then I have yet to see it. The use of the term in this context is blurry and remains satisfactorially unexplained.
Whatever, even if there is a definition particular to audio then I have yet to see it. The use of the term in this context is blurry and remains satisfactorially unexplained.
Industry related bias is a real phenomena. Just another form of group bias.
Last edited:
And as for this business of MQA-style shenanigans being hard, or it being difficult to develop a perceptual codec... it's not. It could easily be one person in a bedroom type stuff.
I'm no mathematician, but I know how to process overlapping windows of sampled data with an FFT, do a bit of heuristic rules-based messing about with the results (maintain full resolution here, quantise resolution there, discard some bins) based on human hearing sensitivity curves and the much praised 'masking'. Then do the reverse FFT on the simplified data. I could cobble something together in an afternoon, I reckon. Maybe I'll try it...
One thing I know: if the main thrust of the task is compressing the ultrasonics i.e. the bits people can't hear anyway, it won't be too hard. I can then throw together some gibberish gobbledigook in a 'paper' to impress the masses, give some ego massaging interviews to the salivating audio press and I'm on my way...
I'm no mathematician, but I know how to process overlapping windows of sampled data with an FFT, do a bit of heuristic rules-based messing about with the results (maintain full resolution here, quantise resolution there, discard some bins) based on human hearing sensitivity curves and the much praised 'masking'. Then do the reverse FFT on the simplified data. I could cobble something together in an afternoon, I reckon. Maybe I'll try it...
One thing I know: if the main thrust of the task is compressing the ultrasonics i.e. the bits people can't hear anyway, it won't be too hard. I can then throw together some gibberish gobbledigook in a 'paper' to impress the masses, give some ego massaging interviews to the salivating audio press and I'm on my way...
From my unlofty and simple-minded POV. I have to seriously ask: What does MQA solve as a problem?
For me the answer shouldn't lie in what Mr Stuart says but to peer-reviewed analysis.
Anything from the AES in which MQA treated things are proven superior to non-MQA treated things?
I am also rather taken aback by the curves that purports the superiority of R2R and other to good ol' digital ... It screams of money grabbing IMO...
my $0.02
For me the answer shouldn't lie in what Mr Stuart says but to peer-reviewed analysis.
Anything from the AES in which MQA treated things are proven superior to non-MQA treated things?
I am also rather taken aback by the curves that purports the superiority of R2R and other to good ol' digital ... It screams of money grabbing IMO...
my $0.02
Don Hills
Addicted to Fun and Learning
... in spite of repeated attempts by the industry. There's also "MQA CD."
Didn't this all start to a lesser degree with Dolby Noise Reduction?
Don Hills
Addicted to Fun and Learning
It is for now. The option exists to also scramble part of the baseband stream, leaving just 4 bits "in the clear". The decoders already have the code to handle this. The problem with this is not technological. It can be reverse engineered. It just can't be done legally without paying a royalty to MQA.
The process is more or less like this, taking a 192k master as example:
1) measure the master's innate noise in the audible band (20kHz), this is Ln
2) downsample from 192k to 96k with the minimum-width (i.e. 'laziest') anti-alias filter that fullfils these criteria
-it puts a null at the original mastering ADC's ringing frequency
-the aliasing it causes (and it will cause this, since it is 'lazy') in the audible band does not exceed Ln
3) take the resulting 96k file and split it into its baseband and ultrasonic band. This requires
quadrature-matched filters. Downsample the ultrasonic band by unfiltered decimation. Now there
are two 48kHz signals, the baseband Bb and the ultrasonic band Bu.
4) observe Ln again and determine a suitable bit budget for Bb. Requantise Bb to this (this will be 15-17 bit or so).
The remaining bits from 24 are the budget for Bu.
5) Scale Bu according to the 7 or so bits it will have, note the scaling factor in the embedded control stream.
6) Requantise Bu to its bit budget.
7) Scramble/encrypt Bu and the control channel into pseudo-random noise and tack this below Bb(*). This gives a 48kHz 24 bit MQA encoded file.
Replay is roughly the above in reverse, with another set of quadrature filters.
(* in reality the scrambled data are partially placed in the upper 16 bits, so that MQA can survive a CD-resolution channel, or indeed be
sold on a CD.)
The hard part is 3), specifically the design or choice of the QMFs. These QMFs must guarantee a mathematically lossless
band split and join. As such, it is highly unlikely that they are at the same time optimised for the sound quality of
an undecoded MQA file. Why is this important? Because when listening to undecoded MQA you are directly listening
to the anti-aliasing filter used in creating Bb, which is one of the QMFs.
I suspect that the CD-rate replay filter of MQA DACs, which according to Stereophile measurements are the same for all MQA DACs, is nothing more or less than one of the replay QMFs. If so its performance is attrocious for CD-rate use.
Many thanks for the description of MQA. I think my suspicion, though, is that they really could have come up with any tortuous scheme they liked and still been able to sell it. As was often pointed on a BBC reality show called Dragons' Den, ideas are important, but business acumen is the really valuable commodity when it comes to reaping profits. Commercial strongarm tactics are what can turn a mediocre product into a market-dominating cash cow.
Because I understand the industry? You want me to forget what I know and believe fantastical theories here. There is no possibility whatsoever of that happening. You honestly think companies like Apple and Amazon will license MQA and pay royalties? And that after years of offering music in the clear, the labels will switch to an unknown format among consumers and even industry like MQA?
There are larger odds of an asteroid hitting you than anything like this happening.
I know what it takes to create a new format and MQA both lacks the value proposition to get there and financial wherewithal to remotely make this happen.
FUD is not a good thing to cling to.
I am not sure what you would do with 4 bits.
But here is the main problem: the main value proposition is its backward compatibility with PCM. Take that away and MQA loses just about everything that it is about. If someone wants fully encrypted content, you would use a system designed to do that, e.g. the many DRM systems out there. They are professionally designed and can withstand attacks. Nothing remotely like that is in MQA. It will be torn open with ease and stay broken. There is no evidence at all that such a proposition is sold and accepted by the labels.
Under the scenario of "anything can happen" we can manufacture all kinds of scenarios. Like Apple phones only playing some new audio format they invent. We don't go there even though Apple has far more power than MQA has. But somehow willing to accept all of these theories against MQA.
Well, we're not alone.
https://arstechnica.com/gadgets/201...hing-you-need-to-know-about-high-res-audio/8/
Warner and UMG support is for high-res music which MQA helps with. It has nothing to do with any form of copy protection which MQA lacks anyway.
As to the speculation from the author on "better control," here is the bio of that writer:
Andrew Harrison specialises in audio, and has been testing equipment and writing on hi-fi and sound technology for nearly 20 years for magazines such as Hi-Fi News, Hi-Fi World, and Hi-Fi Critic. More recent postings include technical editor at Macworld UK and PC Advisor, adding expert hardware reviews for The Register and BBC Focus magazine. He now enjoys the possibilities of high-end computer audio while never forgetting the magic of vinyl played on a good turntable.
Nothing there gives him any experience to make such predictions. The Sony which produced SACD, is not the Sony music we have today. All of those people who were advocating those formats were fired after Apple offered download service and they realized that was the future, not pushing high-res discs.
As to the speculation from the author on "better control," here is the bio of that writer:
Andrew Harrison specialises in audio, and has been testing equipment and writing on hi-fi and sound technology for nearly 20 years for magazines such as Hi-Fi News, Hi-Fi World, and Hi-Fi Critic. More recent postings include technical editor at Macworld UK and PC Advisor, adding expert hardware reviews for The Register and BBC Focus magazine. He now enjoys the possibilities of high-end computer audio while never forgetting the magic of vinyl played on a good turntable.
Nothing there gives him any experience to make such predictions. The Sony which produced SACD, is not the Sony music we have today. All of those people who were advocating those formats were fired after Apple offered download service and they realized that was the future, not pushing high-res discs.
But...
https://arstechnica.com/gadgets/201...hing-you-need-to-know-about-high-res-audio/4/
It seems to me that it is quite the swiss army knife when it comes to DRM, and can be as secure as they want, especially seeing as it requires a dedicated chip.
Here is the actual quote from record labels when they announced their support: https://www.billboard.com/articles/...ra-riaa-announce-support-hi-res-streaming-mqa
See? They simply see it as a way to bringing streaming of high-res content to users. They hope this means more royalties as distributors license high-res content in the form of MQA.
See? They simply see it as a way to bringing streaming of high-res content to users. They hope this means more royalties as distributors license high-res content in the form of MQA.
They can throw all the ideas they want in a patent application. Creating a secure content distribution system is hugely difficult and labor intensive against all the attacks it will receive. And that is on top of having to build the thing in the first place which they have not. It is also a patent land mine and will attract litigation like nobody's business.
And what dedicated chip? Tidal decodes MQA purely in software today. A simple firmware update enabled Audioquest Dragonfly to do the same. No chip is needed or required in either.
The purpose of the secure signatures is to detect licensed devices. Not content.
There is a reason there are just a handful of companies in the world that can create such secure systems.
MQA is what it is today and no more. It is on that basis that it has support. A backward compatible, in the clear baseband music format, with enhancements to produce high-res content from that. It enables bandwidth efficiency for providing high-resolution content and that is the reason it is getting adopted. No more. Nothing else is there. What dreams or fantasies patent writers had is just that: dreams.
TIDAL surely just delivers MQA - it doesn't decode it..? If that was the case, we could have 'room correction' and DSP crossovers today just by directing TIDAL's output to the appropriate driver...?
The high res digital stream stays inaccessible until final decoding in a chip, surely..? Isn't that the big deal about "end-to-end"? It may not be MQA's chip, but they could ensure that licensees use secure, approved hardware..?
(I may be missing something about how the scheme works...)
svart-hvitt
Major Contributor
- Joined
- Aug 31, 2017
- Messages
- 2,375
- Likes
- 1,265
- Status
Similar threads
